www.nature.com/natureneuroscience
EDITORIAL OFFICE neurosci@natureny.com
345 Park Avenue South, New York, NY 10010-1707
Tel: (212) 726 9319, Fax: (212) 696 0978
Editor: Sandra Aamodt
Senior Editor: Kalyani Narasimhan
Associate Editors: Cara Allen, I-han Chou, Annette Markus, Charvy Narain
Copy Editor: Dorothy Moore, Lavanya Reddy
Senior Production Editor: Nicole D. Fournier
Production Editor: Jessica Iannuzzi
Cover Design: Kimberly Caesar
Editorial Assistant: Jessica Chen
MANAGEMENT OFFICES
NPG New York
345 Park Avenue South, New York, NY 10010-1707
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience

Tel: (212) 726 9200, Fax: (212) 696 9006
Publisher: Beatrice Renault
Executive Editor: Linda Miller
Chief Technology Officer: Howard Ratner
Global Head of Advertising: Fabien Savenay
Head of Nature Research & Reviews Marketing: Sara Girard
Assistant Marketing Manager: Amy Maurer
Production Coordinator: Diane Temprano
Associate Director, Web Publishing: Timo Hannay
New Technology Director: Greg Suprock
Head of Web Services: Anthony Barrera
Web Production Editor: Edwina Hay

NPG London
The Macmillan Building, 4 Crinan Street, London N1 9XW
Tel: 44 207 833 4000, Fax: 44 207 843 4996
Managing Director: Annette Thomas
Publishing Director: Peter Collins
Editor-in-Chief, Nature Publications: Philip Campbell
Marketing Director: Della Sar

NPG Tokyo
Chiyoda Building 2-37 Ichigayatamachi, Shinjuku-ku, Tokyo 162-0843
Tel: 81 3 3267 8751, Fax: 81 3 3267 8746
Asia-Pacific Publisher: Antoine E. Bocquet
Manager: Koichi Nakamura
Senior Marketing Manager: Peter Yoshihara
Asia-Pacific Sales Director: Kate Yoneyama
Asia-Pacific Sales Manager: Rinoko Asami
DISPLAY ADVERTISING display@natureny.com (US/Canada)
display@nature.com (Europe)
nature@naturejpn.com (Japan)
US Head of Display Advertising: Stephen Schwartz, Tel: (212) 726 9256, Fax: (212) 696 9481
Global Head of Display Advertising Sales: John Michael, Tel: 44 207 843 4960, Fax: 44 207 843 4996
Head of Display Advertising—Europe: Gerard Preston, Tel: 44 207 843 4960, Fax: 44 207 843 4996
Business Development Manager: Claire Hines, Tel: 44 207 843 4960, Fax: 44 207 843 4996
Asia-Pacific Sales Manager: Rinoko Asami, Tel: 81 3 3267 8751, Fax: 81 3 3267 8746
Western Region Sales Manager: George Lui, Tel: (415) 781 3804, Fax: (415) 781 3805
Sales Executives:
New England: Sheila Reardon, Tel: (617) 399 4098, Fax: (617) 426 3717
New York, Mid-Atlantic, Southeast: Jim Breault, Tel: (212) 726 9334, Fax: (212) 696 9481
Midwest: Mike Rossi, Tel: (212) 726 9255, Fax: (212) 696 9481
Northwest: Mathieu DesJardins, Tel: (415) 781 6422, Fax: (415) 781 3805
Eastern England/Scotland/Italy/Spain/Israel: Matthew Wilkinson, Tel: 44 207 843 4960, Fax: 44
207 843 4749
South/West United Kingdom/Scandinavia/Holland: Marianne Boulakas, Tel: 44 207 843 4969,
Fax: 44 207 843 4749
Northern Germany: Gerard Preston, Tel: 44 207 843 4960, Fax: 44 207 843 4749
Southern Germany/Austria/Switzerland/France/Belgium: Sabine Hugi-Fürst, Tel: 41 52761 3386,
Fax: 41 52761 3419
NATUREJOBS naturejobs@natureny.com (US/Canada)
naturejobs@nature.com (Europe)
advertising@naturejpn.com (Japan)
Publisher: Ben Crowe, Tel: (212) 726 9245, Fax: (212) 696 9482
US Sales Manager: Peter Bless, Tel: (212) 726 9248, Fax: (212) 696 9482
Japan Sales Manager: Rinoko Asami, Tel: 81 3 3267 8751, Fax: 81 3 3267 8746
naturejobs Sales Director: Nevin Bayoumi, Tel: 44 207 843 4961, Fax: 44 207 843 4996
SITE LICENSE BUSINESS UNIT
Americas: Tel: (888) 331 6288 institutions@natureny.com
Asia/Pacific: Tel: 81 3 3267 8751 institutions@naturejpn.com
Australia/New Zealand: Tel: 61 3 9825 1160 nature@macmillan.com.au
India: Tel: 00 91 11 324 4186 harpal@nature.com
ROW: Tel: 44 207 843 4609 institutions@nature.com
CUSTOMER SERVICE www.nature.com/help
Senior Global Customer Service Manager: Gerald Coppin
For all print and online assistance, please visit www.nature.com/help
Purchase subscriptions:
Americas: Nature Neuroscience, Subscription Dept., 303 Park Avenue South #1280, New York, NY
10010-3601. Tel: (866) 363 7860, Fax: (212) 689 9108
Europe/ROW: Nature Neuroscience, Subscription Dept., Macmillan Magazines Ltd., Brunel Road,
Houndmills, Basingstoke RG21 6XS, United Kingdom. Tel: 44 1256 329 242, Fax: 44 1256 812 358
Japan: Nature Neuroscience, Nature Japan K.K., MG Ichigaya Bldg. 5F, 19-1 Haraikatamachi,
Shinjuku-ku, Tokyo 162-0841. Tel: 81 3 3267 8751, Fax: 81 3 3267 8746
India: Harpal Singh Gill, Macmillan Magazines Ltd, 5A/12 Ansari Road, Darya Ganj, New Delhi,
110 002 India. Tel: 00 91 11 324 4186, Tel/Fax: 00 91 11 327 2010
REPRINTS reprint@natureny.com
Nature Neuroscience Reprint Department, Nature Publishing Group, 345 Park Avenue South, New
York, NY 10010-1707, USA.
For commercial reprint orders of 600 or more, please contact:
UK Reprints Sales Executive: Christine Fothergill, Tel: 44 207 843 4967, Fax: 44 207 843 4749
US Reprints Sales Executive: Sharda Tulsie, Tel: (212) 726 9631, Fax: (212) 679 0843

VOLUME 8 NUMBER 11 NOVEMBER 2005

E D I TO R I A L
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience

1413 Taking addiction research into the clinic

BOOK REVIEW
1415 The War of the Soups and the Sparks
Addiction is pervasive, affecting By Elliot S Valenstein
millions of people around the world.
The progression from recreational
Reviewed by Nicholas C Spitzer
drug use to drug dependence and
addiction is influenced by many
factors, including the nature of the NEWS AND VIEWS
drug, the personality of the user, and
1417 How neurons keep in touch
environmental stressors. In this issue,
we present reviews and opinion pieces Fekrije Selimi & Nathaniel Heintz  see also p 1534
on the neurobiology of drug abuse, 1418 Synaptic plasticity and self-organization in the hippocampus
decision making and habit formation,
as well as a commentary on how the
György Buzsáki & James J Chrobak  see also p 1560
neuroscience of addiction should 1420 Glial cells under remote control
guide public policy and treatment.
Klaus-Armin Nave & Markus H Schwab
This special focus is sponsored by the
National Institute on Drug Abuse and 1422 Senseless makes sense for spinocerebellar ataxia-1
the National Institute on Alcohol Abuse
Vikram Khurana, Tudor A Fulga & Mel B Feany
and Alcoholism.
(pp 1427–1489) 1424 How the brain recovers following damage
Yalçin Abdullaev & Michael I Posner  see also p 1603
1425 Time to smell the roses
Cara Allen  see also p 1568

I N T RO D U C T I O N : N E U RO B I O LO G Y O F A D D I C T I O N
1427 Neurobiology of addiction
I-han Chou & Kalyani Narasimhan

S P O N S O R S ’ F O R E WO R D : N E U RO B I O LO G Y O F
ADDICTION
1429 The neuroscience of addiction
Nora Volkow & Ting-Kai Li

Semaphorin controls cerebellar
granule cell migration
(p 1516)

Nature Neuroscience (ISSN 1097-6256) is published monthly by Nature Publishing Group, a trading name of Nature America Inc. located at 345 Park
Avenue South, New York, NY 10010-1707. Periodicals postage paid at New York, NY and additional mailing post offices. Editorial Office: 345 Park Avenue
South, New York, NY 10010-1707. Tel: (212) 726 9319, Fax: (212) 696 0978. Annual subscription rates: USA/Canada: US$199 (personal), US$1,809
(institution). Canada add 7% GST #104911595RT001; Euro-zone: €271 (personal), €1,558 (institution); Rest of world (excluding China, Japan, Korea):
£175 (personal), £1,005 (institution); Japan: Contact Nature Japan K.K., MG Ichigaya Building 5F, 19-1 Haraikatamachi, Shinjuku-ku, Tokyo 162-0841.
Tel: 81 (03) 3267 8751, Fax: 81 (03) 3267 8746. POSTMASTER: Send address changes to Nature Neuroscience, Subscriptions Department, 303 Park
Avenue South #1280, New York, NY 10010-3601. Authorization to photocopy material for internal or personal use, or internal or personal use of specific
clients, is granted by Nature Publishing Group to libraries and others registered with the Copyright Clearance Center (CCC) Transactional Reporting Service,
provided the relevant copyright fee is paid direct to CCC, 222 Rosewood Drive, Danvers, MA 01923, USA. Identification code for Nature Neuroscience:
1097-6256/04. Back issues: US$45, Canada add 7% for GST. CPC PUB AGREEMENT #40032744. Printed by Publishers Press, Inc., Lebanon Junction,
KY, USA. Copyright © 2005 Nature Publishing Group. Printed in USA.

i

VOLUME 8 NUMBER 11 NOVEMBER 2005

C O M M E N TA R I E S : N E U R O B I O L O G Y O F A D D I C T I O N
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience

1431 Neurobiology of addiction: treatment and public policy ramifications
Charles Dackis & Charles O’Brien
1437 The role of neuroadaptations in relapse to drug seeking
Yavin Shaham and Bruce T Hope
1440 How do we determine which drug-induced neuroplastic changes are important?
Peter W Kalivas
1442 Plasticity of reward neurocircuitry and the ‘dark side’ of drug addiction
George F Koob & Michel Le Moal

P E R S P E C T I V E S : N E U RO B I O LO G Y O F A D D I C T I O N
1445 Is there a common molecular pathway for addiction?
Eric J Nestler
1450 Genetic influences on impulsivity, risk taking, stress responsivity and vulnerability
to drug abuse and addiction
Mary Jeanne Kreek, David A Nielsen, Eduardo R Butelman & K Steven LaForge
1458 Decision making, impulse control and loss of willpower to resist drugs:
a neurocognitive perspective
Antoine Bechara

R E V I E W S : N E U RO B I O LO G Y O F A D D I C T I O N
1465 Nicotine addiction and comorbidity with alcohol abuse and mental illness
John A Dani and R Adron Harris
1471 Laboratory models of alcoholism: treatment target identification and insight into
mechanisms
David M Lovinger & John C Crabbe
1481 Neural systems of reinforcement for drug addiction: from actions to habits to
compulsion
Barry J Everitt & Trevor W Robbins

B R I E F C O M M U N I C AT I O N S
1491 The cerebellum communicates with the basal ganglia
E Hoshi, L Tremblay, J Féger, P L Carras & P L Strick
1494 Shift of activity from attention to motor-related brain areas during visual learning
S Pollmann & M Maertens
1497 The essential role of stimulus temporal patterning in enabling perceptual learning
S-G Kuai, J-Y Zhang, S A Klein, D M Levi & C Yu
1500 COMT genotype predicts longitudinal cognitive decline and psychosis in
22q11.2 deletion syndrome
D Gothelf, S Eliez, T Thompson, C Hinard, L Penniman, C Feinstein, H Kwon, S Jin,
B Jo, S E Antonarakis, M A Morris & A L Reiss

ARTICLES
1503 MPS-1 is a K+ channel β-subunit and a serine/threonine kinase
S-Q Cai, L Hernandez, Y Wang, K H Park & F Sesti

NATURE NEUROSCIENCE iii

I Mody & F Bezanilla N AT U R E N E U R O S C I E N C E C L A S S I F I E D See back pages NATURE NEUROSCIENCE v . L Li. T Miyazaki. Z Pang. L P van den Boom. C Kealey. R H Frank & E A Phelps Down Syndrome Critical Region 1 TECHNICAL REPORT and mitochondrial function 1619 A hybrid approach to measuring electrical activity in genetically specified neurons (p 1577) B Chanda. L de Hoz. Y Rong. A Z Snyder & A Sapir  see also p 1424 1611 Perceptions of moral character modulate the neural systems of reward during the trust game M R Delgado. H-C Lu. K G C Westphal. M Goulding & Q Ma Activity-dependent decrease 1516 The transmembrane semaphorin Sema6A controls cerebellar granule cell migration in global excitability G Kerjan. VOLUME 8 NUMBER 11 NOVEMBER 2005 1510 Lbx1 and Tlx3 are opposing switches in determining GABAergic versus © 2005 Nature Publishing Group http://www. D Bao.nature. J Mukai. M J Kincade. A Friedman & U Heinemann  see also p 1418 Binocular rivalry in the 1568 Encoding a temporally structured stimulus with a temporally structured lateral geniculate nucleus neural representation (p 1595) S L Brown. C Lewis. distribution and synaptic plasticity N Rouach. S A Siegelbaum. H Adesnik. D Fricker. G M Elias. P Pavlidis. C Haumaitre. D H Brager. J Joseph & M Stopfer  see also p 1425 1577 Drosophila melanogaster homolog of Down syndrome critical region 1 is critical for mitochondrial function K T Chang & K-T Min 1586 Transcriptional and behavioral interaction between 22q11. R Mizuguchi. B Olivier. S Karimzadegan. W-S Lai. X Chen. H Zhang. R S Petralia. J Parris. K J Mitchell & (p 1542) A Chédotal 1525 TARP γ-8 controls hippocampal AMPA receptor number. J Dolan. M Karayiorgou & J A Gogos 1595 Neural correlates of binocular rivalry in the human lateral geniculate nucleus K Wunderlich. S S Zakharenko. M Yuzaki & J I Morgan  see also p 1417 1542 Activity-dependent decrease of excitability in rat hippocampal neurons through increases in Ih Y Fan. K Byrd. S Shirasawa.2 orthologs modulates schizophrenia-related phenotypes in mice M Paterlini. R A Chitwood & D Johnston 1552 Fine-scale specificity of cortical networks depends on inhibitory cell type and connectivity Y Yoshimura & E M Callaway 1560 Induction of sharp wave–ripple complexes in vitro and reorganization of hippocampal networks C J Behrens. P Luo. F E Schweizer. O Abdel Samad. J Qin. S Schneider-Maunoury. L C Faria. E Miura. D Sulzer. S Tomita. D S Bredt & R A Nicoll 1534 Cbln1 is essential for synaptic integrity and plasticity in the cerebellum H Hirai. H Fujisawa. R Blunck.com/natureneuroscience glutamatergic transmitter phenotypes L Cheng. K A Schneider & S Kastner 1603 Neural basis and recovery of spatial attention deficits in spatial neglect M Corbetta. M Watanabe. Y Xu.

) Each der very closely. J. largely loss. Nat. For example. an anticonvulsant that acts at AMPA/kain. In addition. but unfortunately they are too reduces the costs of bringing them to market. remove the roadblocks to development and drug before it reaches receptors in the brain. (Acamprosate affects Some animal models of addiction mimic the als for addiction typically involve drugs that GABA and glutamate receptors. In contrast. pharmaceutical compa. mimicking relapse in human addicts. Psychiatry 58. receptor 1 antagonist that reduces the expression nies lack incentives to develop medicines for 3. complicated for use in large-scale drug screen. but later exposure to stress.P. Poling. 158–164 (2005). Instead clinical tri- in the US and Europe. Other therapies affect not only drug addiction. which nism of action is not fully understood. Mitchell. E. diagnostic characteristics of the human disor. Neuropsychopharmacology 30.. responses to food.M. ing. their own trials. Gonsai. such as the identification of neural adaptations that problems—diet and exercise to heart disease. K.nature. Biol. However. and com. & O’Brien. even for the many alcoholics and therapies. now in clinical trials in humans. suggesting that it may affect Beyond the practical problems involved in 205–211 (2005). For example.. For alcohol.. drug moderately reduces the amount and fre. This entire assay takes 10 to 12 weeks. Treating addiction to illegal drugs (2005). B.R. with more complex behaviors. H. for testing in addiction.. For addicts). our understanding of the neu. and has no abuse potential or overdose again. Governments and charitable foundations could agonist buprenorphine is available in Europe ing the lever. as the brain circuits that respond to relapse—and such neural changes might be we do not refuse these patients medical atten- natural rewards partially overlap with those that the most directly relevant to addiction treat. most drug targets proposed by quency of drinking in clinical trials. over the counter) reduces cocaine addiction scale development of new drugs. in the US.. are already approved for other uses. modafinil (an atypical stimulant intermediate cue that predicts the drug’s future companies that own the compounds have not that is approved for other uses) seems to availability (as money might do for human made them available to researchers. lated). For these reasons. have jobs or medical insurance. we need to enough antibodies.M. However. either because drugs to target more effective than either alone. despite Preliminary results from clinical trials sug.com/natureneuroscience A s the focus articles in this issue demon. compliance. like schizophrenia or hypertension. Doctors and insurance com- than 30 patients. federal regula. N-acetylcysteine (a food supplement available panies are often reluctant to undertake full- Some medications are available. rats are trained first to press a lever erly in the clinic. Vaccines against cocaine2 or nicotine seem to companies often restrict the availability of ful in translating this knowledge into effective reduce drug intake in patients who produce treatment. Kampman. C. Rimonabant is a cannabinoid identifying targets.. 8. compound is also in clinical trials as a treatment addiction as a character defect rather than robiology of addiction has progressed for obesity. For example. Martell. If this behavior fails to produce the What can be done to improve the situation? heroin addicts. C. which then intercept the nicotine addicts who have coverage. and is addiction. & O’Brien. K. who most need treatment are least likely to having some of the best animal models in neu. the provide better incentives for drug develop- and the US. the µ-opioid receptor partial drug. the opioid receptor antagonist naltrexone without affecting the response to food in rats. by rodents produces distinct effects in the brain go a long way toward improving their care. and cocaine addiction. can cause the drug-seeking behavior to start amount of any effective drug that is developed. increase abstinence during treatment1. the addicts through studies in rodents. This is a potentially serious problem for are specific to compulsive drug seeking and for example. so it too is likely to affect eating behavior. Dackis. ment. E D I TO R I A L Taking addiction research into the clinic © 2005 Nature Publishing Group http://www. drug self-administration A similar attitude toward addicted people would mate. and they unfairly reduce the availability tests. despite difficulties with patient of help for patients who need it most. relations problem. Buprenorphine is as effective as addictive drug. tage of this approach is that it does not allow Behavioral choices contribute to many health but also the response to natural rewards. or cues associated with the drug ment. but it is also effective drug is simply injected by researchers. pharmaceutical com- testing of new treatments. respond to addictive drugs. For cocaine for access to a drug. is used for alcoholism that are not seen when the same amount of 1. Promising hypotheses can then be studied the long term. acamprosate. Pettinati. T. & Kosten. from pharmaceutical companies for testing in tions prohibit physicians from treating more Such models are time-consuming and expen. researchers have been less success.G. food. C. against binge eating. However. are approved and is now being tested in people. and the common view of strate. so investigators typically begin screening panies should begin to think of drug addiction for treatments that target addiction to illegal for drug-induced neural changes with simpler as a chronic disease that must be treated over drugs. but its mecha. ate glutamate receptors. and then to work for an these mechanisms do not exist or because the addiction. Dackis. gest that other approaches may be promising. Such restrictions are common sive. topira. The disadvan. or even to ism. K. release the compounds that they already have and a newer drug. the animal eventually will stop press. 1431–1436 of several addictions in animal models. To solve this problem.A. Insurance roscience. or they could purchase candidate compounds liability. Lynch.A. and seems to promote modest weight a neurobiological disorder3 creates a public substantially over the last decade. 2. in one model of drug seeking basic researchers have not been tested prop- bination therapy with both drugs seems to be and relapse. or smoking to lung cancer—but drug design. the raises legal issues. NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1413 . Neurosci. tion because they are culpable for their illness. such as promising to purchase a certain methadone (a full agonist that is tightly regu. C.

Hirai et al. the authors but have not yielded significant insight into its significant fraction of the remaining synapses produced and analyzed Cbln1/Grid2 double- role in the cerebellum. In this new study. Cerebellin. NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1417 . of a mouse lacking Cbln1.com/natureneuroscience Fekrije Selimi & Nathaniel Heintz Cerebellin 1 is abundant in the cerebellum. of the pre. discovered two decades ago. Any dysfunction during this retrograde signals to the PF bouton (dashed Rebecca Henretta process—which includes formation. they also need to cell–parallel fiber synapses. and the Howard This phenotype is a pleasant surprise. UMR7102. This counteracts the Stabilization and memory formation requires continuous destabilizing effect of Grid2 activation by a small. receptor δ2 (Grid2) knockout mouse. Laboratory for Molecular Biology. NEWS AND VIEWS How neurons keep in touch © 2005 Nature Publishing Group http://www. both synapse stabilization and refinement. which is specifi. Moreover. The Rockefeller because it is similar to that of the glutamate function as part of the same signaling pathway. The Taken together. although an additional role for Cbln1 in e-mail: heintz@mail. Electrophysiological studies confirmed these two genes in a common pathway. spillover? + specific types of neurons is the last step in This process involves as-yet unidentified Cbln1 establishing functional neuronal networks intracellular signaling cascades and possibly Grid2 X X Grid2 in the brain. These findings provide further evidence that there is a molecular pathway devoted to maintenance of synapses. In this issue. Paris. USA.edu not yet fully understood. They are ataxic. New York. detailed analysis of serial Given the phenotypic similarities between proteins and on its biochemical properties6 sections by electron microscopy showed that a Cbln1 and Grid2 knockout mice. when the authors examined the Cbln1 knock. rather than their formation. France. These results provided strong evidence for the 16-mer peptide associated with Purkinje cells5. the development of synapses seems probable. although no inactivation of the Grid2 gene in adult animals glycoprotein Cbln1 is essential for the stability obvious abnormality of cerebellar histology also leads to detachment of the parallel fibers and plasticity of a specific synapse: the parallel is evident in the light microscope. Hirai had mismatched pre. However. poorly diffusible to cognitive disorders1. The actions of Cbln1 are restricted to the Destabilization PSD tion and elimination of synapses—can lead immediate vicinity of the presynaptic terminal of PSD owing to its properties as a large. the actions of Cbln1 and Grid2 in Purkinje cell spines can control pre. Moreover. cally expressed in cerebellar Purkinje cells. Forming and maintain. Hirai et al.4 demonstrate gross phenotype of Cbln1 knockout mice is development7 and from the observation that that the transsynaptic function of the secreted unremarkable. Thus. strong evi. arrow). At sites distal to the actions of Cbln1. learning glycoprotein complex. from the parallel fiber bouton. providing new and PF-PC synapses has accumulated from detailed cules that mediate this particular transsynaptic important insight into Cbln1 function. via the orphan glutamate receptor subunit Grid2. Cbln1 to the C1Q/TNFα family of secreted partner.and postsynaptic elements. was identical to that of the Grid2 knockout. University. now show that this gene is required to maintain parallel fiber-Purkinje cell synapses. However. is lum. phenotype of the double mutant animals Fekrije Selimi and Nathaniel Heintz are at the sion between parallel fibers and Purkinje cells. In this way. Cbln1 is secreted work hard to keep it. aptic properties and stability in response to specific cues. regulating the stability arrows)).and postsynaptic specializations8. matura. Whereas the role of activity in the maintenance of synapses seems clear3. perhaps through its ability to responds to bidirectional signaling between control autophagy locally. knockout animals to test the involvement of izations. a striking phenotype emerged: 78% of the distal existence of a specific pathway dedicated to Subsequent studies have focused on the rela. Cbln1 in synapse stabilization in the cerebel- Recherche Scientifique. Neurons face a classic issue: it is not enough Figure 1 A model for Cbln1 action in Purkinje to find the right partner. demonstrating that these two genes do indeed Hughes Medical Institute. The this result.rockefeller. but its function remains a mystery. is a out cerebellum at the electron microscope level. New York 10021. spines that are the site of contact for parallel governing the structural stability of synapses tionship of the full-length cerebellin protein fibers on Purkinje cells lacked a presynaptic in the brain.and postsynaptic special. synaptic plasticity at both the functional and diffusible ligand (presumably glutamate (blue structural level2. very and colleagues describe in detail the phenotype dence that Grid2 is involved in stabilization of little is known about the identity of the mole. where it acts in combination with Grid2 and other unknown ing the proper synaptic contacts between Glutamate molecules (‘X’) to stabilize the PF-PC interaction. from Purkinje cell spines and mismatching fiber–Purkinje cell (PF-PC) synapse. of synapses must be a dynamic process that activation of Grid2 has a destabilizing effect on postsynaptic structures. these data strongly implicate Fekrije Selimi is also at the Centre National de la function of the Grid2 gene. demonstrating a deficit of transmis. The analysis of the Grid2-null phenotype during action.nature. altering syn.

et al. Danho. B. Cell Biol. T. Indeed. suggested that Cbln1 acts with Grid2 gene encoding choline acetyltransferase in that operates transsynaptically in the brain. 245–252 (1995). 8. mation that is then transferred to the neocor- Newark.. Storrs. 197 University Avenue. tion of the synaptic contact and matching of (2005). Cbln1 and Grid2 are Given the results of Hirai et al. This would explain the phenotypes 13. H. receptors in Lurcher mice results in stimula. Kurihara. Chklovskii. James Chrobak How does this change take place. those same mice. Kishore. et al. have made an important step in identi- evidence that these two genes act in the same phenotype can be rescued by inactivating the fying a mechanism for synapse stabilization pathway. F. stability. Kummer. which interact receptor in PF-PC synapses. 869–874 changes in PF-PC synaptic structure. the pre. extracellular structures that promote synapse the destabilizing effects of nAChR activity at 618–629 (2005). Neurochem. Selimi. Neuron 46. secreted glycoproteins on neurotransmitter not by Purkinje cells. suggests an intracellular mechanism through the cleft. Neurosci. 1534–1541 (2005).B. 95. Hirai and colleagues show—using a receptor activation.. 8. 3. tainly a handy feature if one hopes to remodel ing action of Grid2 receptors. NEWS AND VIEWS Resolution of this important issue must await muscular junction13.chrobak@uconn. 11088–11093 the neuromuscular junction. et al. the demonstration that the neuromuscular junction. U. 1) can structural remodeling of these critical CNS stability and plasticity of the PF-PC synapse. Z. notype of the Grid2 knockout suggests that it (2004). It was thought to promote Although a great deal of additional infor- further experimentation. 2146–2156 lating autophagy11 and that activation of Grid2 ately adjacent to the active zone. Grid2 would retain its destabilizing 11. W. J. Opin. Nat. assemblies transfer information to the neocor. et al. Pang. 782–788 (2004). Rutgers. 9613–9623 (1997). Proc. A very interesting parallel can be made with this new pathway in promoting both stabiliza. 10. Nonetheless. J. Nature 431.L. which is presumably free of Cbln1 9. & Morgan. Synaptic plasticity and self-organization in the hippocampus György Buzsáki & James J Chrobak A new paper reports that long-term potentiation in the hippocampus. lar C1q domains.com/natureneuroscience combination of in vitro studies. This would pro. T. hybridization in vitro studies. Mel. Hirai. a model of learning and memory. USA. Takeuchi. and previous receptors may provide a key function for part of a transsynaptic pathway controlling the studies of Grid2. 25.. synapse formation by clustering acetylcholine mation will be required to understand in The extensive similarities between the Cbln1 receptors. N. and no longer depend on the hippocampus. which the Cbln1/Grid2 pathway could effect actions. 921–933 (2002). Z. including deficits in very few acetylcholine receptor clusters and tural integrity of central synapses. These autophagy suggests a destabilizing role for this 2. Acad. Natl. J. S. are thought to represent stored infor- University of New Jersey. have indicated that the role of junction and at central synapses. Grid2 larly interesting given the biochemical proper. et al. They often are important in is necessary for synapse stabilization and main. 14. J. mainly seen and Behavioral Neuroscience. & Morgan. Hua. which are thought to be critical for the stabilization of memory traces in cortex. T. observed by Hirai et al. signaling pathway at the PF-PC synapse. Hempstead. D. J. J. the mechanisms responsible is at the Department of Psychology. and how do tex. Neurosci. The transsynaptic action of Cbln1 is particu. Sci. Curr. tissue remodeling. Nat. discharge required to induce synaptic change? transfer our phone book to the new gadget cling bin. 1418 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . H.R. Science 302. boutons may be to locally inhibit the destabiliz.Y. sometime after an event. 4. University patterns of activity within hippocampal cell for the induction of these SPW-R complexes of Connecticut. After buying a new cell phone. Neurosci. the State ferred to and consolidated in the neocortex in vivo. 551–561 tion of this important catabolic pathway11. Misgeld. D.nature. can induce sharp wave-ripple complexes. and sug- in situ and a transgenic approach—that Cbln1 agrin is to inhibit the destabilizing effect of the gest that the transsynaptic actions of large. mote stabilization of PC-PF contacts immedi. ties of other C1q/TNFα family proteins10.and postsynaptic specializations. W. USA. 406 Babbidge Road. seems to have a dual function. Trends Immunol. 7.. Lichtman.I. Zoghbi. J. Self-organized population discharges in the memories that initially depend on the activity hippocampus such as hippocampal sharp György Buzsáki is at the Center for Molecular of hippocampal neuronal assemblies are trans. J. & Smith.15. requires associated protease activities—cer. Acad.R. Neurosci. be proposed for the roles of the Cbln1/Grid2 structures. Hirai et cerebellar long-term depression4. as agrin inhibits 6. 12. tex and consolidate it there? And how can cell Now Behrens and colleagues1 report that e-mail: james. Grid2 is tethered to a molecular complex regu. Sci. active zone. wave-ripple (SPW-R) complexes.I. K. as well as the role of (1992). USA 82. J. et al. together with important process at the neuromuscular © 2005 Nature Publishing Group http://www.9 and genetic lack neuromuscular junctions. Its regulation of 1. Furthermore. One role for Cbln1 secretion from PF 5. 7145–7148 (1985). and activation of acetylcholine receptors14. Likewise.T. & Svoboda. thus preventing acetylcholine Their studies reveal a common logic for this However. Slemmon. Cell 81. These results.W. U. et al. 4. Thus. et al. in the postsynaptic density of PC spines. 569–579 (2005). Lin. USA 102. protein secreted by motoneurons at the neuro. proteins form large complexes. H. are unclear. 25. is a glycoprotein secreted by granule cells.W. Kashiwabuchi.J.. J. & Sanes. Neuron 37. McMahan.. Yue. However. agrin knockout mice have detail the mechanisms regulating the struc- and Grid2 knockout mice. 7. 813–829 (2003). Agrin is a glyco. we quickly and relegate the old instrument to the recy. Proc.. Bao. Outside of the (2005). 15.J. whereas the phe. In some cases. et al. this action tenance. a similar model (Fig. 826–830 (2003).12 owing to its limited ability to diffuse out of (2004). New Jersey 07102.Y. 17. Connecticut 06269. 327–332 with a variety of molecules through their globu. Neuron 35. this al.edu assemblies produce the patterns of neuronal they can induce in vitro SPW-R complexes. Natl.

inhibitory connections2. IPSP-EPSP qualifying it as the most synchronous network SPW-R complexes gradually increased in inci. They also provide evidence that the in the activation of SK2 calcium-activated both the CA1 and CA3 pyramidal layer in induction of SPW-R complexes is paralleled by potassium channels11. the estab. a protocol that tude excitatory inputs from a large area of SPW-R complexes in vitro have been observed induces long-term depression of synapses. but they pattern in the brain3 (Fig.13. The observation that were ‘saturated’—that is. the rat dorsal hippocampus is that the density SPW-R complexes show a strong parallel with in contrast to the regularly occurring and of recurrent axon collaterals and the mutual synaptic changes observed in vivo13.15. One possible explanation for the with earlier observations of SPW-R complexes tion of the truncated CA3 collateral circuitry.000 neurons—10–20% of the total neuronal population of the rat hip- pocampus—discharge simultaneously in the have found. vivo counterparts are very irregular in both about a self-organized population burst. they have established a fast-frequency field oscillation (200-Hz EC link between the induction of LTP and the © 2005 Nature Publishing Group http://www. these drugs did not Although these findings show substantial demonstrate that stimuli that induce LTP lead prevent SPW-Rs. However. The the CA3 region. (400 ms at 100 Hz). the consolidation of synaptic plasticity and the parasubiculum (‘Para’) and deep layers of the entorhinal cortex (‘EC’). ence of these drugs. the spontaneously emerging reminiscent of epileptiform activity. once these com.8 and the also attenuated ripple occurrence. Hippocampal Para studied in vitro and the consolidation of output. If so. Figure 1 Self- terns commonly seen in vivo. synaptic connectivity. the in-vivo ripple event in changes in both excitation and inhibition in the lished SPW-R events were reduced or abolished CA1 reflects a convergence of small-ampli- CA3 region. organized burst SPW-R can be induced with stimulation of activity in the hippocampal CA3 protocols known to induce LTP. Importantly. produces similar sharp wave-ripple The hippocampal SPW-R complex has complexes in the features that make it a candidate pattern for subiculum (‘Sub’). memory traces in the intact brain. Behrens and colleagues1 plexes were established. In contrast. However. After approximately the fifth of ripple complexes and may contribute to Although its recruitment dynamics are delicately LTP train. Thus. including how and D-AP5 (a competitive NMDA receptor determined largely by a unique distribution SPW-R complexes emerge and whether this antagonist) could prevent the induction of of synaptic strengths at both excitatory and pattern is actually accompanied by changes in SPW-R complexes. to induce SPW-R. in accordance preted as an artificial augmentation of excita- mouse9. inhibitory plex arises in the recurrent collateral system stimulus. ated event in vivo and population events in of the dorsal hippocampus of the rat. potential (EPSP-IPSP) sequences. They observed extracellular post. uniformly sized SPW-Rs in the slice. Nevertheless. their in- excitation are simply not sufficient to bring Using combined intracellular and extracel. This dal cells during the development of SPW-R enous brain pattern in vitro allows for the is precisely what Behrens and colleagues1 complexes. Trains that failed to induce LTP failed inputs are prominent in the development of the CA3 region and spreads downstream. Importantly. Further. the incidence of estab. Second. between 50. This difference may be inter- the ventral hippocampus of the rat7.10. Moreover. First. but not the expression. Thus. as a phase-related Their report takes one big step in bridging discharge of the Ann Thomson the chasm separating synaptic mechanisms neurons. perhaps because the synapses involved pling14. there are of the NMDA subtype of glutamate receptor) ing neurons during the SPW-R complexes is several missing links in the story. absence of spontaneous SPW-R in slices of in vitro12. investigation of a number of important NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1419 . of that involve decreased calcium influx through out. vitro. the pattern of inhi- a three. dence and amplitude after the third to fifth never found isolated EPSPs. some differences are worth pointing the induction. strengthening the surviving synapses examined intracellular responses in pyrami.nature. it show that the rules of synaptic plasticity govern the emergence and the recruitment of particular cell also has a widespread effect. NEWS AND VIEWS similar to the fast-frequency ensemble pat. this indicated gives further support for the fundamental role Both MK-801 (a noncompetitive antagonist that the discharge sequence of the participat- of these population patterns. probably via mechanisms vitro. During repeated stimulus trains synaptic potential–intracellular postsynaptic CA3-CA1–subicular complex–entorhinal axis. lular recordings. In the approximately groups in these hippocampal output events. the replication of an endog- could rescue the compromised circuit. homology between a hippocampus-gener- to the generation of SPW-R complexes in slices lished SPW-R incidence increased in the pres. Behrens and colleagues1 transfer of neuronal patterns2. Indeed. SPW-Rs are associated with large SPW-R complexes is NMDA receptor– NMDA receptors and a subsequent reduction population events and ripple oscillations in dependent. repeated every 40 s. in turn. Behrens and colleagues also their temporal distribution and magnitude. they reached their remained stable during the development of the SPW-R is shaped by previous experience6 maximum possible strength. Importantly. The authors also demonstrate that dendritic layer of Sub these events induce synaptic change among CA1 and a short-lived CA3 neurons. 1). The SPW-R com.to fivefold gain of network excitability in the generation of the spontaneous patterns bition-excitation in any particular neuron is achieved transiently5. a popular region produces a CA1 neurophysiological model of learning and field potential in the memory. SPW-Rs in individual neurons. by low-frequency stimulation. In short. as well believed to be involved in shaping memories. the incidence of SPW-Rs reached a the temporal precision of EPSP-spike cou- controlled by various classes of interneurons4. plateau.com/natureneuroscience ripple) within stratum CA3 emergence of a physiological network pattern pyramidale. 100-ms time window of a hippocampal SPW. sequences and prominent IPSPs. without synchronous ripple only in the mouse hippocampus and from blockade of gap junctions with carbenoxolone oscillations.000 and 100.

Germany. M.1 now provide experimental evi- brate nervous system. ciated ligand2. Max Planck Institute Nrg1 signaling also contributes to synaptogen. & Heinemann. & Buzsaki. dence that neuregulin-1 type III is necessary for membranes are synthesized by highly special. T. 319–325 (1995). The finding that such modi. Brain Res.) 521. used a coculture system in which in spectacular membrane growth. 385. Y. nevertheless. Colgin. We do not Taveggia et al. 62. of more than 15 membrane-associated and Mice that selectively lack Nrg1 type III die Ever since electron microscopists secreted growth factors that are derived from perinatally and have a marked decrease in demonstrated the complexity of compact a single gene by alternative splicing and pro. C. & Lynch. Like the insulation on electrical wires in your 1 µm are myelinated. & Haas. gave rise to the patterns. © 2005 Nature Publishing Group http://www. and dendrites seem not to in the postnatal nervous system. myelination in the PNS and can instruct imma- ized glial cells. nal signal that regulates myelin sheath thickness dorsal root ganglion (DRG) and superior cer. et al. including the requirement of triggered by axonal signals. King. Cross. 6160– 13. Brucher. van den Boom. B. vivo. & future experiments will be able to reveal if the even more exciting. F. Klausberger. Casale. If so.. 81–89 (2003). For example.. These in vitro findings PNS. the information about the perturbations that (2005).. a regulator of axonal neuregulin-1 and glial ErbB receptors. Physiol. 187–193 (2002). both of which are potentially vitro myelination is initiated by the addition proteins of myelin and the axon-glia junc. J. 844–848 (2003). but the nature of of developmentally regulated signaling pro. 551–570 (1989). & Buzsaki. 873–887 (2003).. J. In of ascorbate8. In the developing nerve. J. J. NEWS AND VIEWS mechanisms that are difficult to explore in fications can be brought about by stimulation 679 (1994). 7. H. their (including proteins named NDF.com/natureneuroscience 2. 5. C.) 555. been verified. Nrg1 type III has a second transmem.. Neuroscience 31. Papatheodoropoulos. Nature 421.. M.. 3–13 (2003). Kubota. & 713–725 (2004). It is unclear why axons larger than and ErbB3 are expressed at the cell surface of Lentiviral expression of Nrg1 type III was Schwann cell progenitors and are essential for sufficient to restore myelination competence. even in the membrane growth process. X. plexes may contribute to the process after all. Friedman. 3. L P. Sensory DRG neurons from tion have been identified. & Kostopoulos. & Buhl. It is therefore surprising that. 10. Maier. L. secreted or shed upon proteolytic cleavage. L. 89. H. 1159–1163 network patterns. protocols that induce LTP and can be altered Bull. N. the requirement in Neuron1. de Hoz. 12. and SPW-R com. Major subgroups are Nrg1 type I sensory nerves7.A. Neurosci. Neurosci. R. Hirase. F.A. One forming Schwann cells. tion in conditional mouse mutants5. Physiol. Pouille. Nimmrich. E.E. Henze. A. Y. 159–167 (1999). mechanisms involved in the consolidation of Buzsaki. In the for oligodendrocytes and Schwann cells. parts of the circuit may identify the elementary 4..L. cellular neurobiologists have motor use.L. 8. seems to operate at all stages of Schwann cell identification of neuregulin-1 type III as an axo- ferentiation of myelinating Schwann cells in development and myelination. (Lond. one signaling system. demonstrated that the PNS. M.. 46–51 (2005). LeBeau. and the myelin growth. J. Numerous GGF) isoforms. G.J. Brankack. 64..de embryo. in or whether these mechanisms are the same ture ensheathing cells to become true myelin- the CNS and Schwann cells in the PNS. Lett. D. A.L. The latter has greatly hampered the seen in ‘rescued’ cocultures. but of Experimental Medicine. neuregulin-1 for the survival of precursor cells these signals has remained unclear. by those that produce long-term depression is 8. U. 14. Colgin.L. the decision of Schwann cells to myelin. impulse propagation throughout the verte. Memories may really be Lynch.. Behrens. termed oligodendrocytes. Science 265. J. at least in and immature Schwann cells4. ation in mice that overexpress Nrg1 type III. would expect neurons and myelinating glia to close a gap between related in vivo findings by ate resembles a cell lineage decision that is communicate through a complex assembly other groups.mpg. A recent study shows that the expression of neuregulin-1 on an axon membrane determines whether immature Schwann cells will differentiate into myelinating Schwann cells. interneurons and cardiac development in the unusually thick myelin profiles were e-mail: nave@em. Taveggia et al. & Scanziani. In agreement with a report of hypermyelin- Department of Neurogenetics. Axmacher. multiple examples of 37075 Göttingen. interact at all with oligodendroglia. Sabatier. G. involve stable recruitment mechanisms. 19. were cocultured with Schwann cells from the driving force of glial ensheathment. Neurophysiol.. 6.. We do not know brane domain and remains a membrane-asso. H. & Miles. heregulin.. using simultaneous intra.M. 6). 1. Schwann cells and degenerating motor and myelin sheaths. Neuroscience induced but otherwise self-generated patterns made in the hippocampus. Mamiya. G. J.. Common to all isoforms is an normal rats. (Lond. Whittington. R. F.A. is required to induce the dif. Neuregulin-1 (Nrg1) comprises a family in mice with altered neuregulin gene dosage6. G. myelin sheaths are essential for rapid are only ensheathed. hypothesis that endogenous patterns preserve A. Yanovsky. Monyer. 14. 676– (2001). G. Neurosci. To study the competence of been hooked on the myelin-forming glia. 274–287 (1999). Klaus-Armin Nave and Markus Schwab are at the their survival and subsequent differentiation3. G. numbers of neurons and modifying targeted 6170 (1994).A. but this has not tyrosine kinases. esis (at least in vitro). 9.D. 11. Hermann-Rein-Str. J. 15. Csicsvari. The multilayered myelin know what signals provide specificity of axon. Bull. these mutant axons to be myelinated. Recently teins. & Draguhn. G. N. Taveggia complex interaction with axons and their and ARIA) and type II (‘glial growth factor’ or et al. Jia. & McNaughton. 1560–1567 Neurosci. but small-caliber axons conventional genetic analysis of Nrg1 function house. Physiol. Wilson. comprised of of glial ErbB2 receptors for normal myelina- axonal neuregulin-1 type III. J. Traub. Brain Res. J. J. Leinekugel. Chrobak.) such findings would provide evidence for the 550. 1420 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . Nat. glia interaction and instruct glia to myelinate. L. Science 293. Glial cells under remote control Klaus-Armin Nave & Markus H Schwab Not all axons in a peripheral nerve are myelinated. Buzsaki.. vical ganglion (SCG) explant cultures. Monitoring large 3. M. the migration of cortical not type I (ref. D. 57. (Lond.. Czurko. A. major contrast. The Nrg1 type III mutant sections suggest that myelin assembly is a spiral EGF-like domain that activates ErbB receptor axons never became myelinated. wild-type and Nrg1 type III–deficient mice questions remain unanswered. cellular recordings from two or more neurons.H.nature. ErbB2 presence of a fivefold excess of Schwann cells.

com/natureneuroscience This suggests that increasing Nrg1 expression in neurons above a certain level may confer myelin- ation competence. differentiation (as determined by myelin peripheral neuropathy in transgenic mice10. and whether it is 1 mm or type III mutant mice. From a neuronal point of view. the importance of this interaction has (B. and possibly modified.6. then how does a neuron know nals—promotes ensheathment and myelina. or alternatively they single out larger axons and differentiate into myelinating Schwann cells (4). Work in (type I and II) are functionally not equiva- several laboratories indicates that the entire path of Schwann cell development and myelination is remote lent to the juxtacrine type III (localized to controlled by the neurons through expression of the neuregulin-1 (5) on the axonal surface. The authors also grouped into so-called Remak bundles. they tion (Fig. extend these 5 observations by showing that neurite mem- branes from wild-type (but not from Nrg1 type III–deficient) neurons activate the PI3 Debbie Marzels kinase pathway. An important next question Mitotic Schwann cell 1 progenitors is whether insufficient Nrg1 expression can also explain why axons smaller than 1 µm in diameter (such as C-fibers devoted to pain perception) are typically unmyelinated.A. The latter cord differs with respect to other growth fac- tested an ectopic juxtacrine source of Nrg1 suggests that Nrg1 type III also signals between tors13. and K. These Most likely. M. This issue of whether axon length normal7. whereas other signaling mol. In peripheral nerves. but only using conditional mutants and transgenic mice.5.S. PI3 kinase and MAP kinase in Schwann cells are activated by GGF. observations suggest that neurons produce will be rederived. Schwann cell ecules underlie the activation of MAP kinases. NEWS AND VIEWS Because Nrg1/ErbB2 signaling is required for myelination in vivo5. Thus. 2 dent thresholds for myelination. but not myelinated). Assuming that the paracrine neuregulins they face a binary choice: they either stay tightly associated with several axons to form a Remak bundle (3). odomain to the coculture system. neural crest–derived the secreted neuregulin isoforms. Later. the authors observed a oligodendrocytes respond to Nrg1 (refs. By transducing cultured SCG neu. 11. glial cell differ- gene expression) could not be induced.H.12). cell bodies rons (with axons that are normally ensheathed by Schwann cells. 1). 1 meter long. In contrast. Schwann cells were cultured under C-fibers and ‘non-myelinating’ Schwann cells. the authors asked whether a type III protein promotes myelination when offered as a paracrine signal. Additionally.10 an unexpected picture its own size. they Axon could show that indeed low levels of Nrg1 type III are sufficient to induce ectopic myelination. 3 Remak bundle What are the effector molecules in neu- regulin-induced Schwann cells? The inhibi- tion of PI3 kinase blocks myelin formation in cocultures9. © 2005 Nature Publishing Group http://www.N. To address that found that the sciatic nerves of adult mice are growth. expression determine Schwann cell fate. protein retained mitogenic activity. also analyzed heterozygous Nrg1 has recruited a neuregulin-based signaling system 10 µm in diameter. they added a recombinant type III ect. In agreement with previous work6. a soluble type II neuregulin. and preliminary evidence suggests a more type III. which cations. as immature Schwann cells (2). It will be most interesting to extend this issue. Like Schwann cells. or by contact with neurite 4 membranes. Axon surface Nrg1 type III emerges as a critical activator of Myelinating glial PI3 kinase. the factor retained perturbation of ErbB receptor function. Now Taveggia et al. These other molecules may or may not include Figure 1 Neuregulin-1 directs Schwann cell differentiation. complex regulation by different Nrg1 isoforms myelinating conditions on a monolayer of Indeed. the axon surface). from other studies3. line of research into the CNS. Again. unpublished CHO cells that stably expressed Nrg1 type been demonstrated before by dominant-negative data). The soluble tion velocity. hypomyelinated and have reduced nerve conduc. in vivo If quantitative differences of neuronal Nrg1 gene multiple neuregulin-1 isoforms. but Schwann cell causes disorganization of Remak bundles and ation (or the lack thereof) in multiple sclerosis. This research may also have clinical impli- III on the surface. emerges already: evolution of the nervous system It should matter whether an axon measures 1 or Taveggia et al. or how much neuregulin to make? tion as a contact-dependent signal. or whether axon size and Immature Schwann cell membrane curvature pose neuregulin-indepen.. Brinkmann.nature. including the fine-tuning of myelin and caliber control neuronal gene expression NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1421 . but did higher number of small caliber axons (C-fibers) Myelination control in the brain and spinal not support myelination. even type III—in concert with other axonal sig. if axonal Nrg1 contributes to remyelin- some mitogenic potential. the conclusions presented here entiation emerges as largely remote controlled. But when the new data1 are combined with those myelin thickness. Schwann cell progenitors (1) proliferate and populate axon bundles. the authors then asked whether normally unmyelinated axons would become myelinated if they ectopically expressed Neuronal Nrg1 type III. which are phenotypically for virtually all steps of Schwann cell differentia. MAPK activation Neuregulin was not impaired in mutant cultures.

Neuron 43. Rev.N. removing a single copy of Gfi1 tion contributes to the progressive demise of toxicity by enabling an interaction with 14-3-3 enhances the neurodegeneration phenotype Purkinje neurons in SCA-1. A. 148. Nature 378.. Gfi-1. et al. A recent paper in Cell by differing polyQ repeat lengths are required to interactions to the human disease. NEWS AND VIEWS may be relevant for many more proteins.K. Eldridge. In a final elegant proof-of-prin- Autosomal-dominant polyglutamine now report that expressing the fly homolog ciple experiment. Neuron 28. The authors these findings are recapitulated in a mouse domain of ataxin-1. Intriguingly. J. Biol. thus teasome system. respectively. further. Taveggia. Tsuda et al. diseases. ataxia in Gfi1 knockout mice. Wolpowitz. recapitulates hAtx-1–induced phe.. thus demon- all caused by the expansion of unstable CAG notypes in different fly tissues.1 in this model. Topilko. the authors show that reducing Gfi-1 levels scriptional assay and a functional analysis results in an enhancement of hAtx-1–induced The authors are at the Department of Pathology in the fly reveal that both Sens activity and neurodegeneration. 6. 154. has no effect on Sens levels. Meyer. In mammalian cells. 7. Neurosci. Significantly. or polyQ- tures. 14–30 (2003). impaired axonal transport geted for death in these diseases. dAtx-1 does cause Purkinje cell loss. of the causative gene2. Importantly. These similarities have motivated research into upon the causative protein. also through tant features that distinguish them from one sequences outside the CAG repeat region in its AXH domain. the authors show that a pro- (polyQ) expansion disorders. The Hiroshi Tsuda and colleagues1 reminds us that initiate neurodegeneration in the different authors show that hAtx-1 binds to the ver- neurodegenerative diseases also have impor. 1245–1258 (2001). despite ubiquitous expression in the nervous of experiments relating the findings in including dysfunction of the ubiquitin-pro. but not a polyQ repeat gressive loss of Purkinje neurons accompanies Huntington disease and a number of SCAs. & Vartanian. expressing hAtx-1 with rodegeneration in flies or mice. et al. J. & Mirsky. with the transcription factor Senseless (Sens) tially a therapeutic target.com/natureneuroscience Senseless makes sense for spinocerebellar ataxia-1 Vikram Khurana.. 386–390 9. Miller. SCAs are debilitating features of ataxin-1 that mediate degenera. P. J. and increasing hAtx-1 stability4. 1023–1034 (1987). P.1.B. The dominant mode of inheritance.harvard.. in SCA-1. 700–703 (2004). preceding Purkinje cell loss and ataxia. as in flies. an expanded polyQ tract reduces Sens levels would further strengthen the case for Gfi-1 e-mail: mel_feany@hms. Beyond 1422 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . © 2005 Nature Publishing Group http://www. and functional interaction between the AXH gation of this protein in neurons. Once again.1 have now taken us a signifi. albeit with strating that decreased Gfi-1 is sufficient to repeat sequences within the coding region reduced severity. et al. D. strengthening the relevance of AXH domain and oxidative stress. system. evidence in animal models that this interac.nature. Maurel. 1035–1046 (2000). they suggest that the tion of the polyQ-expanded proteins in each a domain recently implicated in RNA bind. D. Notably. although by means of this domain. C. 2. 81–90 (2000). C.R. Gfi-1 is maxi- another. R. tion of Ser776 by the kinase Akt is critical to Furthermore. Birchmeier. J. R. a protein of unknown previously showed that overexpressing human model of SCA. Cell Biol. Massachusetts 02115. Tsuda et al. Jessen. cant step closer to understanding the unique group of neurons that selectively degenerate ebellar ataxia (SCA). 105. The authors provide compelling tion3. Nat. These differences strongly implicate tebrate homolog of Sens. including a progressive loss of neurons function mechanism whereby the expanded expanded hAtx-1 with the AXH domain and the formation of proteinaceous aggregates. Nat. AXH (ataxin-1/HBP1) domain with hAtx-1. ing and self-association5.M. 681–694 (2005). Michailov. Tsuda et al. Fernandez. M.A. 13. only certain neuronal groups are tar. are alone. Garratt. An in vitro tran. Bunge. and the transcription factor known as ataxin-1 (hAtx-1) with an expanded polyQ expanded hAtx-1 in Purkinje cells leads Senseless in Drosophila melanogaster and Gfi-1 tract in Drosophila resulted in neurodegenera. 77 Louis Pasteur Ave. The study proceeds with a logical series common underlying pathogenic processes. Falls. Drosophila to a mouse model system. Chan. (2000). Exp. Cell Res. 183–191 (2004)..P. however. Neurodegenerative diseases share many fea.edu more potently than dAtx-1. Tudor A Fulga & Mel B Feany Why are some neurons selectively targeted for death in neurodegenerative diseases? A recent paper combines genetics in the fruit fly and mouse to uncover mechanisms underlying the vulnerability of Purkinje cells in spinocerebellar ataxia-1. 284.R. 79–91 (2000). S. However. USA. Furthermore. where expression of polyQ- function. expressing the polyQ tract alone. S. 5. I. glial cells have evoked important questions. Science 304. Voiculescu. polyQ tract confers new molecular functions deleted. an effect progressive gait incoordination and cerebel. Neuron 25. et al..V. but it 682 (2005). 3. stabilization as a potential treatment. whereas over. et al.F. C. Gfi-1. Gfi-1 neurodegenerative diseases characterized by tion of Purkinje cells in SCA-1. Neurosci. G. Neuron 47. 1186–1193 (2003). K. 11. Furthermore. Bunge. et al.L. Chen. 6. & Wood. one The authors focus on a type of spinocer. out animal models. J. Charnay. P. Park. 6. suggests a toxic gain-of. including the selective vulnerability of disease pathogenesis. J. Boston. Krane.. 4. interaction between the AXH domain of disease. is caused by a polyglutamine expansion in that is post-translational and depends on the lar atrophy.L. including of ataxin-1 (dAtx-1). to an early decrease in the abundance of in vertebrates. 671– 8. T. delineate a physical ataxin-1 and accompanied by nuclear aggre. establishing whether ecto- at Brigham and Women’s Hospital and Harvard protein abundance are downregulated by pic expression of Sens or Gfi-1 rescues neu- Medical School. & 10. ubiquitin-proteasome system. P. 12. Neurodegeneration not contain a polyQ domain but shares an These findings are important at multiple accompanies the intraneuronal aggrega. (1995). Cell 1. Cell Biol. The authors further hAtx-1 and Gfi-1 is important for mediating together with the recapitulation of disease demonstrate that dAtx-1 physically interacts neurodegeneration and that this is poten- phenotypes in overexpression but not knock.. O. they showed that phosphoryla. dAtx-1. R. the Purkinje neurons of the cerebellum. & Salzer. Neurosci. This disease levels are downregulated by hAtx-1. & Birchmeier. D. 4635–4645 has never been recognized as such. C. Most directly. 20. mally expressed in the nervous system within particular groups of neurons. P. levels.

