You are on page 1of 10

125

Ann. N.Y. Acad. Sci. 985: 125-134 (2003). 2003 New York Academy of Sciences.
The Amygdala, Fear, and Memory
MICHAEL S. EANSELOW AND GREG D. GALE
Department of Psvchologv, Universitv of California, Los Angeles,
Los Angeles, California 90095-1563, USA
ABSTRACT: Lesions of the frontotemporal region of the amygdala, which in-
cludes lateral and basal nuclei, cause a loss of conditional fear responses, such
as freezing, even when the lesions are made over a year and a half from the
original training. These amygdala-damaged animals are not hyperactive and
show normal reactivity to strong stimuli such as bright lights. After receiving
tone-mild shock pairings rats normally display an appropriately weak re-
sponse when exposed to the tone. Rats` fear of the tone can be inflated by giving
them exposure to strong shocks in the absence of the tone between training and
testing. This inflation of fear memory is abolished if the frontotemporal
amygdala is inactivated by muscimol only during the inflation treatment with
strong shocks. Based on such findings we suggest that the frontotemporal
amygdala permanently encodes a memory for the hedonic value of the aversive
stimulus used to condition fear.
KEYWORDS: fear; pavlovian conditioning; freezing; memory; amygdala; baso-
lateral amygdala; frontotemporal amygdala; hedonic value; affect; hippocam-
pus; retrograde amnesia
INTRODUCTION
There is near universal agreement that the amygdala contributes to Iear-motivated
learning. Many amygdala manipulations proIoundly aIIect behavioral indices oI Iear
conditioning.
1
Changes in the Iiring properties oI amygdala neurons aIter Iear con-
ditioning have also been observed. However, it is incorrect to think oI Iear as some-
thing localized to a single brain region. Eear represents a complex Iunctional
behavior system, and widely distributed neural circuitry contributes to the percep-
tion and recognition oI danger, the learning and remembering about dangerous ex-
periences, and the coordination oI deIensive behaviors to environmental threat.
2
The
circuit involves many brain regions and distinct patterns oI communication between
those regions.
3
Our laboratory is interested in determining the unique and special
contributions each oI these components oI the circuit makes to the eIIective opera-
tion oI this Iunctional behavior system we call Iear. This chapter Iocuses on one such
component, the Irontotemporal amygdala.
4
Address Ior correspondence: Michael S. Eanselow, Department oI Psychology, University oI
CaliIornia, Los Angeles, 405 Hilgard Ave., Los Angeles, CA 90095-1563. Voice: 310-206-3891;
Iax: 310-206-5895.
Eanselowucla.edu
126 ANNALS NEW YORK ACADEMY OF SCIENCES
The Irontotemporal amygdala (ETA), oIten reIerred to as the basolateral complex,
receives extensive inIormation Irom the neocortex, thalamus, and hippocampus.
Thus, it is in an excellent position Ior sensory integration. It has extensive projec-
tions to the amygdala`s central nucleus and thereby has access to many oI the struc-
tures that generate and coordinate Iear-related behavior.
4
Lesions and
pharmacologic manipulations that selectively target this region have proIound eI-
Iects on both the acquisition and the expression oI Iear.
1,5
Although these Iindings
clearly establish a critical role Ior the ETA, the precise nature oI its contribution to
Iear processes remains an open question. We describe a Iew experiments that lead us
to a rather bold and speciIic conclusion: the Irontotemporal amygdala permanently
encodes the emotional signiIicance (hedonic value) oI the aversive reinIorcer used
in pavlovian Iear conditioning. The experiments are conducted so that this role can-
not be attributed to Iactors such as alterations in general activity or interIerence with
behavioral indices oI Iear.
PERMANENT INVOLVEMENT IN FEAR MEMORY
A number oI Iindings suggest that the ETA plays a relatively permanent role in
Iear conditioning.
6,7
These studies employed a posttraining lesion strategy in which
the amount oI time between training and lesion is manipulated in order to assess how
long a structure contributes to the learned behavior. This strategy has been successIul
at characterizing temporally limited contributions oI the hippocampus to Iear. Post-
training lesions oI the hippocampus are eIIective in blocking expression oI some
Iorms oI Iear when made shortly aIter training but ineIIective when made at longer
intervals between training and testing.
8
To investigate how long the ETA is involved
in mediating a Iear memory, we made posttraining lesions oI this structure.
9
Previ-
ous studies showed that amygdala lesions made up to 1 month aIter Iear conditioning
severely attenuated Iear, as assessed by both potentiation oI startle and Ireezing.
6,7
This pattern oI results appears consistent with a 'permanent role Ior the ETA in Iear
conditioning. However, in some instances the hippocampus showed a time-depen-
dent involvement in Iear conditioning Ior periods longer than 1 month.
10
To test con-
clusively the duration oI the contribution oI the ETA to Iear conditioning, we
interposed an interval oI 1.5 years, which encompasses most oI a rat`s liIe span.
To maximize the power oI our test procedure we used the within-subjects design
developed by Anagnostaras et al.
11
to explore the time-dependent involvement oI the
hippocampus in Iear conditioning (EIG. 1). Adult rats were placed in one distinctive
context and received pairings oI tone and shock. AIter this, they were maintained in
their home cages Ior 1.5 years (the Anagnostaras et al.