Establishing the direct relevance on abnormal interactions mediated by the mediated by the polyQ tracts themselves and of a physiological protein interaction to dis- polyQ tracts themselves. Nucleus rodegeneration in other cell populations vul. downstream of transcriptional dysregulation. The resultant transcriptional system to make such comparisons. For mice and fruit flies. Intriguingly. The obser- vation that expression of an expanded polyQ Drosophila tract alone does not phenocopy certain phe- notypes shared by dAtx-1 and hAtx-1–polyQ Figure 1 Tsuda et al. lates proapoptotic genes14. demonstrate homologous pathways that mediate ataxin-1–induced leads the authors to infer the existence of neurodegeneration in mouse and Drosophila models of SCA-1. In con. to the transcription factor Senseless and targets it for functionally important sequences outside the proteasomal degradation. At domain–Gfi-1 interaction. the present study adopts the nor. NEWS AND VIEWS the direct therapeutic implications. downregu- raising the possibility that such an approach polyQ-dependent interaction could there. interactions in the mouse model system. Flies certainly provide an ideal aggregation. is to try and integrate the multiple pathways the message is loud and clear: resolving the 1 in a manner dependent on polyQ tract downstream of polyQ expansion into a cohe. and distinct changes in different polyQ animal proposed. dAtx-1 (which lacks polyQ repeats) binds.1 also deserves attention. a network influencing common downstream involved in other neurodegenerative diseases. a picture important and thought-provoking study that activation of a physiological pathway. the mechanisms of cell death in SCA-1 would follow from proteasomal degradation only that common mechanisms might oper. but SCA-1 mouse seeming to be p53 dependent (Fig. with neurodegeneration in the of Gfi-1 and transcriptional dysregulation ate in different polyQ-associated diseases. the polyQ tract of models13. there are fly homologs for proteins.com/natureneuroscience other interactions are involved. do the neurodegenerative diseases should guarantee hAtx-1. PQBP-1 is enriched in the cerebellum. not by the protein attaining an entirely aber. including familial Alzheimer and tion factors8. toxic gain-of-function is caused. including the inferior olivary Gfi-1 nucleus or spinocerebellar tracts. 10). Sens The methodology used by Tsuda et al. that are a decrease in basal transcription. transcriptional profiling neurodegenerative diseases for which toxic on abnormal effects on transcription. through its AXH domain. Neurotoxicity flies by proteasomal activation. In Drosophila. For example. by dysregulation mediates neurotoxicity in both model systems. fore contribute to selective neuronal vulner. or whether AXH © 2005 Nature Publishing Group http://www. For example. provided data supporting different toxic other transcription factors converge on apop- genesis of SCA-1 and related disorders? PolyQ gain-of-function mechanisms in polyQ-asso. In summary. similarities and differences among related length9. including disruption of axo. in SCA-1. A homologous interaction. Here. The model proposed in this study would nal transport and downregulation of survival it would clearly be useful to define the molecu- implicate this expansion in the stabilization pathways11. occurs between hAtx-1 and Gfi-1 in mouse Purkinje neurons. but rather from the abnormal Taking these studies together. Because the nuclear the distinct effects that could be attributable ease pathogenesis has implications extending localization of causative proteins is essential to non–polyQ-encoding sequences. Mouse ing sequences outside the polyQ region in the Purkinje neurons selectivity of neurodegeneration in polyQ dis- orders. emerges of both the common mechanisms provides hope for targeted SCA-1 therapies in trast. NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1423 . these studies have focused ing with this idea. ciated diseases. PQBP-1 and RNA polymerase II led to implicated transcription factors function as employment for many generations to come. many previous studies have concentrated in polyQ expansion disorders that could be the future. Further. A clear challenge for future studies Parkinson diseases. tosis also? To guide investigations into events expansion clearly initiates the disease process. be involved12. remain unclear. 1). Direct dysregulation of transcription by recapitulating the biochemical and functional polyQ-expanded repeats may also make an important contribution to neurotoxicity9. and a processes? Gfi-1. potentiated by abnormal polyQ expansion and polyQ region. Wild-type ataxin-3 suppresses lar pathways mediating cell death in these dis- of hAtx-1. the study supports the utility of Drosophila in modeling human diseases. abnormally potentiating the AXH degeneration in multiple polyQ models in eases. Future studies might explore whether AXH PolyQ PolyQ aggregation the hAtx-1–Gfi1-1 interaction mediates neu.nature. Is it possible that might be fruitful for these diseases also. In keep. present us with an rant function. such as tau. that loss-of-function mechanisms might also but not classically apoptotic15. for example. either and microarray studies reveal both common gain-of-function mechanisms have been directly or through sequestration of transcrip. Nucleus mal fly protein as a starting point. these populations Neurotoxicity Proteasome are also vulnerable in Gfi-1–null mice. hAtx1 nerable in SCA1. Other groups have neuronal apoptosis? Could dysregulation of context of what is known about the patho. this study provides critical evidence implicat. implying not present. beyond SCA-1 and polyQ disorders to other for neurotoxicity7. Whereas several models of autosomal-dominant neurodegenera- tive diseases have been made by transgenic AXH overexpression of causative human genes in dAtx1 Ann Thompson Drosophila6. however. Tsuda et al. in addi. In this regard it would be interesting to determine if. whether apoptotic or non-apoptotic. Indeed. A resultant complex forming between sive picture. hAtx-1 binds to the nuclear protein PQBP. Gfi-1 tion to Purkinje neurons. downregulation of Gfi-1 by hAtx-1 leads to How do we place the present findings in the ability in SCA-1 (ref.

the most striking change was is reduced in the chronic phase as compared Yalçin Abdullaev and Michael I. partici. Lipinski. After right-hemisphere stroke. e-mail: mposner@darkwing. These Eugene. 974–981 (2001). C. Nat. & Yuan. Curr. study1 found a dramatic bias from the acute to the chronic stage. Mol. Trends Mol. Cell 95. Cell 122. was visual cortex activation is increased. to recognize their left arm as their own and to eat from the left side of their plate. The dorsal network includes the superior parietal lobe and the fron- tal eye fields. et al. Fernandez-Funez. J. which was to the acute phase. USA. prefrontal cortex1. not activated at all during the acute phase.com/natureneuroscience How the brain recovers following damage Yalçin Abdullaev & Michael I Posner Individuals with neglect fail to process stimuli on the left. Posner are in the that the dorsal right parietal lobe. Neurosci. H. R. attention effects (manifested as increased activ- on the other. nied by reduced activity in the non-lesioned arrow appeared at this central point. Trends Genet. Tsuda. Sugars. Nat. Klement.M. P. ventral parietal and frontal areas may be influ- enced by comparatively restricted lesions of the ventral right hemisphere. Neurosci.1 show that activity in the left visual cortex left. How do these parietal findings relate to what button when they detected a target.M. & Zoghbi. which is crit- successive functional magnetic resonance with neglect show particularly long reac. age.M. NEWS AND VIEWS 1. M.D. & Bellen.Y. The results of Corbetta et their ability to orient and detect stimuli on the al. 37–48 (2005). Opin. Participants were then asked to press a alteration in the pattern of activation in the pari. 11. now strongly activated in the chronic phase. H. K. Neuron 34. Shahbazian. 655 (2004). Jafar-Nejad. Seven months later (in the chronic stage). Cell 18.C.J. H.A. People Therefore. Taroni.L. & Feany. 10. et al.uoregon. 8. F. 15. In the acute stage immediately after acute to the chronic stage. but occasionally it appeared nodes showed no evidence of structural dam. How can a stroke to a local area of the brain be associated with such a mysterious array of symptoms? In this issue. eyes. 101–106 482 (2001). Nature 408. 5. Muqit. they may miss such targets com. H. 641– 7. H. 107–112 (2003). 19. 6. I. and even after many years they have hemisphere. Cell. (2000). 233–238 (2003). unattended side. © 2005 Nature Publishing Group http://www. 14. & Ross.T. directing pletely. D. sible for the observed reduction in rightward right side of the screen. 633–644 (2005). This effect is likely to be respon- the subjects to attend to either the left or the a large deficit in reaction time3. et al.nature. The authors tracked the recovery of this target appears on the side opposite to where activity in similar areas in the left hemisphere. M. H. Oregon 97403. without moving their The Corbetta et al. 5. et al.. Neurobiol. Med. Neuron 34. Cell 113. (a. Margolis. Humbert. 237–243 (2002). In the acute stage immediately follow- ing the stroke.K. Chen. 457–468 (2003). An the stroke. FEBS Lett. 2. J.L. Normal par. The target etal-frontal network four weeks after the stroke is found in visually specific areas of the cor- appeared on the side indicated by the arrow (the acute stage). was the only brain area that showed performed a task. 701–713 (2002).B. Mol. some people see objects in their world as having no left side (Fig. 7. 4. the dorsal parietal area. Rev. 41–53 (1998). usually caused by a stroke. Okazawa. 3. whereas right hemisphere Department of Psychology. may influence the network of areas associated with attention shifts. & Rubinsztein. Dis. 1).A. Functionally. Rev.edu This was in contrast with an actual reduction of Individuals recovering from neglect also 1424 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . 669–670 8803–8812 (2004). 12. 479– 13. et al. findings mirror those in the parietal lobe.1 report that the activity of an interconnected network of dorsal. network in individuals with neglect through they are directing their attention3.b) The pictures demonstrate how they temporal parietal junction and the ventral lateral ignore their left visual field in trying to make simple line drawings of a clock (a) or a house (b). C. and the ventral network involves the Figure 1 Sample drawings by individuals with spatial neglect. 4. ity in visual cortex during stimulus detection) ticipants have longer reaction times when the the participants had considerably improved in lies in parietal areas. et al. 9. side when they are first cued to attend to the increased activity on the lesioned side from the pants were instructed to direct their gaze to right4. Biol. Corbetta et al. & DiDonato. M. 24. even though the individual tex? Several studies5 indicate that the source of most of the time. 551. Pharmacol. Warrick. M. H. 3. (2002). F. 8. In normal people. et al. Orr. In an fMRI scanner. & Saudou. 85–90 (2004). S. imaging studies show that this interconnected network is important for shifts of spatial attention2. S. de Chiara. such individuals with ‘spatial neglect’ may fail to orient to people approach- ing from their left. This was accompa- the center of the screen in front of them. ical for voluntary shifts of attention in normal imaging (fMRI) scans taken while the patients tion times in response to targets on the left people2. University of Oregon. A new paper uses functional imaging to show that a restricted lesion.

Corbetta.I.. mechanisms underlying this strong asymme. 4. E. S. Time to smell the roses Timing is a thorny issue for the chemical senses. M.nature. 2004). Nature 395. Neuropsychology 3. is an asymmetry between the ventral and be gleaned from this work.J. attention system produces neglect because important reason why forcing the increased J. & Elbert. Kanwisher. J. may be consistent with tion on the side indicated by the arrow. NEWS AND VIEWS improve their ability to spot targets on the parietal area alone do not cause neglect7. Neuroimaging of Visual Cognition (eds. A template chosen from any 50-ms window of a response could accurately classify odorant identity for responses in different presentation patterns.. What restricts PNs to repeatedly return to the same response trajectory despite ongoing dynamic activity? Can an ongoing response to one odorant also be reset by the arrival of a different smell? Researchers undoubtedly will continue to sniff around for these answers.D. This finding temporal parietal junction. and made intracellular and extracellular recordings from projection neurons (PNs) in the antennal lobe. the authors considered known features of the locust olfactory system. Psychol. approximately one oscillation cycle. Exp.L. 201–215 (2002).A.. Oxford. may be important for all of neuropsychology. between hemispheres may provide the basis for any case. R. & Driver. Each antennal lobe contains 830 PNs. Nat. Cara Allen NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1425 . Corbetta. the current study. A. J. M. 1863–1874 (1984). hemisphere lesions. T. who show a chronic right. tion of why more dorsal lesions do not cause ward bias. irrespective of the pulse timing.. Rev. Increased right hemi- 1. Rev.. R. Neurosci. Perry.J. may also be helped by inhibiting left 7. Response trajectories for repeated odor pulses overlapped. F. They exposed adult locusts to trains of brief odorant pulses with natural interpulse intervals. With an animal’s movements and the intermittent arrival of odors in the air. G. new odors are likely to interrupt responses to previous ones. J. & Sapir. known to reduce neglect. M. this solution depends on the ability of newly arriving odors to reset ensemble activity in the antennal lobe. whereas most other helping recovery. and Kenyon cells seem to integrate convergent PN input over 50 ms. 3–25 (1980). However..H. In most cases. well. 32. DiRusso. N. could serve as one 2. F.000 Kenyon cells in the mushroom body.. Principal neurons in the vertebrate olfactory bulb and insect antennal lobe have dynamic odor-evoked responses that can long outlast odorant exposure.. responses varied over successive pulses. This area is active only in the right as well. 193–207 (1998). 6. Egly. The temporal pattern of these responses is thought to be important for distinguishing different odorants. C. Neurosci. M. Brain 123. as revealed in many func- unattended side while they focus their atten. Posner. Robertson. Odorant stimulation evokes oscillations in the antennal lobe. in Functional it causes dysfunction of the dorsal system as use of a paralyzed limb can foster recovery9. a brain area important for olfactory memory. Posner. 1603–1610 (2005). Although this was not a specific focus of Among patients with neglect. In contrast to the varied responses in individual PNs.. & Martinez. odor stimuli have their own temporal structure. participants’ performance (Fig.A. Experience tells us that a second sniff of a rose still smells like a rose. the paper provides a strong argument © 2005 Nature Publishing Group http://www. Walker. strong neglect of the left side apparent in the sphere activity through warning signals. Friedrich. dorsal parietal areas: lesions of the dorsal lesions provide some explanation of how local. Remote effects of 9. 3. with successive pulses appearing to partially reset the circuitry such that each response began similarly and followed a similar trajectory. & Shulman. The ized computations. M. These results suggest that the problem of temporal interference may be adequately solved by converging PN input onto Kenyon cells.com/natureneuroscience gets in either direction. study1 reveals that method for improving hemispheric balance8.D. 3. 8. Rafal. Uswatte. but how does the olfactory system sort out these potentially conflicting time courses? On page 1568 of this issue. Lewis. The Corbetta et al.. Q. but in a natural environment. Nat. R. ensemble PN activity showed highly similar responses to multiple pulses. 169–172 (1998). malfunctions of the ventral parietal area and superior parietal lobe activity1. tional imaging studies. The right lateralization in the ipsilesional cortex or by reducing it in the effects of various forms of therapy. Hillyard. The authors suggest that competition apply to aphasia or sensory motor deficits. In hemisphere during shifts of attention to tar. of this area seems to be a major reason for the the contralesional cortex.) 381–388 (Oxford Univ. 228–236 (2002). Apparently. Press. S. damage to the ventral areas of the spatial The need for hemispheric balance could be an 4. there There are more general lessons that might 8. Friedrich. 1). I. & Rafal. Rorden. 5. Kincade. F. The authors pooled 117 PN recordings from multiple experiments and constructed activity vectors over 50-ms windows as well as ensemble activity trajectories over the entire PN responses to different patterns of odor pulses. to be responsible for interrupting attention and sphere activity during the acute stage of neglect The authors also argue that the general prin- allowing a shift of attention toward the target provides some basis for rehabilitation methods ciples of recovery that they have found might location. which also responded reliably to repeated odorant pulses. R. already Neurosci. 3. G. & Beck. This try in the remote effects of these two critical the occurrence of a syndrome with many com- recovery seems to depend mostly on the right areas of the parietal lobe remain unclear. C.B. Taub. A. showing that the temporal structure of odorant presentation did influence the temporal pattern of PN responses.J. Mattingley. a brain area thought Evidence for greater than normal left hemi. Neurosci. either by increasing activity for using imaging during recovery to determine areas are symmetric6. and Stacey Brown more than 100 PNs converge onto each of about 50. plex sensory and motor features. To determine the downstream consequences of this interference. J. Nat. those with right & Duncan.I. D. J. the results raise the ques. thus also produce neglect. & Zeki. 2273–2288 (2000).J. Mark Stopfer and colleagues address this question in the olfactory system of the locust.

connotea. pointing out (9.org/user/NatNeurosci/tag/ drug.5 million Americans aged 12 or older in so many different areas under different cir. crimi. How do Charles Dackis and Charles O’Brien discuss US Department of Health estimates that in we make sense of the multitude of observations social issues that may be hampering develop- 2004. This figure includes circuits that mediate reinforcing effects of drugs. about have the most validity for studying addiction. During this period. health costs. and their relationship to addiction View background material on Connotea at Drug addiction can clearly vary with the vulnerability. we nerable to drug use and abuse? Mary Jeanne I-han Chou highlight the biology of the most commonly Kreek and colleagues discuss the genetic Associate Editor abused substances. that of Drug Abuse and the National Institute of The National Institute on Drug Abuse esti. Only a handful of phar- 5% of the global population. repeat the experience. The results from drug use and drug abuse. 22. N E U R O B I O LO G Y O F A D D I C T I O N INTRODUCTION Neurobiology of addiction © 2005 Nature Publishing Group http://www. cost of addiction. determine why some individuals are more What makes certain individuals more vul- easily addicted than others. addicts cannot resist the urge to engage this complexity. Even in produce physical dependence. even after years of abstinence.nature. but for some. In a commentary. lead. Dackis and O’Brien argue that to strata and age groups. long-term treatments are Nations estimates that 200 million people. others. ment and access to treatment. we are making the content of ing nicotine addiction). public is likely to conceptualize addiction as treatment for their addiction in the US alone. We hope that our readers The progression from initial drug use to immediate and delayed consequences of actions. we must change this perception. Nicotine is one of the most Some users then progress to addiction. this focus freely available on the web for three health consequences from drug abuse and presenting a framework for how occasional behav.1 million people needed but did not get and what specific changes should we exam. to seek a pleasant stimulus. LSD or amphet. recreational and euphoria and is driven by a need to Alcohol and nicotine are legal drugs that use easily slips into dependence and tolerance. and disorder. Cocaine.html. the editorial team of Nature the effects of impaired productivity on these Bechara proposes that volitional decisions involve Neuroscience is entirely responsible for the con- individuals and their employers. ine? In three separate commentaries. Apart from the nal behavior. is the source of its societal stigma. will find this collection of articles useful and addiction is influenced by the drug. Addiction has very high overall changes are likely to be critical to addiction. Barry Everitt and erous financial support for this focus issue. and tobacco addiction the face of harmful consequences to self and on specific receptors and brain areas. even in the face of Alcoholism and Alcohol Abuse for their gen- mated the cost of drug and alcohol abuse at terrible personal consequences. They also discuss the difficulty http://www. enlightening and that it may contribute toward personality. light of the strong comorbidity between addic- after a hard day’s work is familiar to in which the individual feels satisfaction tion and other mental disorders. consumed illicit disorder.000 Americans each year. In this focus. ing to an inability to weigh future consequences understanding and eventually solutions to this mental stressors. peer influences and environ. siveness. can addiction be treated as John Dani and Adron Harris review progress in in the addictive activity. sup. the hallmark of addic- dependence or abuse. marijuana. Senior Editor components of addictive behavior itself. Given kills more than 430. in addictionfocus. risk taking and stress respon. animals (and humans) to alcoholism. choice is no longer volitional. tion. or porting the idea that addiction is a biological few and far between. the user’s He discusses how drugs may tip this balance. These complex interactions and the urge to make impulsive decisions. targets for all addictive substances that could bidity with alcoholism. Trevor Robbins review the cortical and subcortical With their help. A naive 21. but for addicts. Drug use and addiction are pervasive. Drugs also act widely abused substances. Peter Kalivas. once related factors such as heart Taking drugs may begin as a voluntary choice We are grateful to the National Institute disease. cancer and accidents are considered.com/natureneuroscience T he pleasant sensation of sipping a drink amine can create psychological dependence. iors become habits and then compulsions through focus/addiction/index. job loss and pavlovian and instrumental learning. Nestler addresses this issue. NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1427 . affected most. drugs at least once in the last 12 months. explore the genetics of influences on complex personality traits such Kalyani Narasimhan predisposition to addiction. George a character flaw rather than a bona fide brain Although drug abuse cuts across all societal Koob and Michel LeMoal.4% of the population) experienced substance cumstances? What animal models are likely to that loss of control. the young and poor are Yavin Shaham and Bruce Hope discuss which effectively develop treatments for addiction. months at http://www. John Crabbe and David addicts may relapse into drug use under stress be exploited to provide a ‘magic bullet’ for Lovinger discuss the neurobiology of alcohol or when faced with otherwise benign cues that addiction treatment? A perspective by Eric abuse and genetic influences that may predispose remind them of the addicting drug. The Exposure to drugs causes plasticity in neural Despite the enormous social and economic 2005 World Drug Report from the United circuits related to reward and motivation. many people.nature. negligent driving. a balance between neural systems signaling the tent of the focus issue.com/neuro/ their effects on the health care system. critical medical and societal problem. Plasticity (of synapses and circuits) maceutical therapies exist. and examine the as impulsivity. Heroin or alcohol can are prone to abuse. Moreover. about $246 billion in 1992 (without consider. Antoine sponsors’ foreword. in teasing out genetic vulnerability factors. if they do a unitary disorder? Are there common brain understanding nicotine addiction and its comor- stop taking drugs.

brain. socioeconomic class and poor parental support are two other factors that forcing effects of the abused substance6. 40–60% of the vulnerability to addiction can be attributed to genetic fac- tors4. anxiety disorder. novelty-seeking. and addicted individuals have a higher prevalence of mental disorders dence. its normal functioning and how it responds to social stressors Scientists are now able to portray addiction as a medical disease and reinforcers will allow us to develop strategies to engage adolescents with physiological and molecular changes thanks to the scientific and in productive and creative ways that will minimize their chances of technological advances that have occurred over the past decade. and perhaps could also reflect the incomplete devel- adolescents are significantly more opment of brain regions involved in the processes of executive control vulnerable than adults to sub. ADHD and schizophrenia) are at a much further improve prevention and treatment of addiction. genetic and neurobiological fac- sion with physical dependence. Here we highlight some of the compelling neuro. environment interactions that make a person more vulnerable to addic- mous. substance abuse during childhood research is currently focused on finding out whether the sensitivity to and adolescence has the potential to be particularly deleterious. to alcohol. response to peer of drugs and alcohol on children pressure) that increase the probability of someone experimenting with is particularly problematic. because many of the molec. S P ON S O R ’ S F O R E WO R D The neuroscience of addiction © 2005 Nature Publishing Group http://www.14.5. and motivation (for example. Better knowledge of the adolescent years of age or older3. Indeed. it is estimated that in those genotypes that confer increased susceptibility. economic impact in the United States of approximately half a tril. particularly Addiction has a significant genetic component. Furthermore. A more urges to take the drug even at the expense of adverse consequences. The stressors and alternative reinforcers in an individual’s environment7. Indeed. articles in this issue highlight some of the remarkable progress that has revolutionized our understanding of the neurobiology of addiction and Why do addicted people often have other mental illnesses? the way we treat it. and addiction-resistant phenotypes may also reflect sensitivity to the various stress might be a common feature of these environmental factors. increased availability of cocaine and methamphetamine vulnerability for addiction is often suggested to reflect both variability has contributed to the recent epidemics of addiction to these drugs. addiction-prone and are consistently associated with a propensity to self-administer drugs. consequences. The impact cent-specific behaviors (risk-taking. stance abuse and to addiction2. As mechanisms responsible for stress-induced increases in vulnerability to NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1429 . Low in metabolism of the drug and variability in the sensitivity to the rein. neuroadaptations during adolescence generalizes to other drugs and it has been shown that children who begin using alcohol early in child. substance abusers In this editorial we use the term addiction rather than drug depen. Comorbidities tations that result in withdrawal symptoms when drugs such as alcohol may emerge. Similarly. These estimates of heredity include the percentage of the variance What are the neural consequences of environmental risks? attributed to genetic factors by themselves as well as the percentage of the Drug availability is the most obvious environmental factor that influences variance that is attributed to gene-environment interactions. Much involved with brain development. preclinical studies indicate that the neuroadaptations that Also. This could reflect normal adoles- accidents and crime1. higher risk of abusing drugs and alcohol. However. which refers to the loss of control over the intense with depression or schizophrenia use nicotine and alcohol). occur in adolescents exposed to certain drugs such as nicotine or canna- ular targets affected by drugs are binoids are different from those that occur during adulthood10–12. we will be able to tailor interventions for those at high risk. with an estimated annual tion. to reflect overlapping environmental. myelination of frontal lobe regions)9. In fact. to avoid confu.nature. Those are distinct from the adaptations that tal disorder attempt to self-medicate (for example. Why does addiction begin most frequently during adolescence? lion dollars arising from medical Experimentation with drugs and alcohol often starts in adolescence. Individuals suffering from a variety of different disorders (such as scientific questions about substance abuse. in certain instances. The experimenting with drugs.com/natureneuroscience The burden of substance abuse we gain knowledge of the individual differences in genes and the gene- and addiction to society is enor. Physical dependence refers to the adap.8. nico- addicted to alcohol later in life than are those who begin drinking at 20 tine and marijuana early in life13. and whether this phenomenon could underlie the greater hood (ages 14 or younger) are four times more vulnerable to becoming vulnerability to addiction in individuals who start using alcohol. These robust comorbidities are likely Manual of Mental Disorders (fourth edition. Genotypic addiction. tors that influence substance abuse and mental illness. controversial interpretation. as drugs and alcohol. when individuals afflicted by a men- and heroin are discontinued. is the possibility that early exposure to certain drugs of abuse Why do some people become addicted and others do not? might increase the vulnerability to other mental disorders. the answers for which will depression. for which there is still not sufficient evi- dence. DSM-IV). and so does the process of addiction1. loss of productivity. which is the clinical term favored by the Diagnostic and Statistical than the rest of the population. when individuals result in addiction.

Neuroscientist 10. A. (in the press). C. interventions for drug addiction should include strategies that enhance 9. et al. Alcohol Clin. However. Wang.R. D.M. the relearning events triggered during a desensitization session18. McCormick. 336–344 (2005). NY Acad. Hiroi. Grant. N. social neuroplastic changes to drugs and alcohol. and influence our evolving strategies to direct our efforts to modifies the propensity to self-administer drugs. Psychiatry 10. and alcohol can disrupt volitional mechanisms by hijacking the brain However. Neurosci. P. S. fear in phobic individuals through the pharmacological strengthening of 18.R. Crabbe. and this could provide a mechanism by which a social stressor go haywire.. 92–98 (2005). D. 13. Addict. If we understand the prevent and treat substance abuse and addiction more effectively.E. Behav..A. decrease condi. Behav. Psychiatry 56. M.. Cereb. Volkow. despite these advances in understanding the responses to drugs of abuse. 6. W. B. a proof 59. affliction stems from voluntary behavior. Ting-Kai Li ioral manifestations of addiction. showing that D-cycloserine administration facilitates the extinction of 17. et al. National Institute on Drug Abuse addiction that offer new insights into how chronic drug use affects the inner workings of the brain and how this leads to the aberrant behav. Goldman. Morgan. 1039–1044 (2002). T. Gen.E. Volkow. Also. animal studies have shown that an increase in DA the addicted individual should spur a renewed discussion of what consti- D2 receptors in the nucleus accumbens markedly decreases drug con. & Agatsuma. Mol. 372–392 (2004). is the development of medications that act synergistically with an effective 14.D. National Institute on Alcohol Abuse and Alcoholism 1430 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . An interesting approach 389–395 (2004). Abuse 9. B. Kelley.D. 53. inhibitory control. D. affect the behavioral or incarceration. Res. G. Brain Res.. J.A. 1136–1144 (2004). 4. 15. Rev. executive function). 29.. neurobiological consequences underlying the adverse environmental fac- tors that increase the risks for drug use and for addiction. Ressler. 19. Ann. & Fowler.F. Although not yet evaluated for addiction. Genet. 169–174 (2002). Nat. 103–110 (1997). Brook. Oliff. Brook. we will be able 1. 27–32 (2004). 778. 25.W. 86–94 (2004). Psychopharmacology (Berl. Neurosci. M.S. Ranaldi. S.K. J.J. F. Breese.S.G. Kosten. in turn. H. For example. This new knowledge system and the hypothalamic-pituitary–adrenal axis15. & Leslie. Annu. & Landry. et al. A. Pharmacol. Res. The loss of behavioral control in © 2005 Nature Publishing Group http://www.. S P ON S O R ’ S F O R E WO R D drug use and to relapse in those addicted are not yet well understood. et al. long-lasting and implicate multiple brain circuits (reward. learning. Pharmacol. addicted individuals continue status affects dopamine (DA) D2 receptor expression in the brain. & Laviola. in nonhuman primates.. Psychiatry 61. strengthen inhibitory control and executive function. Nat. 318–325 (2000). 521–532 (2005). Adriani. D. Bauco. R. et al. S. the saliency value of natural reinforcers (including social support). This suggests that new 12. P. Arch. 185–195. 1021. J. & Whiteman. 3. P. & Li. Additional pre. Schochet. Pharmacol. J. T. Lee. Belluzzi. 10. Cohen. there is evidence that corticotropin-releasing factor (CRF) might mechanisms involved in seeking natural reinforcement and weakening play a linking role through its effects on the mesocorticolimbic dopamine brain mechanisms that inhibit these processes19. N. 5. et al..F.M.A. even in the face of dire consequences such as loss of a child’s custody tal factors affect the brain and how these. E. N. Eur. T. Psychol. J.) 174. Zhang. 341–352 (2004). Thompson. tutes volition. 12. low to be stigmatized by the pernicious yet enduring popular belief that their status decreases expression and increases the propensity for cocaine self. to develop interventions to counteract these changes 2. & Tossmann. challenge us to identify the neurobiological substrates that sumption.C. F. Sci. Psychiatry behavioral intervention. A.. Biol. We have learned how some drugs Director. Pistis. A. 7.S. Rev. 527–534 (2001). Oroszi.R. & Toga. Gen. G.W. Baumeister. G. & Ducci. 11.. The adaptations in the brain from chronic drug exposure seem to be 6. Cools. R. P. What is volition and how do drugs disrupt it? Nora Volkow Remarkable scientific advances have emerged in the neuroscience of Director.. 5389–5396 (2005). has started to provide explanations of why the addicted person relapses clinical studies have provided tantalizing insights on how environmen. 11. K.16. 435–462 (2002). & Wise. Sowell. 8. C. Subst.nature. Cortex 10. motivation. Exp..com/natureneuroscience administration17. 15. & Dawson. J. Ther. tioned responses and improve mood if disrupted. of principle for such a concept has been recently established in a report 16. 418–429 (1997). How can we repair the brain circuits disrupted by drugs? 5. Arch.

N E U R O B I O LO G Y O F A D D I C T I O N
C O M M E N TA RY

Neurobiology of addiction: treatment and
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience

public policy ramifications
Charles Dackis & Charles O’Brien

In the United States, efforts to treat addiction are hampered by prejudice and a public view that treats it as a disorder of self-
control, not a disease. We highlight select advances in addiction research that, if disseminated to the public, could reverse these
misconceptions and facilitate changes in policy to improve treatment access and care delivery for this highly prevalent disease.

The infrastructure to treat addictive illness, and readily criminalized. Specialized treatment substantiate the biological basis of addiction
when compared with treatment for other for addiction is even viewed as unnecessary and improve treatment outcome should
traditional medical illnesses, is lacking in the (why not ‘just say no’ to drugs?) or misperceived ultimately erode entrenched societal attitudes
United States. This situation is tolerated by a as being ineffective. In contrast, treatment that prevent addiction from being evaluated,
public that views addiction more as a social response does not dictate availability of care treated and insured as a medical disorder.
problem than an actual disease, despite sci- for other medical conditions like cancer, stroke Addiction is best conceptualized as a disease
entific evidence supporting a disease concept and heart failure. Why should it be considered of brain reward centers that ensure the survival
of addiction based on neuronal mechanisms, an appropriate standard for the availability of of organisms and species2. Given their func-
heritability, treatment responses and a charac- addiction treatment? This blatant discrepancy tion, reward centers have evolved the ability to
teristic progressive clinical course1. Pejorative in access suggests that, despite therapeutic grip attention, dominate motivation and com-
views toward addicted individuals also exist advances and improved clinical outcome, treat- pel behavior directed toward survival goals,
and contribute to policies that would be sim- ment parity will not be achieved until addiction even in the presence of danger and despite
ply unacceptable if applied to ‘real’ medical is widely viewed as a disease. our belief that we are generally rational beings.
disorders. These policies have created lim- Much of our knowledge about addiction By activating and dysregulating endogenous
ited access, insufficient capacity and a dearth neurobiology is based on decades of animal reward centers, addictive drugs essentially
of trained providers in most geographical studies that model the dynamic clinical hijack brain circuits that exert considerable
regions, especially for adolescent patients who components of the illness. Elegant study dominance over rational thought, leading to
might avoid progressive addiction with appro- designs assessing self-administration, progressive loss of control over drug intake
priate treatment. Even patients with access to conditioned place preference, reinstatement in the face of medical, interpersonal, occupa-
treatment typically discover that its duration (after cues, stress and drug priming) and tional and legal hazards. There is even evidence
is severely limited by insurance company poli- intracranial self-stimulation have provided that denial, once thought to be purely ‘psycho-
cies (managed care), even though addiction is a tremendous amount of behavioral and logical’, may be associated with drug-induced
a chronic illness requiring sustained aftercare. neurochemical information. Although this dysfunction of the prefrontal cortex3.
Imagine limiting treatment duration for dia- research has identified neuronal mechanisms Loss of control is both the hallmark of addic-
betes, chronic heart failure or hypertension. underlying drug reward, craving, relapse tion and the source of its societal stigma. An
Stigma and misconception create formidable and hedonic dysregulation, the predictive uneducated yet strongly opinionated pub-
obstacles to a more enlightened public policy value of animal models varies considerably. lic does not understand the technical field of
toward addictive illness. Rather than being Naltrexone treatment for alcohol dependence addiction neurobiology and is more likely to
treated as patients, afflicted individuals are often stemmed directly from animal studies showing conceptualize addiction as a character flaw (for
blamed for their illness, discriminated against that opioid antagonists reduce alcohol self- example, addictive personality) than a brain
administration. On the other hand, the robust disease. Therefore, the dissemination of under-
phenomenon of sensitization has received standable information about this brain disease
Charles Dackis is at the Department of Psychiatry, considerable emphasis even though its clinical could change public perceptions and hence pub-
3900 Chestnut Street, University of Pennsylvania, significance is questionable, and it has not lic policy toward addictive illness. Considerable
Philadelphia, Pennsylvania, USA, and Charles produced new treatments. Neuroimaging emphasis has been placed on the prevention of
O’Brien is at the Philadelphia Veterans Affairs may ultimately circumvent the limita- addiction through widespread educational ini-
Medical Center, Treatment Research Center, 3900 tions of animal models and delineate brain tiatives targeting children, adolescents and par-
Chestnut Street, Philadelphia, Pennsylvania, USA. mechanisms associated with clinical features ents, but much less emphasis has been placed on
e-mail: dackis@mail.med.upenn.edu of addictive illness. Scientific discoveries that the disease of addiction. Addiction is a disease

NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1431

C O M M E N TA R Y

+ Figure 1 The cycle of addiction is positively
Addictive agent Drug euphoria reinforced by drug euphoria and negatively
Positive reinforcement reinforced by withdrawal, craving and hedonic
Activated reward pathways dysregulation. Drug-related cues and stress
increase craving, and loss of control may stem
in part from prefrontal cortical dysfunction.
Drug administration Neuroadaptations Neuronal mechanisms for these cardinal
Drug-seeking behavior Withdrawal and tolerance components of addiction have been increasingly
Failed impulse suppression Protracted hedonic dysregulation delineated with animal models and human
neuroimaging studies.
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience


Drug craving
Negative reinforcement negative reinforcement, and alternate with the
Loss of control Dysregulated reward pathways positive reinforcement of euphoria to drive the
Denial/poor decision-making cycle of addiction (Fig. 1). Chronic exposure to
Hypofrontality/low D2 heroin, cocaine or alcohol produces a number
Reduced gray matter density
of common neuroadaptations8, including DA
Drug-related cues hypoactivity, that contribute to a remarkably
Limbic activation similar clinical course in severely addicted indi-
Stress
viduals. The cycle of addiction becomes etched
in midbrain and frontal structures that rein-
force the pursuit of survival-related behaviors
of brain regions that are intrinsically interest- mechanisms, which provides insight into by dominating attention and decision-making.
ing to the general public because they subserve the pathological process and identifies new Addictive illness reminds us that desire and
the human experiences of pleasure, craving and treatments. Discrete clinical phenomena of pleasure can be impervious to rational thought,
motivation. Fascination with this topic could this pleasure-reinforced illness are integrated clashing with deeply engrained cultural values
be exploited by educational initiatives to gain into a dynamic cycle of addiction (Fig. 1) that placed on stoicism and self-control.
ground against moralistic attitudes that stig- becomes progressively more entrenched and Craving is a complicated phenomenon
matize, ostracize and often criminalize patients uncontrollable as the brain becomes addicted6. that can be dramatically amplified by stimuli
with addictive illness. The biological basis of these clinical compo- (cues) that have become associated with drugs
Access to treatment for millions of addicted nents has been increasingly delineated through through conditioned learning. Neuroimaging
patients is a costly proposition. However, there an explosion in addiction research that initially studies of addicted human patients demon-
would be offset savings in the cost of medical involved animal models and has since expanded strate a fascinating link between brain function
care, lost productivity, neighborhood destruc- to neuroimaging studies of addicted patients. and cue-induced craving, which is arguably
tion, crime and prison capacity4. Even though Addictive drugs produce euphoria by acti- the most persistent and insidious clinical
the United States has a disproportionate num- vating brain pleasure centers, and it is notewor- component of addictive illness. Cues associ-
ber of prisoners, and most have been incarcer- thy that diverse agents (for example, opioids, ated with diverse substances (for example,
ated for drug-related crimes, their addiction is stimulants, alcohol, nicotine, marijuana) all cocaine, heroin, alcohol and nicotine) produce
seldom treated within the prison walls or, more increase extracellular dopamine (DA) levels robust activation of limbic structures on PET
importantly, after they are released to a drug- in the shell of the nucleus accumbens (NAc). and functional magnetic resonance imaging
infested environment. Similarly, although med- Drug-induced euphoria in humans has also (fMRI). Images depicting limbic activation
ical complications of addiction are commonly been closely linked to DA receptor (D2) binding during cue-induced craving provide an inter-
encountered in clinical practice, their cause is by several elegant positron emission tomogra- esting and graphic means of demonstrating
seldom addressed and treated5. By fully inte- phy (PET) studies6. Animal studies demon- the neuronal basis of cue-induced craving to
grating addiction treatment into our medical strate that natural rewards (sex, food, water) the general public (Fig. 2).
care delivery and judicial systems, we could dra- also elevate DA levels in the NAc, although to a Another interesting neuroimaging finding
matically improve medical care and justice. lesser extent, and human neuroimaging studies associated with addictive illness is that of hypo-
Here we highlight select areas of addiction report that DA release in the dorsal striatum frontality (reduced baseline metabolism in the
research that illustrate brain involvement correlates with meal pleasantness7. These stud- prefrontal cortex)6. Baseline hypofrontality
and would probably stimulate public inter- ies link drug euphoria to natural reward centers involves the same frontal regions that
est if conveyed in an understandable fashion. that have evolved to ensure survival. become hypermetabolic during cue-
Perhaps the dissemination of these and other Once experienced, drug euphoria promotes induced craving, and the exaggerated change
examples of current knowledge could begin to the repeated use of an addictive drug, espe- (∆ metabolism; peak minus baseline) in frontal
reverse popular misconceptions about addic- cially if genetic traits enhance the pleasurable metabolism might contribute to the remarkable
tive illness, increase compassion and tolerance experience. For instance, there is considerable salience of drug-related cues3. In addition to
and facilitate changes in public policy that evidence that individuals with a genetic predis- hypofrontality, cocaine-addicted individuals
improve treatment access and care delivery position toward alcoholism experience more show reductions in frontal gray matter density9
for this highly prevalent disease. pleasure from this drug because it produces an and poor performance on neuropsychological
exaggerated β-endorphin response. Over time, tests assessing prefrontal cortical function3.
The cycle of addiction addictive drugs disrupt reward circuits and As the seat of executive function in the brain,
Addiction neurobiology ties clinical phe- produce dysphoric states such as withdrawal, the prefrontal cortex is involved in decision-
nomena of the illness to specific neuronal craving and hedonic dysregulation that provide making, risk/reward assessment, impulse

1432 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE

C O M M E N TA R Y

the ventral striatum that correlates with the
severity of their reported craving for alco-
hol15. Thus, both genetic and alcohol-induced
alterations in β-endorphin are important in
the neurobiology of alcoholism. The involve-
ment of endogenous opioids in alcoholism led
directly to the development of naltrexone as an
approved treatment for this condition.
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience

Cue-induced craving and limbic activation
We have known for decades that environmental
stimuli (people, places and things) associated
with drug use can trigger intense craving in
addicted patients. Aside from perpetuating
active drug use, cue-induced craving triggers
relapse after protracted abstinence because it
persists for months or years, and even perhaps
indefinitely, as a direct avenue to recidivism.
Seeing a syringe in the doctor's office, smelling
a cigarette, or glancing at a vodka advertisement
Figure 2 Cue-induced craving (produced by a cocaine video compared to a nature video) is associated
are innocuous experiences for most of us but
with significant limbic activation on PET, which graphically demonstrates the neuronal basis of this
important clinical component of addictive illness. can be painfully compelling for vulnerable indi-
Reprinted with permission from the American Journal of Psychiatry, Copyright 1999. American Psychiatric Association.
viduals. Largely unknown to the general public,
neuroimaging studies have demonstrated dra-
matic limbic responses to drug-related cues that
control and perseverance. Functional and endogenous opioids and influenced by genetic correlate with the degree of reported craving.
structural abnormalities in the prefrontal factors affecting opioid function. One of the This phenomenon graphically demonstrates
cortex might therefore contribute to clinical earliest reports10 pertaining to this topic was the biological nature of addictive illness and
characteristics of addicted patients (such as published in 1980, showing that naltrexone provides one of the most fascinating examples
poor impulse control, lack of resolve, faulty pretreatment extinguishes alcohol self-admin- of the mind/brain interface.
decision making) that are viewed prejudicially istration in rhesus monkeys. Several lines of Neuroimaging studies of patients addicted
by the general public. Hypofrontality is animal research subsequently demonstrated to various substances demonstrate the activa-
associated with reduced D2 receptor availability that alcohol acutely increases opioid activity, tion of similar frontal regions as a common
on PET, which may be a marker for reduced DA especially in animals bred to prefer alcohol, pathway of cue-induced craving. Cocaine-
function in addicted patients. and that alcohol is not self-administered by dependent patients have been studied exten-
The next sections review specific neu- µ-opioid receptor knockout mice11. sively in PET and fMRI experiments, and they
robiological findings that are likely to be of Human studies also demonstrate involvement consistently show activation of the amygdala
great interest to the general population and of endogenous opioid systems in alcohol and anterior cingulate cortex that correlates
convey the biological basis of addiction. It is reward. Compared with normal subjects, closely with their reports of craving severity3.
not widely known that the brain produces opi- individuals with a genetic predisposition for Alcoholic subjects also show activation of the
oids and opioid receptors, that heroin binds alcoholism have low baseline blood β-endor- anterior cingulate, medial prefrontal cortex
to these receptors, and that alcohol pleasure phin levels and enhanced β-endorphin release and striatum in response to alcohol-related
involves opioid function. Also unappreciated and euphoria after alcohol administration12. cues on fMRI16,17, and the intensity of the cue
is the importance of cue-induced craving, its Enhanced release of β-endorphin against reactivity correlates with their likelihood of
basis in limbic activation, and evidence that low baseline levels constitutes a surge in the relapse18. PET studies of heroin-dependent
addicted individuals have impairments in concentration of this rewarding endogenous subjects also demonstrate a strong correla-
executive and hedonic function. These findings opioid that may explain why these individuals tion between cue-induced opioid craving
should be disseminated to the general public experience more pleasure from alcohol. Other and hypermetabolic responses in the inferior
in understandable and interesting forums to studies corroborate an interaction between frontal and orbitofrontal cortex19, and patients
promote the disease concept of addiction. β-endorphin levels and alcohol consump- with nicotine dependence show increased
tion. Cerebrospinal levels of β-endorphin metabolism in the anterior cingulate on fMRI
Endogenous opioids are three times higher in normal subjects during exposure to cigarette-related cues20.
Endogenous opioid pathways activated by than in patients with chronic alcoholism, These studies demonstrate a common neuro-
addictive drugs are involved in pain, plea- and there is evidence that β-endorphin levels nal response to cues associated with diverse
sure, appetite, sexual function and natural might become depleted after chronic alcohol substances, and they justify the commonly
drive states, and it is noteworthy that separate intake13. Alcoholics experiencing withdrawal held notion that various drug dependencies
and antagonistic enkephalin and dynorphin symptoms have plasma β-endorphin levels should be conceptualized as a single disorder.
populations of medium spiny cells in the NAc only half as high as those in normal subjects, Natural drive states are also associated
are involved in addictive illness8. Although yet their levels normalized after several weeks with activation of glutamate-rich cortical
alcohol is ubiquitous in our society, few peo- of sobriety14. In addition, abstinent alcoholics regions. Remarkably, the same frontal regions
ple know that alcohol reward is mediated by show increased µ-opioid receptor binding in that are activated by cocaine-related cues in

NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1433

C O M M E N TA R Y

cocaine-dependent patients are also activated of striatal D2 receptors in patients with Autoreceptor downregulation might also
in normal subjects viewing sexually explicit addictive illness. PET studies using contribute to the controversial finding of
videos21. Furthermore, neuroimaging studies [11C]raclopride, a radioligand that competes sensitization, which has unclear relevance
demonstrate that the subjective report of with DA at D2 receptors, demonstrate to addictive illness despite its considerable
hunger in response to food-related cues is persistently low striatal D2 availability emphasis by many animal researchers36.
temporally associated with marked activation (↓D2) in patients addicted to cocaine26, Whereas tolerance is defined as a reduced
of frontal regions22. These studies link drug- alcohol16, methamphetamine27 or opioids6. dose response after repeated drug adminis-
related craving with natural drive states, and Individuals with morbid obesity also have tration, sensitization involves accentuated
graphically support the idea that addictive ↓D2 that is inversely related to their body responses, classically in the form of enhanced
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience

drugs hijack endogenous reward circuits that mass index28. locomotion with a repeated fixed dose of a
have evolved to ensure survival. It is not known whether ↓D2 in addicted stimulant or opioid agent. Enhanced cocaine-
Prefrontal cortical regions that are activated patients precedes or results from their induced elevations of DA37 and glutamate38
during cue-induced craving receive DA drug exposure, and there is evidence that in the NAc of cocaine-pretreated animals are
projections from neurons originating in the both possibilities may occur. That ↓D2 associated with sensitization and could be
ventral tegmentum. A series of elegant single- persists beyond detoxification from alcohol explained by persistently downregulated D2
cell recording studies demonstrate that these and opiates suggests that it might be a (ref. 3) and mGluR2/3 (ref. 39) autoreceptors,
midbrain DA neurons fire during unpredicted predisposing factor or at least a persistent especially as mice lacking D2 expression (hav-
hedonic activity, but their firing habituates to drug-induced finding6. The possibility that ing no autoreceptor function) show strikingly
predictable reward and shifts instead to cues ↓D2 represents an inherited trait is compelling enhanced striatal DA levels after the admin-
that reliably predict impending reward23. because D2 binding varies considerably across istration of cocaine and morphine40. Thus,
These and other studies reviewed elsewhere3 individuals, and nonaddicted individuals with sensitized DA and glutamate responses to
suggest that DA firing is correlated with ↓D2 report significantly more pleasure cocaine, often invoked as a rationale for test-
cue-induced limbic activation. Interestingly, after receiving stimulant drugs6. Similarly, ing DA- or glutamate-inhibiting agents, may
studies demonstrate that DA firing in the monkeys with ↓D2 are significantly more merely reflect a homeostatic autoreceptor
ventral tegmentum plunges below baseline likely to self-administer cocaine than response to DA hypoactivity. Sensitization in
when anticipated reward is not delivered, those with increased striatal D2 receptor animals has led some researchers to speculate
linking DA hypoactivity to an animal model availability on PET (↑D2; ref. 29). Indeed, that cocaine euphoria actually increases over
of acute deprivation (craving). Furthermore, ↑D2 may protect against addiction because time, even though patients typically report the
animals chronically exposed to stimulants, alcohol intake is significantly reduced in opposite and escalate their daily consumption
alcohol or opioids show dramatic depletion of rats after D2 receptor expression has been of cocaine. This area of research provides an
extracellular DA in the NAc24, and DA deple- increased with an adenoviral vector30. The excellent example of why animal models must
tion might contribute to craving and hedonic importance of genetic factors in addictive be reconciled with clinical experience.
dysregulation in addictive illness8. illness, especially those affecting the intensity In cocaine and methamphetamine abusers,
Because cue-induced craving is associated of drug reward, reinforces the biological basis ↓D2 is correlated with reduced metabolism in
with DA firing and hypermetabolic of this disorder. the orbitofrontal cortex6. As previously noted,
responses in glutamate-rich cortical regions, Although ↓D2 may represent a consti- hypofrontality in cocaine-dependent patients
medications that reduce DA neurotrans- tutional trait and addiction vulnerability, may contribute to poor impulse control, ele-
mission have been widely proposed as it can also result from cocaine exposure ments of denial and compulsive drug use2,6.
potential treatments for this phenomenon. because chronic cocaine treatment Cocaine-dependent subjects with reduced
Glutamatergic neurotransmission is also produces ↓D2 in monkeys. In addition, anterior cingulate and right prefrontal cortical
implicated in cue-induced craving, and glu- D2 varies with social dominance rank in metabolism have concomitant difficulty con-
tamate-releasing neurons in the orbitofrontal cynomolgus monkeys and is reduced with trolling impulses during formal neuropsycho-
cortex (which receive reward-related sensory social demotion, leading to an increased logical testing41. These findings suggest that
input from the thalamus) fire during cues propensity to self- administer cocaine 31. agents that increase metabolic activity in fron-
related to natural rewards and send excitatory There is considerable evidence from animal tal regions, such as modafinil, might improve
projections to the VTA and the NAc25. Cue- studies supporting DA hypoactivity after impulse control in addicted patients3.
induced craving often leads directly to relapse, chronic exposure to stimulants, opioids, Through animal models and human neu-
and an effective treatment for this phenom- and alcohol 24,32,33, and human studies roimaging studies, researchers are eluci-
enon should dramatically improve outcome. also report DA hypoactivity in alcohol-34, dating neuronal mechanisms that underlie
DA and glutamate antagonists should be heroin- 35 and cocaine-addicted patients 6, the dynamic clinical elements of addictive
tested in the laboratory before concluding with the latter group showing evidence of illness. First, this body of research strongly
that they reduce limbic activation during cue DA hypoactivity on neuroimaging and a supports the disease concept by linking the
presentation, especially as addicted patients host of neuroendocrine and autopsy studies activity of reward-related structures in the
may already be DA depleted. Indeed, limbic reviewed elsewhere 25 . DA hypoactivity brain to clinical manifestations of this dis-
activation during drug-related cues provides after chronic cocaine administration is ease. Common neurobiological phenom-
a unique biological marker that should be associated with the downregulation of D2 ena also justify categorizing addiction to
exploited with further targeted research. autoreceptors that are abundant in the diverse agents under a single general disor-
striatum3. Consequently, ↓D2 may reflect der. Diverse agents like cocaine, heroin and
The D2 story autoreceptor downregulation and may alcohol increase striatal DA levels during
One of the most interesting findings in serve as a marker for DA dysregulation in intoxication, whereas chronic exposure to
addiction research is the reduced availability addictive illness. these agents is associated with DA hypo-

1434 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE

alcoholic outpatients prompted laboratory over methadone. ↓D2 and limbic activation during significantly more likely to benefit from naltrex. treatment development. Although naltrexone represents a success and hypofrontality. heroin and nicotine dependence is a industry. sive clinical course across various substances. capacity and quality of oid dependence because naltrexone does not but more often to recapture the experience of addiction services. Acamprosate modu. sometimes in response to side effects improve the access. One of these studies suggests by FDA rules stipulating which physicians can © 2005 Nature Publishing Group http://www. drinking. If ment may also be reluctant to promote treat- genetic determinants and a similar progres. and is greatly underused. most controlled studies have Pharmaceutical companies have just begun that research findings in this technical area reported significant reductions in daily drinking to view alcoholic patients as an important pop. naltrexone is not Animal models with demonstrated relevance significant daily drinking and alcohol craving in widely prescribed to the enormous population to the clinical setting. has many advantages despite popular misconceptions. this finding is replicated. finding that has been reported to attenuate cocaine eupho- craving and even denial. However. semination is now indicated to promote the Alcoholics with this variant are reported to be nists have not been made available for testing disease concept of addiction. of active alcoholics. Still. reported higher alcohol craving at baseline and tigation in three large clinical trials3. have been promoted by the pharmaceutical ing. This issue is being facilitated by public acceptance of the dis- patients often stop the drug and resume heroin addressed by the development of depot deliv. and other promising medi- also depends on their translation into clinical alcohol or money. will likely be identified nist) was originally developed to treat heroin story that stemmed directly from neurobio. the greatest translational ology and continued funding in this essential the development of depot delivery systems that problem likely stems from a curious lack of area of research. dependence by blocking euphoria. the ulation. in marijuana-dependent patients. However. Treatments for nicotine dependence alcohol significantly more often. After these find. attained with advances in addiction neurobi- adherence has recently been addressed with trexone. One of the polymorphisms (Asp40) for the lates N-methyl-D-aspartate (NMDA) receptor involvement in alcohol euphoria. reducing opioid reward with naltrex. One problem has been patient nonadher. the might be screened in the neuroimaging labora- plans. the pharma. and the recent approval of acampro. Therefore. an estab. is seldom prescribed by primary care feature of alcoholism that may be specifically clinical components of addiction. by the fact that cannabinoid receptor antago. exemplified unknown to the general public. craving. Conversely. even though alcohol often ing self-administration and reinstatement. The development of placebo. Cocaine cal outcome should reverse social stigma and after alcohol priming. Naltrexone treatment for alcoholism.nature. NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1435 . even after the priming dose when addi. being tested3. be made available to the public in an under- with naltrexone treatment. Studies showing naltrexone efficacy in of opioid dependence. findings from this research that illustrate brain ics. which fuels the common phenomenon can treat. and their dis- risk of alcoholism45 and heroin addiction46. Buprenorphine. stress-induced craving. practice. including drug-induced effect of naltrexone pretreatment on baseline rently under intense investigation. Pharmacological strategies in which the first drink leads to uncontrollable New pharmacological strategies that target are emerging that target specific clinical com.42. cue-induced gene encoding the µ-opiate receptor produces subunit expression49. We have reviewed select in a subgroup of genetically defined alcohol. Since then. hedonic dysregulation Naltrexone (an opioid receptor antago. topiramate. propranolol and baclofen) are (such as bupropion and the nicotine patch) the naltrexone group reported less crav. including physicians. are often prescribed by primary care tional alcohol was available. cue-induced and alcohol-induced craving48. has not been sponsored by they begin to drink is an important clinical Potential pharmacological treatments for other industry. Public policy implications outcome by weakening the addiction cycle. ological research3. in alcoholics reported a reduction in clinically for alcoholism since 1994. Furthermore. logical research. Addiction also has one than patients without the variant are47. drinking behavior after the priming dose of vaccines that prevent the drug from enter- However. available genotype to match alcoholic patients an effective partial µ-opioid agonist that was These attributes are certainly consistent with with effective treatment. This study also tested euphoria is also being targeted with cocaine justify an expanded care delivery system. toward addictive illness. its use is curtailed Pharmacological treatments for addiction are mediated. especially those assess- active versus placebo groups43. ameliorated by naltrexone. and the placebo group chose cations for cocaine dependence (disulfiram. Animal and human research allow for monthly medication injections. awareness among primary care physicians should be closely coordinated and focused on The initial controlled study of naltrexone regarding naltrexone. standable manner that conveys the biological efficacy of naltrexone might be more dramatic sate might signal a change in industry attitudes nature of addiction. the clinical impact of new treatments alcohol by asking subjects to choose either ing the brain50. clinicians will have an ments for illegal addictions.com/natureneuroscience Several approved and promising treatments for that naltrexone diminishes alcohol-induced prescribe the drug and how many patients they addiction have been identified through neurobi. ings were replicated at another site44. This controlled study evaluated the specific elements of the addiction cycle are cur- ponents of addiction. It is also imperative holism. this treatment for alcoholism lished pharmacological strategy that improves is markedly underused in clinical practice. Our govern- cue-induced craving. recently approved for office-based treatment the idea that addiction is a brain disorder. These changes would be convincingly ameliorate opioid craving. this strategy has been limited in opi. consumed fewer logical target for candidate medications that physicians and are covered by some insurance drinks and drank more slowly. C O M M E N TA R Y activity. These interesting examples are largely and individuals with this variant have increased treatments for illegal addictions. ence. ery systems that will eliminate the need for more effective treatments. through expanding research3. ease concept and through the development of use. Changes in public policy are needed to However. craving and the genetic vulnerability to addic- a receptor with high affinity for β-endorphin. a full agonist with many legal testing to assess how the beneficial effects restrictions. Cue-induced craving in cocaine. opioid. tendency for alcoholics to lose control once tory before being tested in large clinical trials. and patients to make a daily decision about nal. naltrexone produces the very illnesses for which these should be prioritized as a means of guiding gained FDA approval for the treatment of alco. Both goals could be one provides benefits for some patients. on the other hand. hedonic dysregulation. Nonetheless. patients seek medical treatment5. Despite FDA approval developing practical treatment interventions. and alcohol-induced euphoria. Modafinil euphoria. ceutical industry is still reluctant to develop tion.versus naltrexone-treated patients ria in two controlled studies and is under inves- treatments that dramatically improve clini.

et al. 2015–2021 and effective medical care. J. Am. 45. Neurochem. 14. J. Sell. Pharmacol. 19–29 and humanitarian benefit to our society. Fowler. A. et al. & van Ree.. 24. as illustrated by nal. Grusser. 393– © 2005 Nature Publishing Group http://www. & Berger. M. Exp. Drug Des. Bart.. A. S. Pharmacol. T. 18.S. J. Y. 547–549 (2004). Such individuals should (2002). A. et al.. & Fowler. Assoc. Med. Brain Res. et al. et al.) 431–444 (Cambridge Univ. 1221–1247 (2004).R. & Dagher. 924. treatment. et al. Coiro. Kalivas. Altshuler.E. et al. 40. Tupala.. & Passeri. not always been adequately translated into and credibly to our youth. G. 1790–1797 (2004). 284. Neurosci. Wisconsin. & Kreek. C. tant determinant of drug experimentation.X. C. Menominie. et al. Schultz. G. Addict. The training of phy.. A. C. M.L. Biol.) 175. D. they have dangers of drugs should be conveyed accurately 13. 30. Dackis. Psychiatry 51. Gen. Primary-care physi. Rossetti. Primary-care physicians Public-policy changes that improve access. S. Krishnan-Sarin. Volpi. Although medical complications of substance tion. 79–86 (2005). Shannon. & O’Brien. 49. programs. V. 1546– Ending inappropriate criminalization of this System (eds. D. Carboni. debilitating disorder will probably not be 29. J. Wang. (2005). tion to influence training policies. Assoc. P. 111–117 (2001). Given the progressive nature of addic. already integrated into our schools and should Suppl 1. Melis. Jones-Gotman. The judicial sys. Goadsby. J. Sci. J. procurement. & Tiihonen. L.R.R. Active addiction often involves 36. Myrick. for addictive illness.G. Westenberg. and their considerable treatment. 9. 1140–1149 (2003). 28..P. Drug Abuse Network. Heinz. Addicted patients already incarcerated 5. 471–475 (1992).S. 169–174 (2002). Economic (2005). clinical practice. et al. (2002). tion have been identified. Unfortunately. et al. Prog. Belenka. 561– is another area in dire need of guidance and 567 (2001). be integrated into mainstream medicine. Arch. D. Subst. O’Malley. Psychiatry 28. Wang.J. S. 749–760 10.K.L. J. 207–216 (2000). We cannot allow the stigma of addic. P. D. Psychiatry 53. Neuroscientist 7. et al. 1436 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE .. 34. as well as ongoing individual. 50. 293–302 (2001). O’Malley. Neurosci. 247–291 (1993). et al. Giannini.C. C. M. C. Czoty. (1996).A. 48–58 (1990). 44. A. Lu. 1715–1723 (2001). S. edly has. Bart. tem should develop an integrated interface 3. C. Psychiatry 49. J. 1003. 876–880 nated. Volkow. et al. G. R. Ann. & Kleber. et al.. Robinson. 27.. Psychiatry 161. K. F.P. B. 881–887 should receive treatment within the prison 6. & Duffy. & French. Heinz. cians should competently evaluate their patients treatment. Am. found. 13 (2001). M. 8. & Gessa.. addiction and medical treatments are substance abusers is entirely unacceptable. (1992).T. Schools provide a 583–586 (1982). Small.W.W.L. C. Neuropsychopharmacology 28. W. J. P. Med. G. 1094–1103 tion-related medical conditions. 2005). C. et al. D. (2005).. 26. 1325–1329 financial interests.. E. 134–142 (2002). S. We thank A. Education designed to prevent addiction is (2002). Garavan. 1789–1798 (2000). the mainstream of medicine. Neuroimage 19. Alterman.J.P. F. Behav.M. as it undoubt. Sinha. and the opportunity of early intervention. H. expanded in our medical schools and residency a comprehensive assessment by qualified 25. NY Acad. 21. 1689–1695 (2000). G. Psychopharmacology (Berl. 1709–1715 (2003). J.D. 2002). Oslin.. 11017– recovery over incarceration. there are few practi. McDonald. Psychiatry 157. & Wang. Neuropsychopharmacology 24.J.com/natureneuroscience 402 (2004). & Thavundayil. P.. MpcLellan. Rammes. practice. R. P. Brody. Psychiatry 9. 39–53 (2004). McKhann. J. O’Brien. Brain Res. Neuropharmacology 40. S. J. Parthasarathy.A. Patapis. Psychiatry 158. and drug testing of inmates and correc. Lewis. Neuroimage 21. 233–240 (1980). criminal behaviors related to drug use and COMPETING INTEREST STATEMENT 37.W. Mertens. 57–64 nicotine dependence.. 5. Am. S. Morgan.D.S. savings than quality enhancement. especially those with addic. Dackis. trexone treatment for alcoholism. Volpicelli.A.C. M. T. Volkow. will continue to improve and streamline (2001).L. A. Gage. 23. will require significant investment in our 22. 141–150 the extent that addiction treatment is placed in occur in a climate that focuses more on cost (2002). et al.P.P. Dackis. macological treatments as part of their clinical addiction as a medical disorder.. Psychopharmacology (Berl. R.. Neuropsychopharmacology 30. has recently become interested in alcohol and means of identifying students who may require 15. 269–281 seldom coordinated5. 20. Gen. 513–516 (1992). M. & Berridge. et al. Dis. ACKNOWLEDGMENTS 35. Arch. ventions. 250–257 (1996). walls. Cambridge. Biochem. et al. G. Psychopharmacology (Berl. J. are essential in this regard and should have a capacity and quality in addiction treatment 21. Thanos. C. A.S. J. requiring Ann. et al. 296–302 (2004).. A. the Gen. J. Vescovi. H. J. 24. J. C.R. with specialized treatment teams to ensure that Benefits of Drug Treatment (National Rural Alcohol and 43. be made with the same frequency as those to regarding the cost of untreated addiction in 32.. 55.. G.. D. et al. H. Haney. & (1992). O’Brien. Gerrits. Drug Alcohol Depend. Franklin. Am. Georges.. N.. 41. Morgan. Sci.D. & Kosten. N. O’Brien. G. P. G. S. Am. in Diseases of the Nervous 47. & O’Brien. J. Gianoulakis. Endocrinol. 608–615 during recovery. It remains to be seen whether 28. Rouge-Pont.P. 11–18 (2004). H. F. Psychiatry 162. 2. 42. A. & Aston-Jones.. (2004). practitioners. 1423–1431 4.. Asbury.D.R. 1002–1008 (2000). Ther. Am.S. 417–422 tional officers should be applied to eliminate 7. N. W. Krishnan. Clin. the need for 31. Psychiatry Res. Hayashida. P. NY Acad. Kish. Reid. The judicial approach to addicted patients Biol. Di Chiara.. Rev. & McArthur.) 174. 654. Neuropsychopharmacol. C. Thanos. policy change. C. et al. T. Psychiatry 49.J. Genazzani. A. and view phar. and addicted patients typically The authors declare that they have no competing 38.J. much greater role in the assessment and treat. Neuroreport 7.A. Farren. receive innovative judicial interventions that 1.K. 48. Arch. 381– 388 (2004). 3293–3301 H. Addiction treatment should also cent substance abusers in most geographical 17. Expert Rev. H. Xi. disease would produce tremendous financial Press. care delivery in this country will be improved to improvements in our care delivery system will 33. Neurosci. Because perceived risk is an impor. G. M.. The pharmaceutical industry natural setting for these interventions and a Alcohol Alcohol. Moore. M. Mol. 46. 78. if we wish to provide comprehensive pharmacological treatments for addiction 27. J. sicians to assess and treat addiction should be chronic disorder is labor intensive. Petromilli. Nat.. promote treatment over criminalization and 2. 22. J. Neuropsychopharmacology 29. 11022 (2004). Roberts. Neuropharmacology 47 (Suppl. a ‘magic bullet’ for this chronic J.M. Arch. Psychiatry 62. 286. 293. Hester. ment of addicted patients. 303. & arsenal. & appropriate interventions are closely coordi. Discov. Vaccines 3. Neuropsychopharmacology other medical specialties. Treatment for this 23. regions. The general quality of dollars and lives. Brain Res.I. Gen. If and when the public begins to view (2001). C. Abuse Treat. Neuropsychopharmacology 29. et al. Volkow. be further developed with research-based inter. Although 1190–1202 (2004).) 160. J. S. 11. C O M M E N TA R Y Even when effective treatments for addic. abuse are commonly encountered in clinical inadequate access to treatment for adolescent 19. 60. Martinez.P. 18.J. Exp. Childress for contributing Figure 2. 12. Weisner.) 3–13 (2004). Metab. Pharmacol. (2004). et al.E. C. Referrals to addiction specialists should treatment parity could be added to arguments Nader.M. Dackis. H.A. 16. group and family interventions. Ther. Z. engage in activities they would never consider 39. 130. 328–345 (2003). Z. et al. & Garavan. E.. 1552 (2003). P. 1783–1789 resources could potentially expand addiction tioners qualified to evaluate and treat adoles. the widespread use of drugs within our prisons.nature. Synapse 5. N. & O’Brien.D.. health delivery system. M.A.

The effect of drug cues on drug drug-induced neuroadaptations are involved cocaine-induced reinstatement and cocaine- seeking can also be measured in extinction tests in relapse to cocaine seeking. cause of long-term relapse vulnerability in cocaine seeking. Kosten. evaluating the implications of the above Behavioral Neuroscience Branch. in which cocaine-dependent patients were the neuronal mechanisms underlying relapse to neuroadaptations induced by repeated drug treated with modafinil (a drug that increases drug seeking4. Research Program. Hope are in the Sensitization refers to the enhanced psycho. Drive. glutamate transmission). Maryland 21224. Department of occurs after repeated exposure to psycho. Subsequently. induced neuroadaptations in the progressive mens of psychostimulants diminish cysteine- ister drugs by pressing a lever and then undergo increase in cocaine seeking after withdrawal.nature. of drug seeking induced by drug priming. and both behaviors are trigger human relapse also reinstate drug seek.com/natureneuroscience relapse to drug seeking Yavin Shaham and Bruce T Hope One of the most difficult problems in treating addiction is not withdrawing addicts from drugs. Poling in nonsynaptic glutamate is critical for acute injections) or exposure to drug cues or stress on & A. Baltimore. cue. promising results15. T. the effect craving6 that was recently demonstrated in levels in the nucleus accumbens. We also discuss the role of drug. the nucleus accumbens12. this decrease of ‘drug priming’ (acute noncontingent drug humans (T. reinstatement of drug seeking. Sensitizing regi- mice.13. during which lever-presses a phenomenon called incubation of cocaine to decreases in basal non-synaptic glutamate do not deliver the drug. 90. Dependence Annual Meeting Abstracts. including relapse. p. National Institutes of Health. this sensitized drug priming provoked relapse behavior. Reversal of this neuroadaptation with systemic ing (the operational measure of drug seeking) 2005). the available induced glutamate release14. These extinction tests. After cellular and neurochemical adaptations in the are also involved in drug priming–induced prolonged abstinence. Persistent neuroadaptations are thought to underlie aspects of addiction.7. This commentary assesses the degree to which these neuroadaptations. during which that these neuroadaptations are the main in psychomotor sensitization and relapse to rats are exposed to the drug-associated cues. USA. as stimuli that neuroadaptations contribute to reinstatement motor response. Our assessment suggests that although injections of cysteine prodrugs prevents both is measured5. These and related findings permit the characterization of the time course humans. Oliveto. This clinical scenario can be modeled the degree to which specific drug-induced ciated with sensitized drug-induced loco- in a reinstatement procedure. which leads ‘extinction’ training. Intramural motor response (quantified by measuring findings for general treatment strategies. glutamate transporter activity. Kosten. rats or monkeys are trained to self-admin. N E U R O B I O LO G Y O F A D D I C T I O N C O M M E N TA RY The role of neuroadaptations in © 2005 Nature Publishing Group http://www. laboratory animals is often provoked by NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1437 . but preventing relapse. led to the suggestion that pharmacotherapies of relapse vulnerability.nida. Here we assess psychostimulant or opiate seeking is asso- stress2. and many Over a decade ago.A. preclinical data do not allow us to conclude a link between a neuroadaptation involved nence6. The reinstatement and for drug relapse prevention should aim at extinction procedures described above are Neuroadaptations and reinstatement reversing cocaine-induced neuroadaptations7. regarded as valid animal models. drug relapse and craving brain that underlie different facets of addiction. Relapse to drug use is one of the core features According to a popular hypothesis.11 hypothesized that esis.gov months after the last drug exposure10. Several issues should be considered in Yavin Shaham and Bruce T. J. chronic The neuronal systems involved in endur- of addiction and probably the most difficult drug exposure causes long-lasting molecular. primarily identified in preclinical studies on cocaine.nih.3. demonstrating administered after different durations of absti. exposure that produce enduring psycho. drug relapse in both humans and e-mail: yshaham@intra. studies using cocaine-trained rats in which Health and Human Services. and stress-induced relapse to drug seeking. College of Problems on Drug cocaine-induced synaptic glutamate release7.R. National Institute on Drug locomotor activity and stereotypy) that These data were derived primarily from Abuse. drug-associated cues or after cessation of drug use7–9. the authors of two A recent clinical trial based on this hypoth- investigators use these procedures to investigate influential reviews10. drug associated with enhanced dopamine release in ing in drug-abstinent animals4. 5500 Nathan Shock stimulant or opiate drugs. induce relapse. ing psychostimulant and opiate sensitization clinical problem in addiction treatment1. cues or stress. In this model. has led to tentatively motor sensitization also underlie drug-. reinstatement of nonreinforced lever respond. The mag- is often precipitated by acute re-exposure including prolonged relapse vulnerability nitude of drug-induced reinstatement of to the drug itself. response to drugs can persist for many However.

rather than directly mediate. generality of the findings from studies on the stress on reinstatement is not correlated with induced alterations of mesolimbic BDNF role of cocaine-induced neuroadaptations its effect on locomotor activity and nucleus augment.5. Wentzell naive rats. attenuates cocaine seeking. We the critical issue for future preclinical studies is mone corticotropin-releasing factor (CRF) in found that exposure to cocaine cues increases the identification of specific neuroadaptations the ventral tegmental area (VTA. This drug-induced neu. which also potentially contribute to persis. Thus. Campos. experimental manipulations should A. These find. also prevent cue. given that relapse rates chomotor sensitization are also involved in without alterations in BDNF expression23. Gardner. over the first several weeks of withdrawal as measured in preclinical models. We and others identified an anal. Intramural Research Program. are not likely to be involved in amygdala24.R. sal on relapse behavior. increases in the responsiveness of the central L. involved in learning and memory processes26 drugs. activation of the extracellular signal-regulated and thus treatment approaches that involve tent drug priming–induced reinstatement kinases (ERK) signaling pathway in the pharmacological reversal of cocaine-induced (see above). However. We also would like to acknowledge which in turn activates VTA dopamine neu. indicate that glutamate is involved in the Supported by the National Institute on Drug Abuse.28 and may be altered clinical situation of polydrug use. 1438 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . 691. assessing this issue of generality across drugs ings suggest that the neuroadaptations that We also explored whether the time. sensitized response to COMPETING INTERESTS STATEMENT neuroadaptations in VTA neurotransmis. Psychopharmacol. J. Most drug addicts are polydrug users1.L. whether cocaine-induced neuroadaptations and then determine the effect of this rever- shock stress–induced reinstatement was underlie the incubation of cocaine craving. These data mediate the incubation of cocaine craving24. underlying drug seeking induced by stress. in our view increases the levels of the stress neurohor. Intermittent footshock stress addiction and relapse27. (2005). it seems important to determine the unlike drug priming. cues. Abstr. H. which increases basal ciated with increasing peptide levels of the induced reinstatement7.P. time-dependent increases in cocaine seeking Furthermore. an in relapse to other drugs. cues or drug priming are not identical16. and that after 30 d. it has been suggested above support the notion that drug-induced is evidence that the neuronal mechanisms that craving induced by cocaine cues increases neuroadaptations are involved in drug relapse. sucrose-trained rats that drug-induced brain neuroadaptations drug and stressors11. it is too early to On the basis of studies of cross. However. to determine whether specific © 2005 Nature Publishing Group http://www. On the basis of the aforementioned adaptation specific to stress-induced rein. block stress-induced but not drug priming– the incubation of craving that also occurs with 3. Am. Gilbert. However. and blockade of CRF receptors in 1 d of withdrawal. Psychiatry 162. This experimental not affected by systemic injections of an We found that the time-dependent increase approach so far has been applied only to antagonist of the mGluR5 metabotropic in cocaine seeking after withdrawal is asso. reinstatement of drug seeking4. Alternatively. 343–359 (2001). Neuroadaptations and incubation of craving Conclusions administered drug. Our data also ACKNOWLEDGMENTS in cocaine-experienced but not in cocaine. this brain area attenuates stress-induced rein.9. There relapse to cocaine use. VTA increase cocaine seeking during early conclude from the available preclinical data sensitization of locomotor activity between withdrawal23. Psychopharmacology (Berl. pathway. although nucleus accumbens glutamate levels and plasticity-related growth factor BDNF in the there is evidence that specific cocaine-induced blocks cocaine-induced glutamate release VTA. neuroadaptations’ role in cocaine priming– glutamate receptor. and that injections of BDNF directly into the forms of cocaine relapse.29. Ashby & E. opiate. CRF receptor antagonists ness to cues may be a general mechanism for 2. (1996). would also influence medication develop- mediate enduring psychomotor sensitization. R. Exp. We subsequently explored aim to reverse endogenous neuroadaptations Neurosci. Lu for his contribution to the work described in the rons.nature. to cocaine seeking. way in mesolimbic dopamine areas25. including exposure to To account for the persistent propensity for The results from the studies described drug-associated cues1 and stress2. Xi. We concluded that time-dependent comments.21. dependent increase in cocaine seeking involves ment. Preliminary data suggest that induced by exposure to cocaine cues over neuroadaptations are causally involved in drug this might not be the case (Z. A recent study. the cell body ERK phosphorylation in the central. amygdala after 30 d but not after a history of stimulant. time-dependent increases in sucrose seeking humans. is important to verify whether manipulations periods19.or foot. C O M M E N TA R Y conditions other than reexposure to the self.X. the effect of footshock considerations. ongoing incubation process6. the first months of withdrawal (incubation relapse.14. such activation is known to induce amygdala ERK pathway to cocaine cues section on the role of neuroadaptation in incubation of cocaine craving. C. We thank E. nucleus accumbens and amygdala22. Clin. so it and remains high over extended drug-free this conclusion is applicable only to cocaine. it has been suggested demonstrated short-term (several weeks) underlie long-term relapse vulnerability in that neuroadaptations associated with psy. alcohol and system). we suggested that cocaine. Sinha. inhibition nicotine self-administration.) 158. however. of central amygdala ERK phosphorylation statement of cocaine seeking. Cue. neuroadaptations are involved in specific and cocaine-induced reinstatement. In this regard. in humans are similar for different drugs of stress-induced reinstatement10. but not that similarly contribute to relapse in rats with region of the mesolimbic dopamine reward basolateral. 4. specific to stress. 5–10 induced reinstatement of drug seeking17. of craving)20. Data from studies accumbens dopamine release17. exposure to stress or local infusion activation of the ERK pathway by cocaine for editorial assistance and K. cocaine cues may be due to cocaine-induced The authors declare that they have no competing sion are important in stress-induced relapse neuroadaptations of the central amygdala ERK financial interests. C. J. O’Brien. 1423–1431 roadaptation in CRF effects is likely to be increase in the amygdala ERK’s responsive. 2004). neuroadaptations7 are more likely to succeed stress-induced reinstatement of drug seek. non-drug rewards6. de Wit.16. Preston for helpful of CRF induces glutamate release in the VTA. Furthermore. Cocaine activates the ERK path. by chronic exposure to abused drugs8.or stress-induced rein. In the if neuroadaptation-related findings from ing. this pathway is studies using cocaine do generalize to other demonstration of cocaine-induced neuro.com/natureneuroscience statement. that were hypothesized to be involved in drug similarity in relapse rates across drugs and the statement18. Soc. the time-dependent 1. On the basis of these findings and other abuse30. may be the first amygdala and other brain areas. suggest that long-lasting cocaine-induced This time-dependent. the role of neuroadaptations in relapse to that reverse drug-induced neuroadaptations ogous phenomenon in laboratory animals: other abused drugs has not been assessed.

. 455–456 (1971). Psychol. & O’Brien.A.. 223–244 (1991). E. 3565–3571 (1998). F. Adams. 101–108 (2004). P. 135–163 (2002). 7. J. Brain Res. Nat.S. Neurosci. Pharmacol. J. 1–42 (2002). 212–219 (2005).) 1. 119–128 205–211 (2005). Clin. A. & Berridge. Lynch. J. Suto. Nat. Brain Res. 13.nature. L.. Brain Res. Lu. Bossert.W. E. Shalev.Z. Barnett. 1604–1611 (2004). 3–20 (2003). 24. Brain Res. Nestler.) 61–79 (2004). Pharmacol. Wise. Dempsey. J. 22.P. Gawin.J. Gen. Shaham. Abnorm. Brain Res. Arch.. Lorrain.. Pharmacol. 16. & Vanderschuren.. Rev. G. 798–805 30. S. 5. & Shaham.J. 247–291 (1993).J. Sun.. 12. K. Y. NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1439 . Nat. Grimm.W.W. & Branch. Mulder. H. J.. P.D. Psychol. Neisewander. et al. B. & Chao. J. Neurosci.. & Shaham. K. Wolf. 4707–4715 (1996).D.M. 197–234 (2004). Dempsey. M. Shalev..W. T. 54.M. Y. (2004).W. et al. N. Neuropharmacology 47 (Suppl. P. J. H.A. J. 18. Symp. B. Schoffelmeer. (2003). Neurosci. 50. Dackis. H. Nebr.. Neurosci.E. D. B. 28. Neurosci. Rev. J..A. Arnold.. 13–33 (2000). N. Rev.J.com/natureneuroscience Rev. J. & Psychopharmacology (Berl. 25.A. J.J. 14. Dickinson.. 26. T.T.. Kalivas. 27. 8. U. Y. Stewart. Curr. & Stewart. Erb. J.W. & Sweatt. 4. Y. 42. 33. 5389–5396 (2005). 6. Eur. Everitt. M. 16.. 25. A.H. Grimm. Opin. & Kleber. (1988). Neuropsychopharmacology 30. & Nestler. K. 29. Austin. (2000). Brain Res. Shaham. Rev. Mangiavacchi. 6. 10. 214–226 (2004). Annu. 8. C. 107–113 (1986). Neurosci. 17. X. 11. 742–747 (2003). J.) 168. W. 118–132 10. Kalivas. Neurosci. Rev. 21.Y. & Stewart. 129–138 (2001). Y. R. Toxicol. Vezina.M. Lu. 921–949 (1996). L. R. L.. 22. et al. Binnekade. Baker.. 16. Hope. Motiv. J. Hunt. D. 20.. A.. 23.M.. Liu. S. Lu. Psychopharmacology (Berl. Y.. J. J. J. De Wit. Lu. Addiction 91.T. Neuropharmacology 47 (Suppl. L. U. Berhow. 18.E. 19.. 36. C. J. 4654–4662 (2002). J. & Robbins. 20.P.G. Hiroi.. Brain Res.L.) 176. Psychiatry 43. et al. & Stewart. 743–749 23. Wang.H.. Lu.G. Grimm.M. J. Neurosci. Robinson. Kampman. 27. J. Neurosci.. et al. L. T.W. Brain Res. Rev.N. Grimm. 23–29 15. L. White. & Shaham. J.D. Neurosci. & Shaham. N. De Vries. 4. L. L.C. © 2005 Nature Publishing Group http://www. S. S. (2001). 9. 2. 24. Pettinati. 97. J.W. J.

of these changes are antecedents to habitual tion of genetic disposition.edu products such as c-Fos. memories of related By designing experiments to investigate drug use because they have not been validated events and neuroplasticity induced by the drug-induced neuroplastic events that are in animal models of the acquisition of drug event1. mice with a deletion/induction of the relevant use and is characterized by changes in proteins cotherapeutic targets for treating addiction. Charleston.8. progress has been slow in identifying the ones that actually mediate addiction. in drug-induced neuroplasticity7. This hypothesis has led to a del. of drug seeking represents neuroplasticity whereas the induction of Homer1a is without tions that mediate addiction from those unre. Identifying changes that are specific to particular elements of the transition from initial to habitual to relapsing drug use may be a fruitful strategy for pinpointing which forms of drug-induced plasticity are critical for addiction. Adaptive behavioral responses to important eled in animals and is paralleled by a transition and Narp. NAC-1 of drug-induced plasticity. it is preferable to locally stimulate or in this category are transcription factors such tic events results from the sheer enormity of inhibit protein expression using in vivo DNA. (siRNA) or Tat-fusion proteins. small interfering RNA ∆FosB is a well-characterized neuroplastic event work for integrating animal models of addic. use. However. Unfortunately. induction of immediate early gene (IEG) ral category is not yet validated in animal models e-mail: kalivasp@musc. Medical University of South models of synaptic plasticity. The induction of However. outlined above show that ∆FosB is necessary gative efficiency. South Carolina. protocols. the acquisition of drug self. cific neuroplastic changes in the three stages The second temporal category of drug- not only for understanding the neurobiology of of addiction can be validated directly using induced neuroplasticity mediates habitual drug addiction but also for identifying new pharma. istration are generally short-lived. and many of tion with the discovery and interpretation of shown below of how temporal categorizing can the gene deletion and transfection techniques drug-induced neuroplasticity hampers investi. N E U R O B I O LO G Y O F A D D I C T I O N C O M M E N TA RY How do we determine which drug-induced © 2005 Nature Publishing Group http://www. our inefficiency in recognizing the plasticity is often selective for a particular brain then disappear during withdrawal. and conditioned place preference8. These processes use. it does not efficiently distinguish neuroadapta. patterns of daily self-administration ioral plasticity.nature.11. the Neurosciences. from the changes required for habitual drug a form of enduring behavioral plasticity in ing drug-related memories3–6. the absence of a procedural frame. Although this research has been use. However. nucleus. habitual drug use. lant administration produces a progressive pathological behaviors such as habitual drug The following three phases of neuroplasticity increase in locomotor activity3. induced by psychostimulant administration of drugs serve as important environmental ticity related to the acute effects of the drug have been linked to behavioral sensitization. This less- use and the overwhelming drive to get drugs can be inferred using drug self-administration than-optimal model of addiction-related (as in relapse). A fundamental hypothesis in addiction specific to a particular part of this temporal self-administration. neuroplasticity has revealed that rather than uge of experiments identifying cellular changes administration represents acute neuroplasticity. Homer1a. the reinstatement of cocaine-induced locomotor sensitization9. the induction of NAC-1 in the addiction. viral transfection. induced neuroplasticity in addiction. cues for inducing neuroplasticity and creat. measurable consequence10. a few IEGs research is that the pharmacological properties sequence. This distinction is critical. it is unclear which environmental events result from a combina. including the relevance of most cellular changes in this tempo- Carolina. as most 1440 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE .com/natureneuroscience neuroplastic changes are important? Peter W Kalivas Although many drug-induced neural changes are known. which repeated noncontingent psychostimu- are thought to mediate the development of the ing vulnerability to relapse. The involvement of spe. USA. investigators can distinguish plas. Importantly. RNA or protein transfer techniques such as and glutamate transmission. for the development of behavioral sensitization sition from initial to habitual to relapsing drug The cellular responses to acute drug admin. Included most important addiction-related neuroplas. Similarly. from the permanent adaptations mediat. as ∆FosB and proteins regulating dopamine deciphering the biology of neuroplasticity. Kalivas is in the Department of are similar to plasticity identified in in vitro induced behavioral plasticity12. Each stage of this transition can be mod. Examples are in terms of relevance to addiction. First. Addiction develops as a tran. underlying relapse. promoting psychostimulant-induced behav- that may cause neuroplastic events underlying Second. because drug-induced neuro. represent neuroplasticity associated with the nucleus accumbens inhibits the development ful in cataloguing drug-induced neuroplasticity. gene. be used to determine the relevance of drug. that emerge with repeated drug exposure and In part. are closely increases in GluR1 in the ventral tegmental area tied to the molecular site of drug action and are necessary for the development of cocaine- Peter W. lated to addiction. However.2. Third.

it reinstatement of cocaine seeking contrasts COMPETING INTERESTS STATEMENT is also unknown if many drug-induced adap.19. & Seamans. R.J. Robinson. neuroadaptations in dopamine transmission used to target specific temporal stages of addic. Carlezon. K.. & Berridge. Ahmed. & Self. without objective measures of model of relapse has led to identification of 4. J. not necessary for the reinstatement of cocaine addiction and. This linkage is a necessary inter. dopamine transmission in the expression of This lack of information is a critical shortcom. C. 625–626 (2002).. tion to the accumbens are critical in the rein. 3. G. disappearance. Keitz. Neurosci. Natl. For Over the next decade. Baptista. Nat.. E. the validation between drug. 1208– in animal models endures for weeks of drug of the neurobiology of addiction. M. Plast. that more accurately model addiction than do tocol. Physiol. Wise. Nat. W.npp. (2005). 27. 6.16. drug-induced plasticity with the reinstatement 10. 6.F. Jr. © 2005 Nature Publishing Group http://www. 647–650 (2005). N. with the critical role identified for enhanced The author declares that he has no competing financial interests. Kelley. A. N. Mackler. Proc. Neurobiol. 24. Nestler.nature. 1215 (2003). 20–25 distinct behaviors are emerging or the behaviors the cysteine-glutamate exchanger and gluta. P.H. Neurosci.1038/ ity that has been most successfully evaluated mediary toward an integrated understanding sj. L. tations are elicited by self-administered drugs. in this way. J. 83. Canales. This approach will delineate an integrated 1403–1413 (2005). example.. systematic comparison of the cellular plasticity ogy of neuronal networks have been another Sci. USA 98. and animal models should be 16. et al. Science 305. L. Opin. J. D. Neurosci. 610–615 (2002). & Everitt. Neurosci.A. & Volkow.A.A. Curr. Koob. 17. & Weiss. Barrot. 12.J. Kauer.C. A molecular neuroplasticity with the physiol. In contrast. more common limited access protocols13. 6. 11042–11046 (2001). 2. et al. Nat Rev. many new addiction. 66. 212–219 (2005). Trends Neurosci. 8.20. Psychol. protocols could clarify this issue. behavioral sensitization3.J. & Bonci. 5.com/natureneuroscience cellular plasticity. Rev. in the nucleus accumbens that are critical for tion.E. statement of drug-seeking3.A. 25. plasticity in Erk signaling in the related neuroplastic changes will be discovered. Neurosci.L. S. Neuron 44. be identified and catalogued according to the 14.16. 25–53 (2003). D. Attempts to link 7. 5. Martin-Fardon. induced by the limited. Lu. model15. Neuron 45. Kalivas. 13. Rev.A. J. Neural self-administration induce behavioral changes in the self-administration/reinstatement pro. A. Am.E. J. Annu. 8. Kenny. Neurosci.W. et al. Neurosci. 1017– 1019 (2004).J.. Rev. identify the neuro. abstinence and may underlie relapse. 19. C O M M E N TA R Y C O M M E N TA R Y studies use investigator-administered drugs. Szumlinski. seeking17. Volkow. 6210–6217 (2000). (2004). & Leenders. 20.J. K. maintaining cocaine self-administration7 are sequence of neuroplastic changes underlying 18. P. Koob. E.. B.A. Mem.14. highlighting the ing for validating claims that extended periods of importance of validating neuroadaptations 1. Jones. Biobehav. amygdala and proteins regulating glutamate These drug-induced neuroadaptations should Nat.J. K. S. E. 739–749 higher daily self-administered drug intake. Unfortunately. et al. Baker. transmission in the prefrontal cortex projec.1300890). (2005). 5. it remains unclear whether new pharmacotherapeutic targets such as 5. 11. Pharmacol. are dose-dependent extensions resulting from mate receptor subtypes18.W. A.D. Moreover. Vanderschuren. Kalivas. S.A. & Markou. 483–494 (2004). induced neuroplasticity and the reinstatement 54. 20. R. 93–103 of dopamine in the nucleus accumbens in the therapeutic targets for treating addiction. temporal characteristics of their appearance and 15.versus extended-access important outcome of integrating enduring 9.K. Martin-Soelch. C. 4723–4727 (2004). 10. M. Neuropsychopharmacology The form of drug-induced neuroplastic. Learn. & Nestler. 121–128 (2003). 447–475 (2004). McClung. et al. Psychiatry 162. & Nestler.J. published online 14 September 2005 (doi:10. Acad.W. The relatively reduced importance plastic events that can best serve as pharmaco. 161–179 (2004). Annu.F. G. P.15. T.. M. J. Nat. F. Neurosci. 743–749 (2003). NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1441 .

ders experience an increasing sense of tension to the pathological state known as addiction. Neurotransmitter/ impulsive stage. the behavioral repertoire to drug seeking. neural circuits mediating two different moti. and the transition to drug addiction. France. neuropeptide Y and glucocorticoids which removal of the aversive state increases the validity (resembles the human condition) and of the hypothalamic-pituitary-adrenal (HPA) probability of the response. in and of itself. tion. the lateral hypothalamus. then relief from the aversive states. the acute reinforcing effects of drugs of abuse act. studies of or arousal before committing an impulsive Our thesis is that addiction involves a long. withdrawal/negative a general feature of biological systems3. Important anatomical circuits include reproach or guilt after the act1. Institut and is currently under test for construct valid. Different as naltrexone for excessive drinking). 2). ing neuroadaptive changes in the reward circuits National de la Santé et de la Recherche Médicale. drug-taking behavior progresses from systems in place to limit reward (see footnote the bed nucleus of the stria terminalis and a impulsivity to compulsivity in a three-stage in ref. based on the hypothesis that there are brain comprises the central nucleus of the amygdala. In addic. For the binge-intoxication stage. nucleus accumbens and a major projection to affect and preoccupation/anticipation2. an ‘opponent process’ concept that is transition area in the medial (shell) part of the cycle: binge/intoxication. recruitment of anti-reward systems that drive projecting to the nucleus accumbens and the pulsive repetitive behavior.nature. theoretical positions favor models from each of a progression from impulsive to compulsive imposed on the stages of the addiction cycle2. self. has construct validity for drug intoxication and for GABA and opioid peptides in the acute rein- elements of drug binges in humans. for opioid peptide receptors work). psychostimulants. and Michel Le Moal is at the Laboratoire de stress-induced reinstatement has face validity Such studies provide a framework for examin- Physiopathologie des Comportements. ending in chronic. In the From a neurobiological perspective. of the transition to dependence have the most taking. In contrast. in models have been developed that have face glutamate. addiction. drive transitions to negative reinforcement. gratification or relief at the time term. California. Patients with impulse control disor.com/natureneuroscience the ‘dark side’ of drug addiction George F Koob & Michel Le Moal Drug seeking is associated with activation of reward neural circuitry. La Jolla. and framework) and neurobiology (counteradap. Université Victor Segalen Bordeaux 2. As individuals move to the compulsive stage. 92037. and they Unite 588. Different theoreti. There is strong evidence for a role of cal perspectives from experimental psychology power) for all three stages of the addiction cycle dopamine in the acute reinforcing actions of (positive and negative reinforcement frame. relapsing drug These stages are thought to feed into each other. The concept of anti-reward is extended amygdala. vational systems: decreased function of brain the mesocorticolimbic dopamine system patients with compulsive disorders experience reward systems driven by natural rewards. USA. although in our view. γ-aminobutyric acid (GABA). substrates for each of the stages have already medications to treat excessive drug intake (such e-mail: gkoob@scripps. narrowing the these neuroanatomical sites include dopamine. strates.edu been identified using these models4. neural also provide a valid model for development of Bordeaux. social psychology (self-regulation failure administration of drugs (intravenous and oral) in the acute reinforcing effects of opioids. The extended amygdala stress by performing the behavior1. Drug addiction has been conceptualized as tive and sensitization framework) can be super. the drive for the drug-taking sion through the three stages of the addiction neuromodulator systems implicated in the behavior is positive reinforcement. and originating in the ventral tegmental area and anxiety and stress before committing a com. the study George F. models behavior. Koob is in the Molecular and Integrative avoid a location previously paired with an aver. progres. Here we argue that drug addiction also involves a ‘dark side’—a decrease in the function of normal reward-related neurocircuitry and persistent recruitment of anti-reward systems. N E U R O B I O LO G Y O F A D D I C T I O N C O M M E N TA RY Plasticity of reward neurocircuitry and © 2005 Nature Publishing Group http://www. forcing actions of alcohol. Understanding the neuroplasticity of the dark side of this circuitry is the key to understanding vulnerability to addiction. Although acute drug ulation and place conditioning (learning to use is not. drives compulsive drug taking. 1442 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . with the development of addiction. ity. Cue-induced or has had heuristic value in two major domains. the three stages. tive to the human condition is needed. Self-stim. in which cycle induces plasticity in neural circuitry that acute reinforcing effects of drugs of abuse in stimuli increase the probability of the response. some construct validity (possesses explanatory axis5. Although more construct validation rela. persistent plasticity in the activity of per se have identified key neurobiological sub- of committing the act. and then regret. The Scripps sive stimulus or state) are sensitive measures the acute reinforcing effects of drugs of abuse Research Institute. Animal opioid peptides. Acute self. of the neuropharmacological mechanisms for Neurosciences Department. becoming more intense and ultimately leading heuristic value for the human condition. pleasure. of ‘motivational’ withdrawal.

malaise. which ultimately leads thresholds during withdrawal from all major manifested by motivational signs of withdrawal to the pathological state of addiction. Recently. We believe process of attempting to maintain apparent as critical for the acute reinforcing effects of that research in this domain has been largely reward function stability by changes in reward drugs of abuse. do protracted abstinence. Significant plasticity occurs in in humans is produced in part by recruitment specifically. The observation reinstatement of responding for saccharin—or and acute stressors acquire even more power that CRF receptor antagonists in the amyg. drawal. glucocorticoids and norepinephrine. the motivational aspects of drug-seek- in the plasticity of the extended amygdala in that constitute the dark side of drug addiction ing and genetic vulnerability to addiction the development of addiction. Both animal and human neuro. there is evidence of nucleus accumbens. sug.and cue-induced drawal from alcohol17. in which addicts return to increased CRF function and is hypothesized dopamine system as measured by electrophysi. CRF.nature. cortex system7. mainly because of an and anti-reward system neurocircuitry5. We further hypoth. Similar observations with the animal studies. system associated with the development of drug19. We hypoth. to date involves activation of the CRF-HPA releasing factor (CRF).20 and the extended amygdala. These neurotransmitter/neuro. there A critical problem in drug addiction is in the anti-reward system is highlighted by is decreased activity of the mesocorticolimbic chronic relapse. producing aversive reinstatement neurocircuits now driven by a oped. of the aversive emotional state that drives the stability of the reward system through change. sition to addiction. In contrast. medial prefrontal. glutamate and neuropeptide Y lined above. as gesting that there is reactivation of both reward condition. and maintain addiction2. residual dysregulation of the HPA axis16 and of of addicts during withdrawal or protracted limbic/cingulate) and the basolateral amygdala the brain CRF system weeks after acute with- abstinence give results that are consistent are key mediators of drug. pre. GABA. In this context. to be particularly slow to return to homeosta- ological recordings and in vivo microdialysis. side potentially more important for driving peptide. These models have strong face valid- tion and recruitment of anti-reward systems motivational aspects of withdrawal is a major ity for the Diagnostic and Statistical Manual of NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1443 . hypoactivity of the orbitofrontal-infralimbic text-induced craving again include dopa.11. The most prominent functional dependence and withdrawal develop. dependence than decreases in natural reward in elements of the extended amygdala and/or imaging studies show that the prefrontal cortex function. similar changes in brain excessive focus on psychostimulant drugs and Neuroplasticity in the natural reward sys- reward threshold occur that temporally precede reward pathways (largely misattributed to the tem is highlighted by decreased dopaminergic and highly correlate with escalation in drug mesolimbic dopamine system). implicated in stress-induced relapse include only serves in the long-term to dysregulate the tioning of the HPA axis8. high-incentive stimulus such as a reward state not only produces the motiva- are activated during the development of exces. system further. Most extended amygdala system. A reward allostasis model is proposed drug despite adverse consequences19. Although the relapse models have face validity. In animal models of the tran. source of potential allostatic changes that drive withdrawal/negative affect stage can be defined forcement that defines compulsive drug-seek. Neurotransmitter withdrawal-induced decreases in dopaminer- reflect hypodopaminergic functioning) and systems implicated in drug. dysphoria. a control rarely explored tional symptoms of acute withdrawal and sive drug taking. but other neuro- extracellular CRF in the extended amygdala reinstatement studies to date have been done adaptive processes associated with behavioral is increased during acute withdrawal from with nondependent animals and may be of responses to stressors also may have potential drugs of abuse.10. The development stasis’ is defined as the process of achieving drawal in humans: chronic irritability. emo. These neurotransmitter systems a nondrug. This corresponds to the preoccupation/ sis. futilely in the short term. This makes the system that drives the dark and there is also decreased activity in opioid anticipation stage of the addiction cycle. norepinephrine there remain serious concerns about construct axis and subsequent activation of the CRF– and dynorphin are recruited. ‘allo- as the presence of motivational signs of with. attempts to misregulate these drug-induced modulator system changes may persist during Neurotransmitter/neuromodulator systems neuroplasticities by taking more drug. value—parallel those of the neural substrates Clearly appropriate. allostasis is the © 2005 Nature Publishing Group http://www. Neuroplasticity intake6. the provides a powerful source of negative rein. It is termed the ‘dark side’ of addiction. which protracted abstinence and include hypofunc. alexithymia negative reinforcement of addiction is here An allostatic state is a state of chronic deviation and loss of motivation for natural rewards. in drug addiction. dynorphin and antagonists block excessive drug taking during than studies of reinstatement of responding for norepinephrine. out. which include escalated drug intake or stress-like states8. drug cues and protracted abstinence).com/natureneuroscience the neurotransmitter circuits identified above of the brain anti-reward systems. the drug addict. saccharin solution. mine. We further of the reward system from its normal (homeo- characterized in animals by increases in reward hypothesize that this chronic aversive state static) operating level. During such acute withdrawal. For example. In other words. construct-validated ated with the development of dependence for nondependent doses of cocaine or heroin? animal models for the stages of the addiction provides a compelling example of a key player We also hypothesize that the dysregulations cycle. in dopamine D2 receptors (hypothesized to and relapse in humans. the combination of to explain how dysregulation of the reward a narrowing of the behavioral repertoire for decreases in reward neurotransmitter func. any other nondrug reinforcer of high incentive to elicit drug-seeking behavior. dala can block excessive drug intake associ. For example. Thus. glutamate and GABA. esize that anti-reward circuits are recruited as the tone for vulnerability to ‘craving’ by acti. validity relative to the human condition. brain and anti-reward systems during relapse12–14. C O M M E N TA R Y For the purposes of this opinion piece. persist during protracted abstinence to set are critical to test these hypotheses. animal models with excellent face validity for between-system neuroadaptations9 during the vation of the drug-. ing behavior and addiction.and stress-induced the transition to addiction have been devel- development of addiction. and this activation is manifest in reinstatement studies. neglected by the field. Human imaging studies system (orbitofrontal. responding for esize that within the motivational circuits of system15. CRF receptor little more relevance to the study of addiction roles such as neuropeptide Y. including decreases reinstatement in animal models and craving are made in human addicts18. increase of the anti-reward system identified anti-reward systems such as corticotropin.15.20. opioid peptides. More drugs of abuse. activity and hypofrontality. but also provides the when the drug in removed (acute withdrawal the neuropharmacological substrates for the background by which drug priming. reorganized and hypofunctioning prefrontal driven by dependence6. leading to a worsening of the More importantly for the present thesis. compulsive drug taking long after acute with. The allostatic dysregulated dependence8. and critically.and cue. cue. gic function are relatively transient10. tional pain.or con.

Rasmussen. standing the neuroplasticity of the dark side of 111. Volkow. under.E. Coyle.J. 5. 21.E. tion. tems (mediating natural rewards) and increased 206–225 (Williams and Wilkins. G. E. & Le Moal. Brain Res.. G. Fowler. vulnerability to addiction.J. G. R. American Psychiatric Association. Koob. 294–307 (2003). & Wang.. Classification of Diseases and Related Health Problems. & Zhu. A.V.F.D. Classification of Diseases21 criteria for addic. Clin. J. We thank M. Rev. Kreek. Diagnostic and 14. & Everitt. Neurosci.. & Le Moal. Bloom. 51. & Bloom.. financial interests. preparation of this manuscript. & 23–47 (1998). J.E. that provides the core element of the motivation 5. not a hyperactive or sensitized 6. Exp. C O M M E N TA R Y Mental Disorders1 and International Statistical ACKNOWLEDGMENTS 9.com/natureneuroscience 1994). a long-term persistent decrease in function of (American Psychiatric Press. reward system)22. B.) 283–294 (Raven.) 1381–1397 (Lippincott Williams and 19. Ledford. Wikler. Belin. 232–243 22. Alcohol. See. K. J. Valdez.F. N. M. Res. New York. Rev.. & Stewart. 13. in our view. et al. Koob. & Kupfer. Science mal reward. Ann. Druhan. Drug Alcohol Depend.F. validity and show promise for measuring the NY Acad. Shippenberg. World Health Organization. 1992). Shaham. produced by excessive drug taking.. Alcohol.L. 1968).A. Science 305. & Markou.. G. 81. 715–723 (1988). Le Moal.J. 36. 1444–1451 (2003). Annu. Alcohol. Vanderschuren. 1017– 24. A. 486–498 (1999). this circuitry will be the key to understanding 8. & McLaughlin. C. Geneva. M. Kenny. M. Baltimore. 13–33 (2000). 877. R. M. J. J. Koob. Sci.D.F. 17. D. Charney. a perspective often over.F..T. & Piazza. 625–626 (2002). In our view. Piazza. G. 119–128 (2001).. Koob. Deroche-Gamonet. Washington. Nestler.J. C. 1995). Y. G. V. Pharmacol. Solomon.S. looked in the drug abuse field is that there is Statistical Manual of Mental Disorders 4th edn. 1494– anti-reward systems (recruited as an opponent 4. Nat. Davis. 2002). & Le Moal. G. M. 26. 2. J. Y. Neurosci. et al.J. 16. International Statistical reward state for drugs per se. Clin. 1444 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . Clin.C. F. Rev. Aston-Jones. Nemeroff. L. sources: decreased function of brain reward sys. Rev. Neuropsychopharmacology 20.D. D. Res. 15.C. Erb.. Ann.S. G. in Psychopharmacology: The Fourth © 2005 Nature Publishing Group http://www. P.. P. D. It is the deficit state for nor. 27. S. 1. Martin. 24. F. (2003). Res. normal motivational systems driven by two 2. Delfs. Progress (eds. (1997).J. 33. & Koob. 3. J. 97–129 (2001). are currently under test for construct 11. 52–58 D. Brain Res. in 1501 (2002). Ahmed. Wilkins. to seek drugs. D. G. Arends for his assistance with the 10.L. & Koob. Invest.F. Philadelphia. & Corbit. P.) 1836–1849 (2000). 10th revision (World Health Organization. T. W. Fuchs.M. Exp. NY Acad. 359–378 (1996). J. Nat. 305. 7..nature. Generation of Progress (eds. process to excessive activation of the brain Neuropsychopharmacology: The Fifth Generation of 18. 1019 (2004). Psychol. R. Exp.R. 985.F.R.J. Koob. Science 242. 119– individual vulnerability to addiction.V.H. Sci. S. Clin. Science 278. G. in The Addictive States (ed. genetic and environmental contributions to The authors declare that they have no competing Toxicol. 145 (1974). 1014–1017 (2004). Thus. COMPETING INTERESTS STATEMENT 12.

more heterogeneous.1038/nn1578 part. 1 determine whether such common underpinnings of addiction legend) seem to activate the brain’s endogenous opioid and cannabi- can be exploited for the development of more effective noid systems within the VTA-NAc pathway. Texas 75390-9070. such as food.2. However. although called ‘extended amygdala’). These same regions have also been implicated circuitry in the brain’s limbic system1–5. that important aspects these shared rewarding and addicting actions of drugs of abuse. and Neuroscience. The question addressed by this Perspective is are part of the brain’s traditional memory systems. On the basis of these common acute actions. that all drugs of abuse converge on a common social interactions14. this has led to the whether there are common neural and molecular pathways underlying notion.5. regions of frontal cortex.com/natureneuroscience Eric J Nestler Drugs of abuse have very different acute mechanisms of acute rewarding effects of all drug of abuse. regardless of its distinct producing a series of common functional effects after both mechanism of action. circuit-level actions1–8. to addiction—a loss of locomotor activating and rewarding effects9. Indeed.11–13. examples of effects upon acute administration. can induce cross- drugs are all acutely rewarding. Dallas.15. N E U R O B I O LO G Y O F A D D I C T I O N PERSPECTIVE Is there a common molecular pathway for addiction? © 2005 Nature Publishing Group http://www. converges on the VTA and NAc with common acute and chronic administration. These reward and chronic changes in reward associated with addiction. Growing evidence indicates that the VTA-NAc pathway and the other Common actions on brain reward circuits limbic regions cited above similarly mediate. all drugs of abuse mechanisms of addiction are the observations that certain drugs of cause certain common effects after both acute and chronic exposure. which promotes repeated drug intake tolerance and cross-sensitization to one another with respect to their and leads eventually. The University of Texas Southwestern Medical Center. the clinical syndromes of the natural addictions are likely to be much 5323 Harry Hines Blvd. from animal models and more positive emotional effects of natural rewards.4. a prolonged period of sensitization. which further underscores shared acute mechanisms of drug action1. Some of these areas siderable controversy.10. numer- ery and elicits a distinct combination of behavioral and physiological ous common chronic adaptations have been described. yet dopa- progress in understanding the molecular and cellular basis of mine-independent effects on the same NAc neurons.10–13. adaptations are thought to contribute to the intense drug craving and These regions include the amygdala (and related structures of the so- relapse that can persist even after long periods of abstinence. as exemplified by reduced treatments for a wide range of addictive disorders. Accordingly. by common neural and molecular mechanisms. sex and recently from humans. compulsive consumption to the mesolimbic dopamine pathway. Most attention has been given in the so-called ‘natural addictions’ (that is. at least in Published online 26 October 2005. under particular experimental conditions. Nevertheless. Each drug activates dopaminergic occur for natural rewards as well. In addition. drug effects in cannabinoid and opioid receptor knockout mice.edu possibility that the similar behavioral pathology that characterizes drug addictions and certain natural addictions may be mediated. doi:10. early findings in the field raise the e-mail: eric. chronic functional changes in the VTA-NAc pathway. All drugs also produce similar negative emotional More recent work has established that several additional brain areas symptoms upon drug withdrawal. of addiction involve powerful emotional memories2. among others1. NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1445 . hypothalamus and several the relative contribution of each mechanism remains a subject of con. A major goal for future research is to indirect. which includes dopaminergic of natural rewards) such as pathological overeating. the acute There is now considerable evidence. Some similar actions acute functional effects (Fig. now supported by increasing evidence. hippocampus. despite these disparate which are discussed in the next sections. Preliminary findings suggest that shared targets in the limbic forebrain. one would expect Drugs of abuse are highly diverse chemical substances. especially the nucleus accumbens (NAc). control over drug use. pathways may be involved: two examples are cross-sensitization that This VTA-NAc pathway is one of the most important substrates for the occurs between natural rewards and drugs of abuse16 and similar abnor- malities found in brain imaging scans in drug and natural addictions4. 1). each that chronic exposure to drugs of abuse would also cause common drug binds to its distinct initial protein target in the brain and periph.4. and research over the past action but converge on the brain’s reward pathways by several decades has delineated how each drug. Consistent with common mechanisms of action and pharmacological effects. in many cases via these common effects.. several drugs (see Fig. at least in part. in vulnerable individuals. it must be emphasized that the mechanisms underlying natu- ral addictions are much less well understood than those underlying drug Eric J. Researchers are making transmission in the NAc and many produce dopamine-like. However. pathological gam- neurons in the ventral tegmental area (VTA) of the midbrain and their bling and sexual addictions. that interact with the VTA and NAc are also essential for acute drug and associative learning toward drug-related environmental cues. Nestler is in the Department of Psychiatry and Center for Basic addictions because the animal models are far less straightforward. The abuse.nature. USA.8.nestler@utsouthwestern.

PPT/LDT. only a few illustrative examples are included (Fig. Alterations in GABAergic regulation of VTA 1446 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . Cannabinoid mechanisms seem complex. Opiates ing memory. These adaptations seem to be In contrast. Stimulants directly increase dopaminergic transmission in the NAc. are Gi linked) on glutamatergic changes in these frontal cortical regions and and GABAergic nerve terminals in the NAc. and they any of several drugs of abuse causes complex involve activation of CB1 receptors (which. ity (being driven by irresistible inner forces to take a drug) that characterizes a state of addic- tion4. Alcohol.nature. opiates. together the following scheme1.9. along with the extent to which they mediate common behav- dopaminergic transmission occurring in response to the drug in question ioral abnormalities associated with drug and natural addictions. PERSPECTIVE Nicotine + Alcohol the negative emotional symptoms as well as Opiates Opioid many of the somatic symptoms that occur Glutamate inputs peptides upon drug withdrawal. Finally.. routes glutamatergic transmission to the NAc parallel the changes reported in of administration and dosing regimens. Abrupt with. This sensitized state is mediated via central corticotropin releasing factor (CRF) systems. they also produce common adaptations in presumably underlies the hypofrontality noted in human brain scans. At the same time. drug-taking ceases and may relate to drug craving and relapse. it is possible to piece dopaminergic transmission to the NAc discussed earlier (Fig. These characterized for their involvement in fear and other aversive states. After chronic drug use. including work- they inhibit GABAergic interneurons in the VTA. Fig. tolerance. which presumably Nicotine DA involves adaptations within amygdala neu- rons or their inputs (see below).9. Phencyclidine (PCP) may their glutamatergic outputs. Nicotine seems to activate VTA dopamine neurons directly glutamatergic projections from these regions via stimulation of nicotinic cholinergic receptors on those neurons and indirectly via stimulation of its to the NAc and VTA. like D2 and opioid receptors. from cortex) to drug craving and relapse as well. alcohol or nicotine. 2). Ann Thomson Glutamate + Cannabinoids – Another common adaptation to chronic drug inputs (e.11. which are impli- act by inhibiting postsynaptic NMDA glutamate receptors in the NAc. classically intracellular trafficking of AMPA receptors in these neurons21. including cocaine. chronic drug able.g.23. the two mechanisms converge within some NAc neurons. and opioid receptors. which may occur drawal from virtually any drug of abuse leads to activation of CRF. These regions (for example. The chronic drug-treated state is Common circuit-level adaptations associated with reduced basal activity of cortical pyramidal neurons and Just as all drugs of abuse increase dopaminergic transmission to the NAc a reduced sensitivity of the neurons to activation by natural rewards.via induction of the GluR1 AMPA glutamate receptor subunit and altered containing neurons in the amygdala17.10. attention and behavioral inhibition also directly act on opioid receptors on NAc neurons. 2).3. the behavioral consequences of each of these adaptations remain exposure seems to sensitize the dopamine system. Opiates do the same indirectly: tex) control executive function. This after acute administration. It may similarly inhibit glutamatergic terminals that innervate NAc neurons.com/natureneuroscience + focusing on the molecular basis of this hyper- + DA functional CRF system. dopamine function after chronic exposure. These neurons. Impressive evidence from receptors on glutamatergic nerve terminals that innervate the dopamine cells. We now know that sensitized behavioral responses to drugs of abuse.increases in AMPA glutamate receptor responsiveness. despite diverse initial actions. as inferred from brain imaging studies4. adaptations in glutamatergic transmission have been related directly to innervate many forebrain and brainstem regions. and to drug-associated cues5. Current work is © 2005 Nature Publishing Group http://www. these neurons are hypersensitive to activation by drugs of complex in that different effects have been reported by numerous labora. and may contribute – (e.g. may contribute to the negative emotional symptoms observed between However. hence. baseline levels of dopamine tations at the cellular and molecular level in the VTA-NAc and other function are reduced. may inhibit GABAergic terminals in VTA and hence disinhibit VTA dopamine rodent models and from human brain imaging neurons. induce a long-term potentiation (LTP)-like Chronic drug states are also associated with common changes in state in VTA dopamine neurons19–22.This sensitization can last long after Numerous types of drugs of abuse. Many additional studies demonstrates that chronic exposure to mechanisms (not shown) are proposed for alcohol.. by promoting GABAA receptor function. partly because of differing drug doses. Figure 1 Highly simplified scheme of converging acute actions of drugs of abuse on the VTA-NAc. which can be viewed as a Common cellular and molecular adaptations homeostatic response to repeated drug activation of the system (in other Chronic exposure to drugs of abuse causes numerous common adap- words. produce some common effects on the VTA and NAc1–8. 2). Chronic exposure to any of several drugs of abuse causes an impaired dopamine system. from use is cortical ‘hypofrontality’: namely. in part via of the other drugs remain more conjectural. there is some evidence cated in the profound impulsivity (acting on that nicotine and alcohol may activate endogenous opioid pathways and that these and other drugs of abuse (such as opiates) may activate endogenous cannabinoid pathways (not shown). and normal rewarding stimuli may be less effective brain reward regions. There are too many adaptations to describe here at eliciting typical increases in dopaminergic transmission. CRF can Alcohol therefore be viewed as an example of ‘oppo- VTA Opiates nent process’–like changes that drugs induce GABA ? interneuron PCP – in the brain that serve to counteract drug Alcohol – effects and drive withdrawal symptoms when ? Stimulants the drug is discontinued1. amphetamine.5. These changes comprehensively. with greater increases in uncertain. sudden urges to take a drug) and compulsiv- peduncular pontine tegmentum/lateral dorsal tegmentum.18. like D2 dopamine (DA) and are important in controlling an individual’s receptors. The actions response to environmental stimuli. anterior cingulate cortex and orbitofrontal cor- Drugs of abuse. which disinhibits VTA dopamine neurons. whereas the occurrence of such shared adaptations is indisput- drug exposures or upon drug withdrawal. and on NAc neurons themselves. reduced amygdala baseline activity of several regions of frontal PPT/LDT) VTA NAc cortex.13. signal via Gi. These drug-induced changes in tories even for the same drug.abuse as well as drug-associated stimuli. although aspects of this activation of these neurons during drug withdrawal partly mediates model remain controversial23. Nevertheless. prefrontal cortex.

drug exposure NAc induces TH and increases AMPA glutamatergic VTA responses (Glut) via regulation of glutamate receptors. after chronic exposure. The NAc is biochemical. under normal conditions. The net effect of these adaptations on glutamatergic transmission remains uncertain. xc– Glut Several. such as high levels of wheel running biosynthesis. effects mediated in part by upregulation of AGS3 in cortical neurons and downregulation of the cystine-glutamate transporter (system xc–) in glia.36. After chronic drug administration (lower panel). The top panel (Control) shows a VTA neuron innervating an NAc neuron and glutamatergic inputs to the VTA and NAc neurons. but whether of the transcription factor ∆FosB. reduce the size of cell bod. decreased TH levels or and sucrose drinking. responses to cocaine and opiates as well as to sucrose and wheel-run- Reduced amounts of neurofilament proteins seen within the VTA ning. Overexpression of ∆FosB in NAc increases behavioral the behavioral plasticity associated with addiction27–29. Our hypothesis is that by these drugs25. when induction of the other family members shows © 2005 Nature Publishing Group http://www. which is also activated in unique stability at the protein level. decreases behavioral responses to cocaine. in the VTA25. amphetamine. There is also evidence that VTA dopamine neurons decrease in size. explain the disparity between the higher alcohol and nicotine reduce CREB phosphorylation (an indirect marker levels of TH seen in VTA and the lower levels seen in NAc. Stimulants and nicotine also induce dendritic outgrowth of NAc neurons. Conversely. transport from the VTA to the NAc observed after chronic opiates.41. nicotine. which may then AGS3 mediate some of the shared aspects of addiction Glia Addicted by regulation of numerous target genes. Induction of CREB activity and alterations in neurotrophic factor (NTF) signaling may partly mediate these Basal glut effects.26. Repeated exposure to cocaine.30. addictions by regulating a set of common target genes34–36. PERSPECTIVE dopamine neurons have also been implicated for opiates24.com/natureneuroscience (see above). several adaptations occur. TH Basal DA cAMP-CREB Several additional changes have been found for ∆ NTF Stimulated DA stimulant exposure. These changes in postsynaptic glutamate responses are associated with complex changes in glutamatergic innervation of the NAc. cannabinoids and cocaine. infusion of any of several neurotrophic factors into in numerous experimental systems. one can where)40–42. This amphetamine or opiates induces CREB activity in this region37–39. cocaine or alcohol exposure may be a biochemical blockade of ∆FosB function in the NAc by overexpression of a dominant marker of common morphological changes to VTA neurons induced negative antagonist causes the opposite effects. alcohol. all drugs of abuse induce the Stimulated glut transcription factor ∆FosB. the rate-limiting enzyme in dopamine consumption of natural rewards.39. In contrast to ∆FosB. including Chronic administration of any of several drugs of abuse.43. which. but not all. amphetamine. a Fos family protein. activation of CREB seems to produce similar behavioral effects: cells. cFos and other Fos family activity in VTA nerve terminals in the NAc are reported under certain members are induced in NAc after acute exposure to drugs or natu- experimental conditions. whereas latter finding could. One of the most dramatic examples is induction molecular mechanisms governing drug regulation of CREB in NAc remain Figure 2 Highly simplified scheme of some common. increased CREB activity in the NAc the VTA prevents these morphological changes and also produces sen. ies and the caliber of proximal processes of VTA dopamine neurons. including increased incentive drive for these rewards. also increases levels phencyclidine34. after chronic opiate. There is now considerable evidence VTA by several drugs of abuse after chronic administration. This latter adaptation could mediate the ral rewards. although opiates produce the opposite action. Recent evidence implicates the transcription factor CREB desensitization.35. opiates and alcohol. Interestingly. ∆FosB accumulates during chronic exposure owing to its (cAMP response element binding protein). In VTA. in mediating that ∆FosB accumulation within NAc neurons contributes to a state of drug induction of GluR1 and TH in this region27 as well as in some of sensitization34. It is also induced in this same region by chronic of tyrosine hydroxylase (TH). CREB is of endogenous neurotrophic factors exerts the opposite effects30–33. Stimulants decrease AMPA glutamatergic responses in NAc neurons. whereas ∆FosB accumulates in this region uniquely after reduction in dopaminergic signaling seen after chronic drug exposure chronic exposure. although Reduced neurofilament levels could also account for the impaired axonal it is not as universal as induction of ∆FosB. alcohol and nicotine. also induced in the NAc by natural rewards (such as sucrose) and simi- Most evidence for common chronic effects of drugs of abuse in the larly reduces an animal’s sensitivity to sucrose’s rewarding effects39.nature. in turn. opiates. In NAc.35. which accumu- similar changes occur with other drugs of abuse is not yet known. such changes have been documented only induction of ∆FosB mediates many shared aspects of drug and natural for opiates. To date. it is not yet known whether CREB they generalize to other drugs. lates in the NAc after chronic exposure to all drugs of abuse. Concomitantly. Activation of CREB is another common adaptation in the NAc. which may be Glut ∆FosB Ann Thomson mediated by upregulation of the cAMP pathway. possibly by regulation of glutamate receptors or postsynaptic density proteins (such as PSD95 and Homer-1). including cocaine. drugs of abuse induce CREB activity in this region. NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1447 . chronic actions of drugs of abuse on Glutamate Control inputs the VTA-NAc. Although of CREB activity) in the NAc (although they induce CREB activity else- the functional consequences of these changes are not known. whereas blockade decreased CREB activity increases such responses28. Despite this opposite regulation of CREB in the NAc by differ- speculate that they reflect a fundamental impairment of the dopamine ent drugs.44. whereas sitized behavioral responses to several drugs of abuse. opiates.

of addictive disorders. Drugs that target the brain’s dopamine.and postsyn- with D1 and D2 dopamine receptors exerting opposite effects: D2 receptor aptic elements as a function of chronic drug administration. This is probably This could occur in part via ∆FosB-mediated induction of GluR2 due in large part to the very different effects the drugs elicit outside the (which reduces AMPA receptor conductance and Ca2+ permeability) in brain’s reward circuitry.7. cAMP pathway activa. inhibition of the pathway facilitates drug responses. drugs of abuse. Changes in postsynaptic responses could also be mediated tives owing to their actions on opioid receptors. which are located in by altered AMPA receptor trafficking or by adaptations in the neurons’ brainstem and spinal cord. An interesting with certainty. One general strongest stimuli only.41.46. one can presume that the hypofrontality demon. which are Homer or increased levels of F-actin18. whereas D1 receptor activation favors activation by the one drug of abuse that are not seen with the others2.45. Hence. chronic administration of cocaine. These findings.com/natureneuroscience dopamine neurons and their nerve terminals44. nicotine and cannibinoids48. Implications for treatment of drug and natural addictions sion to the NAc when the cell bodies fire in response to cocaine and Because common mechanisms seem to contribute to at least some aspects associated cues13. Clearly. protein signaling-3). frontal cortical pyramidal neurons. dopamine has a profound effect on activity of prefrontal cortical neurons. as well as natural rewards. opiates. along with located in heart and vascular tissue.46. an effect reported to date for cocaine. This could explain the functional observations cited above that example. changes in synaptic transmission associated with the altered morphology. This is unfortunate. One possibility is that the different morphological effects characterized only for stimulants and suggest an interesting pathophys. which may be partly mediated via ∆FosB and some addiction process. For example. although this has not yet been demonstrated underlying molecular adaptations in amygdala neurons. with this view is that in most behavioral assays. This type of information is available for stimulants like cocaine and amphetamine Drug-specific actions in brain reward circuits but has not been examined sufficiently for the other drugs or natural The argument that acute and chronic drug effects are mediated by rewards. may still lead to a common behavioral endpoint owing to the more specific iology13. PERSPECTIVE unknown but could involve upregulation of the cAMP pathway observed tant findings now highlight the need to characterize the effects of other in this region after repeated exposure to any of several drugs of abuse.35. and such differences may help explain why profound dysfunction in cortical control over the NAc18. For example. opioid dampened. For example. 1448 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE .46. More work is needed to better understand in these cortical regions (Fig.13. which is a negative regulator of Gi and hence of D2 whereas more common adaptations may be expected more distally. as mentioned earlier.6. including reduced levels of PSD95 and system because of its actions on monoamine transporters. on these same endpoints. causes very similar strated in rodents and humans after chronic exposure to several drugs sensitization to the behavioral effects of the drugs. perhaps further exaggerating glutamatergic transmis. after chronic administration. To date. GABAA receptors and nicotinic cholinergic receptors. A major need for future investigations is to determine whether of abuse comes from morphological changes observed in NAc and pre- similar changes occur with other drugs of abuse or natural rewards. One target gene in opiates. which requires further analysis. Stimulants cause a long-term depression (LTD)-like state in common mechanisms is inconsistent with the knowledge that drugs NAc neurons and also reduce postsynaptic responses to glutamate21. The most striking example of different chronic cellular actions of drugs tion. and possibly to natural addictions as well. According to this scheme. chronic opiate administration causes opposite candidate is CREB. might this effect contribute to common abnormalities or natural rewards also elicits common changes in the electrophysi. alcohol or nicotine baseline activity of these neurons and their responses to natural rewards is is specifically associated with changes in dopamine transporters. amphetamine. in numerous brain regions4. 2). amphet- Fewer studies have searched for common drug-induced molecular amine or nicotine causes long-lasting increases in dendritic arborization changes in other brain areas (outside the VTA and NAc) associated with and in the density of dendritic spines of these neurons49. which decreases dopaminergic may be critical for the formation of new memories.47. these changes are complex and interact with of all drug addictions.18. mon consequence of chronic drug administration raises interesting ques- A major gap in our knowledge is whether chronic exposure to drugs tions. opi. whereas tate gyrus of adult hippocampus. ates and alcohol. These mor- addiction. This opposite action of opiates is surprising. glutamate. opiates and alcohol2. alcohol.38. respec- are enhanced. the finding that reduced neurogenesis is a com- © 2005 Nature Publishing Group http://www. could contribute to the state of sensitization seen of central CRF pathways upon precipitation of drug withdrawal17 reflects after stimulant exposure. this transporter alter dopaminergic transmission3 and induce ∆FosB34. although this remains tone and opposes reward via activation of κ opioid receptors on VTA unproven. Nevertheless. The function of adult hip- the NAc through which CREB may exert these effects on drug and natural pocampal neurogenesis is a subject of great debate: the new nerve cells rewards is the opioid peptide dynorphin. Chronic exposure to cocaine. administration. opiates are analgesics and seda- the NAc34.35 in the NAc. In any event. including cocaine.nature. Induction of AGS3 is accompanied by decreased levels of tively. turn relate to the impulsive and compulsive features of stimulant addic.9. promotes release of glutamate from prefrontal cortical glutamatergic nerve terminals. Nevertheless. whereas responses to cocaine and cocaine-associated stimuli receptors. much more detailed examination is needed of pre. these impor. it might postsynaptic adaptations in glutamate receptor function in NAc neurons be possible to develop treatments that would be effective for a wide range in ways that remain incompletely understood. For signaling. such adaptations have been this paradox.18. like chronic stimulant administration. because it is induced in amygdala by stimulants. even though all of the drugs the cystine-glutamate transporter in glial cells in the NAc. changes in the dendritic arbor of NAc and prefrontal cortical neurons49. as these other regions are also key to the phological changes. of abuse can readily be distinguished from one another. In contrast. and the CRF gene is regulated by CREB37. one can presume that the hyperactivity of its target genes35. and even cross-sensi- of abuse is mediated via common molecular and cellular adaptations tization. Consistent We know that several drugs of abuse. there are likely to the evidence of abnormal glutamatergic innervation of the NAc from be very important differences in the effects of each drug of abuse on frontal cortical regions (discussed in the next section) would suggest a reward circuitry as well. in memory or other cognitive functions seen in many addicts? ological properties of NAc neurons or of their synaptic inputs. There are also large numbers of molecular adaptations reported for ronmental stimuli. given that chronic opiate As another example. which could in most addicts have their preferred drug of abuse. activation tends to promote the neurons’ responsiveness to diverse envi.46. Chronic stimulant administration seems to cause principle is that such drug-specific adaptations may increase in likelihood a shift toward the D1 state in part by induction of AGS3 (activator of G with increasing proximity of a protein to the immediate drug target. whereas cocaine activates the cardiovascular postsynaptic densities (PSDs). converge to reduce neurogenesis—birth of new neurons—in the den- tion also causes reduced behavioral responses to drugs of abuse.

Robinson. Lu. Am. T. 358. are critically important to help us understand cocaine in heterozygous brain-derived neurotrophic factor (BDNF) knockout mice. 17–20 (2003).W. C. 233–250 (2003). & Diamond. J. B. Adult neurogenesis: implications for psychiatry. CAMP-dependent protein kinase type 5. Sci. A diet promoting sugar dependency causes behavioral cross- sensitization to a low dose of amphetamine. 3312–3320 (2002). Walters. 11. et al.E. C. & Berridge. et al. no treatment aimed at common drug mechanisms has yet been Med. responses to emotional stimuli.J. 483–494 I regulates ethanol-induced cAMP response element-mediated gene expression via activa- (2004). Bonci.. R. Rev. Mol. 1208–1215 (2003). tion of CREB-binding protein and inhibition of MAPK.M. A. In the more arousal and affect regulation. 138. 22. and behavioral plasticity. X. 69–83 (2004)..A. Synapse 48.J. D. target genes. Everitt. opiate addiction.. C. Acamprosate. 12. Neuron 41.D. Rev. Proc. M. & Zieglgansberger.Z. M. Neurosci. Psychol. Grimm. W. 557–569 (2004). 16.E. Li. et al. 427–440 (2004). NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1449 . 48. Preparation of this review was supported by the National Institute on Drug Abuse. Littleton. as a unitary disorder. & Bonci. 35. Nat. Neurochem.C. Hall. 3306–3311 (2002). Sci. (2003).J. Nat.N. & Nestler. Pharmacol.. Sci. Neuron 37. Lond. 23–29 (2002). Neurosci. Pandey.A. Rev.nature. E. The role of neurotrophic factors in psychostimulant-induced fully validated across of range of addictions to multiple drugs of abuse and behavioral and neuronal plasticity. & Self. 9. N.J. 44. Godfrey. B. Such 33. action of drugs of abuse and natural rewards raises serious red flags about 22.J. In vivo nicotine treatment regulates meso- Neurosci. P.. et al. Koob.. 24. perspectives on alcohol actions. NY Acad. Involvement of cAMP-dependent protein kinase in the nucleus accumbens J.J. whereas nicotine patches and chewing gum are effective 27. 1032. J.W. 21. 12. W. 28.146–154 (2004). Jr. 41. K.A. 5. T. distant future. Sci.J. Dopamine. an agonist. buprenorphine and naltrexone. Prog. II et al. L. respectively. Brain Res.J. tion. Regulation of drug reward by CREB: Evidence for two functionally distinct for nicotine addiction.nature.M. R.. Addict. 1581–1602 (2004)..W. Fiorillo. Annu. Science 307. healthy responses to natural rewards)? Thus 24. 1431–1441 3. Fowler. E. (2004). Dopamine in drug abuse and binding protein gene promotes alcohol-drinking behaviors. & Nestler. learning and motivation. Neurosci. M. Neuropharmacology 47 (suppl. 315–342 13.R. J. et al.Z. T. G. A. & Nestler. Neuromol. J.D. & Koob. COMPETING INTERESTS STATEMENT 36. ∆FosB. 84. F. H. E.com/natureneuroscience Phil. Appetitive behavior: impact 47. J. 50. Ther. Eisch... which may act by reducing 199 (2005). R. Neurotrophic mechanisms in drug addiction. Annu. mon palliative effects in individuals addicted to a wide range of drugs or 17. M. Biol. E. Neurosci. 647–650 (2005). subregions of the ventral tegmental area. Pharmacological mechanisms of naltrexone and acam- 15. N. gamma-Aminobutyric acid A receptor subunit mutant mice: new (2004). B. & Uhl. J. Acad. D. & Xu.. Nestler. For example. M. which are. Ann. dysregulation of reward. 26... Kuo. Neurosci. 39. Different requirements for cAMP response element binding 7. E. et al. Natl.T. Heinrichs. 2. & Picciotto. Drug addictions. A high priority for current research should be to focus 32.S.L. of the individual (that is. Barrot.. ∆FosB: A molecular switch for long-term adaptation. 49. G. 25–53 (2003). A. G. Drug addiction.. Chem. 42. B Biol. Molecular neuroadaptations in the accum- partial agonist and antagonist at the µ opioid receptor. © 2005 Nature Publishing Group http://www.A. (2004).. Impey. Neuropsychopharmacology 28. 31.A. J. 12 (suppl. (2004).R. Asher. Russell. 30. Hyman. Brain Res. 447–475 dampen common mechanisms of impulsive and compulsive consump. & Seamans.N. 14. K. The neuroscience of natural rewards: relevance to addictive Neuropharmacology 47 (suppl. Corticotropin-releasing factor in brain: a role in activation. Structural plasticity associated with exposure to drugs of abuse.C. G. Barrot. & Malenka. 1.A. A. Natl. Studies are underway in each of these areas. Molecular basis of long-term plasticity underlying addiction. all established treatments for addiction are drug-specific and are morphine. Kalivas.com/natureneuroscience/ 37. D. 695–703 (2001).C.W. 1485–1490 (2003). A.J. C. Kelley. R. and its persistence. Addiction. NY Acad.. D. Neuron 31.J. McClung. 12. Neurosci. & Hoebel.E. 45. I.L. Brain Res. D. T. 1848–1859 (1998). 2. S. Borgland. 1642–1645 (2005). USA 99. 29. Neurosci. Parkinson. Cardinal. Trans. 40. P. Neuron 45. 25. A single infusion of brain-derived neurotrophic factor into the ventral tegmental on bringing some of the most promising common anti-addiction mecha. & Malenka. ∆FosB: A molecular switch for addiction. our knowledge of shared mechanisms of 21. R. Psychiatry 9. J. ment of alcohol and nicotine addictions.C. 3663–3672 (2002). of common molecular adaptations to drugs and natural rewards. Neuron 46. N.. & Blendy. 436–445 (2005).S. & Schultz. J.. 11435–11440 (2002). in cocaine self-administration and relapse of cocaine-seeking behavior. Shaham. Ann. Howlett. Neurosci. Adaptive coding of reward value by dopamine prosate in the prevention of relapse in alcohol dependence.. Wise. J. E. A. R. & Bari. are effective for bens and ventral tegmental area during the first 90 days of forced abstinence from cocaine self-administration in rats. O.M. 5553–5562 (2005). Mol. Drugs of abuse and stress trigger a common example. M. 610–615 (2002). tive). J. D. Naltrexone can also be of some use in the treat. J.C. although this approach remains highly speculative. natural rewards. Pharmacol. Goeb. & Nestler. 25. Nat. Glutamate systems in cocaine addiction. M.nature. 132.N.J. Kauer. 815–819 (2003). J.. C. Constantinescu. Learning mechanisms in addiction: synaptic plasticity in the ventral teg- the ultimate safety and effectiveness of such treatments. Walters. in the cAMP-CREB pathway. J. Dopamine and drug addiction: the nucleus accumbens shell connec. it might be possible to exploit our increasing knowledge 18.J. Thus. R. W. McClung. 66. R.. Unmanageable motivation in addiction: a pathol- ogy in prefrontal-accumbens glutamate transmission. corticolimbic CREB and ERK signaling in C57Bl/6J mice. Yao. Biochem. to natural rewards. S. 10–17 (2003). Storm.F. Pharmacol. The many faces of CREB. correlates in individual rats. Volkow. Kalivas.E. Neurochem.C. 24. Bonci. 24. Duman. G. Dani. J.D.. Lu. and ∆FosB. 5022–5030 addiction: results from imaging studies and treatment implications. Boehm. J. Wang. Shaw-Lutchman. and thermoregulation in CREB-deficient mice. at least in part. & Williams. P. 34. A. 1604–1611 (2004). 8. mu-Opioid receptor and CREB activa- tion are required for nicotine reward. & Wolf. V. E. Shaw-Lutchman. A. Rev. Rev.. while clearly difficult. 23. & Kolb. Regulation of gene expression and cocaine reward by CREB The author declares that he has no competing financial interests. Trends Neurosci. Published online at http://www. S3–S11 (2003).L.A. & Robbins. Elevated levels of GluR1 in the midbrain: a trigger for tion of drug or natural rewards without affecting the normal functioning sensitization to drugs of abuse? Trends Neurosci. CREB activity in the nucleus accumbens shell controls gating of behavioral Neuropsychopharmacology 24. 43.C. Rev. Identification of PSD-95 as a regulator of dopamine-mediated synaptic of amygdala-dependent mechanisms of emotional learning. 85. Neurosci. Malenka. Opin. Y. area induces long-lasting potentiation of cocaine seeking after withdrawal. & Le Moal. Alterations in morphine-induced reward. C. 11042–11046 (2001). 279. J. metha. 20. J. 193– is modest in most individuals.) 227–241 (2004).) 345–358 (2004). Regulation of CRE-mediated transcription in mouse brain by amphetamine. however. 933–943 (2005).L. J. 25.C. Psychomotor stimulant addiction: a neural systems perspective. S. Robinson. 22.. Olson. Increased probability of GABA release during withdrawal from far. Cleck. Pierce. 9438–9444 (2001). et al.J.A. Y. Zhang. opioid or cannabinoid systems might well be expected to exert com.A. Soc. L.. 311. 95–110 (2001). E. 17. Synaptic plasticity in the mesolimbic dopamine system.W. Sci. M. 985. USA 98. S.H.S.) 33–46 (2004). W. drugs. Neuroscience 122. 54. S. Partial deletion of the cAMP response element- 4. Tobler. R. Exp. Neurosci. Regional and cellular mapping of CRE-mediated transcrip- Reprints and permissions information is available online at http://npg. Cannabinoid physiology and pharmacology: 30 years of progress. 7482–7490 (2004). 10. M. D. & Hope.J.A. C. Molecular and cellular endpoints. Volkow. Nestler. or any of their numerous synaptic adaptation in dopamine neurons. Nestler. Everitt. 2439–2450 aimed at the acute protein target of the drugs46. 796–803 (1997).G. S. 1604–1613 (2003). J. Reduced behavioral effects of trials. whether addictions can be treated.C. J.L.. Is it possible to mental area as a result of exposure to drugs of abuse. Wu. & Malenka. Y. J. and allostasis. Kreek. & Blendy. 4. Neurosci. reprintsandpermissions/ 38. Neuropharmacology 47 (suppl. 22. 349–352 (2001).W. Neurosci. Carlezon.. 5.G. J.C. B. J. Walters. & Nestler. done. J. et al. 97–129 (2001). 68. Neurosci. & Zhou.C. B. Curr. W. Roy. 28. protein in positive and negative reinforcing properties of drugs of abuse. Jr. ACKNOWLEDGMENTS Acad. et al. 43321–43329 6. Neurosci. A. is reported to be effective for some alcohol addicts50. 6. Drgonova. Self. Molecular mechanisms of drug addiction. Thomas. T. PERSPECTIVE CRF. NMDA glutamatergic receptor function (although this remains specula. 27–49 (2001).) neurons. J.. Proc. Di Chiara.F. & Blendy. 625–638 (2004).A. J. 577–582 (2003).. M. 937. J. or in the complex molecular constituents of the glutamater. Brunzell.T. Ji. Carlezon. T. nisms into the clinic for broad trials across several addictive disorders. J. Dong. Acute and chronic cocaine-induced potentia- tion of synaptic strength in the ventral tegmental area: electrophysiological and behavioral gic synapse. Physiol. for 19.A. Neurosci.com/ tion during naltrexone-precipitated morphine withdrawal. Synaptic plasticity and nicotine addiction. Addiction and the brain: the neurobiology of compulsion 46. 18. At the same time. F. but the magnitude of its effect locomotor activity. Avena. Saal. 119–128 (2001). Nat. Bolanos. (1992). 2. & Berridge.E.S.A. & Swanson.

part. Nielsen. including the physiological consequences of receptor activity) is particularly well documented for opiate addiction4. including alteration of pharmacodynamics (a drug’s effects at a studies4. of 30–60%.edu classification of personality dimensions and to tease out the influence Published online 26 October 2005. regular drug use. Addiction has some of the highest overall medical health costs of any medical dis- Vulnerability to develop a drug addiction is influenced by a combina. Moreover. Variation in personality dimensions. 1). It is regular drug use. ism. the transition from abuse to addiction. 1). Predisposition to addiction may be that do not readily self-administer drugs (implying genetic resistance). there or pharmacokinetics (a drug’s absorption. order.5. Addiction heritability was first demonstrated with alcohol- there are inbred strains and selectively bred lines that readily self-admin. involve complex genetic The genetics of addiction encompasses heritable factors that influ- contributions based on multiple variants of multiple genes and dif. including changes in expression of genes or their protein products. These challenges also must be faced in the study to the study of genetic variation in these personality and of other complex genetic disorders. once comorbid disorders such as HIV/AIDS. be studied.nature. A number of inventories have been developed for the description and e-mail: kreek@rockefeller. New York. including severity of dependence or withdrawal and the most studied diseases. Loss of productivity. and the propen. interdiction and drug-induced effects to influence the progression from intermittent to the criminal justice system incur additional economic costs. stress responsivity—as well as a substantial proportion of comprise only a small proportion of the estimated genetic contribution. in neural networks. physiological traits. its own genetic basis. For some of to drug addiction. In rodents. a genetic variance to drugs3. Genetic factors may also be involved in direct drug-induced shared by multiple classes of drugs of abuse is demonstrated in twin effects.1038/nn1583 of genetics on personality. defined chronologically as phenotypic characterization of individuals and the characterization of initiation of drug use. the identified variants. distribution. including genetics. the costs to society of drug with addictive diseases. with the goal of improving primary Chronic exposure to drugs of abuse causes persistent changes in the prevention.2. including cancer. and alcohol addiction are too enormous to ignore. However. Steven LaForge are in the Laboratory of the Biology of Addictive Diseases. early intervention and chronic treatment. addiction/dependence ethnic/cultural backgrounds (as different backgrounds yield differences and potentially relapse. therefore imperative that all components contributing to addiction sity for repeated relapse after achievement of a drug-free state1. metabolism and are different influences of environment versus genetic factors on the excretion) of a drug of abuse or of a treatment agent. such as impulsiv- ity. all of which ultimately affect behavior. in K. brain. risk taking and novelty seeking. These include precise the various stages of addiction. risk taking. David A. may contribute to the initiation of drug use as well as the transitions from initial use to regular use to Mary Jeanne Kreek. New York. Each of these personality dimensions may have. doi:10. transitions from initiation of drug use. personality Studying the genetics of complex psychiatric or behavioral disorders and physiological traits themselves may differentially affect such as addiction poses additional challenges. to regular drug use. David A Nielsen. such as certain cancers. USA. Furthermore. Eduardo R. hepatitis C and tion of genetic and environmental factors. The Rockefeller University. which is influenced by distinct genetic factors such as the alde- ister drugs of abuse (implying genetic vulnerability) as well as strains hyde dehydrogenase 2 genotype. with estimates of heritability and synaptogenesis. as receptor. obesity and heart disease.com/natureneuroscience responsivity and vulnerability to drug abuse and addiction Mary Jeanne Kreek. ence the different stages in the trajectory of initiation and progression ferent combinations of these variants in different people. the specific genetic risk of relapse. and their influence on and interaction Despite the complexity of the problem. and in neurogenesis to the vulnerability to addictive disease. in their entirety. vulnerability to addictive diseases. Here we focus on recent approaches in allelic frequencies). due both to genetic variants that are common to all addictions and to Different strains show differences in the cellular and molecular response those specific to a particular addiction. addiction/dependence and then potentially to relapse6 (Fig. risk taking and by multiple studies. N E U R O B I O LO G Y O F A D D I C T I O N PERSPECTIVE Genetic influences on impulsivity. However. Eduardo R Butelman & K Steven LaForge Genetic variation may partially underlie complex personality contributions and genetic variants have been identified and verified and physiological traits—such as impulsivity. Four instruments often used in genetics 1450 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . Family and twin epidemiological studies show that genes contribute in protein-protein interactions. stress © 2005 Nature Publishing Group http://www. but most. Both factors couple with direct lung cancer are factored in. some genetic variance is specific to drug class. Butelman & addiction (Fig. to drug Many medical disorders have some genetic component. For example.

family of addiction pathophysiology. disequilibrium (which occurs when there is a non-random distribu. tor D2 variant. in the human genome. As SNP panels become more inclusive of the common variants This is increasingly the experimental approach of choice for identify. ↓↓↓. Of the traits measured by these tests. Other variants may alter the amount of gene expression (for example. which measures notably the Collaborative Study on the Genetics of Alcoholism spon- neuroticism.nature. out- elty seeking. which measures nov. of underlying traits.000–80. Cocaine also increases gene expression and promotes pretation of different studies. Linkage studies investigate the transmission of genetic markers in identified chromosomal regions is then required. to develop cocaine addiction2. 0. research to quantify personality dimensions are the Tridimensional studies in illicit drug addiction are difficult to conduct because of the Personality Questionnaire (TPQ) or the more complete version. However.com/natureneuroscience Initiation of Intermittent to Addiction responsivity (top) and the three domains drug use regular use and relapse (bottom) interact to influence the progression to addiction. panels.000 nucleotides of the genotyped variant.9. By contrast. sored by United States National Institute on Alcoholism and Alcohol ness) and the Barratt Impulsiveness Scale. in a haplotype) is relatively prodynorphin promoter region variants). NEO-PI-R provide a broad and more time-intensive these factors1. the Barratt and KMSK genes governing direct and downstream molecular events altered by scales provide a relatively rapid evaluation of a particular phenotype chronic exposure to a drug of abuse. However. Variants of the instruments may also be advantageous for the study of more refined preprodynorphin gene (PDYN) have been associated with vulnerability phenotypes relevant to a particular clinical situation. including stud. ies. agreeableness and conscientious. DRD2). including the GABA receptor subunit large number of individual questions contribute to a smaller number A2 (GABRA2) and muscarinic acetylcholine receptor M2 (CHRM2). medium relative influence. Until recently. TCI. Genetic factors for addiction ↓. in which a involvement of several genes. abuse and Impulsivity* (genetics?) addiction. and (iii) if the effect sizes are moderate to high. no influence. and the difficulty in ascertaining family members. It is frequently useful to inves- seeking. the NEO Personality Inventory-Revised (NEO-PI-R. small relative influence. of the diseases and/or addiction and based on specific hypotheses about The TPQ. ↓↓. Further fine mapping used. with or without aggression One general approach for identifying specific genes involved in a or suicidality. In studying drug addiction. to detect linked variants involved in a disease (i) if they are within Variants in the coding region of genes may change the protein prod- 40. which the identification of more defined regions for fine mapping in a far ask whether a particular gene allele is more prevalent in patients. over the last few years. ↓. some addic. Although family-based studies may be feasible for personality Such variants can affect both normal physiology and specific aspects traits such as impulsivity. greatest relative (~ 100% of cases) influence. harm avoidance. (ii) if linkage uct. risk taking and stress responsivity. as depicted. family-based linkage studies have been most widely associated with a specific disorder or addiction. We expect that these personality factors Comorbidity (genetics) contribute less to addiction and relapse after substantial changes to the brain. extroversion. consisting of. all of which can contribute to functionality2. and yet other variants may high. Association studies are able associated with a phenotype more quickly. the disruption of involved families Temperament and Character Inventory (TCI. For example. which is often associated with novelty disease is hypothesis-oriented selection. These atypical (genetics) two personality factors. in alcohol dependence8. simpler manner. openness. a surge in extracellular dopamine by blocking the action of the dopa- Use of common questionnaires does facilitate the comparative inter. In addition. preferably. which measure duration and magnitude of drug use. standing family studies have been done in the field of alcoholism. respectively). Another strategy is to use positional approaches—conducting Identifying genetic factors in personality traits and addiction genome-wide scans to identify chromosomal positions that may be Until recently. tion research has focused on impulsivity. for example. Some of these questionnaires Abuse7. a more refined approach than microsatellite marker compared with control subjects. This is a much alter the rate of mRNA degradation (for example. effected by Stress responsivity chronic exposure to the drug of abuse. as in the A118G variant of the µ opioid receptor gene (OPRM1). cocaine produces (impulsiveness and degree of exposure to a drug of abuse. two or more generations.2. reward dependence and persistence). However. minor relative Ann Thomson (30-60% of cases) influence. NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1451 . *Lifelong or identified Environmental factors in early childhood. the enormous stigma of addiction. mic- on specific genomic regions of interest and phenotypes in pedigrees rosatellite marker panels were used to scan the whole genome. addictions can be defined with scales such as the tigate specific genes involved in diseases based on a prior understanding KMSK. comorbidity and stress © 2005 Nature Publishing Group http://www. PERSPECTIVE Figure 1 Diverse contribution of genetic influences to initial drug use. These ratings reflect our estimates and opinion based on current Drug induced effects (with some genetic factors) (~ 100% of cases) information. various approaches using single nucleotide poly- ies of affected sibling pairs (more powerful when both siblings and morphism (SNP) arrays or other panels of single SNPs have allowed parents are included). it should be possible to examine the variants ing genes responsible for complex traits. The alternative is association studies. We suggest that impulsivity and risk taking contribute most to the initiation of Risk taking* (genetics?) drug use and the progression to regular drug use. and risk taking. a dopamine recep- smaller distance than is possible with family-based linkage stud. one can initially consider characterization of personality traits. tion of allele combinations. Strong evidence has been provided by these studies for the and variables are based on the concept of factor analysis. than would be expected by chance. optimization of more focused release of the κ opioid ligand dynorphin in the striatum. mine transporter.

2 – – – – + Alc D/Ab CYP2D6 Cytochrome CYP450 Drug metabolism 22q13. Moreover. trade-off that most geneticists have to make—whether to study large Moreover. D/Ab ADH1B Alcohol dehydrogenase 1B Ethanol metabolism 4q22 – – – – + Alc D/Ab ADH1C Alcohol dehydrogenase 1C Ethanol metabolism 4q22 – – – – + Alc D/Ab ALDH2 Aldehyde dehydrogenase 2 Ethanol metabolism 12q24.nih. This perspective is not intended to be a decrease in numbers of subjects studied. vigilance for error of any type in through the potential impact before carrying out actions. or to do very careful phenotyping of the subjects. Statistical genetics methods involve techniques that are evolving.ncbi. however. a or negatives. Population genetics can also be influenced by additional ‘genetically resistant’ individual (or strain of rat) may self-administer a factors—for instance.3-p14 + – – – + Alc D/Ab TPH2 Tryptophan hydroxylase 2 Serotonergic 12q21. Alc D/Ab OPRK1 κ opioid receptor Opioid 8q11.1 – – – – + H/O. strongly without which one runs the risks of generating more false positives influence the development of drug abuse or addiction. because of the time and expense. primarily from established research teams using acceptable addiction. and progression to importance. Such assessment entails the use of such as methods for statistically determining inferred haplotypes . Alc: alcohol. D/Ab FAAH Fatty acid amide hydrolase Cannabinoid 1p35-34 – – – – + Alc CSA NPY Neuropeptide Y Neuromodulatory 7p15. which would in turn give the study and the statis. H/O D/Ab HTR2A Serotonin receptor 2A Serotonergic 13q14-q21 – – – – + Alc CSA.2 – – – – + C/S D/A TH Tyrosine hydroxylase Dopaminergic 11p15. C/S: cocaine or stimulants. and rigorous state-of-the-art statistical genetics anal- all studies of addiction genetics because poor or inadequate phenotypic yses10.nlm. There is therefore an inherent comprehensive review. Johns Hopkins University. bAssociation with drug addiction in two or more studies Ultimately. comorbid we have included only a selection of studies that we consider of potential disorders. desire: for example. there are significant ethnic/cultural differences drug of abuse under specific environmental conditions (Fig. molecular techniques personnel. D/A: dependence/addiction.gov/omim) as of July 2005. R: risk taking. CSA: continued substance abuse.5 – – + – + Alc D/A DRD2 Dopamine receptor D2 Dopaminergic 11q23 – – – – + Alc D/Ab DRD3 Dopamine receptor D3 Dopaminergic 3q13.1-q35 + – – – + Alc D/A GABRB1 GABA receptor subunit β-1 GABAergic 4p13-p12 + – – – + Alc D/A CHRM2 Muscarinic acetylcholine Cholinergic 7q35-q36 – – – – + Alc D/Ab receptor M2 CNR1 Cannabinoid receptor 1 Cannabinoid 6q14-q15 – – – – + Alc. H/O D/Ab GABRA1 GABA receptor subunit α-1 GABAergic 5q34-q35 + – – – + Alc D/Ab GABRA6 GABA receptor subunit α-6 GABAergic 5q31. rigorous phenotypic assessment is essential for the molecular work. acting on the spur of the moment. C/S D/Ab © 2005 Nature Publishing Group http://www. Here a diverse battery of instruments to evaluate personality traits. Alc D/Ab DBH Dopamine β-hydroxylase Dopaminergic 9q34 – – – – + C/S D/A DAT (SLC6A3) Dopamine transporter Dopaminergic 5p15.1 – – – – + Alc CSA.com/natureneuroscience DRD4 Dopamine receptor D4 Dopaminergic 11p15. detailed histories of initiation of drug use.2 + – – + + Alc. assessments lead to incorrect results. H/O: heroin or opiate. These must be controlled for or analyzed using a variety of newly developing techniques Personality traits and addiction: impulsivity involving primarily combinations of SNPs or other variants. it can lead to a and statistical genetics analyses. Precise phenotyping takes time. factors. D/Ab SERT (SLC6A4) Serotonin transporter Serotonergic 17q11. we must emphasize that evidence of enhanced genetic vul- numbers of subjects.1 – – – – + H/O D/A ANKK1 Ankyrin repeat and kinase Signal transduction 11q23. nerability to addiction does not imply that addiction will occur.3 – + – – + Alc.3 + – – – + Alc D/Ab TPH1 Tryptophan hydroxylase 1 Serotonergic 11p15. Alc CSA. Many tics greater validity. C/S CSA. or making 1452 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE .2 – – – – + H/O D/A PDYN Preprodynorphin Opioid 20pter-p12. D/Ab HTR1B Serotonin receptor 1B Serotonergic 6q13 – – – – + Alc.2 – + – – + Alc D/Ab domain–containing 1 (predicted) I: impulsivity. Baltimore (http://www. defined as acting suddenly in an unplanned manner to satisfy a plex disorder. There are Impulsivity is a personality trait characterized by behavioral disinhibi- innumerable further challenges to molecular genetics studies of any com. aGene map locus: Online Mendelian Inheritance in Man. E: environment. tion. Alc D/Ab MAOA Monoamine oxidase A Catecholaminergic. S: stress responsivity. 1)3.nature. such as environmental influences or availability of drugs. including other diseases that are present at the same time. and requires highly trained or optimal study designs. C/S. A: addiction. Conversely. not thinking the specificity in diagnosis of subjects.5 + + – – + H/O. in allelic frequencies of variants of many specific genes. Xp11.23 + – + – + Alc D/A serotonergic COMT Catechol-O-methyl transferase Catecholaminergic 22q11.1-q12 + – + – + H/O. phenotypic assessments. PERSPECTIVE Table 1 Genes having one or more variants that have been reported to be associated with one or more addictions Chromosomal Gene Protein System locationa I R E S A Drug Status OPRM1 µ opioid receptor Opioid 6q24-q25 – – – + + H/O.

ylindolacetic acid. frontal cortex and amygdala. often defined as one metabolism. no apparent DRD4 studied for association and linkage with candidate genes (Table 1) in binding is detected in the nucleus accumbens. which codes for a extrapolate to all people. serotonin metabolism. Four psychiatric conditions (depression. PERSPECTIVE statements without thinking in advance about what is to be said. dorsomedial thalamus. Other genes such as SERT.12). and the noradrenergic system (dopamine β-hydroxylase Notably. dopamine receptors. to over 50% of people with alcoholism. the dopami. with a propensity to experiment with novel stimuli. In contrast. that require an evaluation of relative risk versus reward (for instance. a major metabolite of serotonin and an indicator of in the choice of career opportunities or selection of automobiles). including those Loss of impulse control may be due to impaired inhibitory control result. results reported for the Taq1A variant ments used. and its genetic variants. substance abuse does not occur as an isolated dis- Thus. Future studies on the role of mon comorbid conditions are depression and anxiety.nature.17.com/natureneuroscience prolactin release after fenfluramine challenge. Hence. inhibitory system.000 nucle- Although a substantial body of knowledge is accumulating from these otides downstream (3′) from the DRD2 gene and has recently been genetic studies. Novelty seeking. relapse and recovery. Experimentation Novelty seeking may be correlated with progression from abuse to with addictive drugs and the onset of drug abuse generally occur in addiction for several drugs. demonstrates a relationship between low serotonin metabo. this association has yet to be solidly and dopamine receptors D3 and D4 (DRD3 and DRD4) are related to documented. [COMT]). and the DRD4 receptor distribution pattern suggests roles in attentional. 20% therefore shed light on neurobiological mechanisms underlying clini. GABAergic and nitric oxide systems. Although several studies have identified associations of different ated with alcoholism or some other addiction (Table 1). the ethnic/cultural populations studied and the statistical may be potentially ascribed to the action of the ANKK1 gene product. Notably. In impulsive violent offenders. phenotyping instruments used thus. Some studies report ism and prescription opiate addiction. which codes for the rate-limiting enzyme in the production Other molecular targets involved in monoaminergic function have of serotonin. with possible progression to addiction. ing from drug-induced changes in the frontal cortex. with or without any is not an indicator of pathology. these findings have not the neurotransmitter systems coded by these genes are interactively been consistently replicated16. impulsivity per se with or without inherent negative consequences. anxiety. prefrontal and entorhinal cortex. Most of these candidate genes code personality disorder and attention deficit/hyperactivity disorder) are for proteins that control major neurotransmitter systems. Interestingly. possible or probable harm to oneself or others. 11. with the rare exception of some reported cases of alcohol. nergic system (tyrosine hydroxylase [TH].15.14. However. between high Tridimensional Personality Questionnaire changes during adolescence and early adulthood may modulate impulse novelty-seeking scores and a particular allelic variant16. with conflicting meta-analyses from different groups19. among other issues. Behaviors characterized by a deficit in impulse control have been lateral septal nucleus and hypothalamus18. Low levels of cerebrospinal fluid 5-hydrox. Acts Personality traits and addiction: risk taking of impulsivity may include aggression.14. aggression and various literature and popular press for its association with alcoholism and vari- forms of suicidality. which may occur later. with unipolar impulsivity. These heterogeneous findings may involved in the acute and chronic effects of most drugs of abuse and. as tions and decision making. taking. Risk taking may be operationally measured in tasks ous forms of impulsivity. For many addicts. or opiate addiction have depressive and/or anxiety disor- NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1453 . Each of these genes is reportedly associ. In epidemiological studies. is related to impulsivity. on TPH1. a biomarker of serotonin the South Oaks Gambling Screen. the elderly. occurs on a continuum. lism and impulsive behavior13. tryptophan hydroxylase 1 and 2 which are major sites of D2 receptor binding and mediate the direct [TPH1 and TPH2] and serotonin transporter [SERT]). the results gleaned from these investigations may not reported to reside in the neighboring ANKK1 gene. In addition. antisocial impulsivity have been reported. Neurodevelopmental processes and reproductive hormone for example. In human control. the prefrontal cortex is a site of cognitive and executive func- [DBH]). DRD4 binding is found in brain regions that include the drugs of abuse.17. dopamine transporter [DAT]). Furthermore. the Taq1A variant is located approximately 10. DRD4 receptors are members of the ‘D2- adolescence. cocaine and other stimulant cally defined stages in the process of addiction. violence and suicide. meta-analyses of these studies Comorbid disorders may be fraught with pitfalls. a TPH1 gene variant was been related to novelty seeking and drug abuse (Table 1). and without robust con- Early work implicated low serotonin levels and its metabolites in vari. for this reason. detected in certain risk for impulsive personality traits in first-degree relatives13. the results of apparently similar studies cannot be directly compared and. well as other genes)11. aspect of risk taking. In addition. It is part of a constellation of traits observed in individuals diverse regions. Risk taking is characterized by behaviors performed under uncertainty. with potentially high reactivity to novel stimuli. Pathological gamblers and addicted patients may exhibit signs of risk as well as to early-onset alcoholism (reviewed in refs. as a trait. The DRD2 associated with reduced CSF 5-HIAA and suicidal behavior12. several major candidate genes with variants associated with order. impulsivity itself is or physiology of addiction to opiates and alcohol22. Hence. emotional and mnemonic processing. impulsivity. like’ family of Gi-coupled dopamine receptors. on the basis of (monoamine oxidase A [MAOA] and catechol-O-methyltransferase some major functions thought to be mediated by these brain areas. underlie addiction as well as the initiation of drug use. addiction. impulsivity (Table 1)11. the serotonergic system (for example. and ethnic composition of patient populations. ous psychiatric disorders. which may contribute to vulnerability to experimentation with brain tissue.14. hippocampus. DRD3. psychostimulant and reinforcing effects of drugs of abuse. aggression and depression. the monoamine metabolism pathway motivational. result from differences in age of subjects. 5-HT2A. in The earliest candidate gene studies on impulsivity were conducted different studies16. which can be assessed by specific clinical questionnaires such as © 2005 Nature Publishing Group http://www. The most com- associated with specific addictive diseases. These studies vary in their assessment instru. produced by drugs of abuse. serine/threonine kinase21. methods applied. even in an association between novelty seeking and DRD4 receptor variants. psychometric instruments (such as the Temperament and Character Serotonergic neurons project from the raphe throughout the brain to Inventory). including the hippocampus. which is also associated with an increased can alternatively be considered a personality trait. at specific stages of addiction could depression being the most common. However. thus. tingency planning. MAOA. caudate or putamen. TPH1 Taq1A polymorphism is widely studied and reported in the scientific variants are also associated with impulsivity. for which commonly present in and probably are involved in psychopathology a wealth of basic data is available. DRD4 polymorphisms with novelty seeking.20.

addictive disorder. who discovered that various noxious stimuli caused what he abolishing a putative glycosylation site in the N terminus2. Cortisol or corticosterone exert negative self-administration studies. 23). including former heroin addicts in methadone proopiomelanocortin (POMC) mRNA and opioids ACTH maintenance treatment with ongoing dependence on cocaine.com/natureneuroscience induces synthesis of Endogenous with cocaine dependence. recently abstinent cocaine-dependent sub- mon. ACTH acts on ACTH receptors in the adrenal other components of stress responsivity in the brain. Animal studies document physiological and corresponding molec- ular alterations in components of the HPA axis caused by acute or chronic administration of drugs of abuse. thereby Selye. acute and chronic stress affect the HPA axis. + stressors. For example. prolactin and norepinephrine. initiation and maintenance of addictions. the prevalence of comorbidity in people when they of cocaine in a ‘binge’ protocol for 1 or 2 days to rats causes an increase first try a drug of abuse is not well defined. This heroin addicts have a hypo-responsive HPA system and that patients _ POMC β-endorphin © 2005 Nature Publishing Group http://www. and induce reinstatement of self-administration. to determine which of the gene variants contribute to the psychiatric The endogenous opioid system. PERSPECTIVE Figure 2 Stress causes between HPA axis function and addiction? In addition to the classical increased mRNA _ Hypothalamus feedback regulation of the HPA axis by corticosteroids. receptor is the primary target of addictive opioid drugs. It has been established that genetics are somewhat cortisol. Stressors can influence the and guinea pigs) or corticosterone (in rats and mice). In addition to this classical circadian negative feedback regulation by glucocorticoids. phin bound the 118G (Asp40) receptor variant with threefold greater An important component of the stress-responsive system is the hypo. as well as increased plasma levels of ACTH. However. or to both.κ-) peptide in the anterior + hyper-responsive HPA axis1. it is difficult system is involved in cocaine craving during abstinence29. rather than feedback and circadian inhibition. disease. as well as from processing of POMC. as is the case with glucocorticoid regulation of the axis25. specifically µ and κ opioid receptors. the endogenous opioid system. to the addictive disease. adrenocorticotropic In animal models of conditioned place preference and drug self- hormone (ACTH). scripts evoked increased craving. 2). extinc- feedback regulation at both the hypothalamus and the pituitary to inhibit tion and reinstatement. which are primary rewarding properties of drugs at each of the stages in laboratory animal mediators of the stress response. The most common coding region polymorphism is the A118G Stress responsivity SNP with allelic frequencies that vary widely among populations (allele The modern concept of stress and its importance for many human frequencies from 0. show a (µ-. Thus. thus investigating genetic variation in addiction sion in response to opiates. stress can enhance acquisition. anxiety and if the psychiatric disorder preceded or followed the development of the cardiovascular measures. which is significantly attenuated follow- who are already addicted to illicit drugs potentially could be classi. physical dependence and respiratory depres- independent variable. clinical stud- synthesis and release ies with opioid antagonists demonstrate that the endogenous opioid of hypothalamic CRF corticotropin releasing _ system. Mice lacking ing focused studies of its influence. contributes to the con- which acts on CRFR1 Anterior tinuation of specific addictions.01 to 0. and may increase cortex and induces release of the stress hormone cortisol (in humans the reinforcing effects of drugs of abuse. 2). we found that the endogenous opioid peptide β-endor- itary and adrenal glands. reward. with a particular focus on the HPA axis. which are thought to model human states of the synthesis of POMC and release of ACTH and β-endorphin. as well as to relapse once the brain has _ pituitary receptors in the undergone plasticity due to addiction1. This is appar- The role of comorbidity in the genetics of addiction remains an ently tonic inhibition. requir. especially in people who are dependent on cocaine and other jects were read individually tailored scripts designed to provoke stressful. demonstrate inhibitory control over the HPA axis (Fig. such as morphine (reviewed in ref. but not neutral. administration ders22. withdrawal and relapse. deficit/hyperactivity disorder in the childhood or adult form is com. tion. of the HPA axis. Stressful and drug-cue important to use more refined psychiatric diagnostic tools to determine related.nature.26. in the presence of comorbidity. Also. Corticotropin releas- fied as having antisocial personality disorder because of their criminal ing factor mRNA in the hypothalamus is also increased following 1 or 2 activity pertaining to the acquisition and use of illicit drugs. many people in plasma corticosterone levels. Thus. ing chronic 14-day binge cocaine administration. with and without comorbidities. into the circulation guinea pig of underlying physiological and ultimately genetic conditions) or was of β-endorphin and b Cortisola Glucocorticoids Mouse. The µ opioid cally important in the vulnerability to or severity of addiction. taking comorbidity into account as an absent analgesia.11. vitro studies.48) and results in an asparagine (Asn) to diseases was developed by the pioneering neuroendocrinologist Hans an aspartic acid (Asp) substitution at amino acid position 40. one of the important intracellular signaling systems of this 1454 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . Conversely. also tonically inhibits the factor (CRF) into the CRF HPA axis. but also suggesting that the sympatho-adreno-medullary eases discussed here. comorbidity may be mechanisti. via both µ and κ opioid receptors. In in called a ‘general adaptation syndrome. relaxing experiences. β-endorphin thalamic-pituitary-adrenal (HPA) axis. studies are needed in the µ opioid receptor gene (OPRM1) show dramatically reduced or addictive disease populations. In these clinical studies. increase resistance to extinc- tonically inhibits this axis.25. including both µ and κ opioid receptors.30. activation of G protein–coupled inwardly rectifying potassium (GIRK) tion24–27. For each of the comorbid disorders. area of controversy. HPA axis activation or suppression influences addic. 2). it is drug-cue related or neutral. The question can therefore be posed: is there a genetic link channels. stimulants (reviewed in ref. We have found that active anterior pituitary. affinity than the prototype 118A (Asn40) receptor30. In a series of clinical studies.’ mediated in part by the pitu. which are derived administration. Numerous polymorphisms of the OPRM1 gene have been identi- fied. For example. including initiation. Attention days of administration but is significantly reduced after 14 days28. rat Corticosteroneb caused by long-term self-administration of opiates or cocaine. maintenance.24. Certainly. it is not possible Ann Thomson pituitary and release Adrenal to distinguish whether this atypical stress responsivity preceded (because a Human. not only indicating involvement involved in each of the psychiatric disorders just as in the addictive dis. We have proposed that an atypical responsivity to stress and portal circulation. in general. Exposure to stress activates the binding to the Asp40 receptors showed threefold greater potency in HPA axis (Fig.

which causes immediate activation of the HPA axis who had experienced stressful life events and had two copies of the short © 2005 Nature Publishing Group http://www. defects in the MAOA gene functional A118G SNP was associated with heroin addiction in two have been linked to aggression. the variant receptors30. Although the environment contributes relatively non-admixed populations.39–41. Linkage studies have been conducted to identify genetic determinants Overall. as this polymorphism also affects HPA risk-taking. of stress responsivity. anxiety.33. that are involved in addictions to multiple substances41. Childhood abuse also increases reported in some studies32. for the OPRM1 A118G SNP. genetic factors account for 30–60% of the overall methionine for valine at residue 158. New techniques allow association that stress is a precipitating factor in relapse. if they lacked ‘life events’ such as employment.11. that is. the activity of the HPA axis seems to undergo extensive of addictive diseases2. HPA of Alcoholism (COGA). The Collaborative Study on the Genetics plasticity as a result of exposure to drugs of abuse. Furthermore. be due to the specific gene-environment interaction that must occur for This difference in response to treatment may be mediated by differences an effect to be observed. In twin studies. Furthermore. which encodes an enzyme that catalyzes the degradative metab. in people one or naltrexone. these studies may have to be in HPA axis activation owing to receptor genotype. All the Environmental factors affected subjects had polysubstance abuse. cortisol than do people with one or more high-activity valine alleles Many of the genes for which there is evidence of association or linkage (Val/Met or Val/Val)35. and comorbid psy- orders. The methionine form has greater variance in the risk for the development of drug addictions. promising candidate is the µ opioid receptor gene (OPRM1). response to opiate antagonist challenge. repeat polymorphism expressing higher levels of SERT mRNA. of these factors. along with potential gene variants involved in each risk for alcoholism in several studies. so the regions identified may contain genes tional on exposure to environmental determinants. However.31. These results suggest that common genetic variants maintained measured by serum ACTH and cortisol levels2. utable risk for the 118G allele was 21% for Swedish individuals with NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1455 .nature. but there thermolability and a three. The low-activity methionine form is associated with increased chiatric conditions. On the other hand. genetic studies of addiction to alcohol. opiates and cocaine and other After administration of naloxone. A common Genetic factors directly associated with addiction guanine-to-adenine transition34 in exon 4 causes the substitution of As noted previously. abuse4. in these children. Genetic linkage and association studies suggest that ferent stages4–6. suggesting that there may be factors also contribute to abuse of or dependence on other drugs of general genetic factors for addiction40. these results point to the studies to be done on thousands of genes using microarray technology. Several cial problems in adulthood as children with the low-expression vari. MAOA metabolizes a variety of neurotransmitters. as subjects. in part. Along with the finding in alcoholism. olism of the catecholamine neurotransmitters dopamine. individuals were administered a µ opioid receptor antagonist. PERSPECTIVE receptor30. 42 chromosomal regions that may be involved in vulnerability to drug abuse in African-Americans and European-Americans. people at a high frequency in the population promote resistance to environmental with the 118G variant receptors had a more favorable clinical response stressors. these polymorphism had similar depressive symptoms and episodes. Although the years of age. condi. as well predictions have been borne out in clinical studies in which healthy as suicidal ideations. those with the MAOA variant Genetic variants may also contribute to opiate addiction. Subjects SERT alleles.com/natureneuroscience by blocking the µ opioid receptor.494-SNP microarray identified axis and drug addiction. One that directs high expression levels were not as likely to develop antiso. all subjects were A/A homozygous are those already discussed as potentially contributing to impulsivity. We will now highlight direct Genotype at this polymorphism may influence HPA axis function. was an early project7. In this study. Another association study investigated why stressful events may or GIRK activation of. nalox. lead to depression in some individuals but not in others37. No other agonist tested showed differences in binding to. whereas an intermediate increase was observed in heterozygous opioid antagonist than did subjects with the only prototype receptor. environment as leading to associations with addiction.to fourfold lower enzymatic activity than may be different influences of environmental or genetic factors at dif- the valine form11. stress response and addiction is point to the critical interaction between specific genetic variants and the COMT. The influence of family and non-family environmental of alcoholism and nicotine addiction. one of Han Chinese and the to the development of antisocial traits. people with the long or short form of the SERT promoter molecular or cellular mechanisms have yet to be fully elucidated. The potential influences of the personality traits of this polymorphism may be involved in several different psychiatric dis. including with opiate dependence2.11. In the latter study. importance of more extensive studies of genetic variants in the HPA A multiple pooling technique with a 1. by disinhibition. This difference in of the SERT promoter polymorphism that affects expression38. as modest activation revisited in light of the recent report of an A → G variant in the long repeat of this axis is desired by at least some alcoholics27. including family environment. depression and stress-responsivity (Table 1). depression and suicidal ideation increased at a much higher heterozygous for the 118G allele showed a greater HPA response to rate.11. a multi-center effort to identify genes involved responsivity is affected by genetic variants.11. influence the polysubstance abuse showed that at least 15 large chromosomal regions development of alcohol dependence in individuals with a relatively high were shared with regions identified in one or more other linkage studies genetic risk. relationship or health stressors from age 21 to 25. This study of environmental factors. the lack of replication in many genetic studies may to treatment for alcoholism with the opioid antagonist naltrexone2. these studies focus on genetic variants and addic- Met/Met genotype have greater increases in plasma ACTH and tive diseases without analyzing the personality traits mentioned above. impulsivity and risk-taking. as well as hydroxylated estrogens. the risk of developing alcoholism or other drug addiction. norepinephrine and dopamine. individual variants and haplotypes at the OPRM1 locus are associated ants36. subjects with the homozygous stimulants. Additionally. have been discussed above.5. These two HPA axis responsivity may be a factor in the possible contribution of specific variants (the MAOA and SERT promoter polymorphisms) are this variant to the risk for developing opiate addiction and alcoholism each associated with alcoholism (Table 1). That is. At 26 ity may be altered in people expressing the variant30. Among maltreated children.31. with the long form of the opioid receptor led us to predict that HPA-mediated stress responsiv. including nicotine and The expression of a genetic predisposition may be. SERT has a The in vitro findings of changes in responses of the 118G variant µ repeat polymorphism in the promoter region. alcohol abuse or addiction. norepineph- rine and epinephrine. The 118G allele of the common serotonin. the resultant other in central Sweden2. These studies A second gene linking the HPA axis. the population-attrib- antisocial behavior was moderated by genetic factors.

as well as of A pharmacokinetic gene product. This enzyme bio- linkage to an addiction with this locus. reports of associations of these alcohol-metabolizing gene variants induced increases in perisynaptic dopamine levels in reward-related with protection from alcoholism are diverse. in a family-based haplotype relative-risk study. rays. such as with cocaine-induced paranoia46. differences in greater opioid potency. population substructure or the use of different assessment criteria. Several important challenges remain for the near future. Meta-analyses focus on ies of addictions. drug-induced neurobiological changes. pharmacotherapy. Phenotypic assessments. variants of genes involved in specific neurotrans- has been reported in Swedish individuals from central Sweden. Other associations of addictive status with µ and κ opioid receptor systems studies focus on the association of genetic variants with personality can be viewed in the context of the importance of these two systems traits such as impulsivity and risk taking. Hence. TPH2 and MAOA and genes new and previously studied genes is of critical value in the discovery of encoding serotonin receptors 5-HT1B and 5-HT2A have all been associ. encoded by the FAAH gene. environmental factors. the refinement of phenotyping in the addictive diseases. the genes have variants that are protective against alcoholism11. SERT. 2). in CNR1 identified a haplotype in an intronic 5′ region of the gene that is ethnic/cultural group studies and statistical methods used must be sim- associated with substance (cocaine. or aggregate patient data. further mitter systems are implicated in vulnerability to alcoholism. with heroin addiction44. but this finding was not replicated in other studies specific pharmacotherapeutic treatments. COMT). and response to morphisms in SERT. whereas amphetamine caused deterioration of cortical the µ opioid receptor. Another association of the OPRM1 118G allele with alcohol dependence As detailed above. have not identified association or CYP2D6. We found that a 68-base repeat polymorphism in reviewed elsewhere11. cocaine major focus for the future could be integrated studies on the role of dependence. the results Fatty amide acid hydrolase. the cytochrome CYP450 gene other polymorphisms in this gene. involved in biotransformation or degradation of alcohol are also We found an association between a single SNP and also a specific implicated11. A func. 2). Alcohol dependence is associated with variants laboratories in phenotyping and statistical approaches (and the shar- of the GABRA2 gene. personality traits (for example. pared. with alcoholism11 and. robust and exhaustively areas of the brain1. TPH1. age of initiation. amphetamine or heroin)47. tional variable nucleotide tandem repeat in the 3′ untranslated region The advent of more modern technologies. The alcohol-metabolizing enzymes alcohol dehydro- © 2005 Nature Publishing Group http://www. possibly owing to differences transforms codeine and several of its congeners into metabolites with in the genetic makeup of the populations under study. A trinucleotide repeat polymorphism in the 3′ flanking particular questions (such as an association between a genetic variant region of the cannabinoid receptor 1 (CNR1) gene is associated with intra. Variants in SERT. the promoter of the dynorphin gene was associated with cocaine abuse Some studies in Table 1 were primarily designed to test for or dependence and also with cocaine-alcohol dependence2. Without such relative standardization. Functional magnetic resonance imaging shows that individuals with the multiple genetic variants that affect these factors may work in high activity valine/valine genotype of the COMT gene have enhanced concert to affect vulnerability and severity of addiction. although these findings have not always been replicated12. tional SNP that alters the sensitivity of the enzyme to protease in vitro is Additional information. is an enzyme that of meta-analyses of apparently similar studies may not be directly com- metabolizes endogenous ligands of the cannabinoid receptors. comorbidity. including phism is associated with polysubstance abuse43. enhances our capacity to study genetic influence on the addictive ated with amphetamine-induced psychosis2 and with alcoholism11. 11 for review). severity of Two studies reported an association of heroin dependence with poly. in A functional polymorphism in the promoter region of DBH that particular. and a personality trait or an addiction) and combine results from mul- venous drug abuse (cocaine. associated with alcoholism8. associated with drug and alcohol abuse50. Addiction is a complex disorder with interacting factors. alcohol or other drug) abuse49. genetic variants of potential interest. genes indicating the importance of ethnic/cultural background33. Summary and conclusions The high-activity Val158 allele of the COMT gene V158M polymor. speed of progression to regular drug use. is available in the Supplementary Note online. Analyses are based on individual coding region SNP in the 3′ untranslated region is associated with symp. vulnerability to opiate addiction and alcoholism with methamphetamine abuse45. Prodynorphin is the precursor to dynorphin peptides. personality traits and stress responsivity. this ing of these data) is desirable to assess more directly replicability and gene is located in a region of chromosome 4p. vulnerability to relapse. toms of delirium in alcohol withdrawal48. Molecular resequencing of (reviewed in ref. and meta-analyses of disparate studies may be misleading. opiate. The CYP2D6 gene is highly polymorphic. such as SNP microar- of DAT (reviewed in ref. ilar to decrease heterogeneity in the combined data.nature. causes lower plasma dopamine β-hydroxylase activity is associated which may focus on clinically relevant aspects of this disorder. 1456 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . and aldehyde dehydrogenase (ALDH) addiction42. a functional SNP in the OPRM1 gene (A118G) influences memory task. as defined by molecular and cellular studies efficiency in individuals with the methionine/methionine genotype (see and in human studies. PERSPECTIVE two Swedish parents32. meta-analyses of studies using The endogenous cannabinoid system is also implicated in genetic stud. Some personality trait variants in addictive diseases. diseases. which codes for the α2 subunit of GABAA. A synonymous tiple studies into a coherent summary. as well as in response to a specific addiction Cocaine-induced psychosis is associated with a potentially func. Clearly. One Alleles of the DRD2 gene are associated with alcoholism. As a concrete prefrontal cortex function when given amphetamine during a working example. The endogenous ligands of the κ opioid receptor that can attenuate cocaine. in defined populations. Some studies of the A118G SNP. TPH2. highly disparate methodologies are difficult. which is linked and generalization across different populations. The specific associations between genetic variation and addictive diseases. resulting in large differences in enzyme activity11. Many of these variants are in the neurobiology of reinforcement and reward by different drugs of associated in independent studies with both addictive diseases and abuse (including opiates and psychostimulants1). psychostimulant abuse or polysubstance abuse2. has been studied in codeine dependence. and results in clinically observable changes in ref. Alleles of the DRD4 and COMT genes also interact stress responsivity. A study of 22 polymorphisms although the use of the latter is more common. A relative standardization across ated with alcoholism11. studies report the DRD4 gene to be associated with opiate addiction or alcoholism.com/natureneuroscience haplotype of variants of the κ opioid receptor gene (OPRK1) and opiate genase (ADH1B and ADH1C). Variants of this gene have also been associ. including references for additional reading. dependence or withdrawal. with the former being the preferred type.

.. Science 301. 34–38 (2004). Gen. 74. F. C. Genes associated with addiction: alcoholism. M.J. G. 161–168 (1997). A tryptophan hydroxylase gene marker for suicidality and alcoholism. Med. et al. Gen. & Naiman. H. DRD4 and novelty seeking: results of meta. Kreek. E. et al.M. M. Med. significant linkages for a cocaine-related trait and cocaine-induced paranoia. 10. & Kreek. K. et al. E.J. 13. et al. Gerber. 933–945 (2000). Horowitz. Pharmacotherapy of addictions.A. J. J. (1994). Sci.J. and association of haplotypes with opiate addiction. et al. Assoc. F.G. Medical safety.) COMPETING INTERESTS STATEMENT 160.. 120–124 (2004). G. A. ase polymorphism alters hypothalamic-pituitary-adrenal axis responses to naloxone: a 7.) 170. B Neuropsychiatr.. opiate and cocaine addiction. 439–442 (1997). Gen.J. Wang. & Butelman. X. LaForge. Identification and characterization of regions. I. et al. Trends Neurosci. Med. Sipe. Nielsen.T. R. Harv.. R. Rodriguez.. Arch.. Association analysis of the DRD4 and COMT genes in methamphetamine 17. associated with low plasma dopamine-β- 18.C.. et al. 46.F. Genomewide linkage scan for cocaine dependence and related traits: for disease gene discovery. Cannabinoid receptor gene (CNR1): Association with IV drug use. 497–499 (2001). Neuropsychopharmacology 24. and the New York State Office of addiction. Psychopharmacology (Berl. Single nucleotide polymorphism in the human mu opioid receptor gene 1. polymorphisms.C. M. & Kreek. Blum. Mutat. C. Zhou. & Walton. 318–324 (1994). Exp. 34. Psychiatry 55. K. & Lachman. B. J. 171–174 (1972).S. analyses. Genet. 1020–1027 (1997). K. T. 41. 221–224 (2002).. 136. Kreek. Hum. (2000). Choi. A. Drug Alcohol Depend. Mol. 96.. 967–972 (1998). Naltrexone the preparation of this manuscript. et al. M. Uhl. Schmidt.I. Arch. 44. R.. National Association for the This work was supported in part by US National Institutes of Health-National Prevention of Addiction to Narcotics (NAPAN)-NIMH. K. P. R.A. 25. 30.. Beutler. Psychiatry 55. et al. 11.S.F. & Cravatt. candidate regulatory 21.A. Issues in association analysis: error control in case-control association studies 39. Diagnosis and symptoms of depression in opiate addicts. & Kleber. 1025. For. environmental risk factors for use and abuse/dependence of cannabis. Res. Sci. Liu... M. receptor. Crits-Christoph.R. et al. 851–854 (2002). D. L. Psychiatry 9.E. 568–575 (2001). Natl. 243–250 (1996). D. Arch.. Hum. Alcoholism and Substance Abuse (OASAS). Genet. Association between dopamine receptor D3 gene BalI polymorphism ity catechol-O-methyltransferase allele is more prevalent in polysubstance abusers. G.F. J.C. 36. with novelty seeking (NS) and substance abuse: the saga continues. K. D. C.S. 45. 473–479 (1999). 916–931 (2004). S. High-activ- 15. Bart.M. F. Acquisition and maintenance of intravenous cocaine self-administration in Lewis tor gene polymorphism in association with heroin addiction in central Sweden.. Mol. An expanded evaluation of the relationship of four alleles to the level muscarinic acetylcholine receptor M2 (CHRM2) gene with alcohol dependence and of response to alcohol and the alcoholism risk. Levin. Psychiatry 39. Rounsaville. Med. Bart.A. Redefinition of the human kappa opioid receptor (OPRK1) structure 14.J. genes. M. 40.A. 5.. Nestler. D. H. 5. Co-occurrence of abuse of different drugs in men: the role of drug. tionships and implications for treatment. V. & Moffitt. Clin. 114.J. 102–105 (2004). et al. J.J.S. Lavoie for assistance in 27. et al. Eur. Bart. LaForge. Oswald. The authors declare that they have no competing financial interests. Am. & Haile. Brain Res. For. Yang. 50. 3. 599–603 (2000). L. et al.S. side effects and toxicity of methadone. Noble. Kreek. 47. Genet. Zhang. Neuromolecular Med. S. P-W. et al. Exp. J. Goldman.H.E. 56–63 (2000). Proceedings of the ACKNOWLEDGMENTS Fourth National Conference on Methadone Treatment. Sinha. N. Specificity of genetic and Neuropsychopharmacology 30. J. et al. Increased attributable risk related to a functional mu-opioid recep- specific and shared vulnerabilities.C. J.T.. opioids and addiction. Res.L. 705–714 (2004). Genet.A.nature. Proc. Schluger. et al. Am. M. Neville. Psychiatry 20. & Wand.1. 417–422 (2005). A haplotype at DBH. also associates with cocaine-induced paranoia. et al. cocaine. Opioid receptor and peptide gene polymor- human molecular genetics of opiate and cocaine addictions and their treatments. 5. Weissman.J. Substance abuse vulnerability loci: converging 12.. D. M. Familial correlates of reduced central serotonergic system function vergence of genome scan results.. J. Clin. Hum. 58.M. Pharmacol. B Neuropsychiatr. Hypothalamic-pituitary-adrenal axis and sympatho-adreno-medul- lary responses during stress-induced and drug cue-induced cocaine craving states. K. 778. Chiang. Association of a CB1 cannabinoid receptor gene (CNR1) polymor- 20. J. D. Rev. Foroud.J. H. Coccaro. Sugden. genome scanning data. 9608–9613 (1998). Psychiatry Genet. Borg. 33.C. 1903–1911 (2004). Tsuang. H. Am. NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1457 . 269. J. Genet. C. Li. and opiates in male twins. Cubells.. Kreek. NIH-NIDA Research Center Grant P60-DA05130. decreases craving and alcohol self-administration in alcohol dependent subjects and © 2005 Nature Publishing Group http://www. II.. Wilber. M. are associated with alcohol dependence and with brain oscillations in the 37. J. 23. phisms: potential implications for addictions. 593–602 (1998). Tsuang.T. Influence of life stress on depression: moderation beta frequency. and cognitive impulsiveness in alcohol-dependent men. sample and further mapping. Pharmacogenetics 6. Ther. Psychiatry 51.C. Lachman. Letsch.. Mol. Variations in GABRA2. et al. Pharmacol. Comings. Hu. Course and relationship to Sci. Role of genotype in the cycle of violence in maltreated children.N. T.J. Psychiatry 1. Rev. Alcohol. K. Y. & Ball. A missense mutation in 22. Human catechol-O-methyltransferase pharmacogenetics: descrip- lucinogens. disorders. 6. Genetic and environmental influences on transitions in drug use. Alcoholism susceptibility loci: Confirmation studies in a replicate preliminary report. Pharmacol. 710–726 (2002). M. 19. sedatives. Gelernter. B Neuropsychiatr. Ott. R. J. Confirmation of an excess of the high enzyme activity COMT val 16. Lilly. 547–549 (2004). 74.A. D4 receptor-selective ligand [3H]NGD 94–1. Natl. in patients with personality disorders. et al. & Smith.S.. Am. 282. 4. Bond. X. 2. 49. DA00049. by a polymorphism in the 5-HTT gene. 232–254 (1996).com/natureneuroscience activates the hypothalamo-pituitary-adrenocortical axis. 246–258 (1995). Ann. Nielsen. Yuferov.. 48. 1–26 (2005). Dopamine D4 receptor gene (DRD4) is associated allele in heroin addicts in a family-based haplotype relative risk study. Vandenbergh. D. 386–389 (2003).J. et al. Hum. Jacobson. D. K. PERSPECTIVE Note: Supplementary information is available on the Nature Neuroscience treatment outcome. Psychiatry 2. Psychiatry 160. G. Kosten. et al. Opiates. O’Malley. 29. et al.M. & Risch. J. McCaul. 171–174 (2004). G. V. Exp. & Crawford.. Mol. E. 32. 29. S. (2005). Pharmacogenetics 14. H. G. 65. haplotypes and association with polysubstance abuse. E. Am. Ho. Am. 83–88 (1993). Substantial attributable risk related to a functional mu-opioid recep- E. Psychiatry rat and human brain by use of the novel. C. 8–16 major depressive syndrome.. Genet.J. & Nielsen.L. Chandler. Uhl. E. Acad. L. 304–306 Am. A. Uhl. 16.J. Exp.com/natureneuroscience/ administration and chronic withdrawal. J. stimulants. 410. NY Acad. T. Human cannabinoid receptor 1: 5′ exons. Pharmacogenetics and 31. M. & LaForge. 1673–1677 (1993). L. Substance abuse vulnerability and D2 receptor Mol. Psychiatry 51. Corticotropin-releasing factor and type 1 corticotropin-releasing factor receptor messenger RNAs in rat brain and pituitary during ‘binge’-pattern cocaine Published online at http://www.J. 9. K. Limosin.. & Kreek.A. 35. et al. T. Am. et al. encoding the α2 subunit of the GABAA 297. A. et al. Mol. K.M.. Mol. and Fischer inbred rat strains.. abuse. J. Caspi. Schinka. Gen. Kendler. Genet. J.S. C. ANKK1: a novel kinase gene closely linked to DRD2 on chromosome band 11q23. 1–13 (2004). 1. hal. 18. Psychiatry 9. Thanks to K.R. 279. Evidence of common and specific genetic effects: association of the 38. 24. Attention deficit hyperactivity disorder and substance abuse: rela- website. Genet. Edenberg. Med. reprintsandpermissions/ 29.. Caspi. & Kleber. tion of a functional polymorphism and its potential application to neuropsychiatric 687–695 (2003). Miller. 8394–8399 (2002). 151–156 (1982). Psychiatry 55. E. 42. Alcohol. USA 99. Ther. G.A. 28. 418–429 (1997). & Neale. For. et al. Primus. 643–648 (2002). Eur. Suicidality and 5-hydroxyindoleacetic acid concentration associ. J. Science 8.. et al. Q. Farren. 43. Trends Genet. B.. 420–425 (2002). 249–268 Pharmacol. Catechol-O-methyltransfer- Behav. J. Genet. M. Med.J. Institute on Drug Abuse (NIH-NIDA) Research Scientist Award Grant K05. J. M. 351–358 Reprints and permissions information is available online at http://npg. Psychopharmacology (Berl. Gen. 85–108 (2004). 45–52 (2005).. Jatlow. Uhl. 6. Genet. 57. M. and alcoholism: a reappraisal. alters beta-endorphin binding and activity: Possible implications for opiate addiction.. Rev. L.. Altered HPA axis responsivity to metyra- pone testing in methadone maintained former heroin addicts with ongoing cocaine General Research Center Grant MOI-RR00102. M. Prescott. Arch. Hered. Proc.R. Sinha.com/ (1996). Krishnan-Sarin. Lusher. Drug Discov. NIH-GCRC 26. human fatty acid amide hydrolase associated with problem drug use. Kreek. tor gene polymorphism in association with alcohol dependence in central Sweden. USA 95. 540–545 (2004). G. Arch. Acad. The A1 allele at the D2 dopamine receptor gene phism with severe alcohol dependence. DeCaprio. 62–72 (2003). Yuferov.nature. (2005).T. 23. D. Nielsen. Psychiatry 2. Miserendino. Gelernter. Localization and characterization of dopamine D4 binding sites in hydroxylase activity. LaForge. Johnstone. J. 793–804 ated with a tryptophan hydroxylase polymorphism.nature. C. 19–29 (2002). Biol. 24. Molecular genetics of substance abuse vulnerability: remarkable recent con- 13. 129. Nat.

subjects receive feedback telling them mechanisms. but interacting. 32 and medial aspect of 11. 12 and 10. and I losses. thus leading to loss of are not aware. including loss of reputation. loss in social standing. emotion and feeling8. Los dala and VMPC are critical for triggering somatic states. This linkage energized a new line of offering you alcoholic drinks and drugs. prefrontal cortex system for signaling pain or pleasure of the Iowa Gambling Task2. or endure sacrifices now in order to obtain benefits later. but there was no laboratory probe to detect and measure for signaling pain or pleasure of immediate prospects. normal operation of the reflective system and for exercising the Early on. In order for somatic signals to influence 1458 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . Both the amyg- Creativity. but emotion. neural systems making and failure to learn from repeated mistakes was obvious in their that control decision making: an impulsive. but as I will Angeles. instead of focusing on one specific process of addiction. normal within the impulsive system can override the reflective system. doi:10. California 90089-2520.nature. even at the risk of incurring future negative disruption of the VMPC leads to loss of self-directed behavior in favor outcomes. to present a broad conceptual framework that brings together several as defined by the Encarta World English Dictionary. outcomes1. and a their impairment. abnormalities in the VMPC region were observed in cocaine willpower to resist drugs. this control is not absolute. thus encompassing Brodmann’s areas The somatic marker hypothesis is a systems-level neuroanatomical and (BA) 25.5–8 for reviews). a gross picture of how multiple brain mechanisms come together in thing despite the difficulties involved. This mechanism enables one to addiction. the decision making consequences. In this task. explain shortly. In contrast. bias or even hijack damage and drug addicts persist in making disadvantageous choices the goal-driven cognitive resources that are needed for the despite the rising losses associated with their choices2. The main purpose is to provide determination and self-discipline that enables somebody to do some.and brain-related responses Antoine Bechara is at the Institute for the Neurological Study of Emotion and that are hallmarks of affective and emotional responses. The patients begin to make choices that often lead to financial drug-use choices on the basis of long-term outcome. Department of Psychology. The key idea of this hypothesis is that the pro- other cognitive functions.) After injury cognitive framework for choosing according to long-term. lower 24. affect and social behavior change cess of decision making depends in many important ways on neural substrates that regulate homeostasis. rather than to this area. the amygdala responds to events that occur in the envi- e-mail: bechara@usc. each with a different potential payoff. they appear obliv. hyperactivity how much money they won or lost. When faced with a choice that willpower to resist drugs. This complements previous proposals that brings immediate reward. propose a neural framework that explains this myopia for future When this syndrome was initially described1. The aim would you do? This is a hard decision. addicts3. of more automatic sensory-driven behavior3. In the back of your mind. subjects choose from four of future prospects. Willpower. but you are the one who will of this perspective is to highlight the key role of choice in addiction. impulse control and loss of © 2005 Nature Publishing Group http://www. with a clear sense of exercising free will. and even loss of family and friends. and ultimately decide. Both often deny. 2. decision-makers learn to avoid decks that yield high immediate gains I propose that drugs can trigger bottom-up. ious to the consequences of their actions. USA. and family. The view I present here is that addiction is a There are similarities in behavior between patients with ventromedial condition in which the neural mechanisms that enable one to choose prefrontal cortex (VMPC) damage and drug addicts. patients with VMPC originating from the amygdala that modulate. These deficits were linked to the decision making impairments of VMPC patients when cocaine addicts were shown to make poor deci- Imagine yourself at a party during the first year in college. What abuse and poor decision making (see refs. that they have a problem. your friends sions on the Iowa Gambling Task4. N E U R O B I O LO G Y O F A D D I C T I O N PERSPECTIVE Decision making. is a combination of disparate lines of research on addiction. involuntary signals but larger future losses down the line. amygdala system everyday lives. After each choice. job. (For the purposes of this piece. the decks of cards. University of Southern California. one specific brain region.edu ronment. or according to long-term outcomes are weakened. whereas the VMPC triggers somatic states from memories. This challenge was overcome by the development reflective. VMPC is defined as the ventral medial prefrontal cortex and the medial A neural system for willpower sector of the orbitofrontal cortex.1038/nn1584 knowledge and cognition. After an individual learns social rules.com/natureneuroscience willpower to resist drugs: a neurocognitive perspective Antoine Bechara Here I argue that addicted people become unable to make completely. memory and short-term. Through this feedback. However. or vice versa. to maximize their reflective system controls the impulsive system via several monetary gain. The term ‘somatic’ refers to the collection of body. you research aimed at understanding the relationship between substance hear the voice of your parents warning you against such activities. Published online 26 October 2005. patients tend to recover normal intelligence. I suggest that addiction is the product of an deficit seen in these patients was puzzling because their poor decision imbalance between two separate.

such as a snake) or a cue once. The reflective system I will suggest. but acquires them with learning.g. memories of a previous experience reactivates the pattern of affective tive/emotional responses through the amygdala system11. numerous and conflict. can be genetic or environmentally induced. and a reflective system. Subsequent evocation of predictive of a fear object.g. a winner takes all: in other words. periaqueductal gray (PAG) and evidence that recalling the experience of a drug reactivates the pattern other brainstem nuclei that produce changes in facial expression and of affective state belonging to the actual previous encounter with that specific approach or withdrawal behaviors1. (iii) memory..nature. associated with drug use. insula. hippocampus (Hip) and anterior systems to modulate cognition and behavior. Provided that representations to the somatic marker framework.g. (ii) triggering of affective states (e. According state belonging to an original experience. automatic and exaggerated autonomic responses thinking that people may be equally vulnerable to addiction once drugs are triggered by cues related to the substance they abuse. Figure 1 A schematic As I will explain. as well as through behavior. Affective reactions can also be generated from recall of personal—or imagination of hypothetical—affective/emotional events. amygdala (A) and VMPC). the amygdala links the features of of these affective state patterns develop normally. hypothalamus and autonomic brainstem nuclei that produce changes This hypothesis is based on evidence from patients with lesions in in internal milieu and visceral structures. the VMPC11. thus allowing the reward value addresses one important question in drug addiction: of the millions of money to be compared with that of food. reflective system Insula ate and future prospects of an option may trigger numerous affective (blue). cingulate (AC). inferior frontal of affective signaling emerges that then can act on appropriate neural gyrus and dorsolateral prefrontal (DLPC). somatosensory amygdala are short lived and habituate very quickly10. After an affective state has been experienced at least an object that induces fear (a ‘fear object’. thereby resulting in exaggerated processing of the incentive individuals who have used drugs from losing control and succumb. they must act on appropriate neural systems.. an overall. values of substance-related cues13. PERSPECTIVE cognition and behavior. mon scale12. but interact- ing. I propose that willpower motor area). Several lines of direct and indirect behavioral to addiction. This mechanism should also bring up the negative consequences Unlike food and water. during the process of pondering decisions. survival of the fittest)9.. This framework that encodes the value of different options. there are several target sites through which somatic diagram illustrating key (affective) signals modulate cognition and behavior. affective state patterns of these negative consequences become repre- dala damage fail to show such responses11. and representing patterns weak ones are eliminated. and I will propose structures belonging to the impulsive DLPC AC that this modulation is in fact mediated by neurotransmitter systems system (red) and the Striatum (Fig. This ascribes a functional role to the ing to addiction.g. more dominant. money does not initially have affective proper. as drug use can induce neuronal changes that lead effects of monetary gains2. the insula and somatosensory cortices). Rather. tion because the process that enables one to inhibit actions elicited by the impulsive system is dysfunctional. such that exposure to monetary ply aversive experiences resulting from the actual consumption of the reward triggers affective signals through the amygdala system. trigger quick. The source of this dysfunction. An emergent (somatic) responses that conflict with each other. automatic and obligatory affec.. a neural pattern for this state is formed. as patients with bilateral amyg. For some individuals. 1). Thus. Affective state The impulsive system patterns develop in brainstem nuclei (such as the parabrachial nuclei) Physiological evidence suggests that responses triggered through the and in somatosensory cortices (for example. family shown that autonomic responses to large sums of monetary gains or losses or finances) and psychological harms associated with drug use. thus suggesting that there may be a common neural ‘currency’ tive/emotional signals of long-term outcomes (Fig. neural systems: an impulsive system. in which the amygdala is a critical neural structure involved in triggering the affective/emotional signals of immediate outcomes. it is consistent with the currently proposed framework of addiction. I argue that before one gets to the stage where a certain evidence have supported the view that conditioned approach behavior pattern of drug use can cause changes to the brain. We have drug. why do only Similarly. of people who drink alcohol or experiment with drugs. The depend on the integrity of the amygdala. fast. this decision making striatum in the motivational and behavioral aspects of drug seeking. These negative consequences are not sim- ties. 5-HT: serotonin. and mechanism is relatively weak. positive and negative signals that are strong are reinforced. ing signals may be triggered simultaneously. striatum and supplementary On the basis of this neural framework. emerges from the dynamic interaction of two separate. drugs may acquire powerful affective and emotional prop- about 10% become addicted? The view I present here challenges the old erties. in which the showing that the brain can encode the value of various options on a com- VMPC is a critical neural structure involved in triggering the affec. the immedi. drug. In other words. and (iv) behavioral actions (e. We have proposed that affective signaling can modulate activity of DA the mechanisms that determine the valence of the dominant pattern of VMPC A Hip © 2005 Nature Publishing Group http://www. lateral orbitofrontal. DA: dopamine. Therefore. reflective systems. This process can be very fast. similar to the are made available. they relate to social (such as trouble with the law. the end result is that dominant pattern of an overall positive or negative signal emerges. or not to use the drug. there is a decision by to drug cues relates to abnormal activity in the amygdala–ventral stria- the person to use.com/natureneuroscience several components affective signaling are consistent with the principles of natural selection 5-HT Ann Thomson of the impulsive and (that is. but stronger ones gain These include selective advantage over weaker ones. sex or other rewards. Such individuals are vulnerable to addic. Over the course of pondering a regions involved in (i) decision. pattern impulse and attention control (e. However. This is consistent with research sented in the brain when individuals learn from parents or society about NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1459 . substrate in the neural system necessary for triggering affective states tional response is evoked through visceral motor structures such as the from recall or from imagination11. 1). however. In addicts. we and posterior cingulate cortices) from prior experiences of reward have suggested that pleasant or aversive stimuli. This mechanism protects most tum system. The affective/emo. it is also reasonable to suggest based on this related structures such as the striatum. and ultimately of affective states (e. the VMPC is a critical the stimulus to its affective/emotional attributes. such as encountering and punishment1.

Although the differences in cognitive impairments brought up one’s mind quickly or have problems concentrating21. Patients with dam- may relate to addiction in general. which are critical for memory11. of the reflective system (Fig. a tendency to live for the moment with no in these inhibitory control mechanisms6. Poor performance a degree of abnormality pre-existed the addiction state. Alcohol. the tendency to make tasks2. as opposed to group. However. frequently under conditions of negative affect22. decision making depends performed like addicts and VMPC patients on the Iowa Gambling Task. However. There may be more than one mechanism by which the reflective insular cortex17. the tendency to experience strong breakdown and financial ruin down the line presents a dilemma to an impulses. addicts who have deficits in working memory also show with the Cambridge Gamble Task26.8. we have obtained prelimi- order to fear their consequences. but another subgroup did period involves the dorsolateral sector of the prefrontal cortex. versus 27% of Normal functioning of the VMPC is contingent upon the integrity normal controls) performed within the range of VMPC patients. My view is that reflects an instance of weakness in this mechanism. One system involves the insula and other whereas the rest performed within the range of the majority of normal somatosensory cortices. Besides decision mak- inferior frontal gyrus18. Individuals with a weakness in show poor performance on tasks requiring the inhibition of pre-potent this process (that is. it is difficult to determine whether these abnor. dorsolateral prefrontal cortex and lateral orbitofrontal/ system exerts control over the impulsive system. and a decision has to be made. the remainder of the addicts making tasks7. Difficulties before engaging in that act22. the Iowa Gambling Task versus patients with VMPC damage showed and vulnerability to succumb to addiction. and surprisingly.14. addicts show functional ity of addicts (the 37% of addicts who performed normally) matched abnormalities in these parietal regions when performing decision normal controls in all respects. maintaining an active representation of memory over a delay SCRs when they pondered risky decisions).25. This pattern of abnormal physiological patients with damage to this structure show compromised decision responses when making risky decisions in addicts was also obtained making14. PERSPECTIVE the dangers of drug use. this mechanism27. reflect instances of Cambridge Gamble and Risk Tasks14. Addicts addict. opioid and methamphet. that are critical controls25. cocaine. that a significantly high proportion of addicts (63%. even before experimenting with drugs.23. those who do not reflect on the consequences of motor responses. and a critical neural region for this mechanism Impairments in decision making are evident in addicts. matched the VMPC patients in all respects (that is. in some manifestations. Thus. thus facilitating their escalation of drug use. any subsequent deficits in this mechanism of impulse control27. showed that this small minor- show impairments in decision making11. that attention to individual.8 responses27.5.23. and functional neuroimaging studies in addicts with their decisions) may be similar to individuals with the personality trait inhibition deficits reveal diminished activity in neural systems involved of ‘nonplanning impulsivity’. These studies sug- systems together. the tendency to act without another drug use episode and the potential of losing a job. as Top-down control mechanisms of the reflective system patients with lesions in this area demonstrate signs of disinhibition and Decision making reflects a process in which a choice is made after poor impulse control11. Patients with using skin conductance response (SCR) measures as indices of affective right parietal damage (encompassing insula and somatosensory cortex) states during performance of the task. A critical neural region for this weakness in this mechanism. rather than the effects of one specific age in these areas make perseverative errors and have difficulties type of drugs.7. Further characterization of these decision making deficits. especially on the right side. these negative consequences should nary evidence suggesting that chronic use of methamphetamine may be be there. Abnormalities have also been detected in white ing. differences in these Several voxel-brain-morphometry studies of brain scans of addicts decision making deficits is the key to understanding the nature of the found varying degrees of structural abnormalities in main components addiction problem. one does not need to use drugs in by the use of different drugs remains elusive. The VMPC region links these addicts who did not match the VMPC patients2. Convergent results have the ability to deliberately suppress dominant. cannabis. The choice between the personality trait of motor impulsivity. and therefore when it is damaged. shifting attention28. and not match the VMPC patients. including the VMPC. such as thinking about decision making processes.16. or the trait 1460 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE .20.com/natureneuroscience for representing patterns of emotional/affective states1. The other system involves the dorsolateral sector of the (the 63% who performed abnormally) had two profiles: one subgroup prefrontal cortex and the hippocampus. 3. and shifting attention to something else. anterior cingulate. However. including the Iowa Gambling Task and the drugs. some of them matched the profile any of these systems compromises the ability to make decisions that of VMPC patients. a tendency to think and reflect on the consequences of an act intrusion of information that is unwanted or irrelevant27. Poor performance on tasks requiring mechanism is the VMPC region. The remainder of the controls were more like the are advantageous in the long term. family thinking21. automatic or pre-potent also been obtained from functional neuroimaging studies (see refs. acting quickly without an intention to act (as for reviews). normal controls. Therefore. by facilitating the on several laboratory instruments requiring response inhibition reflects progress from experimentation to addiction. including alterations of emotional/affective experience.6. However. Damage to and with additional SCR measures. or individuals that lack the trait of ‘premedita. which includes the anterior cingulate and the basal forebrain. A critical neural region excessive and chronic use of drugs can exacerbate these abnormalities. in the case of acting impulsively and using a drug without thinking) malities preceded or were the consequences of drug use. are not uniform across all individuals. My view is poor decision making and abnormal social functioning11. there are other mechanisms of inhibitory control. regardless of appears to be the lateral orbitofrontal and dorsolateral (inferior the type of drug they abuse. for this mechanism seems to be the more posterior area of the VMPC region. For instance. one of which is matter pathways connecting these structures19.nature. 1). © 2005 Nature Publishing Group http://www. which suggests that poor decision making frontal gyrus) regions of the prefrontal cortex. on systems for memory as well as for emotion and affect. Disturbances in this mechanism may relate to amine abusers show impairments in decision making on a variety of the personality trait of ‘cognitive impulsivity’. Disturbances in this mechanism may relate to reflecting on the consequences of that choice. or the trait of ‘urgency’. its prognosis and possible treatment. of other neural systems. A minority of normal controls compromised decision making15. Another mechanism of impulse control is the ability to resist the tion’. but other neural components outlined internal inhibition of intrusive information reflects weakness in earlier are also important11. having poor more harmful to decision making than use of other drugs24. Several tasks are now used to study this inhibiting particular thoughts or memories. mechanisms of decision making renders individuals oblivious to these Direct comparison of the decision making impairments in addicts on negative consequences. there are many gest that decision making deficits in addicts. regard for the future21. they had abnormal Indeed.

such as those made in the gering of affective states (for example. Third.. In contrast. stimuli) suggested that substance-related cues trigger bottom-up mecha- ity in several components of the impulsive and reflective systems.30 for reviews). but we have suggested that the Other studies using tasks in which subjects were required to respond vagus nerve is the most critical11.25. 2003). system in addicts is hyperactive in response to monetary reward. Although addicts Bottom-up influence of the impulsive system show blunted affective responses to affective stimuli that are not drug The reflective system may generate affective states through top-down related34. Ascending signals from this overall affective state can the majority of normal controls. combined with SCR measures. The proposed arrangement provides a way for affective signals to future) engages more anterior VMPC (such as frontal pole)9. changes in representation of patterns of affective states (for such as motor impulse and attention control38. the SCR defect is specific to the anticipatory phase when they dopamine. that synapses on cells and/or terminals throughout cortex.25 for reviews). Soc. complexity the signals triggered by the impulsive system were relatively strong. the basis of the finding that major advancement in the size. Rather. Addicts show deficits in this mechanism of Hyperactivity in bottom-up mechanisms of the impulsive system can impulse control. 1). Abstr. neu- NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1461 . inferior frontal gyrus and dorsolateral prefrontal. The control function of the reflective system is complex. over the impulsive system.com/natureneuroscience mechanisms. Anatomically. nisms in substance abusers. into simpler component processes of decision lower the threshold for triggering subsequent affective signals related making. Van den Wildenberg & A. 2. 33. Using different versions of the be triggered simultaneously through both the impulsive and reflec. The result is emergence of an overall positive or negative subgroup of addicts that were different from both VMPC patients and affective state. several factors can modify the strength of affec- the molecular mechanisms by which neurotransmitters affect synaptic tive signals triggered by the reflective system. impulse control31. noradrenaline and acetylcholine are located in the are pondering which option to choose11. to monetary gains (which are more frequent in men than in women6) and anterior cingulate) can strengthen thoughts about drugs and thus providing indirect evidence for the hypothesis that the amygdala make shifting attention to other thoughts more difficult. C.A. Cutshall. One of the component processes is the tendency of a subject to drugs. the ability to remain focused on a task that may Hyperactive impulsive system be boring or difficult22. Neurosci. Direct autonomic responses to wins and losses are blocked bottom-up signals modulate synaptic activity at telencephalic targets in patients with bilateral amygdala damage. Iowa Gambling Task. the axon terminals of these neurotransmitter neurons make suffer from the opposite condition of amygdala lesion patients. serotonergic. when one is pondering Task2. reward—could be implemented in the VMPC9. Finally. to targets (drug-related stimuli) but not respond to distracters (neutral Changes in neurotransmitter release can modulate synaptic activ. E. Perhaps money represents a special case. Therefore. W.nature. 1). influencing top-down cognitive mechanisms First. If. Addicts trigger exaggerated autonomic responses to cues related to the substances they abuse (see refs. Thus. in VMPC is pharmacological9. thus influencing its control activity in addictive states and that explain how these activities influ. Second. This is supported by both the amygdala and VMPC have direct access to these neurotrans. in lateral orbitofron. signals to these neurotransmitter nuclei.M. changes in regions involved in behavior (striatum and supplementary motor area) can translate into drug use (Fig. a number of channels can convey their response generalizes to monetary reward37. numerous affective signals that conflict with each other may automatically linked to buying drugs. Furthermore. hippocampus. or noradrenergic systems to explore the neural basis of these factors.427. mainly because there are many excellent reviews that describe normal circumstances. alterations in impulse control and the inhibition of to pay more attention to gains or losses encountered on previous trials in unwanted memories or thoughts (for example. and attention bias. ral framework for how factors that affect decision making—such as the main purpose here is not to detail the processes and mechanisms time delay. they and connectivity of the frontal lobes in humans has occurred in relation would have the capacity to hijack the top-down goal-driven cognitive to Brodmann area (BA) 10 (that is. This suggests that addicts brainstem. serotonin. Indeed. to reward. This is on exert a bottom-up influence on the reflective system. tions leading to hyperactivity in this system include hypersensitivity. is. functional neuroimaging studies showing increased amygdala activity mitter cell bodies in the brainstem. in the insula and other somatosensory cortices) can lead to studying these attention biases has been to use cognitive models6 that an increase in the reward utility of the drug. We have suggested trate (i) how one can relate molecular and pharmacological studies that information conveying immediacy (the near future) engages on drug addiction to neural systems concerned with mechanisms of more posterior VMPC (including anterior cingulate. decision making research is how humans assign value to options. Modulating factors The outline of these pharmacological systems given here is very sim. irrespective of the losses that were reflective systems (Fig. in amygdala and VMPC) can Iowa Gambling Task. this subgroup of addicts was drawn then modulate activity of several components of the impulsive and to choices that yielded larger gains.32 and time delay33. Crone. order to make future decisions. Addicts show patterns of high attention tal. and nucleus accumbens). The cell bodies containing the neurotransmitter patients. one of the fundamental questions in ence cognitive systems such as memory (see refs.P. for instance. Evidence suggests that condi- requiring the internal inhibition of an intention to act28 (E.25. we identified a tive systems. PERSPECTIVE of ‘perseverance’. but then ascending signals from these affective states can responses when exposed to monetary reward in the Iowa Gambling exert bottom-up influence on cognition. deconstruct complex behavioral decisions. resources needed for the normal operation of the reflective system and the more posterior areas of the VMPC (such as BA 25) are directly exercising the willpower to resist drugs. and research has begun specific roles of dopaminergic. changes in trig. 29. their amygdala is overresponsive to reward. basal forebrain affect and emotion and (ii) the influence of drug addiction on cogni. we have shown that addicts trigger exaggerated autonomic © 2005 Nature Publishing Group http://www. whereas information conveying delay (distant tion. in that it may be a decision. the frontal pole)39. connected to brain structures involved in triggering (autonomic. For affective states and homeostatic in response to drug-related cues35. as they demonstrate poor performance on tasks weaken control of the reflective system. encountered. Recknor. We have proposed a neu- in decision making. and they generated exaggerated SCRs when they won We have previously proposed that the key mechanism by which these money2. Other excellent lines of research have attempted to differentiate the Several factors affect the value of a choice.B. the goal is to illus. Another approach for example.36 and that this exaggerated brain signals generated in the body. and even under plistic. the probability of the outcome or the tangibility of the of any one specific pharmacological system.

and repre. unless their choices bring increasing social. and they provide testable hypotheses that could be addressed ends up on this down-spiraling path. This raises the Thus. my proposal is that not every individual who tries drugs prognosis. Therefore. One subgroup of addicts appeared normal and did not show behav. later rewards23. drug withdrawal can be viewed like hunger43 in that once it is ing adolescence. It follows that cou. in that it does © 2005 Nature Publishing Group http://www. and the control over powerful temptations. and (iii) quality of mechanism for processing probabilities is also relevant to addiction. but in the systemic and longitudinal studies on decision making in young adoles- profiles of their physiological responses. as choice. These differences may have implications for tion. sooner rewards over larger. which then leads to behaviors reflecting uncertainty) that they have about receiving their drug at the end of a poor decision making. tion of their everyday lives has shown that they have suffered minimal COMPETING INTERESTS STATEMENT social and psychological harm as a consequence of their drug use: for The author declares that he has no competing financial interests. PERSPECTIVE rotransmitter nuclei) or representing (sensory nuclei in the brainstem. is still taking physical withdrawal signs are neither necessary nor sufficient for taking place46–48. Therefore. people described as addicted to coffee. as opposed to specific. and instead they are guided by immediate prospects. the serotonin transporter gene) (ii) the able options (see ref. This level of neural activity in specific neural circuits. brain scanning session3. as addicts not capture all instances of addiction. mation conveying lower certainty engages anterior VMPC9. 7 for a review) are supportive of this view. the have evolved to subserve natural motivational functions. This suggestion is consistent with the such a time. they prefer smaller. drug addictions44. However. described these addicts as ‘functional’ addicts. thereby increasing maturity could be harmful to decision making. However. and strong affective sweets. mere exposure to drugs to become addicted. they are oblivious to future positive or negative that describe a progressive dysregulation of reward brain circuitry con- consequences. other Finally. comitant with a spiraling path from controlled drug use to addiction49 Addicts who partially match VMPC patients are suggested to be hyper. This suggests The research described in this article was supported by the following grants that not every drug user has impaired decision making. This a limitation of the proposed somatic marker framework. and those with normal decision mak- time recovering from addiction and remaining abstinent in comparison ing capabilities are more resistant. persist in escalating their drug use in the face of rising adverse conse- pling of information to representations of affective states via posterior quences. prefrontal cortex may not develop fully until the age of 21. example. the fact remains that not capacity of bottom-up homeostatic signals to hijack control mecha. an alternate possibility is that the lack tional imaging studies addressing how the perceived delay to receiv. exposing the prefrontal cortex to drugs before its drugs. whereas the signaling via more anterior VMPC is relatively sion making. for addiction than individuals with normal decision making to influence decision making and hijack cognition in the direction of capabilities. Future Most addicts show behavioral signs of poor decision making. tend to exhibit a higher temporal discounting rate than normal people. According to this view. Finally. it increases the utility of drug reward. individuals to multiple. every adolescent who tries drugs ends up addicted. that is. Functional Future research using functional imaging methods could focus on imaging studies implicating the parietal cortex and anterior cingulate relationships between (i) genotypes related to specific neurotrans- cortex in computing the probability of outcomes on the basis of avail. or at higher now as opposed to the delayed consequences) have a stronger capability risk. Given the view that neural systems supporting drug reward for neurotoxicity resulting from drug use remains questionable45. clear predictions that can be addressed in future research. not all predisposing factors are necessarily genetic. DA12487. reward values are processed by the VMPC region. such as feed. This suggestion is reasonable in light of the evidence that short-term outcomes. general fashion (such as the serotonin transporter gene) to predispose tainty (or higher probability) engages posterior VMPC. 1462 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . Indeed. so that the prospect of drugs outweighs the prospect hijacking of their willpower as they move from controlled use to addic- of future consequences. effortful and weak. one predisposing factor to addiction is heredity. rather. and genes can act in Similarly. VMPC is associated with relatively fast. they manage to keep their jobs2. as shown by complex laboratory tasks of decision making.nature. the fetus to drugs during pregnancy. just like exposing the the motivation to use drugs42. and some do not (see above). Therefore. physiological profiles that matches VMPC patients. Thus in the presence of withdrawal. effortless. We have from the US National Institute on Drug Abuse (NIDA): DA11779. cocaine addicts show abnormalities in the activity of neural structures This will reveal whether genetic factors lead to suboptimal function critical for decision making in proportion to the degree of certainty (or in specific neural systems.com/natureneuroscience discounting mechanism of time is also relevant to addiction. However. However. it takes more than nisms of the reflective system is increased. and should examine whether drug users undergo a slow and gradual sensitive to reward. my view is that insular and somatosensory cortices) affective states. of evidence for decision making deficits in this subgroup of addicts is ing a reward modulates activity in reward-related brain areas33. mitter systems (for example. they exaggerate the incentive impact of drugs. These are testable hypotheses with with addicts who partially match the VMPC patients. This view is supported by recent func. some addicts match VMPC cents should test this hypothesis and determine whether neurocognitive patients. the internet and so on do not necessarily have impaired deci- signals. it influ. we have suggested that information conveying higher cer. physical or slowed. who match VMPC patients are characterized by insensitivity to future This research should also take into consideration models of addiction consequences. We have suggested that addicts development can serve as a marker predictive of addictive disorders. whereas access of poor decision making in addiction is evident only when individuals more anterior areas is polysynaptic and indirect40. because a closer inspec. poor decision making is what leads to addiction. those with poor decision making in future research: addicts who match VMPC patients may have a harder capabilities are more vulnerable. the development of neural connections that underlie deci- incentive motivational view of drug addiction proposing that although sion making. in doing so. or the sentations of these values are modulated by homeostatic factors such as product of gene-environment interactions. and. events that are more immediate in time (such as having the drug question of whether these individuals are predisposed. ACKNOWLEDGMENTS ioral or physiological signs of decision making deficits. factors could be environmental (such as drug neurotoxicity). psychological harms. potential for harm remains relatively higher if drugs were abused dur- ing42. and DA16708. one subgroup of normal controls shows behavioral and that is. evidence suggests that the functions of the present. and until ences the decision to use drugs. Although the evidence hunger41. whereas infor. my hypothesis is that poor decision making in addiction is not the product of drug Implications for treatment and directions for future research use.

G. 1999). J.D. Everitt.H.. A. Brain systems mediating aversive 36. A.. 11. Clarke. PERSPECTIVE Published online at http://www. A.D. Inhibition and the right inferior frontal cortex. Psychiatry 156. Eslinger.A. et al.F. L. 265–284 (2002). Whiteside. Decision making and drug of choice in substance-dependent Reprints and permissions information is available online at http://npg. E.Z. K.M. Psychiatry 159. Bechara. T. 25–40 (2005). E.. Neuron 36. et al. neurocognitive and clinical perspective. & Bhagwagar. Experience of emotions in substance abusers exposed to images containing neutral. Patton.J. A. Effects of beta-adrenoreceptor Sci. The addicted human brain viewed in the light of Trends Cogn.C.T. E. Psychol. Neurobiological constraints on behavioral tissue composition in abstinent cocaine abusers: a magnetic resonance imaging study. M.W. Sanchez-Barrera. J. after prefrontal serotonin depletion..J. 6. & Volkow. & Roberts. 41–53 (2004).R. and negative affective stimuli. Games Econ..P.. Am. Franklin. E. Neuron 12. 13. A. P.. functions: A latent variable analysis.R. W. 85–114 (1997). M. Lancaster. & Wang. 40. Science 278.. (2004). Int. Disturbances of emotion regulation after focal brain lesions. P. N.P. A.. The five factor model and impulsivity: using a structural model (2004).C.P. Retraction. Stud. Rahman. Montague. Neuropsychology 17. Laibson. J. J. S. & Friston. D. & Van Hoesen. Addiction: A disease of learning and memory. and the Human Brain (Grosset/Putnam. Yuan.J. 195–201 (2005). 205–216 (2001). T. Grant. Factor structure of the Barratt impulsiveness 47.. Drug Alcohol Depend. Morris. Functional imaging of Neurobiol. E.. Childress. T. Reduced frontal white 46. (2004). N. 23–51 (2003). Bechara. 28. Clin. Rev. P. Prefrontal 14. N. & Childress. A. J. Cognitive inflexibility and neuropsychiatry. Kringelbach.. 62. X. 51.A. 1180–1187 (2000). 877. J. A. monkeys and humans. S. Matochik. 768–774 (1995). 283–288 (2003).J.D. cortex: Evidence from neuroimaging and neuropsychology. 878–880 (2004). Cognitive performance and autonomic reactivity in abstinent drug 1. R. The neuropsychology of ventral prefrontal cortex: deci. Wise. J. tion. Ricaurte. 133. 619–639 (2001). & Weinstein. imaging studies: brain circuits and treatment strategies. J. S.. Cord.D. Bechara.A. Frontal cortical 43. processing as reflected by phasic heart rate changes. Clin. Fishbein. (2002). Sahakia. Biol..D. Breiter. & Le Moal. K. Associative processes in addiction and reward: the role of amygdala control and of shifting associated with cognitive bias in polysubstance abusers with and ventral striatal subsystems: the role of amygdala-ventral striatal subsystems.W. 170–177 (2004). A. B. A psychomotor stimulant theory of addiction.A. Neurobiology of decision making: risk and reward. Ernst.W. Clark. & Bozarth. B. Neural economics and the biological substrates of valua. R. 72. Bechara. Molecular basis of long-term plasticity underlying addiction... 94. 11–18 (1999). & Miyake. Garavan. 669–689 (2001). et al. K. 38. J. 159–193 (2004). Robbins. and temporal cortices of cocaine patients. Damasio. Z. Schleicher. Ernst.S. conditioning: an event-related fMRI study. 503–507 (2004). 469–492 (1987). J.) Psychiatry 51. J. 7. Neurocognitive insights into substance abuse..com/ individuals: a preliminary report. Impaired decision making related to working memory deficits 40. AJNR Am. 15. Biol. A. Peralta. tion. G. E.. A. G. A. Baltimore. Verbanck. Neuropharmacology 47. 52. making tasks in cocaine-dependent patients: Do they measure the same construct? 49. describing. The functional neuroanatomy of the human orbitofrontal HIV-seronegative polydrug abusers. Psychol. Drug addiction and its underlying neurobiological basis: 34. 152–162 (2004).H. & Poldrack.R. 1825–1837 (2001). 78. K.D.J. & Price. Crone. Neurobiol.. Drug abuse: Hedonic homeostatic dysregulation. 206–219 (2000). Overman. Ongur.. 19. Cereb. Descartes’ Error: Emotion. models of motivation. Delayed nonmatch-to-sample performance in HIV-seropositive and 41. Prog. U. Behav. in Attention. Neuroimage 19. A. G. Gen. Developmental change in feedback scale.. 52–58 (1997). Kahneman. 964–965 (1998). The somatic marker hypothesis: a neural theory of economic 35. H. Martin. 10. Addiction 96. & Bechara. Sci. Eldreth. Sci.A. G. 48. Arch. 9. 9. I. D. J. Brain Cogn. Neuropsychology 18. Am. G.S. Nestler. 1479 (2003).D. & Perez-Garcia.S. 1095–1102 (2003). C. Risky business: Emotion.M. & Paulus. J. 1302–1309 (2005). T. Volkow. 44. A. R. Am.C. Brooks.W. Robbins. Exp. A. value immediate and delayed monetary rewards. H. Bartzokis.M. & Lynam..F. 16. Hyman. 20. Rogers. 271–277 (2001). Decision making 31.2005. Psychiatry 55.. Neuroimaging evidence for the involvement of the frontal cortex. R. Biol. A. Psychopharmacol. Monterosso. 101–135 (2004). Ehrman. M. J. Limbic activation during cue-induced cocaine craving. alcoholism. 1628–1635 (1999). 97S (2003). Cortex 10.D. Science 304. Deficits of inhibitory 13. McClure.. D. 336–372 (2005). D. J. Psychol. N.com/natureneuroscience/ 24. Memory and Executive Function (eds. Neuropsychologia 38. Psychiatry 162.. The relations among inhibition and interference control 2. 8. Aron.. Neurobiology of decision making: a selective review from a 157–164 (2004). 159–167 (2004).. 1228–1238 22. M. M.. & Bergen. 20. & Bechara. Conceptualizing. Bechara. J. 30. B. Trends Cogn. K. Zilles. Semin. Cortex 10. J.R.L. Cools. of personality to understand impulsivity. Lim. decision making and addiction. R. Grant. R.com/natureneuroscience 4. 8. Pelc.. Annu. sion making and reversal learning. 30. O’Brien. Nader. Martin. positive. 27.R. Separate neural systems 2005 (doi:10. 334–342 (2000). neural responses to expectancy and experience of monetary gains and losses. 3. Three decision Neuropsychologia 42. M.E.nature. J. Dolan. 39.P. 37. 1642–1652 (2002). M.. G. Orbitofrontal cortex and decision.. Hatzidimotriou. et al. Psychol. General and specific inheritance of substance abuse and 19. & Barratt. Psychiatry 53. blockade on components of human decision making. & Martin. Van Der Linden. Am. New York. Crofts. & Rolls. T. D.. M. Reason. K. Wakeley. Exp..I. H. 224–241 (2001). S. frontal cortex of rats. & London. Clin.O. A. C. Rev. 114. E. 119–128 (2001). Cereb. & Rotorsen. 32.S. Performance on the IOWA card task by adolescents and adults. 947–957 (1998).004). reprintsandpermissions/ 25.biopsych. orbitofrontal. Napier. M. M. cortex in humans and apes: A comparative study of area 10. 2. abusers and nonusers. Ann. & Krasnegor. E. D. & Berns. Dalley. N.W. human drug abuse: functional imaging. 3–13 29. et al. laboratory test of decision making. Dif.L. & Van der Molen. J. 26. Contoreggi. Sci.. 1994). cingulate. Biol.A. Psychol. S. 55. 134–142 42. The incidence of T2-weighted MR imaging signal abnormalities in alcoholism.J. Lyon. A. Goldman. Verdejo. Fowler. & Bolla.) 172. Wolkin. Psychiatry published online 10 August 33. S.R. & van der Kooy. Gambl. Neuropsychiatry 6. Rev. 23.L. Psychiatry 51. Noel. 420–441 (Paul H. Drug abusers show impaired performance in a 30. N. P.. & Robbins. 5. Dev.J. K.. A. D.W.R. & Cohen. London. M. Dale. M. Semendeferi. 48. F. D.R. Bonson. Büchel. Phys. R. Loewenstein. Cadet..D. NY Acad.. & Robbins. The organization of networks within the orbital and medial pre- in substance addicts. Jennings. Choi. NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1463 .. K.R. E. & Damasio. H. Koob. Pomara. 1838–1851 (2004). Goldstein. R.. 890–895 (2002).. Bechara.J. et al. London. and measuring components of executive func- matter integrity in cocaine dependence: a controlled diffusion tensor imaging study. Nat. Armstrong. J.06. Rev. Neuroradiol. et al.nature. & Stout. Science 301. G. C. Decreased gray matter concentration in the insular. 6.S. Neuron 20.A. © 2005 Nature Publishing Group http://www. Rogers.A. 18. Psychopharmacology (Berl.C. J.nature. the brain of cocaine-dependent patients is age-related and region-specific.. J. Rudolph. E. Anthropol. A. Science 306. & Shizgal. Gen.I. Pers. 412–438 (1999). I. Friedman.1016/j. Neurosci. 21. & McCann.I. Stanford. S.. 45. G. 341–372 17. Trends Cogn.. 1414–1422 (2005). A.. E. M. Individ. Addiction 100. Aguilar de Arcos.. Aharon.

The conditioned association of the NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1465 . In addition. cues for smoking are almost unavoidable parts of smokers’ lives. but fewer than 10% succeed. Once addicted. they often experience more severe withdrawal symptoms. tine is administered (place preference). Austin. particularly vulnerable groups within the countries1. Smokers adjust their dose precisely to avoid discom- individuals may be more susceptible to nicotine addiction because fort while achieving the most desirable impact5. Nearly one-quarter to cigarettes within weeks of leaving the hospital. report pleasure. The behavioral conditioning occurs more frequently. such as wak- John A.6. death resulting from tobacco use continues to rise. the cigarettes that are smoked12. and R. Nicotine is a mood leveler in humans and other animals. the main addictive component that motivates continued tobacco use synaptic changes and neuroadaptations to nicotine occur after only despite its harmful effects4. nicotine is they establish a routine of smoking on a monthly basis. US. In the smokers wish to quit and try repeatedly. These drug delivery system. but the addic- those hospitalized for other serious smoking-related illness relapse tion process proceeds quickly in adolescents5.tmc. USA. initiates synaptic and cellular changes that underlie the motivational and behavioral alterations that culminate in addiction.000 deaths and $50 billion in medical costs3. The addictive Smoking begins in adolescence power of tobacco is exemplified by the difficulty in quitting4–6. exposure to smoking often highlights the aversive impact21–23. e-mail: jdani@bcm. and nicotine exposure in adolescent rats increases self- Published online 26 October 2005. as demonstrated by intravenous brain stimulation and reinforces preference for the place where nico.10. Nicotine addiction progresses rapidly in adolescents and is most highly expressed in vulnerable people who have psychiatric illness or other substance abuse problems. Therefore. Psychiatry & Behavioral Sciences. of the adolescents report symptoms of addiction at about the time Of the roughly 3. adolescents experience aspects of dependence after only a few cigarettes. drug.1038/nn1580 administration tested later in life24. and about one-third ers attempt to quit each year. In young rats. relaxation. and overall population consume a disproportionately high fraction of all about half of those who do not quit eventually die from smoking. Thus. self-administration and conditioned place preference19.6–10.11. adolescent rats show hypersensitivity tobacco. more than 60% of young people try smoking. relief from hunger and eventually relief from withdrawal it more difficult to quit. Despite to half of them become daily smokers15. the drug-tak- ing behavior is associated with common events of the day. and it supports self-administration. the main addictive component of tobacco. reduced anxiety. Nicotine is addictive in the absence of one exposure17. related diseases2. smokers tobacco provides desired positive mood influences14.000 ingredients in cigarette smoke. and there is a particularly strong correlation causing arousal during fatigue and relaxation during anxiety. About one-third of smok. between smoking and abuse of the other most commonly abused Smoking is a learned (conditioned) behavior reinforced by nicotine. Baylor College of Medicine.16. Dani is in the Department of Neuroscience. Tobacco use is most highly prevalent and is more intense in psy. N E U R O B I O LO G Y O F A D D I C T I O N REVIEW Nicotine addiction and comorbidity with © 2005 Nature Publishing Group http://www. 50% of heart attack survivors and of cigarette consumption escalates over a couple of years.nature. the associations that become Addiction Research. Smoking commonly begins during adolescence. Nicotine. More severely dependent drinkers smoke more and Cigarettes are excellent vehicles for the conditioning because the dosing via puffs is precise and repeated very often5. however. Menninger Department of ing in the morning. imperative medical reasons.edu In summary. A great majority of those who abuse other symptoms5. report fewer aversive effects and more positive effects than adults chiatric patients and drug abusers12. smoking annually causes over 400. enhances reward from to the reinforcing actions of nicotine.6. Austin. Adolescents. Furthermore. substances also smoke. University of Texas. improved attention. for cigarettes than 77030. Houston. It also produces a withdrawal Tobacco can have positive effects on behavior and mood. The comorbidity with mental after their first smoking episode16. doi:10. USA. Tobacco use is the leading cause of preventable death in developed are less likely to quit.18. Most The vast majority of those who initiate tobacco use are young. alcohol. arousal. cigarettes are an ideal illness is particularly high for schizophrenia and depression. In those who initiate smoking. Adron Harris is at the Waggoner Center for Alcohol and for any other addictive drug.13. Furthermore.com/natureneuroscience alcohol abuse and mental illness John A Dani and R Adron Harris The World Health Organization estimates that one-third of the global adult population smokes. Because tobacco use is on the rise in developing countries.20. Texas and is associated with more common everyday events. making relief from stress. Texas 78712. but the first syndrome that is relieved by nicotine replacement6–8. In the United States alone. Furthermore.

The PPT and LDT contribute to events associated with drug the prefrontal cortex as well as limbic and striatal structures.29.10. desensitize (Fig.18. blocking DA release in the NAc with antagonists or lesions is mainly onto DA neurons that project back to the cortex. nicotine influences cellular events and glutamatergic projections from the prefrontal cortex and GABAergic produces neuroadaptations in many brain areas that are directly or projections from the NAc and ventral pallidum10.37) subunits. Microdialysis studies in rats show. as shown in rodent studies. Because The VTA receives massive convergent afferent inputs. and GABA interneurons decrease their inhibition (downward arrow) onto VTA DA neurons. Rather.29. source of innervation into the midbrain DA areas arises from the nearby Although many areas of the brain participate. lung and brain5–7. tegmentum) provides excitatory drive onto GABAergic interneurons. For excitatory glutamate input from the prefrontal cortex.29. not to the attenuates the rewarding effects of nicotine. Another major indirectly important during the addiction process23.10.8.39. whereas ing and persist for hours.30. the mesocorticolimbic pedunculopontine tegmentum (PPT) and the laterodorsal tegmentum dopamine (DA) system has a vital role in the acquisition of behaviors (LDT). The nAChRs on the VTA DA neurons are mainly composed of α4 (ref.40. however. the PPT and LDT provide the main glutamatergic excita- self-administration11.27.28. (b) The more prolonged presence Upon smoking.7.enhances glutamatergic afferent excitation onto DA neurons while tive learning of adaptive behaviors as an animal continually updates nicotine concentrations are elevated10.26. DA may participate in the ongoing associa. smoke.29.31.39.18. as indicated by reduced NAc41. Thus. The PPT projects mainly ing nicotine4.com/natureneuroscience Nicotine obtained from tobacco reaches the brain in 10–60 seconds. A wide range of evidence supports the nicotine self-administration43. particularly of subtypes nAChRs throughout the brain. only α7 subunits commonly form homo-oligomeric nAChRs.tion to the DA neurons projecting to the NAc (Fig. and is initially at a Ann Thomson LDT/PPT Glu ACh LDT/PPT Glu ACh concentration of roughly 100–500 nM in the arterial blood.9. as shown by the observation that lesions in the PPT reduce the nucleus accumbens (NAc). depolarization of the DA neurons. REVIEW a Nicotine initially b Nicotine later and memory9. As a consequence.30.32–34. Thus. 1b). (a) Initially.nicotine elevates DA in the NAc for hours8. Thus. which decreases Mesocorticolimbic dopamine system or terminates the direct stimulation of the DA neurons by nicotine32. more sophisticated theories explain the mounting data that (α7*) nAChRs have a relatively low affinity for nicotine. ing that addictive drugs (such as cocaine. and in the mammalian brain longed DA signal in the NAc. The cholinergic input from the LDT/PPT (laterodorsal tegmentum and pedunculopontine in the body (and brain) as the day progresses. nAChRs composed of α7 subunits10. α4β2* α4β2* α4β2* α4β2* – – initiation and execution of behaviors that serve VTA VTA DA GABA DA GABA a beneficial goal8.with GABAergic and glutamatergic neurons42. addictive drug with common daily events motivates progression along 35) and β2 (refs.17. This coincidence of presynaptic deliver a small pulse of nicotine each time they glutamate release and postsynaptic firing increases the likelihood of synaptic potentiation (for example.9. α7* α7* Nicotine influences synapses of the VTA © 2005 Nature Publishing Group http://www. Subsequently. includ. heroin and amphetamine) Glutamatergic afferents onto DA neurons commonly have presynaptic elevate DA in the NAc8.nature. 1466 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE .41. 1)42. in particular. After a few minutes. the activity of presynaptic α7* nAChRs cation of reward. the high-affinity α4β2-containing (α4β2*) receptors. that information is used in the preparation. Clearly. Although the VTA receives a strong role of the mesocorticolimbic DA system in nicotine addiction4. DA concentrations in the NAc are not a direct indi. Although the nAChR those on VTA DA neurons (Fig.42. The elimination half-life of about 2 hours allows Figure 1 A simplified illustration of several synaptic connections and nicotine-induced events that control nicotine to accumulate with ongoing smok- DA release in the NAc. and the LDT projects mainly to the in the ventral tegmental area (VTA) of the midbrain and projects to VTA. nicotine causes some activation of most nAChR subtypes. which are a loose collection of cholinergic neurons interspersed that are inappropriately reinforced by psychostimulant drugs. The enhanced presynaptic a construct of environmental saliency. including taking27. the low concen- contradict the simplest notions about the rewarding properties of trations of nicotine achieved by smokers do not strongly desensitize the DA8. in combination with other nAChR the path to daily cigarette use and spurs relapse during abstinence. and nicotine accumulates and lingers LTP). An important dopaminergic pathway originates to the substantia nigra compacta. nicotine initially activates of nicotine causes some desensitization (indicated by nAChRs in white text). Nicotine binds selectively to nicotinic acetylcholine receptors (nAChRs). The active presynaptic α7* nAChRs enhance glutamatergic excitatory drive (upward arrow). Neuronal nAChRs are pentameric. subunits38. including containing the β2 subunit.29. The hypothesis suggests that glutamate release is paired initially with the increased firing of the post- addictive drugs act upon mechanisms that normally underlie learning synaptic DA neurons caused by α4β2* nAChRs before they desensitize. 1a). 36. containing combinations of circuitry that controls the firing of VTA DA neurons produce the pro- α and β subunits or exclusively α subunits. Synaptic changes in the line25. causing an increase in burst firing and overall firing rate10.9. direct nicotine excitation of VTA DA neurons ceases. including nAChRs are widely distributed. The distribution half-life of approximately 8 min- utes dictates the initial actions of nicotine. that excitation example. smokers often active α4β2* nAChRs directly excite DA neurons (upward arrow). they are the main subtypes mediating nicotinic influence induced activity of nAChRs produces a direct at the given locations.α7* nAChRs (Fig. Rather. Because α7-containing Recent. That result is consistent with the general find.28. 1)32. Nicotine- subtypes are not perfectly segregated as shown.6–8. the functions of DA nAChR activation Some desensitization within the overall mesocorticolimbic system NAc β2* NAc β2* involve the learning and integration of salient environmental information. that a single injection of which are ligand-gated cationic channels that normally bind acetylcho.40.

and nAChRs influence those sensory events13.com/natureneuroscience tion from GABAergic interneurons in the midbrain. norepinephrine and those nAChRs desensitize. 1b). smokers deliver a small pulse of nicotine with each fort come to elevate reward thresholds on their own58. Additional influences on reward.38–40. depression and anxiety and DA release. However. The striatum is richly innervated throughout by cholinergic phrenic patients. chronic nico- postsynaptic DA neurons is sufficient to produce LTP when paired with tine use upregulates µ-opioid receptors and alters the transcription the boosted glutamate release caused by presynaptic α7* nAChRs. nico. It is hypothesized that among the tion removes the direct excitation caused by nicotine and decreases neuroadaptations induced by chronic nicotine. nicotine addiction. By itself. which in turn influences the long- excitation of DA neurons. A fundamental role of does not improve most symptoms65.40. nicotine desensitizes the α4β2* nAChRs on the rats57. 7% of DA terminals (Fig 1b). thereby altering DA signaling. altered mGluII receptor the endogenous cholinergic drive onto the GABAergic interneurons function decreases glutamate levels and contributes to the discomfort of arising from the PPT and LDT40. homeostatic regulation of glutamate. Although other minority subtypes are Environmental cues linked to withdrawal also may stimulate relapse. 1a)44. When nicotine is applied in vivo. there is greater cholinergic innervation and endogenous excit. such as serotonin.26. but then the prolonged low levels of nicotine favor significant symptoms. the deficits episode of smoking.61. this desensitization would decrease DA the overall population—those who have a psychiatric disorder and release—particularly release evoked by low-frequency action potentials are nicotine dependent—consume 34% of all cigarettes. because mice lacking β4 show much milder symptoms when nicotine As a result of these pharmacodynamics. in reward pathways normally caused by nicotine withdrawal eventually That situation initially causes some activation of most nAChR sub.29. withdrawal57. it desensitizes nAChRs on twice as commonly as the general population12. the α4β2* nAChRs10. tion to daily smoking62. As a consequence. children with ADHD often initiate smoking earlier and nicotine alters the relationship between afferent activity along DA fibers have more trouble quitting13. Adolescent (that is. Acting in the target region. there is other intrigu- NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1467 . gluta- of the glutamatergic afferents (Fig. Remarkably. atory drive of GABAergic interneurons in the VTA (Fig. In addition. more frequently34. withdrawal and relapse Nicotine normalizes several deficits in sensory processing associated Although there is strong support for a role of DA and the overall meso. related mechanisms decrease the inhibi. Nicotine also acts at the fibers early behavioral problems are linked to a greater risk of later tobacco and terminals in the target neuron to alter DA signaling so as to favor use. smoking is particularly important because early tobacco use is asso- by acting on the midbrain source of DA. Often the withdrawal symptoms and the and synaptic events in the VTA that enhance excitation and decrease ‘priming’ cues arising from internal states are more severe for smokers inhibition to the DA neurons. term behaviors associated with addiction55 (also see the perspective © 2005 Nature Publishing Group http://www. Most notable are schizo- projections. factor CREB (cAMP response element binding protein)53. It is commonly argued that psychiatric patients use tobacco for self- medication.64. Although a small by Nestler56 in this issue). An evaluation of female twins suggested that thus modulating information processing along reward pathways51. Nicotine elevates glutamate levels.nature. patients with acting mainly through presynaptic non-α7 nAChR subtypes on DA mood. Comorbidity with mental illness Although most research has focused on the midbrain DA centers. and nicotine accumulates as the day progresses. glutamate (mGluII) receptors serve in a negative feedback capacity for tory GABAergic activity declines rapidly because nAChR desensitiza. Enhanced release has been seen for GABA. with schizophrenia. Stimuli repeatedly paired with such withdrawal discom- In summary. That hypothesis applies most readily to schizophrenia.29.13. conversely. making abstinence more difficult.46. tonic single-afferent pulses along the DA fibers)48–50. nicotine causes DA neurons to ciated with higher risk of later psychiatric problems and. The β4 nAChR subunit is likely to have a role in withdrawal desensitization of most non-α7 nAChR subtypes (such as α4β2*). In a US study. it roles for other neurotransmitters and peptides. Although attention-deficit/ release in response to tonic. 1)45.54. Nicotine dependence is much more prevalent among psychiatric nicotinic mechanisms are also important in the target areas of the DA patients than in the general population12. In that way.58. Addictive While those nicotine-induced mechanisms enhance glutamatergic drugs commonly alter CREB activity. those events are important for subset of DA neurons receive cholinergic inputs from the PPT and the reinforcing influences of nicotine53. α4β2* nAChRs make up the majority of the nAChR subtypes Withdrawal from nicotine decreases the sensitivity of reward systems. Even though nicotine directly mate and dopamine and also for other neurotransmitters that influence activates α4β2* nAChRs on the DA neurons for only a short time before mood and emotional balance. it is likely that such nicotinic mechanisms in largely from familial factors (likely genetic) that predispose individuals the target areas will be better appreciated as important contributors to to both smoking and depression63. anxiety or personality disorders show nicotine dependence terminals.54. evidence also indicates Although nicotine seems to improve attention in schizophrenia.40. corticolimbic system in reinforcing nicotine use. associated with the reinforcing properties of nicotine regulate CREB54. Thus. and this cholinergic activity regulates DA release47–50.18. hyperactivity disorder (ADHD) does not increase smoking prevalence tine boosts DA concentrations in the NAc. on midbrain GABAergic interneurons10. However. Associated with opioid influences. and group II metabotropic GABAergic interneurons in a matter of minutes (Fig. Likewise. who have smoking rates of 70% to 90% compared interneurons. that brief enhanced depolarization of the endogenous opioids. nicotine initiates cellular withdrawal is induced59. arise from conditioned stimuli that then cue smoking to relieve the types. present. the prevalence of psychiatric disorders is about 70% DA release in response to phasic bursts while simultaneously depressing in adolescents who are daily smokers60. fire more bursts of action potentials32–34. REVIEW The combination of enhanced presynaptic drive and strong postsynaptic nAChRs in the brain is the presynaptic enhancement of neurotrans- response favors the production of long-term synaptic potentiation (LTP) mitter release25.46. in all studies. That and thus cues linked to smoking become conditioned stimuli that initi- endogenous cholinergic activity has a significant excitatory influence ate molecular events contributing to the craving and relapse of absti- over the firing of inhibitory VTA GABAergic interneurons through nent smokers. The inhibi. to about 25% for the general population. For example. single action potentials. Environmental cues that are LDT. Nicotine also acts in are associated with higher risk for smoking initiation and for transi- the target region to alter intrinsic GABAergic feedback mechanisms. DA neurons fire with psychiatric illness. the relationship between lifetime smoking and major depression arises As research progresses. As is the case as detected by elevated thresholds for intracranial self-stimulation in for the DA neurons. and the concentration of DA is elevated in the NAc for a prolonged time8.52.

those animals also have differing reactions to nicotine. nicotine and alcohol on the response to acoustic startle and to the sever- The prevalence of nicotine dependence is very high among alco. GABAA receptors and may reduce the severity of alcohol withdrawal79. and α7 expression is reduced in schizophrenics13. anhedonia often accompanies mental ill. and similar effects are produced by stimulant drugs used responsiveness to alcohol is a risk factor for the development of alco- to treat ADHD. This influence Emerging results from human brain imaging suggest that neuroad- is only part of the motivation for smoking. smoking and alcohol reinforcement82. One genetic factor influenc- and dependence. During abstinence. because even first. particularly for abstinent smokers with psychiatric illness. suggesting common ness and is anticonvulsant. Animal studies show that chronic nicotine administration does increase nesses such as schizophrenia and depression. and that action requires the © 2005 Nature Publishing Group http://www. schizophrenics who is indicated by animal studies of motor activity and body temperature smoke have a lower incidence of neuroleptic-induced parkinsonism. stress can stimulate reinstatement limbic DA neurons. and nicotine can overcome this action by enhancing DA drug actions on a common effector protein.70. diminishes alert. the overall action demonstrates a convergence of receptors. 1b). smokers become conditioned to expect nicotine to provide par. and chronic nicotine use may decrease the effects because an unknown ingredient inhibits brain monoamine oxidase. Smoking reduces upregulation of to the influences of stress. it does render alters the function of neuronal nAChRs. the characteristics of the withdrawal syndromes also ing nicotine and alcohol modulation of acoustic startle in mice is a differ markedly71. anxiety and tobacco use also tors. Depression and anxiety often accompany nicotine the development of alcohol dependence and other substance-abuse withdrawal. Although nicotine does not act directly psychiatric medications. Genetically selected lines of mice and rats Comorbidity with alcohol show different specific behavioral responses to ethanol. nicotine17. ethanol does not influence the subjective effects of nicotine. When alcoholics stop drinking. including their legal status or threonine at position 529 of the α4 subunit influences the effects of and wide use. it is possible ness. Although a past history of alcohol dependence molecular interactions between nicotine and alcohol. Craving and relapse can of GABAA receptors as well as acting acutely to decrease GABAergic be elicited by the drug itself or by environmental cues that become inhibition by desensitizing nAChR subtypes that help drive GABA salient through their repeated association with previous use. REVIEW ing evidence linking nAChRs and schizophrenia. thereby increasing alcohol consumption and dependence. Twin studies indicate a substantial genetic alcohol binding site. Although unproven. Low attention. Just as interneurons (Fig. depression. In addition. and of alcohol. Initiation of smoking at an early age is a risk factor for ing internal state69. Indeed. possibly by binding to a site nicotine more reinforcing than for those who have never been alco. would reduce development of alcohol abuse later in life. Nicotine increases the release of DA and improves and that result is not due to differences in alcohol metabolism75. powerful predictor of drunk driving offenses at age 32 (ref. A key tial relief from stress and depression as it does from the symptoms of issue is whether exposure to nicotine during development increases withdrawal67. Regular smoking at age 14 was the most Thus. If alcohol enhances GABA inhibition of the meso- to drug reward. whereas alcohol affects multiple receptor types. Further convergent action release and reducing unwanted side effects. ability for drug use67. Studies with mice lacking hol abusers12.67. monoamine oxidase inhibitors have antidepressant actions66. disorders80. environmental events associated with nicotine withdrawal serve as The link between nicotine and alcohol use is of particular importance priming cues for drug seeking58. ity between forms of mental illness and nicotine addiction. The presence of alanine able commonalities between the two drugs. followed them until age 32. the amount of tobacco smoked is positively the β2 subunit have shown that those actions of nicotine and alcohol are correlated with the amount of alcohol consumed and the severity of mediated by α4β2* nAChRs85. from the observation that smokers report less intoxication from the The self-medication hypothesis also may have some validity among same amounts of alcohol than either nonsmokers or former smokers ADHD patients. The chromosomal correlation between nicotine and alcohol dependence. sitivity to the locomotor effects of nicotine83.nature. Nicotine and alcohol also both relieve pain. alcohol self-administration. Even though the A529T polymorphism is in development of nicotine and alcohol dependence (also see Crabbe and the intracellular loop of the α4 subunit and distant from the proposed Lovinger73.84. Commonalities between nicotine and alcohol also occur at the molecular and genetic level. ity of an alcohol withdrawal sign in mice86. Although both drugs produce tolerance genetic determinants for those actions. that show cross-tolerance for nicotine and alcohol78. the stress response mimics a motivat.64. Cigarettes also may provide a medicating influence hol dependence76. antipsychotic drugs block DA on GIRK channels. There also may be shared genetic influences over the and glycine receptors87. it is possible that the polymorphism influences 1468 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . lines sensitive to the hypnotic actions of alcohol show a different sen- sant. There also is evidence that increased GABAA receptors result in are contributing factors for comorbidity. For example. polymorphism in the α4 nAChR subunit85. analogous (but not identical) to the alcohol site proposed for GABAA hol dependent72. Furthermore. 81). however. There could be psychosocial influences produces synaptic plasticity in the VTA similar to that produced by as well as a pharmacological role for smoking in later substance abuse.com/natureneuroscience Another factor that may exacerbate nicotine addiction arises from GIRK2 potassium channel77. By boosting DA release. promotes alertness and is proconvul. This hypothesis has mechanistic support because stress alcohol use and dependence. A large Finnish study gathered data on 14-year-olds and and relief from specific aspects of those symptoms motivates relapse. Rat and mouse Nicotine has specific receptors. aptations and neurotoxicity produced by chronic alcohol abuse may episode patients who have not received antipsychotic drugs have a high also be altered by smoking. there are remark. During the acquisition phase. a time-dependent increase in the number of neuronal GABAA recep- The links between stress. for adolescents. and some of Nicotine and alcohol seem to share few pharmacological similarities. and haloperidol increases smoking in schizophrenics21. Depression sensitizes smokers more severe withdrawal symptoms. Indeed. A potential mechanistic link arises schizophrenia. with only a mod- region containing the nAChR α7 subunit is linked to genetic risk for est environmental contribution74. chronic nicotine may counter the upregulation of drug seeking and drug self-administration. stress increases the It is possible that nicotine boosts DA signaling that is diminished during sensitivity to addictive drugs. Despite these clear distinctions. supporting a possible causal link between nicotine may ameliorate particularly the anhedonic aspects of the ill. in this issue). making the individual more susceptible alcohol withdrawal. These studies provide hints of possible alcohol dependence. which increases the motivation and vulner. they show incidence of smoking68. It is reasonable to conclude that common underlying mechanisms that preventing or delaying initiation of nicotine use in adolescence influencing motivation and behavior contribute to the high comorbid.

A. Subunit composition and pharmacology of two classes of striatal 13. & Yurgelun-Todd. Nature 436. R. Schilstrom. M.A. Frequency-dependent modulation of dopamine release by smoking and depression.E.A. D. & Sesack. 24. J. 23. 3176–3185 (2003). et al. Neurosci. J. Nomikos. ing actions of alcohol in humans90. J.R.R. Gen. Chevy Chase. Liang. K.com/ 35.M.R. dopamine release in the striatum. Y. Maskos. B. Nature 391. Stolerman. reprintsandpermissions/ tolerance. 29. Kuhar. & Dawson. 583–584 (2004). 2279– 18. & Swartzwelder. dependence suggests that a more complete understanding of nicotine Nicotine Tob.. Wonnacott. The authors declare that they have no competing financial interests.S. E. Nat. Keath.89. Neurosci.D. McGehee. Nomikos.D. 12. Res.. D. Nature 393. Klink.. Psychiatry 61. 1029–1032 (2004). Molecular and cellular aspects of nicotine abuse. M. Cholinergic modu- NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1469 . in tegmental area: target specificity in the synaptic associations with mesoaccumbens Principles of Addiction Medicine 3rd Ed. Chow. F.E. 31. T. Neurol. 70. & Brussaard. T. Pharmacol. A. Biological aspects of the link between 48. A. & Zhou. N.I. 555–561 (1999). Sidhpura.92. 1.S. & Svensson.M. Regulatory strategies to reduce tobacco 49. & Changeux. Prim. presynaptic nicotinic acetylcholine receptors mediating dopamine release in mice. Abuse.. W. Opin. 77.P. Curr. Laviolette. 225–248 (American Society of Addiction and mesocortical neurons. M. Rev. 11–20 (1999).. Boreham.P.G. Omelchenko. Dani. R. V. L. REVIEW 19.B.. J. (2004). 6. Varlinskaya. Nat.. Mol.J. Opin. A. Adamson. 1–11 (2003). & van der Kooy. Nicotine-induced condi- nicotine and alcohol at the α4β2* nAChR. The neurobiology of tobacco addiction. Kits. Pidoplichko. 349–357 (2000). A. M. V. J. 52. R. C. & Coleman. Laviolette. De Biasi. Vastola. Drug Alcohol Depend.L. & Shoaib. Stinson. E. Cellular and synaptic mechanisms of nicotine 44. D. Lanca. O’Dell and the members of the Dani laboratory for comments. & Dani. Tob. the role of the dopamine projections to the nucleus.. 50–59 (2003).D. Tobacco or Health: A Global Status Report (World Health Organization. 21. Neurosci. 43. Grant. Nicotinic receptors in the development and modulation of CNS synapses.) details are not yet conclusively known. Maryland. 96. 4.J. Jones. The pharmacology of nicotine and tobacco. Biochem. et al.. Levin. Nicotine reinforcement and cognition restored by targeted expression in developed countries: indirect estimation from national vital statistics. E. Epping-Jordan. Williams. and psychiatric disorders in the United States: results from the national epidemiologic 46. Physiol. Peto. 14. M. 15. Although the mechanistic onset nicotine self-administration modeled in female rats.J.K. 22. 7.C.D. The mechanisms under. Curr. 410.A. 38. 56. 295–314 (2000). addiction will have a broad impact on society... Role of dopamine in the behavioural actions of nicotine related to addic.W. Eur.R.R. © 2005 Nature Publishing Group http://www. Pharmacol.. A.) i14–i24 (2003).A. Trends Neurosci. 16. Borghese.nature. 1526–1535 (2004). M.. Zhang.. S. M. S.D.com/natureneuroscience nonselective nAChR antagonist mecamylamine reduces the reinforc. Synaptic plasticity and drug addiction. Mol. 1452–1463 (2001).nature. 7. Neurosci. 561–570 (2001). 41. J. 581–582 (2004).F. istration in the rat: a correlative neuroanatomical and behavioral study. G.H. D. 393. Endogenous nicotinic cholinergic activity regulates addiction in youth. Pharmacol.C. L. J. J. S. Behav. P. Mansvelder. H. S. & Benowitz. 3. Schultz. D.. prevention for particularly vulnerable members of the population.. 33.. W... Psychiatry 8. A. & Svensson.A. F. 1593–1599 (1997). E. Corrigall. Collins. & Markou. Care 26. J. & Yakel. 17.E. A. 51. Lodder. (2005). Nicotine: from molecular mechanisms to blocking nAChRs that may contribute to alcohol consumption.J. Baird. P. The neurobiology of nicotine addiction: bridging treat nicotine and alcohol dependence or treat alcohol dependence by the gap from molecules to behaviour. Dual effects of nicotine on dopamine neurons mediated by different nicotinic receptor Published online at http://www. Pharmacol. of nicotinic receptors. Cholinergic axon Sci. Molecular and physi- 2. 173–177 (1998). Lopez. Thun.. Pidoplichko. Neuron 33... Jones. V. Picciotto.P. Long-term potentiation–a decade of progress? Science addiction. Nisell.R. Westman. Adriani. Vaughan. Neuronal systems underlying behaviors related to be useful in treating both nicotine and alcohol dependence91. & Sulzer. nuclei. Nicotine amplifies reward-related dopamine signals in stria- Control 9. Neurosci. Uhl. Neurosci.M. W. 99–110 (2000). Synaptic mechanisms underlie nico- survey on alcohol and related conditions. D.M.. 285.F. & Spear. nicotine. H. 1268–1278 (1992). Psychopharmacology (Berl.S. U. Nicotine Tob. brain reward areas by nicotine. R. Tapper. S. J. 99–107 (1999). aspartate receptor antagonism in the ventral tegmental area diminishes the systemic nicotine-induced dopamine release in the nucleus accumbens.S. commonalities between the two addictive drugs85. Nicotine self-administration in animals as a dependence model.E.P. Lancet 339. Neuron 27. nicotine addiction: neural circuits and molecular genetics... Nicotinic cholinergic synaptic mechanisms in the ventral teg- (2004). The behaviour.I.. & Zeller. W. 905–919 (2002). J. & Heinemann. Coen. Neuron cell populations in the rat ventral tegmental area. 735–742 (2000). 107–114 influenced. Balfour. 60–69 (2004). L. 11. N. B. J. 217–235 16. F. Arterial nicotine kinetics ACKNOWLEDGMENTS during cigarette smoking and intravenous nicotine administration: implications for The authors are supported by the National Institute on Alcoholism and Alcohol addiction.. & McGehee. S. the National Institute on Drug Addiction and the National Institute of 32.. Tob.R. 5.R. J. & Heath. S. tum. M. J. Sudweeks.A. 3864–3873 (2000). G. 467–473 (1991). J. possibility of pharmacotherapies that could either simultaneously 22. J.. Differential desensi- brain reward function during nicotine withdrawal. J. 11. Smoking and mental illness. 22. H. Dramatic decreases in 39.H.P. 6. Nicotine activation of alpha4* receptors: sufficient for reward. Nat. M.) 227–241 (2004). Cholinergic 5.. G. 5. tization and distribution of nicotinic acetylcholine receptor subtypes in midbrain 4. 483. P.. D. Reprints and permissions information is available online at http://npg. M. Science 275. Synaptic plasticity and nicotine addiction.M. I. Quattrocki. Salminen. R. Res. J. S. Schilstrom. D. & Brussaard. & Dani. 47.H. mental area contribute to nicotine addiction. ological diversity of nicotinic acetylcholine receptors in the midbrain dopaminergic 1997).R. Behav. Eur. Science 306.T. N-methyl-D- M. F. 1224–1229 (2001).S. pontine tegmental nucleus and the role of cholinergic neurons in nicotine self-admin- 9.L. W. Comp. Dopamine and drug addiction: the nucleus accumbens shell con. COMPETING INTERESTS STATEMENT 781–789 (1998). & McGehee. Leonard. Hasin. Garzon. Dayan. M. Mansvelder. A neural substrate of prediction and reward. D. J. S. Schoffelmeer. A. M. Evidence for enhanced neurobehavioral vulnerability to nicotine during periadolescence in rats. & Sesack. 8. H. D. 349–352 (2001). Blockade of mesolimbic dopamine transmission dramatically increases sensitivity to the rewarding effects of nicotine in the ventral The substantial interactions between nicotine and alcohol raise the tegmental area.. Projections from the rat prefrontal cortex to the ventral 6.B. 103–107 (2005). & Pickel. Nicotine activates and desensitizes midbrain dopamine neurons. 23. 10.. 20.L. Broide. modulation of nucleus accumbens medium spiny neurons. 313–319 (2000).G. Nat. 53. Mansvelder. Benowitz. A. 36. A. & Bonci. M. H..S. O. Balfour. 1. We thank D.nature. & van der Kooy. Neurosci. S. S. Neuroscience nection.R. Mansvelder. 45. Rose.B. 50. L.R. 65. 4712–4716 (2003). Rev. Role.. 76–79 (1998). H.. et al.N.S. J. 197–210 (1999). & Nicoll. J. 34. Hertel. J. S. terminals in the ventral tegmental area target a subpopulation of neurons expressing 12. Behm. Harv. Henningfield. Mortality from tobacco 37. 20. J. 401–404 (1997). Initial symptoms of nicotine dependence in adolescents. Int. Neuron 16. 899–912 40. Dani. Wooltorton. J. Geneva. Neurol. Neurological Disorders and Stroke. 21.. Zoli.A. Zhou. there are interactions between 169.. 16. 2003). Mem.E. K. & McGehee. D.. J.. B. 20–25 (2005). N. J. C. J. 55–65 (2004).com/natureneuroscience/ subtypes. 23. CB1 receptor antagonists and the anticonvulsant drug topiramate may 27. A. Rice. 53–59 (2005). Douglas. De Rover. Pharmacol. and additional complexities are likely to contribute to the (2002). J. that do not directly affect nAChRs show some promise: cannabinoid 26. et al. Moolchan. 5. Several drugs physiology with function.. Rezvani. Carr. & Montague. Rawal.A. & Corrigall. 3338– The common comorbidity of tobacco use with mental illness or drug 3341 (2002). Dani. Ji.R. N. Neuron 31. Pidoplichko. K. Nicotinic receptors in the brain: correlating because mecamylamine produces autonomic side effects. & Berg. G. Trends Pharmacol. Watkins.C. Control 12 (Suppl. Di Chiara. Neurobiol. et al. Neuropharmacology 47 (Suppl. G. et al. Picciotto. B. The neurobiology of tobacco dependence: a preclinical perspective on dopamine areas.D. Nicotine addiction. et al. 1870–1874 (1999).. 905–908 (1996).S..A. et al.S. Nature 390. 606–617 (2002). DiFranza. lying nicotine addition may indicate common modes of treatment and 30. & Balfour. 28. Jr. D.A. but this treatment is not practical 25.. D. Montoya.T. B. Comp. de Rover.I. Nicotine dependence low levels of the dopamine transporter.I. G. R. D. Neuropsychopharmacol. Mameli-Engvall. Rose. 1077–1085 (1996). Karan. Neurosci. D. N. 611–631 (1999). & Cragg. Chou. V..A.J. Neuroscience 82.. Malenka. Neurosci. Laterodorsal tegmental projections to identified 7. and sensitization. Long-term potentiation of excitatory inputs to 2290 (2002). Di Chiara. M. Learn. R.D. D. Other subtypes also are tioned place preference in adolescent and adult rats. E.. S. & Dani. A.J.S.. 141–149 (2003). Neurosci. Medicine. DeBiasi. & Corrigall. Adolescent- the alcohol sensitivity of the α4* nAChR88. 1107–1115 tine-induced excitability of brain reward areas. et al. 42. Acetylcholine receptors containing the beta2 subunit are involved in the reinforcing properties of nicotine. Koob. The pedunculo- tion. M. Arch. de Kerchove d’Exaerde. Psychiatry 8.L. L.

Dong. Proc. Alcohol Alcohol. & Ridley.S. Pharmacol. Biochem. sensitivity (HAS) and low alcohol sensitivity (LAS) rats differ in sensitivity to nicotine.A. S. Hernandez-Avila.A. Probabilities of alcohol high-risk drink- activity of brain reward systems.G. Gasparini. S. 27. Gen.M. A. 42–52 (2003).. 44. Adolesc. D..J..A. Brooks. Sci. Am. 24. J. R. H. 117–123 (2003). Y. Clark. Acad.P. Alcohol.. 289. 591–599 Neuropsychopharmacology 29. Exp. R. Genet. C. Exp..A. USA 93. P. Acad. tors in long-sleep and short-sleep mice. Staley. A causal analysis. Changes over time in the self-reported level of response 55. D. Borghese.B. K. White. (CHRNA7). Biochem. 78. Natl.P.J. N. Alcohol tolerance and nicotine cross-toler- 57. 74. 54. 29. CB(1) receptor antagonists for the treatment of 72. Heath. G. Jarvelin.S. 58. Behav. et al. Clin. ance in adolescent mice. N. Kenny. G. 1633–1638 (2000). R. W. Exp. Van Kirk. G. The selectively bred high alcohol over 3 years. J. Arch. Nestler. & Lovinger. H. Butt. Pharmacol. & Collins. Nat. et al. et al. & Maldonado. A.R. 6. Pieri. et al. USA 100. J. 1120–1121 (1999). C. Psychiatry 56. S. Kuo. Psychiatry 41. 85. Med. Harris. Eur. & Little.A.F. U. sensitivity and smoking history in men and women. Stoffel. Am. Cortical gamma-aminobutyric acid type A-benzodiazepine receptors © 2005 Nature Publishing Group http://www..C. 245–254 (2004). Bleck. J.R. Kessler.N. J. C. L. and smoking initiation: a prospective study 84. Lindgren. N. Rumpf. 8. Henderson. & de Fiebre. 577–582 (2003). S. addiction. Clin.K. J. Addict. & Brown. De Fiebre. 86. Neurosci. Exp. 20–22 (2001). Deas. abuse or dependence estimated on grounds of tobacco smoking and nicotine 59.E. & Griebel. J. H. 1445–1449 (2005). Crabbe.. Neurosci. Alcohol. Effects of nicotine on cognitive deficits in schizophrenia. Public Health 88. Sites of Natl. Pharmacol. Lopez. & Markou. C. 63. Walters.. for the beta4 nicotinic acetylcholine receptor subunit.. Exp. Salas.. Sci. 83. A. Exp. J. & De Biasi. 27. Cleck. Blomqvist.. et al. 71. M..L... et al. 19. J. R. & Hapke. Nicotine withdrawal versus other drug with. Stitzel. J. J. Exp. B. 81. Watson. Neuron 46.R.M. nicotine. Group II metabotropic and alpha-amino-3. R. J. Topiramate-induced neuromodulation of cortico-mesolimbic dopamine Res.C. 105. activation are required for nicotine reward. Addict. C. Y. 326–331 (2002)..A. 36–43 (1993).J. C. V. Psychiatry 156. J. Upadhyaya.R. 1518–1522 (1998). Johnson. A polymorphism in the alpha4 nicotinic receptor gene (Chrna4) (2000). 24. Alcohol J. Ther. activation of GIRK2 channels. Rose. Alcohol. 1378–1385 (2004). M.J. 307.) to alcohol. 14065–14069 (1996).A. 277–282 (2003). Exp. P. Starmer. Drug Alcohol Depend.C. graphic correlates of symptoms of last year dependence on alcohol. 89. Kenny. & Grant.. & Randall. Neurosci. Dawson. J.. mu-opioid receptor and CREB 1111–1120 (1995).T. Clin. E. 39. rewarding effects.A.J. T. Riala... A. nicotine can increase operant self-administration of alcohol. Res. 655–661 (1999). Brady. Cigarette smoking and psychiatric 108–117 (2001). and dependence in mu-opioid receptor knock-out mice. Decreased signs of nicotine withdrawal in mice null dependence. Behav. Warner. K. Nestler. Cohen. 480.A. & Bickel. Res. Brain monoamine oxidase A inhibition in cigarette smokers. B. 119–127 (2001). 1465–1479 (2004).R. ing. Isohanni. 10935–10940 (2002). Chen. Teenage smoking (2004). R. J. Child Adolesc. M.K. 774–780 (1999). M. Alcohol. & Malenka. Behav. Stress and cocaine addiction. Meyer. F. 87. Exp. L. R. Owens.C. ciception. J. Am. 92. C. R. 79. Exp.L. function: a new vista for the treatment of comorbid alcohol and nicotine dependence? 73. 70. Biol. Bonci.com/natureneuroscience late the deficit in brain reward function associated with nicotine withdrawal in rats.. & Smith.C. 306. 91. Chronic infusion of and cocaine in the U. 26. W. Alcohol.. 308. Kendler. 68. Effects of ethanol on recombinant human neuronal nicotinic 69.J. Hakko. Addiction 89. & Collins. 493–501 (1987). Conditioned nicotine withdrawal profoundly decreases the 80. Trudell. E.. Alcohol 4. A. B. Neuropharmacology 47 (suppl. hydroxy-5-methyl-4-isoxazole propionate (AMPA)/kainate glutamate receptors regu. Interaction of the nicotinic cho- 65. Neurosci. H. M. drawal syndromes: similarities and dissimilarities. John.C.. 733–742 (2003). Res. C. 1867–1875 (2003). O. Hughes. & Callas.E. N. Proc. A. & Kranzler. & Harris. 66. 11–29 (1997). A pervasive mechanism for analgesia: 56. Nicotine response and nicotinic recep- 1294–1305 (2002). P. D.. Alva.M. 805–814 (2003).. Nicotine is more reinforcing in smokers with nicotine addiction.nature. Psychiatry 50. True.. comorbidity in children and adolescents.M. J. D.. Kandel.. A.L. Freedman. REVIEW lation of dopaminergic reward areas: upstream and downstream targets of nicotine tion and insight into mechanisms. Acad. 61. (2002). et al. H. King. & Markou. 433–438 (2004).R.M. Am.A. in young adulthood. 8. L.L. de Fiebre.. R. 79–85 88. marijuana 82. Ther. Saal. Linkage disequilibrium for schizophrenia at the chromosome effects of ethanol and nicotine on the acoustic startle response. Jr.L.M. Kodas. Higgins. Fowler.L. Alpha 4 beta 2* nicotinic acetylcholine receptors modulate the 64.. Animal models of alcoholism: treatment target identifica. Gen. (2004). Psychiatry 62. Smoking in first-episode patients with schizophrenia. 765–772 (2002). 22. Arch.T. anxiety.. linergic system with ethanol withdrawal. K. Addiction 98. 785–789 acetylcholine receptors expressed in Xenopus oocytes. & Rasanen. 1470 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . Attenuation of nicotine-induced antino. 66. 387–395 (2005). 6208–6212 (2005). 301. M. Rose. 15q13–14 locus of the alpha7-nicotinic acetylcholine receptor subunit gene Res. & Blendy. et al.. Neuropharmacology 41. population. H. Health 35. K.. J. C. G.. J. D. et al. Neurosci..A. P. a past history of alcoholism than in smokers without this history. Clin. Exp. 25. Common genetic vulnerability for nicotine and alcohol dependence 53.C. 24–32 (2004).M. & Kruesi. Schuckit. P. & Martin. U. Res. 26. Alcohol. in men. Drugs of abuse and stress trigger a 90. Ther. J. Clin. The effects of nicotine on neural pathways implicated in tylcholine receptors.. Depression. Kieffer. 1068–1076 (2003).M.C. and cigarette smoking. W. C. Ther. Patton.P..S. 1461–1470 (1994). Balfour.W. J. Butt. E. depression: a factor in nicotine addiction? Pharmacol. Blednov. common synaptic adaptation in dopamine neurons.A. Arch. M..C. Prevalence and demo. J. D.G. J. Pharmacol. excitatory and inhibitory actions of alcohols on neuronal alpha2beta4 nicotinic ace- 67. 75. Y. J. Mecamylamine modifies the pharmacokinetics and reinforcing effects of alcohol. 877–888 (2005). Goeders.. J. N. Smoking and major depression. 77. Neuron 37. Whitfield. 76. Cardoso. Is there a common molecular pathway for addiction? Nat. Gen. Hughes. 62. et al. Ther. Medhurst. Berrendero. Pharmacol. in recovery from alcohol dependence: relationship to features of alcohol dependence J..J. 1471–1480 (2005). J. 933–943 (2005). Pharmacol. Molecular mechanisms of drug addiction.J. F. Pharmacol. 10035–10039 81. Madden. & Harris. Clin. and substance use as predictors of severe alcohol problems in late adolescence and 60. J.P. F. modulates enhancement of nicotinic receptor function by ethanol.A. McEvoy.

result from molecules. including animal tissues and live animals. may reflect its increased value as a reward. Chronic alcohol exposure causes neuroadaptations that foster opment of an intense desire to consume the drug (sometimes called continued alcohol abuse5. these experimental approaches are likely to contribute to the development of new therapies for alcohol abuse and alcoholism. proteins and neural systems that are targets for other abused substances Rockville. cellular and genetic basis of and the neuroadaptive changes that take place in response to chronic these neuroadaptations. and John C.9. with other drugs of abuse. decreased ‘craving’). Other physical withdrawal and increased desire for the drug. The changes in ronmental experience interact to alter both direct alcohol actions and reward value and the consequences of withdrawal result from the molecular mechanisms that indirectly affect ethanol-related behaviors.6. However. viduals. Molecules in sity for certain alcohol-related neuroadaptations is also influenced both groups contribute to the acute and chronic stages of alcohol use.nih. Genetic makeup and envi.4. Alcohol abuse and alcoholism involve interactions among have a strong inherited component and may involve polymorphisms a number of neural mechanisms. Portland. Attempts to understand Published online 26 October 2005. National Institutes of Health.1038/nn1581 the common and unique aspects of alcohol addiction have spurred NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1471 . Unfortunately. 1). sensitivity to intoxication) and dependence on it.com/natureneuroscience target identification and insight into mechanisms David M Lovinger & John C Crabbe Laboratory models. Lovinger is in the Laboratory for Integrative Neuroscience. USA. by genetic factors and gene-environment interactions (Fig. neuroadaptations brought about by chronic alcohol. evidenced by both cation and influence responses to chronic drinking1 (Fig. and we review studies in which animal models were used to examine excessive alcohol drinking and to discover genes that may contribute to risk for alcoholism. Oregon Health & Science University. and devel. or both. 2).gov neural processes distinct from other addictions. dala circuitry12). the escalation that leads to the desperate craving susceptibility to later alcoholism2. Oregon 97239. development of tolerance to and dependence upon the drug. including acute sensitivity to alcohol. Tolerance to ethanol (that is. Initial sensitivity to alcohol differs among different indi- as is the loss of judgment and control after continuing to imbibe alco. doi:10. Animal and human studies suggest for alcohol. alcohol abuse and alcoholism also involve e-mail: lovindav@mail. USA. its ability to reduce the ity to alcohol abuse and alcoholism (Fig. The propen- participate more generally in addictive processes (Fig. 2). alcohol seeking in dependent individuals Inherited factors contribute a great deal to an individual’s susceptibil. alcohol abuse. Here we consider strengths and weaknesses of laboratory models used in alcohol research and analyze the limitations of using animals to model a complex human disease. is also experienced by far too that differences in alcohol sensitivity. 1). the Portland signaling proteins such as CREB10. As influencing the likelihood of abuse and the expression of alcoholism. Addiction to alcohol shares common neural substrates with other addictions at the molecular. Despite some limitations of the laboratory models used in alcohol research. National alities include alcohol and drug actions at similar subsets of neural Institute on Alcohol Abuse and Alcoholism. The neuroadaptations underlying these less directly by influencing expression or function of molecules that are behavioral adaptations to alcohol involve molecular mechanisms that direct ethanol targets or by altering the function of neural circuits that are affected both directly and indirectly by alcohol. interact with ethanol repeated exposure to alcohol. cellular and circuit levels. in genes encoding proteins involved in intoxication3. These common- David M. and there is some evidence that these differences contribute to hol. brain glutamatergic systems8. undesirable effects of alcohol withdrawal. destroying lives and families. stimulation of VTA dopaminergic Alcohol Research Center and the Department of Veterans Affairs Medical neurons11 and involvement of mesocorticolimbic and extended amyg- Center. seeking and drinking behaviors many people. We describe targets for the neural actions of alcohol. N E U R O B I O LO G Y O F A D D I C T I O N REVIEW Laboratory models of alcoholism: treatment © 2005 Nature Publishing Group http://www.nature. nuclear Behavioral Neuroscience. The euphoria that follows the first drink at a party is familiar to many. Maryland 20852. have proven useful for discovery of molecular targets of alcohol action as well as for characterization of genetic and environmental factors that influence alcohol’s neural actions. most notably several neural proteins. Alcohol has direct actions on molecules that lead to intoxi. and thus it is Such interactions influence the acute sensitivity to ethanol intoxication important to understand the molecular. Crabbe is in the Department of (such as the GABAA receptor7.

Most behav. edly paired with specific cues. Second. and these much about direct alcohol targets and affinity of alcohol at these targets. The molecule ism-related research that most distinguish this line of research from does not form ionic or covalent bonds. alcohol (direct alcohol targets) and to identify proteins that are indirectly biophysical biochemical.com/natureneuroscience Preclinical drug screening testing are then the focus of clinical testing for safety and efficacy. but until these types of experiments can be per- motor coordination (which turns out to be a very complex domain of formed on a wide variety of proteins. 30% change in any given measure of cellular or molecular function even at concentrations that are near the higher limit of the sublethal range Alcohol’s difficult pharmacology: how do we identify a ‘hit’? (for example. and the animal can be tested on the choice Methodologies for examining direct alcohol interactions have been it makes between a saline-paired or alcohol-paired cue13. which only adds to of other drug targets. These molecules can then be targeted in preclinical analyses screens designed to determine if pharmacotherapeutic agents or other Direct alcohol targets Alcohol-associated proteins target-specific treatments alter responses that are predictive of alcohol abuse or alcoholism (such as alcohol intake. invertebrate models. and the short carbon backbone. biophysical studies of specific ethanol mal models. Attempts this field. This may be a blessing for the person drink- Animals. Pharmacological and genetic techniques are used alone and in combination to uncover molecular targets that are directly affected by Functional/ Pharmacological. animal tissues and molecules arising from experimental ing alcohol as an intoxicant. and many animals—most notably invertebrates—have nervous to be certain that a particular moiety is truly an ‘alcohol binding site’. Before examining some of the latest findings in neuropharmacologist. Furthermore. altered by alcohol or that influence neural responses to alcohol (alcohol- analysis molecular biological QTL analysis. abuse and alcoholism15. cells and molecules. knockouts. Without measures of direct molecular interactions. Laboratory tion of direct alcohol target molecules has lagged behind the discovery animals cannot verbally report their subjective states. etc. Numerous complications dependent changes in the experimental system might be mistaken for arise when using these approaches. First. REVIEW Figure 1 Use of animals to identify direct and indirect alcohol targets Acute Chronic exposure Forward and can lead to development of pharmacotherapies for alcohol abuse and pharmacology (neuroadaptation) reverse genetics alcoholism. in alcohol research. and more recently. in the accuracy and range of variability of the measures that are made. cacy that are difficult to relate to occupancy of a particular molecular The behavioral repertoire of laboratory animals differs from that of site. These experimental indicate molecular targets and genetic underpinnings of alcoholism limitations make it difficult to measure actual affinity of alcohol for and consider the limitations of using animal models as surrogates for potential molecular targets. It is not unusual to observe only a determinants of alcohol intoxication. applied to some proteins (such as the Drosophila melanogaster LUSH ioral assays of alcohol intoxication and tolerance target some aspect of protein. Molecular targets that show promise in preclinical © 2005 Nature Publishing Group http://www. have not been as widely used in investigation of other addictions. 100 mM).nature. associated proteins). only at blood and brain concentrations from ∼5 mM to 100 mM. that can detect sites of direct has traditionally been more widespread in alcohol research than in molecular interactions between proteins and compounds with high investigations of other addictions. Consequently. mental approaches in animal tissues. ogy of alcohol. within the range of physiologically relevant concentrations. such as radioligand binding. which have high rats are offered a choice of drinking water versus ethanol solutions. CNS often use pharmacological approaches involving direct applica. as it ensures that the desired effects occur animals are widely used in studies aimed at finding molecular targets with a large margin of safety from outright toxicity. craving for alcohol or relapse). Ann Thomson Pharmacotherapeutic target candidates investigators to adopt new animal models and research approaches that which is capable of hydrogen bonding. ion interactions do occur in alcohol dehydrogenase)17. for the of alcohol actions. it is best to avoid inferring too behavior genetically14) or dysregulation of body temperature. Contrast this with the pharmacology of opiate drugs. 1472 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . it is difficult humans. and the neural effects ranging from intoxication to anesthesia18 are observed hunt for alcohol targets has spurred investigators to use unique experi. it is worthwhile to consider the limitations of experimental to determine the molecular interactions that underlie relevant ethanol searches for alcohol targets. effects are often confounded by signal-to-noise issues inherent in look- Investigators searching for direct targets of alcohol actions in the ing for changes in a small response. Drug state can also be repeat. and one animals are tested in a protocol of operant self-administration. and changes in molecular or cellular function that result from time- tion of ethanol to neuronal preparations16. binding affinities. methods are also used for studying a broad spectrum of drugs with The foregoing discussion provides some of the reasons why identifica- sedative-hypnotic effects. the This low potency prevents the application of some pharmacological use of genetic animal models to explore mechanisms of ethanol action techniques. or affinity for a well-characterized G protein–coupled receptor19. see below). systems that differ radically from that found in man. In typical studies. such that acute of alcohol action are not shared by most other drugs of abuse. These characteristics generate investigation of other addictions. Small effects are often overlooked. when using these organisms. Given the large role of vertebrate ani. which contributes to weak hydrophobic interactions (although metal Here we focus on two aspects of the use of animal models in alcohol. alcohol reinforcement. the small effect sizes are problematic. Transgenics. interactions with molecules are often hampered by the distribution of we feel it is important to critically analyze the strengths and weaknesses ethanol into many cellular compartments and the weak interactions of of the animal models. However. to infer affective and cognitive states that are likely to be important nearly always produces small effects. Because of the simple structure of ethanol actions. it is difficult Ethanol. the basic molecular mechanisms poor reactivity that results in low potency of the drug. only two reactive sites are present: the OH group. this problem. leaving only measures of potency and effi- understanding human alcoholism. Consequently. Anxiolytic effects are also studied. in which begins to see the problems inherent in the molecular neuropharmacol- a specific response yields access to alcohol. the experimenter must have confidence the ethanol molecule. We highlight findings from animal models that the molecule with many potential target molecules.

than those used with more efficacious drugs. including b Genes Proteins Pathways Behaviors Environments assessment of their interactions with each other and the environments in which they are assessed. pathways and behaviors Proteomics are complex. Taken together. The range of interest of behavioral genomics includes all relevant genes and behaviors. Here we Genomics also show that the relationships between proteins. More commonly. it is especially Although there are many problems in research on the cellular and important to demonstrate reversibility of the alcohol effect to ensure molecular neuropharmacology of ethanol and many complexities that that small changes are not due to ‘drift’ in the measurement over time. Many techniques has been used to characterize ethanol sensitivity of a variety of proteins. great care must be taken in designing pharmacological studies with ethanol.and systems-level pathways that ultimately modulate specific addiction-related behaviors. However. bined with examination of the function of the same protein in neurons hol intake often involve prolonged exposure to alcohol. and a useful model should target certain well-defined features. must be addressed by animal models. for chronic ethanol exposure in reduced neuronal preparations (such including neurotransmitter receptors.21. The only differences between the cases shown in environments 1 and 2 in the example are that the effects of the second and third genes on the first two behaviors differ depending on the environment—a gene-environment interaction. The direct approach: pharmacological methods An alternative approach to identifying molecules with important roles Heterologous expression of molecules allows researchers to examine in the neural actions of alcohol is to examine molecular changes brought the effects of ethanol on the function of proteins expressed in a cellular about by chronic alcohol exposure. ion channels and neurotrans- as dispersed primary neuronal cultures or organotypic slice cultures) mitter transporters1. is beyond the scope of this short review. withdrawal signs and increased alco. and the Interactions between ethanol and GABAergic inhibitory synaptic NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1473 .22. Heterologous expression com- that lead to tolerance. Important variables include the of alcohol actions on GABAergic synaptic transmission illustrates the concentration of alcohol applied. this means that no single animal model will convincingly capture all the features of this complex dis- order. The discussion of all these molecular targets allow investigators to examine molecular adaptations to ethanol in well. Fig. ethanol evaporation at physiological usefulness of this approach. The bottom of panel a depicts the range of interest of genomics. temperatures. In addition. 2 2 Environment 1 3 3 4 4 Gene ҂ number of cycles of exposure and withdrawal to be modeled. Alcohol-induced neuroadaptations context free of many neural proteins. Thus. Most research with genetic animal models has therefore concentrated Sample sizes used in experiments with ethanol often must be larger on a few relatively straightforward phenotypes. proteomics © 2005 Nature Publishing Group http://www. and there are parallels with the pharmacological chal- lenges just discussed. and that they can influence the expression of genes. animal models have been used more Behavioral interactions often in the field of alcohol genetics than in any other area of psychiatric genetics. dependence. This describes what we know of addictive behaviors. Pleiotropic oligogenic or polygenic. These environment 1 1 interaction reduced systems have contributed much to our understanding of cellular 2 2 and molecular neuroadaptations to alcohol exposure. Their common feature is the concentration of the analysis on a single behavior (such as impulsive drug-taking). The final panel shows that behaviors can themselves interact.com/natureneuroscience and metabolomics analyses. the pattern and duration of alcohol exposure. but the following description controlled experimental systems20. This level of analysis will be critical to 1 1 understanding the addictions. These proteins in turn participate in Pleiotropic numerous cellular. the abundance of targets also is. or multiple genes can exert converging Multigenic + influence on a single behavior. 2). there is an abundance of candi- The high concentrations of alcohol used in most experiments bring date molecules that seem to be targets for direct and indirect actions of into play possible artifacts related to ethanol’s solvent properties (that ethanol in the nervous system. Some issues gene expression may be traced to the complex genetics of alcohol dependence disorders Behavioral genomics (Box 1.nature. which is then variously termed multigenic. REVIEW a Genes Proteins Pathways Behaviors Gene effects Figure 2 Complexity of gene-environment-behavioral interactions in the Mendelian neural actions of alcohol. A behavioral difference can be Multigenic determined essentially by a single gene: some complex traits show purely Oligogenic mendelian inheritance. a single gene can influence Polygenic multiple behaviors (pleiotropy). (b) All of the above gene-behavior Metabolomics relationships take place in environments that themselves may differ. (a) Individual genes through their expression lead to synthesis of specific proteins. agents out of tubing or storage vessels). Environment 2 3 3 4 4 Modeling alcohol genetics in animals Largely by historical accident. The wealth of alcohol response data on genetic animal models Effects of gives us a window through which to assess the strengths and limitations behavior on Ann Thomson of animal models in biomedical research more generally. The usual case under investigation must account for both multigenic and pleiotropic effects. the possibility that ethanol may redissolve a hydrophobic compound makes it difficult to determine the most important contributors to the that had previously adhered to plastic tubing or may leach plasticizing neural effects of ethanol.

polymorphism in the rat GABAA α6 subunit may account for differ- the diagnoses are categorical. For this reason. and if a different environment (such thought to mediate tonic GABAergic inhibition28. For example. REVIEW BOX 1 GENETICS OF COMPLEX TRAITS gests that alcohol potentiates GABAA receptor function. Equal attention should be paid to the other part of the understanding the molecular basis for these discrepant results will be gene-behavior nexus. α6 difference31 and this may suggest that there is coordinated regu- Confronted with this daunting complexity.25. they may interact differently. Expression of the α6 subunit is restricted transmission have been the subject of intensive study for over to cerebellar granule neurons30. GABAA receptors exist or severity usually has a small. not simply cosegregate with this polymorphism in the AT and ANT Gradually. such this line of research may prove useful for understanding factors that as depression or anxiety disorders. thanks to tion has been observed in some isolated neuron preparations7. but drug α6-containing GABAA receptors. the development of new ways to measure and manipulate the However. These behaviors may interact as well. it is unlikely that the α6 polymorphism ing GABAergic synaptic transmission23. Ethanol potentiation of GABAA receptor func- behavior are becoming more amenable to analysis. although there is unresolved disagree- not act independently but typically interact with each other (epistasis). rearing conditions. be due to compensations during development.24. However. Thus. This polymorphism genetically influenced diseases. Finally. and thus highly likely to have other (comorbid) diagnoses as well. and a variety of evidence sug. The most recent stud- required to explain the full range of genetic contribution to risk ies carried out in oocytes. However. In However. the behavioral phenotypes. The specific polymorphism leading of the disorder across the life span. there is ongoing controversy as to the prevalence and importance of ethanol The complexities involved in relating genes to addictive actions on this receptor. demonstrate ethanol potentiation of a tonic GABAA-mediated current or both. Furthermore. 2b). Clearly. several acute alcohol sensitivity phenotypes do relevant genes and assess their interactions with other factors. Hence. Finally. Although the severity of any of the various One clue comes from a study28 concluding that a naturally occurring symptoms that in the aggregate lead to diagnosis is continuous. whereas α6. isolated neurons and brain slices indicate for or severity of alcohol dependence. there are developmental shifts in the characteristics sensitivity has been questioned. not less. members of any diagnostic observed in the rats in vivo28. if any.23. genetic knowledge from many studies then can be rat lines that were selectively bred to differ in motor impairment by assembled into a larger system of interactants that enables ethanol32. Indeed. gene-targeted mice lacking the α6 subunit do not us to understand a set of related behaviors. negative results have also been reported in numerous neu- function of genes as well as large-scale applications of classical ronal subtypes24. graded (polygenic) effect on in the plasma membrane as heteropentamers of a variety of subunits that trait (Fig.28 at other behaviors or even things that may have no obvious ethanol concentrations as low as 1–3 mM. 2b). and the polymorphism segregates in rat category (such as ‘alcohol dependence’) are heterogeneous. Addictive suggest some association between this subunit and alcoholism3. The GABAA receptor is one contributes to alcohol sensitivity for most aspects of intoxication. well within the range associ- connection with behavior (pleiotropy). it is premature to conclude that this polymorphism is the only factor contributing to alcohol sensitivity in rats28. Studies in the Xenopus laevis oocyte expression sys- breeding methods. What role. there is little information about the direct effect of ethanol 1474 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . withdrawal may then engender more anxiety. metabolomics. Furthermore. Studies of granule as set of families. alcohol dependence vary continuously over a wide range in Diversity among GABAA receptor subtypes could contribute to the populations. in the same way that any statistic describes ment about the effective concentrations and the shape of the ethanol the specific population from which it is drawn. lines bred for differential alcohol sensitivity30–32. In potential molecular target for ethanol. the involvement of many genes is that can coassemble in various combinations. the role of the rat α6 polymorphism in differential alcohol addition. © 2005 Nature Publishing Group http://www. 2a). where they are on behavior are studied. addiction tem have generally found ethanol potentiation. a key step in resolving this controversy and determining the role of Most behaviors taken to index different aspects of drug or GABAA receptors in alcohol effects. We term this differ in sensitivity to ethanol motor impairment33. although this may perspective behavioral genomics (Fig. Other polymorphisms cosegregate strongly with the diagnostic than a similar level would be in a 65-year-old. studies of human α6 polymorphisms own set of definitional and genetic complexities.com/natureneuroscience Fig. Evidence from in vivo and in vitro studies supports the likely to participate in aspects of intoxication that do not involve motor idea that ethanol produces many of its intoxicating actions by enhanc.29. genes do ated with in vivo intoxication. Anxiety may drive attempts findings indicate the potential importance of direct ethanol effects on to alleviate it through escalation of drug ingestion. that seems to involve this type of receptor28. impairment. Thus. the field often lation of the cluster of GABAA subunit genes in this chromosomal progresses in small steps. β2 or β3 and δ subunits25. 2a). each of which has its contribute to this link. unlike many ences in the alcohol sensitivity of GABAA receptors. Although the field is moving rapidly. very few studies in mammalian cell genomics and its related reductions (proteomics. although there are interesting parallels with the human α6 P385S high consumption levels in 20-year-olds are seen as less polymorphism. A gene that influences that ethanol can potentiate responses to GABA that are mediated by one behavioral contributor to risk is likely to be relevant for receptors containing the α4 or α6. but at relatively high will not be explained by concentrating solely on functional ethanol concentrations26. Differences in alcohol Individuals diagnosed with drug-related disorders are also sensitivity have been related to differences in alcohol intake. diagnosis does not imply a is also associated with a change in acute alcohol motor impairment particular pathophysiology. A study may identify one or two region. influence reflects the environment in which a gene’s effects These receptors are often found outside the synapse. Furthermore. genetic concentration-response functions observed in different laboratories.29. and thus this subunit would seem less 30 years7.nature. addition. neurons in brain slices from the cerebellar cortex and dentate gyrus different genes may be important. does Alcoholism and drug dependence are also complex at the this ethanol potentiation have in acute intoxication and alcohol abuse? behavioral level. and each relevant gene influencing susceptibility differential effects of ethanol in different cells.27. systems have produced evidence for this potentiating effect24. stressors) is studied. any amount of to reduced GABA function is not observed in human or even mouse drinking in an 8-year-old may be considered excessive. (Fig.

including those in the central amygdala. In cerebellum. Powerful molecular genetic methods can be applied in The early promise of invertebrate and non-mammalian vertebrate D. melanogaster mutant may provide one of the best models Invertebrate and vertebrate model organisms. which likely contributes (mIPSCs). cological conditions. Mutant ethanol concentrations and thus could certainly contribute to intoxica. tion of acute movement and egg-laying behavior have a loss-of-func- tion about the molecular mechanisms underlying this potentiation. These include increased paired-pulse facilitation and ethanol action. of the large-conductance calcium-activated potassium channel (BK Here we simply note that ethanol increases GABAergic synaptic input channel)46. several genes that influence acute alcohol sensitivity and toler. els for alcohol research may be complemented by the use of simpler and intracellular signaling pathways involving cAMP and protein kinase vertebrate models that are amenable to genetic analysis and genetic A (PKA). cerebellum and tion in C. melanogaster and C. Flies acetylcholine. elegans neurons46. along with unexpected Behavioral screening methods for zebra fish (a vertebrate model. Studies using cell lines derived from rodents had already if the behavioral outcomes are similar in D. Similarly. Thus. the insulin receptor and insulin-producing neurons. REVIEW in mammalian heterologous expression systems or on isolated neurons lush gene product is putatively involved in olfactory detection of etha- containing α6-δ receptors. The tractability of these organisms for genetic studies such as mutagen- mals. melanogaster is regulated by alcohol responsiveness or alcohol drinking in gene-targeted mice lack.com/natureneuroscience mission arises solely from increased GABAA receptor function is being cates how an alcohol-binding pocket in a protein might look. however. The lush mutant lacks an olfac. dopamine and glutamate are present in the nervous sys- lacking the LUSH protein do not avoid alcohol. intoxication remains to be determined. again raising the specter receptor in alcohol effects on the brain and behavior. The types and receptors that could alter the effect of ethanol in the different NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1475 . an alcohol-binding site identified by NMR and X-ray crystallography44. and it has in a single neuron under tightly controlled physiological and pharma. of problems brought about by alcohol’s solvent properties. ethanol sensitivity in D. other researchers disagree that LUSH is an alcohol binding could contribute to ethanol effects in more intact tissues. Future studies must focus on determining the melanogaster and rodent models. siderable commonalities in neurochemistry. an increase in firing rate of GABAergic inter. This elegant work demonstrates the power of invertebrate electrophysiological measurements indicating an increase in presynap. symmetry is emerging between the findings in D. are beginning manipulation. It remains to be seen if the nervous This finding rules out a host of more trivial interpretations of previous system circuitry of invertebrates can be generalized meaningfully to the findings (such as alterations in muscle function or metabolism). models for analysis of susceptibility genes and molecular targets of tic GABA release. within the nervous system much more rapidly than is possible in mam. Another D. presynaptic ethanol potentiation must The use of invertebrate animal models has also revealed potential be considered when formulating hypotheses about effects of GABAA molecular targets of ethanol action that had not been previously exam- receptor–targeted drugs on alcohol-related behaviors and changes in ined. concordance with previous findings from studies of vertebrates. For example. there is little informa. Mutation ing GABAA receptor subunits. elegans. in rodents. As yet. melanogaster CNS increases sensitivity to intoxication48. fur. but it is not yet clear whether any of comparability of molecular and cellular effects of ethanol in the differ- these proteins are direct targets of alcohol. It will be The spineless approach: invertebrate models interesting to determine to what extent the interactions between insulin Studies in invertebrate models have unearthed molecules with potential and the cAMP signaling system are involved in fruit fly intoxication. Neurotransmitters such tory protein involved in detecting ethanol (from rotting fruit)43. These investigators attribute the avoidance behavior to effects on GABAA receptors and the scope of the involvement of this plasticizing contaminants in ethanol solutions. do share con- for direct ethanol binding to a protein. to be implicated. the effects of ethanol on basic motor control circuits seem chronic ethanol41. Because this neuroendocrine system. because this rules out many indirect effects that However. there are notable differences in neurotransmitter reinstated by reexpression of the LUSH protein in mutant flies43. C. insulin. implicated cAMP-associated signaling proteins in effects of acute and For example. in contrast to wild-type tems of all these organisms. ent circuits found in vertebrate and invertebrate organisms. as well as transcription-associated proteins38. and many neuropeptides are also commonly flies. nantly by GABAA receptors. At many of these synapses. The role of BK channels in mammalian transmission at most GABAergic CNS synapses is mediated predomi. and mammalian brain. tion mutation in the gene that codes for the pore-forming subunit A recent review34 nicely describes this growing area of investigation. Clearly. whereas more sophisticated treatments that affect the cAMP-PKA signaling system in the nucleus circuits such as the basal ganglia and cerebellum contribute to motor accumbens42. ethanol alters several function. The alcohol avoidance behavior typical of wild-type flies can be expressed. Acute ethanol exposure also enhances BK channel func- onto neurons.39.nature. challenged by studies indicating that ethanol potentiates GABA release The nematode Caenorhabditis elegans has also been used to probe in several brain regions. Thus. elegans lines with decreased sensitivity to ethanol-induced disrup- tion and other alcohol-related behaviors. impairment in rodents. roles in alcohol intoxication and alcoholism. Furthermore. the rapidly emerging use of invertebrate mod- ance have been identified in mutant Drosophila melanogaster fruit flies. This issue The long-held belief that ethanol potentiation of GABAergic trans. their utility for understanding human alcohol abuse sensitivity in a subset of putative neurosecretory neurons in the CNS40. alcohol intake can be regulated by to underlie incoordination in invertebrates. Selective overexpression of a PKA inhibitor leads to localization esis screens suggests that further interesting developments may be on of the cAMP-mediated signaling important for regulation of ethanol the way. of the insulin receptor or its substrate in a subset of neurons within the D. protein and have challenged the idea that flies avoid high ethanol con- ther work is needed to determine the factors that contribute to ethanol centrations45. melanogaster as a model for study of ethanol effects behavioral effects of alcohol differ in invertebrates and mammals even on the CNS. However. melanogaster to alter protein expression in only a subset of neurons model systems has spurred intense interest in the alcohol research field. ethanol potentiates BK channels increases in the frequency of miniature inhibitory postsynaptic currents in neurohypophyseal peptidergic terminals47. but the crystal structure of LUSH certainly indi- © 2005 Nature Publishing Group http://www. However. This potentiation is observed at reasonable the genetic and molecular basis of acute ethanol sensitivity. which could alter motor and egg-laying hippocampus35–37. behavior49–51. to alcohol-induced changes in vasopressin secretion and diuresis within neurons also contributes to the increased GABAergic input36. For of course) have also been developed to study effects of ethanol on example. It is always reassuring to observe drug effects nol and other short carbon-chain alcohols such as butanol. and alcoholism is still in question. as it is possible that neural circuits contributing to helps validate D. has yet to be resolved.

Selective breeding has been used since the late 1940s to One of the methods recently brought into play to accelerate the prog- develop lines of rats and mice that differ markedly in voluntary alcohol ress of gene mapping efforts is to synthesize information about both drinking. Significant questions the past few years. technology has improved. and thus it is premature to withdrawal QTL on chromosomes 1 and 11.53. were first identified in 1959. models (standard and recombinant inbred mouse strains.57. thought to model different aspects of ethanol’s reinforcing effects. Methods exploring both gene sequence and of genetic animal model research drawn from quantitative trait locus gene expression variation are proving to be powerful tools to identify (QTL) gene mapping approaches and studies of null mutants. whose inter- Invertebrate models show great promise for alcohol research at present. more complete characterization of the molecular actions of of the close similarity (>85%) between the mouse and human genomes ethanol in multiple nervous systems will be needed to determine which owing to their shared ancestor. Numerous QTL affecting the important for developing better tools to analyze the genetics of alco. The GABAA receptor has been fairly thoroughly studied different drugs of abuse cited in an early review. which responses to alcohol7–9. indi. 64). over half were for using genetic engineering strategies. depends upon the presence of one or more genes within the chromo- ment. Both these traits. a locus very near the dopamine D2 receptor gene strongly affects its expression68. In multiple genetic animal mammals will have a role in invertebrate responses to the drug. or gene expression. despite tially contained several hundred genes. alcohol- on neural circuits and molecules across organisms should yield some related QTL do not act in isolation. REVIEW systems. At least seven different pairs ence. We now know that QTL ‘signals’ may does stimulation of dopamine D2 or D3 or GABAA receptors. the rate of progress has greatly accelerated in tions to norepinephrine in the mammalian CNS. congenic strains). and of the strengths and speculate too strongly on the similarities and differences in the systems. We know of no recent. Until recently. we have stressed similarities between molecular substrates ing populations. tions between the occurrence of the mapped trait and markers that These genetically high and low drinkers have some fairly consistent dif. that the gene’s different protein variants had different neurobiologi- tion of serotonin 5-HT1A or 5-HT2 receptors reduces alcohol intake. It will also be important facilitates coupling of ligands and receptors and interacts with 5-HT2A. It may well be that mammalian brains have codes for a multiple PDZ domain protein. Examination of the effects of alcohol in invertebrate A more detailed description of this project. Not surprisingly. Rather than leading to greater expression of a gene could be found preferentially in review the older material. of alcohol responsiveness in D. segregat- Thus far. The QTL region ini- © 2005 Nature Publishing Group http://www.and gene sequence–based sources of genetic influ- shunned it. mouse map of genomic markers made progress initially slow.nature. and activa. melanogaster38. to determine which invertebrate proteins are direct targets of ethanol 5-HT2B and 5-HT2C receptors as well as cKIT (a membrane tyrosine action (such as the BK channel) and which indirectly influence alco. elegans and rodents. Blockade also derive ultimately from differential expression of the underlying of opioid or 5-HT3 receptors also reduces intake. MPDZ protein are not prominent in the invertebrate brain. A wealth of historical data has been reviewed being pursued55. The influence on acute alcohol consensus on the similarity of intoxication in the different models. systematic review of the trait locus (QTL). hol sensitivity. Comparison of the effects of alcohol increasingly reciprocally informative62. but may have analogous cellular func. a QTL on mouse chromosome 4 has been linked to the severity of However. but as the ter is not found in mammals. and polymorphisms and gene the known importance of these neurotransmitters in mammalian expression data allowed the strong inference that the gene Mpdz. as cal functions affecting the trait. was the only remaining gene evolved alcohol-sensitive circuitry and neurotransmitter systems that in the region that affected the withdrawal response59. melanogaster. a specific genetic variation in a promoter region agents have some efficacy in the clinic with alcoholics56. octopamine is a monoamine neurotransmitter alcohol58. including genes for complex traits65 and are now being used on alcohol-related some gene expression studies. α6.28. In gene mapping. gene findings in mice and humans are mechanisms are most common. High drinkers gene near the markers was followed by functional studies showing tend to have low synaptic levels of serotonin and dopamine. and these classes of gene. we concentrate on a few recent examples high alcohol responders. elsewhere52. there are no findings to date implicating GABAergic or acute alcohol and pentobarbital withdrawal60. QTL mapping efforts simply identified associa- of lines of rats have been bred for high versus low alcohol preference. short-term selectively bred lines. and it is likely that information gained from such models will comple. γ2). The primitive tools then available and the sparsity of the implicated in ethanol actions in D. For example. Because Ultimately. kinase receptor) and p75 (a Trk-associated neurotrophin receptor). weaknesses of QTL mapping approaches. val contains multiple GABAA receptor subunit genes (α1. varied in DNA base pair sequence. a single QTL has been remain as to whether neuronal systems implicated in alcohol actions in definitively traced to the underlying gene59. C. For example. is given elsewhere61. Of well over 100 QTL for many responses to seven mutant studies. individuals’ behavioral responses Genetically engineered candidate genes are first compared with their genotype at many polymorphic markers Mouse stocks genetically engineered to have a null mutation for one scattered across the chromosomes. β2. which is also pursuing models is a relatively new research area. drawal severity are apparent only when considered epistatically with another genome region. To date in addiction research. For reasons discussed earlier (Box 1). as this receptor system has long 1476 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . they are quantitative traits). A pattern of association between a of about 50 genes or to overexpress them have been studied for one or marker and degree of response provisionally identifies a quantitative more alcohol responses. and pentobarbital withdrawal of a chromosome 11 QTL.com/natureneuroscience glutamatergic transmission in invertebrate alcohol effects. Inbred strains of mice that preferred to drink alcohol. tendency of mice to seek to drink or to avoid alcohol solutions are also hol-related behaviors. This neurotransmit. For example. that gained from the continued use of some 1 QTL63. which are initiated in the early 1990s. The chromosome 4 QTL containing the Mpdz Genetic animal models and alcohol responses gene also affects chronic alcohol withdrawal severity in combination Evaluation of the validity and utility of rodent animal models is also with a QTL on chromosome 8 (ref. QTL66–70. as well as with its efficacy Mapping approaches to identify genes affecting alcohol responses were to produce a conditioned taste aversion. some QTL affecting chronic ethanol with- mammalian laboratory animals. but rather are quantitatively distributed in populations (that is. Levels of DRD2 protein are cor- Finding new genes through gene mapping related with ethanol’s motor stimulant effects. Similarly. but certainly not replace. Positional cloning of a candidate ferences in the neurobiology of alcohol responses53–55. show vidual differences in responses to alcohol are not generally all-or-none a strong QTL signal in this region as well.

In other words. and gene deletion of the α1 subunit found that nearly all such studies (at least with mice) showed a marked gene affected sensitivity in one of two such null mutant models. Given the behavioral repertoire of flies and worms. relate the 5-HTTLPR polymorphism to depression or other psychiatric Several behavioral manipulations can overcome these controls in the diagnoses have met with varied success86.org) with the location of QTL that had been or social life). laboratory75–79. they have been able to model simple sensitivity to the sedative effects of allowing the complexities of GABA-ethanol interactions to be reduced alcohol. as well as for the vince some rodents to drink to intoxication on a daily basis. apparent dysregulation of multiple neural systems that normally are the intense social pressures loosely called ‘peer pressures’ experienced by held within bounds by homeostatic feedback loops. including the serotonin transporter (5-HTT) that drinking mentioned above. The effect was seen only in mon- idea that scheduling access to alcohol during the circadian dark78. with three common but they will rarely continue to drink thereafter until their blood levels genotypes termed s/s. but how these have been less convincingly modeled than the biological sequelae of miscalculations play out in the brain’s reward and stress axis circuitry is chronic administration. We are fairly confident that the validity across the biomedical science community. β3. rodent models in place are a reasonable reflection of these psychological including those from which alcohol is subsequently withdrawn. Functional and structural brain imaging studies suggest that chronic alcohol abuse has localized effects on brain circuits72. However. able for analysis is frustratingly incomplete. levels of alcohol before offering it for self-administration83. Mice and rats will drink until they begin to inhibiting 5-HTT. Alcohol dependence adolescents) will never be modeled in non-human species. α2. craving and reduced Studies of tolerance and dependence seem to have reasonable face or increased sensitivity to reward. although some protocols seem and because many different genotypes were used. which is not as straightforward as For example. The specific serotonin reuptake bred for preference. the ultimate utility of these newer This review compared the transgenic behavioral data with those for models will be achieved if it proves possible to selectively breed for expression of the various subunits in brain. an anxious state. or s/l for short (s) or long (l) variants. it difficult to infer anxiety specifically. γ2L. Because the chromosomal very high intakes.80. Because animals cannot self-report. it will be difficult to devise assays tapping such internal states as anxiety. β2. were proposed as potential treatments for alco- periods will develop patterns of repeated drinking that are obviously holism on the basis of animal studies. of available data will serve to identify the lacunae in our knowledge. positive or negative effects of drinking alcohol73. a recent genotypes. Genetic models: challenges and future directions response inhibition or its opposite (impulsivity). including depression and alcoholism. rodents seem to have some internal more short alleles leads to reduced 5-HTT function. Compounds targeted at brain serotonergic systems. α6. ioral sensitivity in the mutants and gene expression patterns taken from Some of the principal defining characteristics of the alcohol depen- a publicly available set of informatics tools and data sets. all the rodent genetic animal models of alcohol by several proteins. Experimental studies seek to con- Integrative Neuroscience Initiative on Alcoholism of the US National strain complexity by using uniform environmental conditions. monkeys with an analogous short variant promoter polymorphism in throughput rodent models in which animals voluntarily self-administer the transporter gene87. display aspects of alcohol dependence. GABA transporter gene. Null mutants and/or limiting access to fluids for a brief period80–82 or establishing chronic transgenic overexpression mutants have been generated for the α1.genenetwork. and Serotonin and alcohol: translating animal research to humans functional MRI signals measured during a behavioral inhibition task The neurotransmitter serotonin (5-HT) is important for normal and may prove to be useful to predict individuals’ expectancies about the dysregulated emotional behavior. However. reduction in general activity of the animals during withdrawal. but these generally require fairly long-duration expo. The s/l heterozygotes show greater alcohol self- ethanol to excess. Results have supported the approach may obscure genetic effects. known to A recent review of this work compiled behavioral data for several drink large quantities of alcohol. morphism (5-HTTLPR) in its promoter region. we must infer their mapped for the same behavioral responses. γ2S and δ receptor subunit genes. However. by QTL mapping efforts. Thus. strongly reduced activity in an apparatus such as the elevated plus maze makes suggests a role for the α1 and α2 subunit genes in this response71. To date. the data matrix avail. However. it was also possible to compare behav. it would be useful to have more high. One or subside somewhat. The α1 subunit gene was differentially expressed in mouse review of rat and mouse studies of anxiety during alcohol withdrawal strains that differed in sensitivity. However. l/l. including SSRIs Some rhesus monkeys given voluntary access to alcohol daily for long and 5-HT3 antagonists. interference with work (http://www.84 can con- α5. and anxiety is reported by alcoholics when they are ing sensitivity to alcohol-induced loss of righting reflex in some mouse withdrawing from a period of chronic drinking. but attempts to controls limiting intake that are not shared by susceptible humans. to work reproducibly.com/natureneuroscience location of these genes is known. withdrawal severity and tendency toward self-administration71. act by like human alcoholism. Because and similar data for the α2 subunit gene on chromosome 5. subjective state from their behavior. keys separated from their mothers at birth and reared with peers after NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1477 . low serotonin transporter activity is also seen in rhesus sure and are labor intensive. it is methodologically tricky to Because many different ethanol-related responses have been studied study withdrawal anxiety in rodents85. Other challenges facing animal model research are more generic. not yet clear. most frequently used to treat depression. including high-dose sensitivity (loss of righting seem to be able to overcome the (unknown) factors that tend to self- reflex). show patterns of drinking that do not look exactly inhibitors (SSRIs). REVIEW been a target of mechanistic studies (see above). most of these studies have focused on the C57BL/6J mouse. and the reduction in that sensitivity (tolerance) with chronic to a more tractable set of possibilities. Although it is encouraging that we ethanol responses. Attempts to create such models are supported by the administration than l/l homozygotes. To date. Dependent animals. Brain levels of 5-HT are regulated excessive74. Activity in the 5-HTT gene is affected by a poly- approach brain alcohol levels that produce clear signs of intoxication. exposure. but this Institute on Alcohol Abuse and Alcoholism. There are many rodent behavioral assays designed to reflect rich chromosome 11 region was found to harbor a gene or genes affect. limit ingestion in these creatures. even those animals intensely selectively removes serotonin from the synapse. in humans obviously must disrupt in some way the overall calculation the feelings experienced by alcoholics at different stages of their disease of an individual of the risks and benefits of drinking a lot. some features (for example. such syntheses Invertebrate models are not exempt from these challenges. In general. WebQTL dence disorders are behavioral (for example. the GABAA receptor subunit– it seems15. This.nature. © 2005 Nature Publishing Group http://www.

dysregulation of alcohol self-administration. including alcohol and drug abuse.. 10. The wide array of potential therapeutic targets raises © 2005 Nature Publishing Group http://www. such as anhedonia or disruption of normal neurophysi- environment interaction between the transporter polymorphism and ological drives. 76S–81S (2001). laboratory animal studies indicated that opiate receptor antagonists and G. 7.).C.. 25. M. Antagonists of the CB1 cannabinoid receptor and the mGluR5 (2002). T. The known they were mother reared89. Rutledge-Gorman for help in preparing the manuscript. AA12714 ral circuitry involved in alcohol ‘craving’ and relapse56. et al. R. 1–26 (2001).C. there is great interest in translating basic findings on tion of basic findings into clinical treatments. This study awaits replication. E. Long-term studies may yet reveal that general 15. 79–94 (2003). S.G. tive antagonists of the NMDA receptor. 11. Cunningham. Elevated levels of GluR1 in the midbrain: a trigger for sensitization to drugs of abuse? Trends Neurosci. brain reward systems. We thank M.. Genetic basis of ethanol reward. Exp. J.. Pharmacol. Rev. Rev. Higley..99.C) and the Division of Intramural Clinical and Basic Research was not developed on the basis of data from animal models. Rev. 13.nature. ethanol. 77. Cellular and molecular neuroscience of alcoholism. which itself is predictive of the hope that effective treatments will be found in the not-too-distant later alcohol-related diagnoses90.C. Ann. monkeys of both sexes showed elevated adrenocorticotropic hormone Two metabolic genes (alcohol dehydrogenase and acetaldehyde dehy- (ACTH) if they were l/s heterozygotes. Nestler. but l/s females showed an even drogenase) have specific polymorphic variants whose unpleasant effects greater increase in ACTH if they were peer reared and no increase if in their bearers significantly protect against alcoholism3. Alcohol. & Gordon. There is a substantial and growing literature implicating the 9. nicotine and even food-related aspartate glutamate receptors and alcoholism: reward. 937. Sci. vulnerability. Weiss. A. & Morris. inhibited by ethanol94.J. the exception of naltrexone. C. 121–146 (2004). 28. behavioral models96. Crabbe. Diamond. Several interesting targets for potential pharmacothera- l/l homozygotes have extremely high levels of 5-HT transporter activity. Ueno.nature.. potential of CB1 and mGluR5 antagonists in human alcoholics. Pharmacogenomics 5. Mason. Heinz. 2003). W. & Goldman. J. An analysis of the genetics of alcohol intoxication in inbred mice. The CB1 and mGlu receptors are just two among many life stress on human depression88.L. Neurosci.). Crabbe. the US National Institute on Alcohol Abuse and Alcoholism (AA10760. Smith. early clinical indicators suggest that drugs tar. Alcoholism: genes and mechanisms. acamprosate and the SSRIs.H. The authors declare that they have no competing financial interests. Alcohol. Trevisan. D. & Phillips. 25.com/natureneuroscience disorders and a low response to alcohol. so interference with brain reinforcement systems might have untoward there is an intriguing parallel with a study showing a similar gene-by.. J. Oroszi. G. Alcohol actions on GABA(A) receptors: from protein structure to mouse The roles of CB1 and mGlu receptors in brain reinforcement and behavior. M. J. Carlezon.. I. as well as identification of new genes and their proteins that are involved in alcohol’s neural Basic-clinical translation: developing pharmacotherapies actions. Cameron. matrix of compounds and behavioral tests is frustratingly patchy93. The search for genes contributing to the low level of response to alcohol: patterns of findings across studies. Neuropharmacology 47. one variant therapeutic agent disulfiram (Antabuse) is directed at this metabolic of the l allele of 5-HTTLPR significantly predicted both alcohol use susceptibility. & Wahlsten. 5. Totowa. Petrakis. control of addictive behaviors are not confined to interactions with 8.D. a neurotransmitter that is directly 3. Annu.L. and suffer studies. Nestler. Res. Gonzales. The role of mesolimbic dopamine in the targeted therapeutics are successful because they tap into generalized development and maintenance of ethanol reinforcement. A different (missense) mutation in the 5-HTT gene itself has been identified in two Summary generations of a family with a high incidence of obsessive-compulsive Much recent progress has been achieved on both fronts highlighted in disorder. For example. N-methyl-D- brain cannabinoid system in opiate.com/natureneuroscience/ alcohol or the apparent involvement of the target in control of alcohol Reprints and permissions information is available online at http://npg. Neurosci.J. dependence.A.. These findings suggest that CB1. Genetic contributions to addiction. exercised in touting these neurochemicals as targets for treatment of any 14.A. the current Physiol. and addictions98.97. & Goldman. G. Is there a common molecular pathway for addiction? Nat. When subjected to isolation stress. 1–20 (1997). J. Pharmacopsychiatry 36 and gene-targeted animals lacking CB1 show reduced intake in some (Suppl. Job. Ther. 103. geted at the CB1 receptor will be useful in the treatment of obesity and Suppl 1. Krystal. Festina lente: Late night thoughts on high-throughput 1478 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . Neuroadaptation. T. Molecular mechanisms of drug addiction. Pharmacol. but that end point seems the neural actions of alcohol into potential clinical applications. Indeed. Exp. 28. whereas (D. E. NY Acad. However. Two closer than ever. 5-HTTLPR polymorphism has pleiotropic effects. A. cessation of cigarette smoking100.M. successfully use information gained from animal models for transla- tive therapeutics. consequences. 2.H.com/ drinking in rodent models. closely related to other addictions that involve these systems comes as 1445–1449 (2005). and conditioning factors.M.J. et al. Clin. & Nestler. noncompeti. A parallel effort is examining the efficacy of reprintsandpermissions/ potential therapeutics in the broader array of behavioral assays men- tioned above.J. pies have emerged from continuing pharmacological and physiological possibly owing to their ‘double hit’ at the two mutant genes. Clin. treatment. are being tested for efficacy in reducing alcoholic 4. W. J. 53. Metten. the finding that the glutamate and cannabinoid Biology of Drug Addiction (ed. P. Biobehav. Ther.C. and in CNS) and naltrexone (an opiate receptor antagonist)—target neu. New genetic animal models are emerging in both vertebrate from multiple disorders. Maldonado. In a human genetics study. medications currently used in the treatment of alcoholism—acam- ACKNOWLEDGMENTS prosate (a taurine analog with no well-established molecular target The authors are supported by the US Department of Veterans Affairs (J. New systems may have generalized roles suggests that caution should be Jersey. R. metabotropic receptors reduce alcohol intake in several animal models. 1037–1048 (2004). reduce alcohol intake before the first tests of naltrexone in alcoholics92.S. Acamprosate and AA13519 to J. & D’Souza. Neurobiological correlates of the disposition and maintenance of alcoholism. Plans are underway to evaluate the therapeutic 6.J. Compulsive drug-seeking behavior and relapse. & Doyon. stress. D.C. I. Psychol. F.C. and with 1. 435–462 relapse95. D. Asperger’s and invertebrate organisms that will lead to greater levels of behavioral syndrome and social phobia91.) S255–S258 (2003). R.L. Res. therapies aimed at molecular targets identified by their sensitivity to Published online at http://www. molecular targets being considered for alcoholism pharmacotherapy93.nature. L. 1449–1458 (2004). one addiction disorder. REVIEW five weeks of nursery rearing. Analogous to the situation with null mutants. McClearn for many previous versions of Figure 2.. Crabbe. The finding that alcohol abuse is likely to be 12. Significant barriers still must be surmounted before we can As the ultimate goal of much alcohol research is development of effec. J. in Molecular no surprise. 99. & Kalmijn. Schafer. Jr. Some promising new targets are emerging.L.M.C. D.A. 8.O. M. 24–32 (2004). E. 785–802 (2005). Peer rearing is demonstrably stressful.) 263–294 (Humana. Krystal.. A. Individuals with the missense mutation who were also this review. Schuckit. COMPETING INTERESTS STATEMENT Researchers have focused on developing and testing new pharmaco. The future. 610–615 (2002).

Neurocircuitry in alcoholism: a substrate of disrup- 40.A. K. Van Patten..S. L. in The Cell pal neurons. et al. G.J. 17.G.. Neurosci.P. Behav. Neurosci. A central role of the BK potassium channel in behavioral responses mice drinking ethanol in a limited access session. Alcohol enhances GABAergic transmission 233–242 (2005). Behav. R. & Tapert. 5. Belknap. Jr. Neuropharmacology 39. 1512–1514 (2004).J. 2000). & Li. Methods in Alcohol-Related Neuroscience Research (CRC.. & Williams. Nat. Soyka. 1581–1602 39. & Weiner. Rhodes. 339–345 (2005). elegans.. quantitative trait Pharmacol. Fehr. Best.H.W. H. 37.A. 64... M. NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1479 . J.J. Neurosci. 25.J.J.N. Trends Genet. Kruse. 363–388 (2002). Nature 436. Davies. α1. J. 3730–3738 (2002).R. D. & Hoffman. M. The role of GABA(A) drinking behavior in rodents. 57. 323–331 (2005).. J.. Philadelphia. D. cynomolgus monkeys (Macaca fascicularis): long-term characterization of sex and tor-stimulated PKA signaling and regulate ethanol consumption. behaviors. 72. Y.. Bloom. J. 10.L.A. J. Hubner. Ther. et al. Murphy. 75. Teratol. J. A. 918–925 Nat. Yeh.. U. Rodd-Henricks. F. New York. Psychopharmacology (Berl. 20. Ann. Bystrykh. High ethanol consumption and low sensitivity to ethanol-induced (fMRI) with adolescents. H.V. Neurosci. tification of genes underlying disease. sedation in protein kinase A-mutant mice. J. J. Res. Browman. Insulin signaling in the nervous system 80. F. Vivian. W. & Seeburg. J. Med.A. J... Pharmacol. Clin.. et al. 733–743 individual differences. J.. 7. Proc. K.L. T. Life Sci. Production of polydipsia in normal rats by an intermittent food schedule. Jansen. Biochem. receptors are a target for alcohol. Neurosci. Direct or indirect actions? Curr. R.L. 5. Lockwood. Kiger. & Bradley. (2002). Zhou. 385–389 (1997). Madras. Mol. Res. Exp. A. Sundstrom-Poromaa. Natl.. Ca(2+)-activated K+ channels in isolated neurohypophysial terminals. Naunyn Schmiedebergs Arch. developing zebrafish: neurobehavior and skeletal morphogenesis. Kettenmann. Alcohol.L. Thomas. & Ryabinin. (2004). J. Struct. D. Exp. S. Chem. 68. Alcohol. Pharmacol. Structure of a specific alcohol-bind.A. Ethanol potentiation of GABAergic synaptic transmis. Otis. Dwyer. Alcohol. E. 655–666 (2003). Neurotoxicol. I.nature. Mittleman. Res. A. 69.M. Mol. S69–S80 869–885 (2002). Alcohol-induced 60. M. et al. J. E. 2. 37. Congenic mapping motor impairment caused by increased extrasynaptic GABA(A) receptor activity. Psychopharmacology (Berl. to ethanol in C. 58. 26. H. 62–65 51. & Heberlein. Repp. Res. 63. Johnson. Examining (2002).C. of alcohol and pentobarbital withdrawal liability loci to a <1 centimorgan interval Nat. 265–275 (2000). S. Mihic. J. Sharpe. Sullivan. cell function using ‘genetical genomics’. N. J. & Mody.D.H. Crit. Alcohol 66. M. B. Am. D. Trans. 8. Behav. Walter. 18–19 simple model of ethanol drinking to intoxication in C57BL/6J mice. & Crabbe. 77. 25. Acute effects of alcohol on larval sites. C. butyric acid type A receptor: lack of effect on responses to ethanol. Genome 14.W.. 2000). Carta. Falk. Chronic ethanol consumption by C57BL/6 mice promotes tolerance to its interocep- sensory responses to alcohols in Drosophila melanogaster. Reilly. USA 100. Molecular genetics of substance abuse vulnerability: remarkable recent motor impairment linked to point mutation in cerebellar GABAA receptor.J. A. et al.O.. J. 218–226 53. P. Nature 361. Franz. Behav. 17.K. Nat. M. Hormonally regulated alpha(4)beta(2)delta GABA(A) (2000). Pharmacogenetic studies of alcohol self-administration 139.. γ2 epistasis in a complex trait: Pentobarbital withdrawal convulsions in mice.. Brain Res. 110. Thiele.E. Identifying genes affecting addiction risk in animal models. 1419–1426 (2005). 38. The hangover gene defines a stress pathway (2004). D. S. J.J. K. Ariwodola. 225–232 (2005). 76. Nat. tonic inhibition mediated by delta subunit-containing GABAA receptors in hippocam. & Heberlein.) (Cold Spring Harbor Press. future. A. & Morrow.C. centrations and repeated deprivations on alcohol intake of alcohol-preferring rats.L. Pharmacol. Weber. 32.J. Behav. Neurosci. J. 539–560 (2004). Carvan. & Wallner. 1892–1900 (2003). (2003). Wolf. & Ait-Daoud. McBride. Peoples. GABA(A) receptor subunit genes associated to a rat nonpreferring ethanol phenotype. Schweinsburg. 356–359 (1993). Hitzemann. 1437–1448 (2004). 74. Rev. Bhave. M.E. 31. J. Alcohols and neurotransmitter gated ion channels: past. Acad. Cell.C. D.. convergence of genome scan results. 721–722 (2002). 54. Roberto.S. Induction and maintenance of ethanol self-administration in 42.. required for ethanol tolerance development. Mammal. et al. RC75. Aguayo. 757–768 (2004).G. Schuckit. Psychopharmacology (Berl. J. D. J. Kobayashi. Genet. A.C. Congeddu. 8379–8382 (2004). Cell 115. & Buck. Genet. III. 66. et al. 453–463 (2004). Genetics 150.K.K.M. Pharmacol. E. Neurobiol. 16. Belknap. J.L. Mameli.. 45. Exp. Tsivkovskaia. & Phillips..V. Alcohol. T. B. present and Genet. H. On the integration of alcohol-related quantitative trait loci and and general anesthetics. & Yevenes. 583–594 (2005). Complex trait analysis of gene expression uncovers polygenic 10679–10686 (2004). Belknap. gaster: evidence for odour recognition and discrimination. Siggins. & Buck. U. Effects of concurrent access to multiple ethanol con- conductance. Stud. 44. G. 1–13 (2004). Nat.E.. K. 67. R. E.J.C. Zhao. Top. Van Brunt. Z. 699–700 (2004). A. 647–654 (2004). 7. & Buck. 8. 15218–15223 (2003). Sites of alcohol and volatile anaesthetic action on GABA(A) and 57. Loucks.A. Revisiting the odorant-binding protein LUSH of Drosophila melano.. J. Exp. U. 19. A. 34. A. M.J.P. J. Bjerke. & Treistman. A. Biology of Addiction (eds. Shirley.K. 37. D. 29. Evaluation of a regulates ethanol intoxication in Drosophila melanogaster. Bergeson. M. B. 107. 1025. Science 303... Pharmacol.P.. Chesler. et al. 1979)... T. & Buck... Science 264. & Olsen. T.E. P.K.A. Behavioral characterization of alcohol-tolerant and alcohol-non. 388–391 (2001). ing chronic ethanol exposure and withdrawal. A. 54.R.F. L.. 711–721 tive cues and increases extracellular dopamine. S. Animal models of alcoholism: neurobiology of high alcohol- 23. 68. zebrafish: a genetic system for large-scale screening. et al. Crabbe. E. M. Use of acamprosate and opioid antagonists in the treatment of ular sites of ethanol action. Lemos.M. Finn. severity. et al. Nat. trait gene for drug withdrawal seizures. S. 30. Middaugh. Brain Res. Exp. and pleiotropic networks that modulate nervous system function.) 149. R. Anderson. Schedule-induced ethanol self- ing site defined by the odorant binding protein LUSH from Drosophila melanogaster. 267–282 (1997). Ataxia and c-Fos expression in 46.E. Marlowe. Y.M. Devaud. Selective breeding. cal loci regulating ethanol sensitivity in Drosophila. Clin. Boehm. 93–100 (2001). Res. Rodan. NY Acad. II et al. M.R. 27.B. Phenotypic and genotypic characterization of the Indiana © 2005 Nature Publishing Group http://www. in press).com/natureneuroscience (2000). Chromosomal loci influencing chronic alcohol withdrawal 32. Genet. Uncovering regulatory pathways that affect hematopoietic stem 3746–3751 (2004). University rat lines selectively bred for high and low alcohol preference. K. W. Wei.D. Yao. 25. Molecular characterization of new polymorphisms at the β2. (2005). Scholz. Liu. 28. T.C. & Morrisett. M.C. Nat. H. Crabbe. Ethanol enhances alpha 4 beta 3 delta and abuse. Neurosci.. Addict. Sci. Physiol. abuse research. & Lovinger.. J. Chemical properties of alcohols and their protein binding 50. Hanchar.L. F. R. L. Nat. Z. D. Crabbe. 1715–1723 (1994). S. 9490–9501 73.W. Biol. Res.. Olsen.M. 12 Suppl 1. (2003). Grobin. Genetical genomics: the added value from segregation. 34. G. J. and gene arrays identify candidate genes for complex drug-related 35. & Chick. pentobarbital.H. 80–98 (2005). Cell 109. 77. Clin. 24. Lovinger. C.J. 22. tolerant rat lines and an F(2) generation.L.B. E. aptic but not extrasynaptic targeting of NMDA receptors. 289–297 (2003). Faria. J.. & Matthews.R. 33. D.. R. et al. 22.C. an effect blocked by naltrexone. & Li. Crabbe. S. A. Matthews. Simantov. in Psychopharmacology: The Fourth Generation of Progress (eds. 51. K. & Nie. Alcohol. 56. Brown. Neuropharmacological treatments for alcoholism: glycine receptors.. 588–596 (1997). Sci. Clin.P. (1998). GABA(A) receptors as molec. & Jones. 327–344 26. S. 195–196 (1961). Invertebrate models of drug abuse. 24. 24. A..J. REVIEW screening of mouse behavior. Neurosci.) 55. J. Uhl. et al. Drug and Alcohol Abuse (Plenum. Crabbe. 339–369 (1998). to cerebellar granule cells via an increase in Golgi cell excitability.J. Ethanol augments GABAergic transmission in the central amygdala via 70.A. Washington. 27. Nature 389.N.L. C. Low ethanol concentrations selectively augment the 22.L. D. 37.. & Heberlein. 40–48 (1996). P. Genetic strategies in preclinical substance (2004). Rodan. Woodward. Mol. O. Gerlai. Exp. P. et al.R. new perspectives on alcohol actions. receptors in the acute and chronic effects of ethanol. et al. Genomewide search for 31. et al. Kim. Neurosci.C.K.C.R. Neurobiol. et al. L.B. G.J. locus analysis. N. L. & Snyder. 71.S.G.. & Smith. Crabbe.N. βγ dimers mediate synergy of dopamine D2 and adenosine A2 recep. D. Mol. & Chandler.A. alpha 6 beta 3 delta gamma-aminobutyric acid type A receptors at low concentrations 59. Functional dissection of neuroanatomi. Shirley. Homanics.P. Neurosci.J. K. & Pfefferbaum. M. Kleingoor.E. K. Gene knockout of the alpha6 subunit of the gamma-amino. 1087–1097 (2001). Genet.. R. Behav. 23. 1140–1150 (2001). R. personality and alcohol expectancies using functional magnetic resonance imaging 41. 356. U. C.. Neurosci. B. R. Radcliffe. of murine chromosome 4: identification of Mpdz as a candidate gene. Integrated transcriptional profiling and linkage analysis for iden- CRF1 receptors. 1175–1179 (2004). 2–19 (1998). Dopico. Mpdz is a quantitative known to affect humans. Corl. Hood. Benzodiazepine-induced 62.K. C.A. Zebra fish: an uncharted behavior genetic model.P. H. 36.. L. Clin. 23–26 Science 133. 24. Genet. J. Genet. 78. R. Szumlinski. Clin.C. and withdrawal. 694–700 (2003). 1–6 (2000).T.) 174. & Kalivas. administration in DBA/2J and C57BL/6J mice. 18. The opiate receptor. 65.. Ethanol increases the activity of large 79. Biochem.W. Biochem. 48. et al. Nie.) (Lippincott. G. I. 49. 243–253 (2005). 43. gene expression analyses. J. 12. The tipsy terminal: presynaptic effects of ethanol. Psychopharmacology (Berl) 180. Genetic animal models of alcohol and drug 27.W. Dynamics of NMDAR-mediated neurotoxicity dur. R. S. Fehr. & Heberlein. Ethanol effects on the (1977). Neurosci. et al. Hanchar. Korpi. Z. 845–847 (2005)... R. Williams & Wilkins. 4491–4498 (2003). Wallner. J. scientific basis and clinical findings. M. J. N. 461–468 (2003).W. LUSH odorant-binding protein mediates chemo. 558. 29. 33. & Kupfer. L. Alcohol.L. 21. & Valenzuela. Smith.E. 20. 49. tion and repair. 47.J. J. Carpenter-Hyland. K. 28. 24. 7859–7868 52. Neurosci.S. gamma-Aminobutyric acid A receptor subunit mutant mice: 161–178 (2003).. 454–463 (2003).. sion may be self-limiting: role of presynaptic GABA(B) receptors. Belknap.E. S. Paulus. & Guo. & Nap. alcohol dependence: a European perspective. S. Chronic ethanol induces syn. FEBS Lett. Dodson. Soc. et al.J. Tabakoff. R. N. 61.

28. Gen. M.M. Psychopharmacology (Berl. Lancet 365. Clin. Scheduled access to ethanol neuropsychiatric phenotype. Bachteler. 81. Exp. Scheen. 1389–1397 (2005). & Goldberg. Anguelova..W. & Crabbe. Hungund. 1146–1152 (2004). G. 837–850 (2005). Science 243. Rissanen. M. Arch. 515–527 87. Lopez. 53–63 (2005). S. Effects of the serotonin transporter gene promoter variant in rhesus macaques. Res. & Spanagel. et al. et al. the cannabinoid-1 receptor blocker rimonabant on weight reduction and cardiovas- Acad. 83. Effect of pattern and number of chronic etha. 8. Endocannabinoid control of food intake and energy balance. Proc. J. Opioids and alcoholism. Le Foll. 8. Exp. N. Hu.M. results in motor impairment and tolerance in female C57BL/6J mice. (2005). Biobehav.L. Neuropsychopharmacology 86.. C. & Rossner. Mol. Psychiatry 8.I. Exp.L. Anxiety-like behaviors following chronic ethanol exposure. 92. Reestablishing an intragastric ethanol 93. A. L. Barr. S. C. Lovinger. 339–358 Biochem.. Influence of life stress on depression: moderation by a polymor. Nagy. Nat. Barr. Di.S. & Basavarajappa. 88. 933–936 (2003). Res. G. et al. Koch. Rev. A. 471–480 (2005).C. B. An expanded evaluation of the relationship of four alleles to the level study. Alcohol. Ann. & Turecki. medications for drug dependence.. Behav.. C. O. 312. Renaissance of NMDA receptor antagonists: do they have a role in the 005-0026-3). cular risk factors in overweight patients: 1-year experience from the RIO-Europe 90. 386–389 (2003). K.) published online 7 July 2005 (doi:10.F. 94. Sci. Egli. T. A procedure to produce high alcohol intake in mice. pharmacotherapy for alcoholism? IDrugs 7. A systematic review of association 29. X. Pharmacol. J. 875–883 (2005). Ziegler.M. 12358–12363 (2004).S.nature. L. T.F. 84. D. M. Mol. studies investigating genes coding for serotonin receptors and the serotonin trans. Interaction between serotonin transporter gene variation and rear. & Cunningham.. 95. Psychiat. 339–350 (2004). et al. Kliethermes. G. Cannabinoid CB1 receptor antagonists as promising new phism in the 5-HTT gene.J. Ethanol inhibits NMDA-activated ion nol exposures on subsequent voluntary ethanol intake in C57BL/6J mice. M. P. 1480 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . ing condition in alcohol preference and consumption in female primates. (2004). Cronise. Behav. NY Acad. H.com/natureneuroscience Neurosci.F. MPEP reduces ethanol-seeking and relapse behavior.. Science 301. & Becker. 921–928 (2004). (in press). F. 29. Metten. clinically effective medications for alcoholism? Addict. J. Affective disorders. Can experimental paradigms and animal models be used to discover self-infusion model in rats. B. & Weight.C. 89. 81. Biol. 96. Finn.S. 99. White. Oswald. 574–591 (2003). of response to alcohol and the alcoholism risk.R. 85. C. (in press).A. 585–589 (2005). D. Clin. Neurosci. Sexual dichotomy of an interaction between early adversity and 100. S. mGluR5 antagonist © 2005 Nature Publishing Group http://www. 8–16 81. current in hippocampal neurons. A. Fidler. et al. REVIEW 84. R. D. Role of endocannabinoids and cannabinoid porter: I. 91.L. P. D. B. Hyytia.L. Benkelfat. USA 101. 98. Ther. Clews. et al. Psychopharmacology (Berl. 97. 943–953 (2005). 895.V. Alcohol. Backstrom. Physiol. Finn. & Wand.. (2004). Serotonin transporter missense mutation associated with a complex 82. 1721–1724 (1989).A. & Matias..S.. Psychiatry 61. 1025. P. Ozaki. Caspi. Sci. Natl. CB1 receptors in alcohol-related behaviors. Van Gaal.. Pharmacol.1007/s00213.) 178. C.

We hypothesize that the change from voluntary drug use to more habitual and compulsive drug use represents a transition at the neural level from prefrontal cortical to striatal control over drug seeking and drug taking behavior as well as a progression from ventral to more dorsal domains of the striatum. automatic or reflexive) elements of approach and consumma- (Box 1). acquired through action-outcome learning between drug reinforcement. These motivational effects of CSs Barry J. increasing rates of respond- ing for food when the CS is presented unexpectedly (called pavlovian– instrumental transfer. 1) to mediate the ‘switches’4. Here we elaborate the hypothesis that these expectancy.1038/nn1579 in response to such factors as unpredictability7. The nucleus accumbens is well known to mediate the reinforcing effects stimuli that are closely associated in time and space with the effects of drugs. (Box 1) that constitutes instrumental drug reinforcement.3. including the shell and core components of the nucleus accumbens. Drugs act as ‘instrumental reinforcers’—that is. also be active. and by an improved understand. such that it becomes habitual and ultimately compulsive. corrected after print 13 April 2006. Drugs produce subjective or in the processes leading first to drug abuse and then to addiction. has led to the concepts of action-outcome and stimu. In particular. In particular. University of Cambridge. Cambridge CB2 3EB. CSs that predict natural reinforcers. dynamic such interoceptive and exteroceptive states. We postulate that any processes2. Stimulant drugs such dent. drug abuse and drug addiction. under other circumstances doi:10. the sense of ‘control’ over behavioral processes can be mapped onto the parallel and serial. which include the sensing of autonomic activity has been stimulated by progress in understanding the dopamine-depen. whether in terms of enhanced Institute. conditioning and the nucleus accumbens can have several effects on behavior. Reinforcement.uk ior. or perhaps even more importantly. of pavlovian conditioning (Box 2). the description of two processes that seem to combination of these effects may constitute the ‘rewarding’ effect of a function partly in parallel. we argue that it is the sense of lus-response (‘habit’) learning. ing of associative learning mechanisms that conceive of behavioral especially those that already predict important environmental events. Considerable evidence now shows that the midbrain dopamine neurons show fast phasic burst firing in response to such CSs6 but may Published online 26 October 2005.com/natureneuroscience drug addiction: from actions to habits to compulsion Barry J Everitt & Trevor W Robbins Drug addiction is increasingly viewed as the endpoint of a series of transitions from initial drug use—when a drug is voluntarily taken because it has reinforcing. novelty.5 arousal accompanying them. in addition to eliciting pavlovian The reinforcing effects of addictive drugs are multidimensional (that is. perceptual impact or incentive salience of environmental stimuli. These neural transitions may themselves depend on the neuroplasticity in both cortical and striatal structures that is induced by chronic self-administration of drugs. they increase tory behavior. output as an interaction between pavlovian and instrumental learning which are known as conditioned stimuli (CSs).nature. locomotor activity or increasing rates of instrumental (operant) behav- e-mail: bje10@cam. (‘feelings’) or distortions in sensory processing. This view ‘discriminative’ effects. UK. involving its dopaminergic innervation. PIT. N E U R O B I O LO G Y O F A D D I C T I O N REVIEW Neural systems of reinforcement for © 2006 Nature Publishing Group http://www. the subjective effects produced by attributions made about behavioral output. including the overall level of functioning of corticostriatal circuitry (Fig. effects—through loss of control over this behavior. CSs self-administration or ‘drug taking’ (defined in Box 2). stress and NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1481 . but more recent research emphasizes the role of the striatum as a of self-administered drugs gain incentive salience through the process whole. Downing Street. Environmental can have motivational effects: for example. Everitt and Trevor W. but with the second eventually dominating drug—that is. which Psychology and the MRC-Wellcome Behavioural and Clinical Neuroscience serves to energize or activate responding. The locomotor stimulation produced by psychomotor the likelihood of responses that produce them. at least in their tonic mode. often hedonic.ac. such as a light that predicts food. the conditioned stimuli. resulting in drug stimulants such as amphetamine and cocaine may arise in this way. Box 2)2. Robbins are in the Department of Experimental can be ascribed to a hypothetical process of pavlovian arousal. Here we discuss evidence that these transitions depend on interactions between pavlovian and instrumental learning processes. serial communication between the various domains of the striatum as cocaine and amphetamine (along with others) also exaggerate the via a cascading loop interconnectivity1.

we have found that unex.com/natureneuroscience (iii) ‘Habits’ depend on interactions between prefrontal cortex and dorsolateral striatum. GABAergic projections. CeN. orbitofrontal cortex. 91. with tentative localization of functions involved in drug addiction. (v) Drug craving involves activation of orbital and anterior cingulate cortex. drug seeking or relapse16–18. to distinguish it from associated CS) has been clearly demonstrated in laboratory studies of general changes (such as in EEG) associated with the term ‘arousal’8. BLA. glutamatergic projections. Acb. selective lesions of the nucleus accumbens core10 or not further be considered here. we have previously used the term neither phenomenon (neither approach to a CS predictive of a drug. (iv) ‘Executive control’ depends on prefrontal cortex and includes representation of contingencies. nucleus accumbens. feelings) associated with drugs. certain circumstances) opiates greatly increase responding with con- pectedly invigorate their efforts to seek and take or ‘want’ drugs as ditioned reinforcement. For this reason. Hipp. These stimuli help to maintain instrumental responding by nucleus accumbens shell potentiates PIT14. in functional imaging studies. and affect Therefore. (vii) Reinforcing effects of drugs may engage stimulant. V. VS.20 and will with these data. bridging delays to the ultimate goal. Ann Thomson globus pallidus (D. possibly including nucleus accumbens but also dorsomedial striatum. DS. (ii) Goal-directed actions involve interaction of prefrontal cortex with other structures. ventral striatum. accumbens core during training11 greatly retard the acquisition of pav. For example. (i) Processing of conditioned reinforcement and delays by basolateral amygdala and of contextual information by hippocampus.92. Green/blue arrows. is well known that drugs such as amphetamine. (b) Limbic cortical- ventral striatopallidal circuitry. anterior cingulate cortex. Consistent behavior and PIT has been reviewed extensively elsewhere19. nicotine and (under tive drug reinforcers. representation of outcomes and their value and subjective states (craving and. as it relates exclusively to studies with infusions of NMDA or dopamine receptor antagonists into the nucleus CSs associated with high-incentive natural reinforcers. and increasing dopamine in the or drugs. However. in terms of drug abuse. dopaminergic Acb shell Acb core Dorsal striatum projections. such as a light. 90). mPFC. OFC. nor ‘activation’ to describe the important role of the mesolimbic and meso. VTA. influence of CSs associated with drugs and natural reinforcers differ pected presentations of drug-paired CSs elicit dopamine release in the fundamentally in this regard19. pavlovian-instrumental transfer and conditioned reinforcement processes in the nucleus accumbens shell and core and then engage stimulus-response habits that depend on dorsal striatum. become reinforcing in consolidation of this response into memory12. it might be feasible in ditions for demonstrating these phenomena in a drug seeking setting certain circumstances to detect the effects of CSs themselves on striatal have not yet been optimized. although both are readily seen in animals Thus. orange arrows. food deprivation8. GP. basolateral amygdala. medial prefrontal cortex. dorsal. (vi) Connections between dopaminergic neurons and striatum reflect ‘spirals’—serial interactions organized in a ventral-to-dorsal cascade. Conditioned reinforcement occurs when stimuli that were D1 receptor antagonists into this region after a training trial disrupt the initially motivationally neutral. SNc. Lesions of the nucleus their own right via association with primary reinforcers such as food accumbens core also abolish PIT13. Modified from refs. REVIEW Figure 1 Representation of limbic circuitry. In testing this hypothesis. amygdala. presumably. (a) Key connectivities in human brain (redrawn from ref. thalamus. The neural basis of pavlovian approach core but not in the shell region of the nucleus accumbens9. such as food or cocaine. © 2006 Nature Publishing Group http://www.nature. in addition to obvious direct influences of responding for natural rewards. It may be that the experimental con- drugs such as cocaine on extracellular dopamine. whereas infusions of NMDA or dopamine forcers. central nucleus of the amygdala. and temporal lobe including amygdala. ventral tegmental area. the nucleus accumbens increases the acquisition of responding for a 1482 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . pink arrows. infusion of amphetamine into emphasized in the incentive salience theory of addiction15. and that drug-associated CSs that occur unex. AC. enhancement of drug seeking by the unexpected presentation of a drug- striatal dopamine systems in behavioral output. dorsal striatum. hippocampus. AMG. but it may also be that the behavioral dopamine function. it might logically be thought that pavlovian approach is the responses to D-amphetamine on a delayed gratification task21. Thal. It involved in maladaptively attracting humans toward sources of addic. ventral). substantia nigra pars compacta. In certain circumstances CSs can also function as conditioned rein- lovian approach responses.

the use of is likely that we have to learn about virtually all of the hedonic the term ‘reward’ for ‘reinforcement’ might have led to the possibly properties of food and that even tastes and smells may not be mistaken view that those structures subserving reinforcement.nature. especially to the nucleus accumbens shell10 mediates the response time interval. rate–increasing or psychomotor stimulant effects of the drug. the hedonic such as the nucleus accumbens and its dopaminergic innervation. amphetamine and nicotine and likely other drugs as well. We assume that these subjective which can become associated with incentive-motivation via responses arise in some way as a post hoc ‘commentary’ on conditioning). For food. have to accommodate these. which would predict enhancement from pavlovian. ‘sensed’ as ‘feelings’ by mechanisms that depend on the insular Similarly. those mechanisms controlling appetitive © 2006 Nature Publishing Group http://www. For example. REVIEW BOX 1 THE SIGNIFICANCE OF SUBJECTIVE RESPONSES IN REINFORCEMENT MECHANISMS Precisely what is a reinforcer. (ii) memory consolidation (for with the view that processes constituting reinforcement can be example.93. This should not at all be construed to mean and it is no surprise that they are reflected in contemporary that we discount emotional subjective responses and the brain theories of drug addiction15. and the consummatory responses associated (appetitive) responses. These theoretical schemes mechanisms mediating them in the analysis of affective behavior. it with reinforcement. including drug seeking under that persists for at least two months without any further experience second-order schedules of reinforcement. of craving for drugs. Defining sensory properties to hypothetical processes of attribution that are associated of reinforcers is more difficult than initially seems. reinforcement (Box 2). Responding in the interim is maintained by the presence of drug-asso- thetically by simulating the behaviorally activating effects of pavlovian ciated CSs that are presented as a consequence of instrumental seeking arousal and affecting the incentive salience of the conditioned reinforcer. or conditioned reinforcers. This seems to contradict the ‘incentive salience’ model of drug seek- The effects of conditioned reinforcers. However. do not. The visual aspects obviously This is in part because they have been confounded with the more gain their hedonic properties through conditioning. Burgeoning evidence links the orbitofrontal cortex to assumed to emphasize the hedonic properties of the reinforcer. This behavior depends on two of self-administered cocaine and that is resistant to extinction of the major influences.94. (textural) and visual elements97. food-related conditioned reinforcer22. effects of food may not arise simply from the exteroceptive stimuli also mediate ‘reward’95. Whereas it is clear described in any textbook of psychology. this is likely to terminated by consummatory behavior (such as eating food be testable only in cases of motivated behavior in humans. their presentation must depend on the animal’s behavior). are pervasive and profound. cocaine. initially involves action- NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1483 . In addition. the identification of ‘hedonic’ responses associated with cortex98. The CSs must be presented as conditioned reinforc- This dopamine-dependent potentiation of conditioned reinforcement ers (that is. the mesolimbic dopamine projec. Drug-associated conditioned reinforc- the conditioned reinforcer10 via its afferent inputs from limbic cortical ers also help to maintain responding under second-order schedules of structures19 (see below). These processes are presumably the product of interactive cortical The incentive-motivational class of theories has often been mechanisms. attributional themselves but from their capacity to evoke visceral changes. There is clearly some truth in each of these accounts. Although these subjectively loaded terms refer the sensory representation itself. olfactory. for example. as ‘unconditioned’ as hitherto believed. the sensory representation of reinforcers as well as their value and especially when there is no obvious deficit or need state: the relative utility of different courses of action producing them. it implicit processes of reinforcement itself. such as instrumental lever pressing and locomotor the expectation of a reinforcer (mediated presumably by its approach behavior that generally occur remote from the reinforcer representation in the brain) evokes appropriate preparatory in space and time. thus more realistically modeling drug seeking behavior24. Early theories of motivation. However. they unexpected.com/natureneuroscience a reinforcer) or (iii) incentive-motivational effects by which behavior. however. as those involved in controlling the reflex itself. aspects. but might also involve nonverbal representations. although an eventual explanation will verbal code. including its subjective. where drugs are provided only after a prolonged tion. However. somatosensory identified or localized to particular brain regions or networks. sucrose or novel objects. and why might a drug of abuse acceptance and rejection reflexes96 may have led to confusion have reinforcing properties? These fundamental questions remain in separating those structures associated with attribution from difficult to answer definitively. the processes themselves have never been must involve a combination of gustatory. refer directly to subjective responses or ‘feelings’ Such responses have to be translated into a (usually subvocal) associated with drug effects. suggested that reinforcers that the latter reside in the brainstem. and it or sexually mounting a female in heat) usually elicited by the is compatible with the often erratic capability of verbal expressions reinforcer. it seems less likely that produce (i) drive (or need) reduction operating according to a subjective attribution does. Subjective responses associated with adjustment that constitutes a sequence of motivated behavior reinforcement are similarly dissociable. It is clearly much harder to test the hypothesis that the such as alterations in heart rate and other autonomic responses. This conveying its hedonic properties remain elusive because the use conceptualization has led to the use of terms such as ‘reward’ of functional neuroimaging procedures thus far has confounded and ‘liking’ that connote subjective responses associated with the sensory properties of a reinforcer with hedonic subjective reinforcers (generally their post-presentational consequences. is a key component of the reinforcing effects of stimulant drugs such as merely presenting them unexpectedly fails to increase drug seeking18. By contrast. the neural mediation of those aspects of the reinforcer stimulation of the brain. This analysis. reinforcers such as intracranial electrical self. is consistent homeostatic regulatory model. nucleus accumbens is implicated in ‘reward’ than ‘reinforcement’. however. presentations of the CS. hypo. support the learning of new drug seeking responses (Box 2). such as approach behavior or physiological with its proximal occurrence. cocaine. responses (Box 2). the nucleus accumbens core mediates the effects of original CS-drug association23. an effect Drug self-administration behavior.88. For example. perhaps especially drug-related ing behavior. responses associated with it. of the association of a conditioned stimulus predicting dissociated: for example.

the animal responds for periods of time and a positive temporal correlation (predictive contingency) (usually 15 minutes. a drug. ‘second-order schedule of drug reinforcement’ (in contrast Pavlovian (or classical) conditioning.e. A does not need to forage. and Discriminative stimulus. This is called of a new instrumental seeking response is measured23.60. properties (positive or negative) by pairings with other. ‘Drug seeking’ can be studied in a number of ways. ‘Drug taking’ is a term used © 2006 Nature Publishing Group http://www. This behavior models aspects of drug seeking in shock. This ‘reinforcer devaluation’. A stimulus can function as a conditioned reinforcer or the real world. motivational influence over instrumental performance. The learning of such behavior is drug seeking responses may actually increase with the duration termed instrumental conditioning. drug seeking responses in the absence of drug100 are widely used procedures because they model a critical aspect of drug Operant. is maintained by drug reinforcement. The subject has been reviewed of obtaining the goal. The actions are sensitive to devaluation of extensively and is not considered in detail here59. sex or electric drug seeking. A stimulus that acquires its reinforcing conditioned reinforcers of the instrumental seeking responses. REVIEW BOX 2 DEFINITION OF KEY TERMS Positive reinforcer. devaluing the reinforcer does not affect stimuli (i. PIT has been interpreted as evidence that CSs exert a increases the probability of the response on which it is contingent. Incentives and conditioned incentives may also need to ‘seek’ the drug. nor does it need conditioned incentive acquires such properties via pavlovian to mediate delays in acquiring it. It is easy to devalue ingestive reinforcers.com/natureneuroscience can be avoided by lever pressing to postpone the naloxone. A depending on environmental contingencies. alcohol greatly enhance instrumental responding for the same reinforcer ingestion increases the probability of licking or drinking. it does not actively conditioning. Such conditioned responses are normally considered to be or heroin under a simple. the decrement in responding event: for example. support the learning of new behavioral strategies directed toward obtaining the primary reinforcer—in this case. For example. and responding during that period is maintained by response-dependent presentations of the CS. This Stimulus-response or ‘habit’ learning. reinforcers such as food. of withdrawal100. A consistent temporal relationship between two The ‘reinstatement of drug seeking’ after extinction of the (or more) events that reduces the uncertainty of the subsequent instrumental seeking response (i. involuntary reflexes. can set the occasion) for behavioral responding that instrumental responding acquired by habit learning. An event that increases the probability of a Pavlovian-instrumental transfer (PIT). their ability to act as goals themselves and thereby administered drugs such as cocaine. namely. drug-associated stimuli reinforce and maintain drug seeking. and discriminative stimulus in the same situation. A stimulus that elicits approach behavior (positive drinking of alcohol (‘continuous reinforcement’). A response on which the presentation of a reinforcer is addiction: namely. infusion of drug. function as reinforcers and conditioned reinforcers. Rats or monkeys that are normally elicited by an unconditioned stimulus. to describe drug self-administration when the drug is readily available: for instance. The pairings require the onset of the CS to each drug infusion being paired with a light CS (simple drug occur before the unconditioned stimulus (temporal contiguity) taking). such as a tone. continuous reinforcement schedule. following each lever press or the simple Incentive. For example. in which devalued either by making the animal ill after ingesting the food. as is drug seeking in drug- instrumental performance is acquired through the association addicted individuals. the actions (lever pressing) are made with the intention seeking measured in this way. Contingency. which act as Conditioned reinforcer. and this defines the pavlovian-instrumental transfer Negative reinforcer. Such behavior is either called prolonged periods of withdrawal (or abstinence). 1484 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . drugs.. generally omission of the contingent CS results in a marked decrease in primary. A stimulus in the presence of which reinforcers primarily serve the function of strengthening the responding is reinforced according to some schedule of stimulus-response association but do not become encoded reinforcement. a situation in which a tone always occurs at caused by non-delivery of the drug)99 or the maintenance of the same time as a shock.nature. procedure models another feature of conditioned reinforcers: but it is much more difficult to devalue intravenously self. or to work. of responses with stimuli present during training. drug infusions stimuli (associated with positive reinforcers such as food) can increase the probability of lever pressing for the drug. behavior is remarkably persistent. drug-associated conditioned reinforcers to maintain or reinstate or else it is a voluntary action. which Drug taking and drug seeking. the ability of a drug-associated CS to support the learning or by pre-feeding it to satiety with the same food. In habit learning. withdrawal symptoms precipitated by scheduled administration of naloxone in morphine-dependent animals. drug cues can act as discriminative as a goal. For example. the goal: for instance. comes to elicit conditioned responses like salivation behavior over relatively prolonged periods24. The subject incentive) or withdrawal behavior (negative incentive). another method of measuring drug seeking has for food will respond much less or not at all for that food if it is adapted the ‘acquisition of a new response’ procedure. when presented unexpectedly (independent of the instrumental response). respectively. Appetitive pavlovian response on which it is contingent.e. in which drugs are not immediately available. The process by which a to continuous reinforcement) emphasizes the role of drug- conditioned stimulus (CS). for the drug. Not only drug-associated stimuli but also Action-outcome learning. the propensity to relapse after sometimes contingent. an animal that has learned to lever-press Recently. but occasionally up to an hour) for each between the two events. such as (and also humans) are initially trained to self-administer cocaine food. that is. It therefore reflects the formation of stimulus-response associations. after a number of associated conditioned reinforcers in maintaining drug seeking pairings. The ability of ‘instrumental’ in obtaining a goal (or ‘outcome’ or ‘reinforcer’). Therefore. such as lever pressing.. Subsequently. When instrumental actions are goal injections of the drug itself and stressors can all reinstate drug directed. An event that when omitted or terminated effect.

including such as satiation or even poisoning in the case of a food goal (Box 2)39. thereby supporting the capability of ventral- the control of the goal20. which also increasingly point or other regions of the ventral striatum29. Cocaine (and important: it applies to instrumental behavior such as intravenous drug other drugs) have positive reinforcing effects when infused response. This conclusion begs the question of where the drug exerts its being the driving influence. Some of the uncon. One possibility is the nucleus accumbens the enhanced significance of drug-associated conditioned reinforcers. as mea- specific regions of the medial prefrontal cortex (mPFC)31. before extended training additionally leads presented as a consequence of instrumental responses—as conditioned to the formation of stimulus-response (‘habit’) associations that help reinforcers. on the interaction of the dorsomedial striatum30 with its afferents from Dopamine release in the nucleus accumbens core and shell. that the habitual behavior is maintained rate–enhancing and locomotor activity effects10. from the output of the core via the learning in rats. Rather. conditioned reinforcement. Indeed. Data for food-maintained habits suggest in the dorsal striatum36. In theoretical terms. though involving mechanisms specifically with psychomotor stimulant drugs38. however. ences over reflexive autonomic and motor responses26. the dorsolateral striatum. Thus. It is possible. shell. definitions of drug addiction (or ‘dependence’ in the Diagnostic and This is consistent with the habit hypothesis and also with the presence Statistical Manual IV). active avoidance behavior in mon- ditioned effects of food reinforcers may be similarly mediated via the keys can have a similarly persistent quality. influencing processing self-administration. such as taste or smell. for example.27 (Fig. An associated with exteroceptive (for instance. which is connected to the full network of descending neural influ. as we have already pointed out that these may merge into one be a persistence or constant reinitiation of such activities. devaluing drugs as reinforcers seems to be quite substantia nigra to other domains of the dorsal striatum1 (Fig. another. in particular in the capacity of CSs to bridge that a rationalization of the ‘out-of control’ habitual behavior rather than © 2006 Nature Publishing Group http://www.42. where the emphasis is on sion of drug-paired CSs in rats extensively trained under second-order slow stimulus-response learning mechanisms with less involvement schedules9. although the acquisition of drug seeking that yet another sector of the striatum. such as the potentiation of conditioned reinforce. it including its dorsal as well as ventral regions6. receptors27. which also function—when they are by corticostriatal ‘loop’ circuitry1.nature. with extended training. but which we nucleus accumbens core. This approach has led to the definition of a role for the dorsolat- bens shell can influence the functioning of the ascending dopamine eral striatum32 and its dopaminergic innervation40 in instrumental habit projections to the core.36. difficult and probably depends on understanding the precise nature of several phenomena. their reinforcing effects (see above and Box 1). is not generally coincident with the provi- training leads to the development of habits. which does not devalue instrumental learning. responding is governed by a suc. significance is that these findings provide further evidence of devolved adaptive stimulus-response habit in which the ultimate goal of the control from the shell and core regions of the nucleus accumbens now to behavior has been devalued so that the behavior is not directly under include the dorsal striatum. These sequential phases of which is part of the ventral striatum. 1). does depend on the shell sistent habit comes from several sources. in part by negative reinforcement37. to-dorsal unidirectional cascades of information processing mediated cession of discriminative stimuli. which is becoming increasingly acknowledged of ‘error prediction’ dopamine neurons innervating the entire striatum. are integrated within the nucleus accum. such as pavlovian arousal and instrumental reinforcement. but such conditioned reinforcers do evoke dopamine release of the goal itself25 (Box 2). which The experiments discussed above and further below show that the has been likened to the subjective state of ‘wanting’15.25. We do not wish necessarily to draw a sharp distinc. as with direct psychomotor stimulant effects of the cocaine.com/natureneuroscience delay. The extent to which instrumental behavior is maintained under these bens circuitry. This especially once the drug itself has been taken18. The analogy with drug addiction would ment. and dependently directly into another region of the ventral striatum. ‘must do!’—although this subjective response could arise post hoc as vision of cocaine. such stimulus-response associative (‘habit’) maintain responding20. Perhaps of even greater seems reasonable to characterize such compulsive behavior as a mal. is under a second-order schedule depends on the nucleus accumbens core. self-administration under a ‘drug seeking’ second-order schedule. habitual responding—evidenced by resis- of CSs in the core. is important in would suggest corresponds more obviously to the subjective state of the maintenance of instrumental behavior involving delays in the pro. This cannot readily be ment by stimulant drugs and pavlovian-instrumental transfer during achieved by simple pharmacological antagonism. that conditioning to interoceptive Evidence for this concept of drug addiction as a maladaptive and per- aspects of reinforcers. Perhaps the most obvious mechanism could stem from conditions reveals the degree of control by stimulus-response mecha- the cascading loop circuitry by which output from the nucleus accum. Thus. such sequential operations may tance to devaluation by gastric malaise—develops more rapidly for a drug result in drug seeking (action-outcome learning) that seems to depend than for a food reinforcer41. By a similar token. nisms. whereas the core is especially to the dorsal striatum as a major contributor to this form of learning. tying one’s shoelace—although it is plausible that such skills result from tion between the unconditioned and conditioned aspects of reinforce. These behavioral patterns are maintained by primary reinforcing effects. as distinct from the shell. could arise from the sequential operation of the a reinforcer so much as remove it.25. However. reduces well-established cocaine seeking under a second-order schedule. Hypothetically. the mixed dopamine recep- transition from initial drug use to drug abuse. It is. especially after treatment nucleus accumbens shell. Two details of this hypothesis are associated with opiate. and finally compulsive tor antagonist α-flupenthixol infused into the dorsal striatum greatly drug taking and drug seeking behavior. of course. Compulsive drug seeking and drug taking are the hallmarks of the yet it has no effect when infused into the nucleus accumbens core43. control over performance may ulti- pavlovian and instrumental learning may be especially relevant for the mately devolve to the dorsal striatum.44. including response obsessive-compulsive behavior. 1). even more extended training. which act as discriminative stimuli for continued drug seeking behavior. eventually dominates behavioral output. This proposed sequence of changes NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1485 . REVIEW outcome learning (Box 2). and similarly. The obvious analogy idea is consistent with evidence that this region is necessary for the is with obsessive-compulsive disorder. learning occurs in parallel with instrumental action-outcome learning but. visual) conditioning19. rather than dopamine. Extended sured by microdialysis in vivo. such as playing the piano or olfactory tubercle28. In studies of oral cocaine and alcohol drug’s impact in the nucleus accumbens shell. operational definition of a habit is that the behavior continues even after Important issues to be resolved include how the contributory factors the controlling influence of the goal is reduced by devaluation procedures. Interacting roles of striatal subregions in reinforcement Crucial to drug addiction is the persisting quality of these habits. the it is not an example of a procedural skill. implicated in habit learning32 (see below). possible. by neuroscientists modeling this behavior33–35.

increase the choice of small. that are accumbens core and OFC and their interactions in the control over goal- primarily focused on the ventral striatum and discrete regions of the directed behavior by discrete CSs acting as conditioned reinforcers. accumbens neurons. Theta-burst stimulation of the hippo- instrumental responding for cocaine is completely unaffected by core campus reinstates extinguished cocaine seeking. the hippocampal formation. along with the nucleus ian conditioning and the dorsal striatum with instrumental learning49. The core is Thus. interconnected. First. the striatum is also implicated in nucleus accumbens core64. whereas AMPA (but Indeed. cocaine seeking under a second-order schedule. They are broadly compatible with functional neuroimaging and overlying anterior dorsal striatum (Fig. McNaughton & B. The OFC and basolateral amygdala are richly associative learning suggested by these studies require further inves.60. T. These data nisms underlying drug taking are dissociable from those underlying are generally consistent with the view that. 1). in the neural mechanisms underlying the ability to From the neurocomputational perspective. This interaction is modulated by mesolimbic receptor blockade bilaterally in the basolateral amygdala impairs dopamine and.E. sistent with a growing body of data that the basolateral amygdala and respectively. similarly. Presumably.nature. Dopamine (but not AMPA) the nucleus accumbens. These attenuated locomotor response to amphetamine and impaired acquisi- 1486 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . critic learns to predict future rewards. the reinstatement of extinguished responding for cocaine by contextual ond-order schedule50–53. and the actor maintains infor. larger. in other words. In dorsal striatum. Studies of the cued reinstatement of extinguished responding for mation about the rewarding outcome of actions. hippocampal lesions impair contextual occurring during the delay (either discrete. The effects of lesions in basolateral amygdala or nucleus ates conditioning to discrete CSs. as well as other with the nucleus accumbens core (Fig.65—but in the context of relapse. unilateral blockade of dopa. in turn. and this modulation influences performance in mine receptors in the basolateral amygdala combined with unilateral appetitive behavioral tasks70. as well as projecting to the nucleus accumbens core tigation. Electrophysiological and in vivo neurochemical stud- the basolateral amygdala and nucleus accumbens core in drug seeking ies show that hippocampal. whereas ventral subiculum–lesioned rats show an seeking as much as bilateral manipulations of either structure43. cellular markers. D1 and D2 dopamine receptors differentially regulate the influ- not dopamine) receptor blockade bilaterally in the nucleus accumbens ence of mPFC versus hippocampal afferents on the activity of nucleus core has a similar effect. immediate rewards over discrete CSs may compete for control over goal-directed behavior3. ences on behavior is unclear. the ventral and dorsal stria. the hippocampal formation underlies accumbens are likely to reflect the interacting roles of these structures conditioning to contextual or spatial stimuli68 and may therefore underlie in conditioned reinforcement19 and also their roles in mediating delays the motivational impact of contextual stimuli on drug seeking. 1).. the hippocampus and the prefrontal cortex. roughly Hippocampus—nucleus accumbens system topographical inputs from limbic cortical structures. of course. The observation that © 2006 Nature Publishing Group http://www. the striatum is only one part of a much more extended network defined by its intimate. amygdala lesions not only impair appetitive behavioral responses also necessary for instrumental learning when there is a delay between under the control of discrete CSs but also result in enhanced control the response and the reinforcer55.W. which is also a major source of glutamatergic afferents to the nucleus accumbens. rather than in response to discrete cocaine cues67. There is general consensus on the functions of the amygdala. The neural basis of such competition between associative influ- Using a disconnection procedure (unilateral manipulation of struc. blockade of AMPA receptors in the core in the contralateral hemisphere Dorsal subiculum–lesioned rats are hyperactive in tests of explor- (neither of which has any effect alone) reduces cue-controlled cocaine atory locomotion. basolateral amygdala and OFC are involved. as well as from neocortical areas. Second. whereas the amygdala medi- drug seeking. translated into goal-directed. it would be very mis.L. like nucleus accum.. consistent with other evidence that this region is not directly transmission in the VTA. has received somewhat less attention in stud- Selective lesions of the basolateral amygdala or the nucleus accumbens ies of drug seeking. B. delayed rewards—indicating greater impulsivity54. These interaction of pavlovian and instrumental learning through the inter. such as the baso. lateral OFC62 and nucleus accumbens core63 as well as leading to imply that it is only the striatum that is implicated in these dopamine and glutamate transmission in the basolateral amygdala and aspects of learning. of contemporary models of reinforcement learning49. accumbens core. unpublished observa- leading to the reward. also emphasize the involvement in drug seeking of the basolateral mediary of conditioned reinforcement. which is a performance. involving the retrieval of appropriate stimulus-response key aspect of drug addiction. drug taking. but may depend upon the projections of tures within a putative neural system. 1). but not by discrete CSs66. to reinforcement. the basolateral amygdala and hippocampus to the nucleus accumbens we have shown the functional importance of serial interactions between core and shell69. drug seeking behavior via its interactions The downregulation of striatal dopamine D2 receptors. the cocaine (Box 2) have been reviewed extensively elsewhere59.com/natureneuroscience evidence in humans that the ventral striatum is implicated in pavlov. However. rules and goal representations. Hippocampal contextual information and amygdala-dependent bens core lesions.R. but on opposite sides of the brain). tions). Thus. which was suggested to mimic the way that rein- implicated in instrumental learning per se: lesion-induced deficits are statement occurs when animals are placed in a context associated with found only when the drug infusions are delayed. especially the nucleus The basolateral amygdala–nucleus accumbens core system accumbens shell26 (Fig. The OFC cooperate to regulate goal-directed behavior58. amygdala and PFC projections interact in sustained by conditioned reinforcers43. seek drugs over long delays bridged by conditioned reinforcers is con- tum could conceivably correspond to the ‘critic’ and ‘actor’ components.J. Most importantly. known to occur in human chronic cocaine abusers45 Selective lesions of the orbital prefrontal cortex (OFC) also impair can also be observed to occur first in ventral and then in dorsal territories the acquisition of cocaine seeking56 and responding with conditioned of the striatum in cocaine-taking rhesus monkeys46–48. These studies also show that continuously stimuli. the mecha. or forming part of the con. conditioning but can also result in enhanced conditioning to discrete text) to act as conditioned reinforcers for instrumental responding. REVIEW caused by drug-self-administration is further supported by observations data indicate that associative information in the basolateral amygdala is in rhesus monkeys self-administering cocaine over an extended period. contrast. Inactivation of the dorsal hippocampus prevents core impair the acquisition of cocaine or heroin seeking under a sec. it acts by allowing CSs by contextual cues. CSs3 (R. acting via glutamatergic lesions. Basolateral amygdala lesions. Ito. can modulate the release of dopamine70–73. but without affecting continuously reinforced cocaine The findings of both parallel and serial cascading mechanisms of self-administration56. amygdala61. reinforcement57. nucleus lateral amygdala.

This hypothesis is plausibly supported in involvement of NMDA. how. Impairments in such processes. Myths about the treatment of addiction.R. and from ventral 6. S. & McFarland. Tobler. Addiction is a brain disease. which has similarly been associated with dysfunctional orbi- in the reinstatement of drug seeking depends on glutamate release in tofrontal-striatal circuitry. Alderson.J. to the shift in balance of behavioral control processes toward those 11. Ito. drug extinction.. C.84. including the orbitofrontal region infusions of D-amphetamine in addition to blocking the potentiative of abstinent addicts45. Miles. 20.A. it may be necessary to postulate a source of the nucleus accumbens core and also on the integrity of the ventral negative reinforcement that maintains responding. 23. Functions of dopamine in the dorsal and ventral stria- tum. frontal cortical to striatal control over responding. the involvement of the mPFC der itself.64.nature.L. indeed. we hypothesize that the transition from voluntary actions 4. Inactivation of the dorsomedial part of the PFC prevents taking history in the face of conditioned or unconditioned aversive the reinstatement of responding elicited by drug cues. 5. Schultz. Seminars in the Neurosciences 4. Ito. activity and con.W. Di Ciano. Cardinal. regions31. responding occurs. 1).. Moreover. rat. S. Science 299. therefore. B. S. 1). Parkinson. Striatonigral pathways in primates form but despite the involvement of this region in reinforcer processing. Hippocampal. © 2006 Nature Publishing Group http://www. F. & Dickinson.C. 125–184 (2002). Whether sensitiza- textual conditioning and also the potentiation of these responses by tion directly affects instrumental drug self-administration seems less psychomotor stimulants—providing. Robbins. & Balleine. & Everitt. M. however. J. in regulating its dopami. to result from any change in conditioned reinforcement75 and may reflect COMPETING INTERESTS STATEMENT instead an impairment in executive control over behavior (including The authors declare that they have no competing financial interests.com/natureneuroscience/ neuroimaging and neuropsychological evidence from human studies Reprints and permissions information is available online at http://npg. Neurosci. J. Dalley. & McDonald.D. K. 45–47 responding in drug seeking behavior represents a transition from pre.W. does not capture the persis- ment75.W. R. Cardinal. On this account. Distinct changes in neuronal plasticity in this region. too. Neurosci. Leshner.nature. 77. A. Whitelaw. the drug effect itself may produce the enhanced drive to responding. A related but distinct view is that logical terms. Parkinson.. Robbins. perhaps question for future research. Olmstead. W. competing for access to response strategies dependent on different lim- ACKNOWLEDGMENTS bic cortical–striatal circuitries70. (governed mainly by their consequences) to more habitual modes of 237–240 (1996).J. Lee. Robbins.P. additional factor seems to be required. A. B. & Everitt. hippocampal mechanisms provide the contextual back. 2369–2382 (2000). Howes. How cortical-ventral striatopallidal circuits59. J. for example. However. 1–18 (1994). Neurobiol. may contribute titive Pavlovian approach behavior and the potentiation of conditioned reinforcement and locomotor activity by D-amphetamine. but not brief. Little. regions mediating the executive control over behavior: for example. Vanderschuren. some the nucleus accumbens. as in the case of obsessive-compulsive disor- injections of drugs or stress59.com/ suggesting that chronic drug abusers show deficits in tests of inhibitory reprintsandpermissions/ control and decision making78–80.. Dalley.J. R. B. and in related processes of goal revaluation by com.. J. in the acquisition of action-outcome learning itself. these effects of mPFC lesions seem unlikely L. J. We hypothesize that. perhaps arising in Dissociation in effects of lesions of the nucleus accumbens core and shell on appe- part as the direct consequence of toxic drug effects. Discrete coding of reward probability and may reflect important changes in the balance of activity in those brain uncertainty by dopamine neurons. 2401–2411 (1999). M. Behav. C. According to The prefrontal executive system the DSM-IV. Y. Fudge. This is consistent with burgeoning Published online at http://www. contexts.J. The research is supported by the UK Medical Research Council.A. 19. T. & Everitt. 1. 1898–1902 (2003).N. J. Pelloux. Neurosci. L. Studies using PET. effect of the same treatment on responding with conditioned reinforce. Lesions of the ventral subiculum neural terms by neuroimaging data in humans showing reductions in also completely abolish the locomotor response to intra-accumbens the activity of the prefrontal cortex. K. Fattore.N. Dissociation in altered instrumental contingencies with associated changes in subjec. 7489– 7495 (2000). behavioral inhibition processes)77. J. T.nature. ule of reinforcement and also enhance the acquisition of cocaine self. P. O’Brien.H. Inactivating this mechanism at the hippocampus sensitization reflects the normal processes of tolerance and inverse or nucleus accumbens shell level reduces exploration.. Hutcheson. B. an additional basis for clear. & Everitt.77.J.81. through pallidum. priming stimuli34. result in increased responding for cocaine under a second-order sched. White. R. A. Haber.J. do result from chronic stimulant self-administration82. high- light changes in metabolism in the OFC in abstinent drug abusers45. Multiple parallel memory systems in the brain of the ever. and it matters. Thomas. R. Motivational control of goal-directed action. in drug seeking measures the reinstatement of cocaine seeking after which continue to seek cocaine after a prolonged. ‘break points’ (the highest number of responses rats will make) for cocaine assessed using a progressive ratio schedule87. REVIEW tion of cocaine self-administration74. experiments have examined its role in controlling drug seeking behavior 2. R. 22. A. In the ‘incentive-sensitization’ nergic tone and mediating the psychomotor stimulant effects of drugs model..W.N. Di Ciano. Dickinson. especially the shell. C. to cocaine cues and during cocaine-seeking behavior in rats. Dickinson.53.81.89. This. Cowell.. Alternatively. Annu. another characteristic of drug addiction is that it per- By far the most detailed investigation of the prefrontal cortex (PFC) sists despite adverse consequences.R. J. ponents of the prefrontal cortex. Weissenborn and R. Fiorillo. induced sensitization of behavior85.com/natureneuroscience ground that defines the motivational arousal upon which goal-directed thus prolonging the duration of a drug taking binge. W. has been modeled in rats. Burns. These data suggest a key role for hippocampal projections to tent. 3. including prelimbic and infralimbic 10. Olmstead. Differential promoting habitual behavior. the detection of 9. P. the potentiated responding is postulated to depend on drug- such as amphetamine and cocaine10. N.. although sensitization can augment responding for conditioned understanding the reinforcing effects of drugs acting on the dopamine reinforcers enhanced by intra-accumbens amphetamine86 and increase and other systems in the nucleus accumbens (see above). A. L. Learn. compulsive aspects of ‘out-of-control’ drug bingeing. Science 278. 8. R. tolerance that modify the effects of many drugs. 7. Some of that transition Neurosci. 20. 119–128 (1992). Learn. AMPA/kainate.60. N. especially. in animals.. R. Rev. B.M. Neuronal coding of prediction errors. in psycho. J.83. Arroyo. Anim. D. and dopamine receptors in the nucleus NATURE NEUROSCIENCE VOLUME 8 | NUMBER 11 | NOVEMBER 2005 1487 . Lancet 347. Pilla. amygdala and such systems interact at a neural level with those inducing the habitual mPFC mechanisms may therefore all influence drug seeking through appetitive behavior associated with drug addiction remains a central their convergent projections to the nucleus accumbens (Fig.. & McLellan. few an ascending spiral from the shell to the dorsolateral striatum. T.I. conditioned dopamine release in the nucleus accumbens core and shell in response tive attribution. We acknowledge Lesions of the mPFC (including the prelimbic and infralimbic cortex) the major contributions of H. (1997). 473–500 (2000). Hall.35.T. administration76. & Schultz. Overall. providing clear evidence of the function of specific limbic opponent motivational systems also engaged by drug abuse88. Habitual responding by itself. M. J. Mem. P. to more dorsal striatal subregions (Fig. Hellemans.

K. C.. 1488 VOLUME 8 | NUMBER 11 | NOVEMBER 2005 NATURE NEUROSCIENCE . of relapse to cocaine-seeking behavior. 2799–2807 © 2006 Nature Publishing Group http://www. J. B. T. B. Fenu. Cardinal.H. Dickinson. et al. Sign-tracking (autoshaping) in rats: A comparison of 47. B.M. Second-order schedules of drug reinforcement in rats of basolateral amygdala and orbitofrontal cortex in impulsive choice. & Gardner. P. The functional divide for primary reinforcement of primates.W. Di Ciano.J. K. The effect of instrumental training con.nature.W. USA 102. the amygdala and hippocampus in aversive conditioning to explicit and contextual tingency on susceptibility to reinforcer devaluation.M.H.E. Schoenbaum. Drug seeking becomes compulsive after prolonged behavior. & Everitt. A.. 952–960 (2003). 66. Daunais.N. Saddoris. B. 44–56 (2003). Involvement of the dorsal striatum central nucleus of the amygdala and nucleus accumbens core in mediating Pavlovian in cue-controlled cocaine seeking. Ostlund. 51. Encoding predicted the accumbens core. 8122–8130 46. K. Neurosci. 37. Cardinal. P.. Dopamine release in the dorsal blockade in the basolateral amygdala attenuates conditioned reward in a rat model striatum during cocaine-seeking behavior under the control of a drug-associated cue. J.R. L. Katz.P. The circuitry mediating cocaine-induced reinstatement ible development of drug-addiction in the rat.W.) striatum in instrumental conditioning.. & Robbins. R. Conditioned reinforcing properties of stimuli paired with 53. 45.R. Opin. 34. associated conditioned reinforcers and discriminative stimuli predictive of drug avail. B. I. Belin.A. & Balleine. & Robbins. G.W. Learn.. J. Theobald. The role of the dorsomedial prefrontal cortex. Eur. Robbins. Effects of cocaine self-administration on striatal dopamine systems Psychobiology 28. Neurosci. Wolffgramm. Robbins. Effects of D-amphetamine on responding of Neuropsychopharmacology 30. but not glutamate. 18.W. Neuropsychopharmacology 27. 692–700 (1981). B. Cocaine self- cocaine and food as unconditioned stimuli. 32. Yin. J. A. 3306–3311 (2002). & Porrino. Qin. Nader.J. Cardinal. 156–162 (2004). J.J.W. & Everitt. C. B. S. Wood.) 84..J.J. A. 7. R. Taylor. 213–224 behaviour. 6600–6610 (2004).W.A. A. Dopaminergic modulation of limbic and cortical drive of nucleus Q. Psychopharmacology (Berl. Jarrard. receptor 36. 181–189 (2004). Bassareo.. C. T.D. See. Everitt. & Dickinson. orbitofrontal but not medial prefrontal cortex disrupt conditioned reinforcement in 28. & Everitt.) 153. 67. Neurosci. without enhanced “liking” or response reinforcement. Psychopharmacology (Berl. T.. 8665-8770 (2005)...E.) 168. of the nucleus accumbens mediates cocaine-induced reinstatement of drug seeking 35. the acquisition of drug-seeking behaviour under a second-order schedule of heroin Psychopharmacology (Berl. Neuron 39. The effects of selective orbitofrontal cortex lesions for limbic structures to reach the motor system? Prog. 50. Cardinal. Brain Res.N. Neurosci. Robinson.W. B. A. J. 6.E. 8655–8663 (2001). 3531–3537 (2003).. P. Theoretical Approaches to Obsessive-Compulsive Disorder (Cambridge Univ.W. A. 301–310 J. 21. 218. & See.N. 35–46 (2002). T. Neuroscience 42. Direct interactions between the basolateral amygdala and and NMDA receptors in the nucleus accumbens. Friedman. H.W. J. Am. T. & Roberts. & See.. of cocaine self-administration in rhesus monkeys.J. & Balleine. 69.the irrevers. M.A. Lapish. M. Brain circuitry and the reinstatement of cocaine-seek- accumbens shell in conditioned taste aversion learning. J.V.. M. The nucleus accumbens: Gateway 56. Nader. 87.A. 117. 61. A role for dopamine D1 receptors of the nucleus 59.J. Psychol. squirrel-monkeys maintained under 2nd-order schedules of food presentation. BMC Neurosci. 111–119 (2000). From controlled drug intake to loss of control . 8. Relapse to cocaine-seek- tric-shock presentation or stimulus-shock termination. J. 129–138 (2001).J. & El Massioui. 27. & Everitt. J. Robbins. S.. Dopamine.C. Exp. cocaine self-administration..W. Groenewegen.. Acad.J. M. B.J. Behav. 25. B. upon input from basolateral amygdala. S. Science 304. 4718–4722 (2004). NY Acad.W. & Balleine. Y.J.E. D. V. 117. C. & Everitt. Groenewegan. R. B. Nucleus accumbens core lesions retard instrumental 25.W. Psychopharmacology (Berl.W. Pharmacol. J. Evans. influences on instrumental behaviour. Dickinson.H. Differential involvement of orbito- outcome expectancy but disrupt habit formation in instrumental learning. D. F. Brain Res. & McFarland.W. T. N. Neurosci. Shaham. 48. 5061–5065 outcome and acquired value in orbitofrontal cortex during cue sampling depends (2005). T. 1984–1992 (2001). Di Ciano. 567–570 learning and performance with delayed reinforcement in the rat.. R. Dickinson. administration produces a progressive involvement of limbic. Oral cocaine seeking by rats: action or habit? J.J. heroin or sucrose: implications for the persistence of addic. Faure. Excitotoxic lesions of the tive learning: links to drug addiction.E. B. (1999).. 20. Ther. 1003. Neurosci. The neuropsychological basis of addictive second-order schedule of reinforcement. F. & Everitt. 13. T.N. Lu. & Robbins. P. 18. Contingent access to stimuli associated with sensorimotor striatal domains.J. 927–938 (2003). 55.L. Dickinson. 1017–1019 (2004). A. 410–411 (2003). 452–454 (2004).. 21.J. R.J.B. & Heyne. 3554–3562 (2004). & Grace. of drug seeking behavior. 52. J. & Di Chiara. L. Lesions of the basolateral amygdala abolish the ability of drug tion but not the expression of goal-directed learning. 17–30 (2000). Involvement of the 44. (2001). Lesions of medial prefrontal cortex disrupt the acquisi. Yin. 68. & Kalivas.J.. K. 383–386 (2000). D. 25..A.W. Behav. H. 24. Science 305. accumbens core and shell lesions on intravenous heroin self-administration and forcers following microinjections of d-amphetamine into the nucleus accumbens.D. et al. 23. B.J. 70. Kruzich. 24. P.) 127.A.. J.. & Everitt. A. de Wit. & Adams. 9 (2005). & Everitt. 32. Prefrontal glutamate release into the core the rat. Spector. The effects of d-amphetamine. Drug addiction and its underlying neurobiological 14.. & Berridge. Robbins. Z.. J. B. F. K.J. U. H. T.. cocaine. Setlow. Selden.R. T. J. Drug addiction: bad habits add up. A. Hayes. Neurosci. accumbens in goal-directed behavior. 1642–1652 (2002). & Robbins.R. 805–812 (2005). Robbins.M.J. 21. 11189–11201 (2003). Neuropharmacology 47 (Suppl. Exp. Neurobiol. Wright. E. ditioned incentive salience of sucrose reward: enhancement of reward “wanting” Psychiatry 159. & Volkow.J. T.R. 26. L. Intra-accumbens amphetamine increases the con. Deroche-Gamonet. et al. H. P. 35B.. B. 23. shell.J. Brain Res. Psychopharmacology (Berl. & Everitt. reinforcement. S.. J. electrophysiological and anatomical data. Rev. basolateral amygdala impair the acquisition of cocaine-seeking behaviour under a 20. 139–148 (1997). Contrasting roles 24. in rhesus monkeys: Initial and chronic exposure. 389–397 (2004). & Grimm. and 17. REVIEW accumbens core in the acquisition and performance of Pavlovian approach behavior. & Nader. R. Knowlton. A. J. 33. J. J. Brain Res. Psychopharmacology (Berl. The effects of nucleus 22. L.W. 14. D-amphetamine lies between the medial and lateral ventral striatum: Is the division of 58. Kelley. Hutcheson. Goto. B.E. J. H.. G. Hutcheson. Science 292.H.M. Neurosci. & Kalivas.A. Progression of changes in dopamine transporter binding site density as a result 15. B.. Smith. 464–472 (2001). R.R.E. B 55. Winstanley. 331–348 (2002). 36. J. V. Nature 398. Brain Res. Robbins.J.E. Fuchs. M. 12. Behav. & Glowa. Haberland. T. J. J. Parker. & Everitt. 296–309 (2005). Parkinson. Whitelaw. 335–350 (1991). L. nucleus accumbens core underlie cocaine-seeking behavior by rats. association.J. & Piazza. 19.. Sci. R. J. (1996). Time-limited modulation of appetitive Pavlovian memory by D1 43. 6897–6904 ing behavior. Ann. 63.E. Natl.J. B..W. (2001). & Everitt. bens as a complex of functionally distinct neuronal ensembles: an integration of mine system disrupts stimulus-response habit formation. 9471–9477 (2001). 168. & Smith. 31. elec. et al. Complementary roles for 39.. J. Dalley. P. J. and dorsal hippocampus in contextual reinstatement of cocaine seeking in rats. & Berridge. Neurosci. 6247–6253 (2002). Sci. Neurosci. 42. Y. 7763–7770 associated cues to reinstate responding during withdrawal from self-administered (2005). 362–375 (2000). 16. 6189–6194 (2005). Neurosci. 1014–1017 (2004). J.. Neurosci. Letchworth. Neurosci. 463–476 (2004). M. Alderson. The neural basis of drug craving: an incentive.) 152. R. A.B. Lesions of dorsolateral striatum preserve 62.L. Liu. Goldstein. Conde.L. & Cheung..J. 1996). Miles. 54. Curr. 77–94 (1995).A.J. 49. cocaine self. The nucleus accum- 40. S. Nat.. McFarland. Prog. R.W. The reinstatement model of 30. Exp. ment of cocaine seeking in rats. N.L. L. Lyons. 25. Di Ciano. Hopewell. J.J.W. 22. B. Ito. Neural and psychological mechanisms underlying appeti. methodology and major findings. Neurosci. 65. 19.. J.A. & Lopas da Silva. 13. B. Brain Res.flupenthixol and behavioural manipulations on choice of signalled the basolateral amygdala on the acquisition of heroin-seeking behaviour in rats. (1983). 21.. Enhanced behavioural control by conditioned rein. P.. J. (1996). Brain Res. Grimm. Cambridge. C. Behav. Shalev. N.B. Ikemoto. 3–20 (2003).J. L. and monkeys: measurement of reinforcing efficacy and drug-seeking behaviour. 42. Knowlton. Differential control over cocaine-seeking behav- self-administered cocaine.A... B.A. Lesions of the addictive drugs. C.W. 25. M.) 153. 1175–1178 (2001).. Rev. 2771–2780 behavioural.. 29.J. Vanderschuren.C. Psychol. 7167–7173 (2004). Fuchs. H. H. M. A. Kalivas. (2000).. The neuroscience of natural rewards: Relevance to 57. J. Robbins. Connor. & Everitt.W. Eur.V. D. alpha.. Wyvell. Everitt. D. Ostlund. Neurosci. 405–412 (1984).C. S.P.. R.J. Pennartz.B.M. 485–511 on the acquisition and performance of cue-controlled cocaine seeking in rats. Proc. J. Barrett. 38. N. The role of the dorsomedial drug relapse: history. tive behaviour.. chlordi. H. Everitt. frontal cortex subregions in conditioned cue-induced and cocaine-primed reinstate- Neurosci. 35–51 cues. McFarland. H. & See. (2001). B. Nicholas. X. Q.) 202–213 (2004).N.. Neurosci. K. Parkinson. Press. J. 513–523 (2005).. R. 247–291 (1993). Lesion to the nigrostriatal dopa. Everitt. D. J. T. Hall. Differential control over drug-seeking behavior by drug. Jakes. Everitt. T. & Robbins.administration is required for reinstatement of drug-seeking behavior. Smith. Neurobiol.I. Psychopharmacology (Berl.W. J.. B.J. 24.com/natureneuroscience sensitization theory of addiction. Neurosci. Psychopharmacology (Berl.Z. Dalley..A. Neurosci. & Weiss. Di Ciano. Eur. & Everitt. J. Neurosci. K. 60. Pears.) 153. 22.P. J. 41. ing after hippocampal theta burst stimulation. The effects of excitotoxic lesions of azepoxide. R. R. & Beijer. Kearns. D. Kruzich. Neurosci. O’Doherty. Neurosci. Meil. basis: Neuroimaging evidence for the involvement of the frontal cortex. W. Evidence for addiction-like behavior in 64. Vanderschuren. 107. 24. 23. Dissociable roles of ventral and dorsal striatum in instrumental ability.C. & Stewart. & Everitt.J. B.J.R. B. B. J. Markou. & Liu. ior by nucleus accumbens core and shell. R. D. Alcohol seeking by rats: Action or habit? 70. basolateral amygdala. P. conditioning. 22. Neurosci. 21. J. Psychopharmacology (Berl. P. & Berridge. & Gallagher. and olfactory tubercle valid? J.. S. Porrino.A. J..C. Science 305. T. & Everitt.W. Nat. Behav.J. U. Ito. 719–761 (2005). A. Vorel. 855–867 (2003). Neurosci.) 154. and unsignalled delayed reinforcement in rats.. (1994). Parkinson.H.

& Robinson. Neural systems 82.) 211–227 (Springer-Verlag. 54–67 (1998). Effects of medial prefrontal or anterior cin. L.. 83. Stewart. T. & Everitt. Orbitofrontal cortex dysfunction in abstinent cocaine abusers perform. & Coutureau. Psychopharmacology (Berl.D. T. 73. The Frontal Lobes and Neuropsychiatric Illness (Salloway.. R. Neuropharmacology 47. T. T. & Le Moal. Opin. R. 79. Malloy. and allostasis. Dopamine. Dalley. Wiens. Cortex 15. & Balleine.W. J.E. withdrawal: a review of preclinical data. et al.E.F.W.T. & ing with conditioned reinforcement and locomotor activity potentiated by intra-accum. 54. Floresco. & Berridge.B. M. Rolls.P. Investigating the neurocognitive deficits associated with 94.) (American Psychiatric Press. Sensitization of midbrain dopamine neuron reactivity and the self- mechanisms of input selection. Neurosci. Biobehav. & Wang.J. Behav.. Washington. Brain Cogn. learning and motivation. dopamine neuron activity by afferent pathway activation: Relationship to patterns of Behav.. Rev. R. 839 (2004). Neuropsychopharmacology 24. R. Psychol. 22.nature. The role of prelimbic cortex in instrumental conditioning. ventral subiculum and medial prefrontal cortex on respond. R. in compulsive drug use. A motivational learning hypothesis of the role of mesolimbic dopamine chronic drug misuse. 97–129