11
hippocampus study only
waited 50 days). At that time, rats were placed in a second distinctive context that
diIIered in terms oI location, shape, lighting, odor, and background noise. There,
they received pairings oI a diIIerent tone with shock. The day aIter this second train-
ing the rats received bilateral lesions oI the ETA made by injecting NMDA into that
region. Although these lesions cause extensive cell loss in the lateral and basolateral
nuclei, nearby regions such as the central nucleus were completely spared. Rats were
given 10 days to recover Irom surgery prior to a series oI tests designed to indepen-
dently test Iear oI the contexts and tones. We tested Iear oI both contexts in the ab-
sence oI tone and shock. Eear to both tones was tested in a novel third context.
127 FANSELOW & GALE: AMYGDALA, FEAR, AND MEMORY
Ereezing served as our behavioral index oI conditioning. The procedure shown in
EIGURE 1 is a sample procedure; all aspects oI the experiment (which tone was re-
cently or remotely trained, which context was used Iirst and second, and the order oI
tests) were counterbalanced across animals.
Data Ior tone testing (EIG. 2) and context testing (EIG. 3) are remarkably similar.
One particularly striking outcome is apparent Irom comparing the Ireezing scores oI
the sham surgery animals Ior the remote and recent memory test. This reveals that
absolutely no Iorgetting oI Iear occurs over a period oI 1.5 years, most oI the liIe
span oI the rat. This is the longest retention oI Iear memory tested in this species and,
we believe, the longest retention oI Iear memory experimentally documented Ior any
species. The lack oI Iorgetting is equally apparent Ior tone and context memory.
The eIIects oI the ETA lesion were very consistent. Ereezing was drastically re-
duced regardless oI whether the test stimulus was tone or context and whether the
memory was old or new. These results can be contrasted with the Anagnostaras et
al.
11
experiment that used the same apparatus and parameters except that only 50
days transpired between initial training and hippocampal lesions. In that study, there
was no Iorgetting in the sham surgery animals either. However, there was no eIIect
oI the lesion on tone Iear, and only recent, not remote, context Iear was aIIected.
In conclusion, Iear memories are permanent, and the ETA is critically involved in
mediating these memories. The ETA plays a general role in that it is critical Ior both
tone and context Iear. It also plays a permanent role, as evidenced by the proIound
FIGURE 1. The design oI an experiment to investigate the duration oI the Irontotem-
poral amygdala`s (ETA) mediation oI a Iear memory. Rats received Iear conditioning in
either the Wilshire Room or the Sunset Room oI the laboratory. Eear conditioning consisted
oI 10 tone-shock pairings, and a diIIerent tone was used in each context. The two experienc-
es were separated by about a year and a halI. What tones were used in what context, what
combination oI tone and context was trained Iirst and second, and the order oI testing were
all counterbalanced across animals. The design allows comparisons oI diIIerent types oI
training to be made in a statistically powerIul within-subjects manner.
128 ANNALS NEW YORK ACADEMY OF SCIENCES
deIicits in Ireezing to stimuli conditioned a very long time ago. Clearly, the ETA`s
involvement in this behavior is not related to consolidation oI a memory in some oth-
er brain region. Perhaps it could be argued that amygdala-dependent consolidation
occurs, but 1.5 years is not long enough Ior it to exert a detectable eIIect. OI course,
this would mean that such consolidation is so slow as to be Iunctionally meaningless
given the liIe span oI the rat.
PERFORMANCE FACTORS
Although the data in EIGURES 2 and 3 clearly demonstrate a permanent role Ior
the ETA in this behavior, they do not necessarily mean that this role is mnemonic.
Because the amygdala is not Iunctioning during testing regardless oI the interval be-
tween training and testing, the lesions may have aIIected perIormance in some other
way. Perhaps the lesioned animals are simply unreactive to signiIicant environmen-
tal stimuli or are hyperactive. Alternatively, the behavioral deIicits may reIlect an in-
ability oI these rats to perIorm the Ireezing response. We conducted a set oI Iollow-
up tests on these animals aIter completion oI the Iour Ireezing tests just described to
address these possibilities.
To determine iI these rats were hyperactive, we placed them in a rectangular open
Iield divided into eight square sections. Our measure oI activity was the number oI
crossovers between the sections. They were Iirst observed Ior 4 minutes, while the
open Iield was dark (EIG. 4, leIt side). Crossovers start out high during the Iirst
minute as a result oI exploration and then decline over 4 minutes. Lesioned animals
show the exact same pattern and level oI activity as the controls. Clearly, these ani-
mals are not hyperactive and display normal habituation oI exploratory activity.
Eollowing the 4-minute dark phase, bright lights situated at both ends oI the open
Iield were turned on Ior an additional 4 minutes oI observation. Normally in this test,
rats react with a sudden increase in movement that is revealed as a small elevation in
crossovers during the IiIth minute, but then activity is markedly suppressed during
FIGURE 2. The results Ior conditioning to the tone. In comparison to 'sham lesioned
controls, excitotoxic lesions oI the Irontotemporal amygdala (ETA) reduced Iear to a mem-
ory Iormed 1 day beIore the lesion (Recent) and 1.5 years (Remote) beIore the lesion,
equivalently.
129 FANSELOW & GALE: AMYGDALA, FEAR, AND MEMORY
the remainder oI the session. Our lesioned rats and their sham controls showed ex-
actly this pattern (EIG. 4, right side). The Iluctuations in activity precipitated by ex-
posure to bright light indicate that the lesioned rats are reactive, in terms oI increases
and decreases, oI activity to a signiIicant environmental stimulus. The proIound sup-
pression oI activity in the lesioned animals during the seventh and eighth minutes oI
the test clearly demonstrates that the lesions do not produce hyperactivity.
Ereezing is a Iar more speciIic response than not moving between sections oI an
open Iield. It is the complete absence oI all movements, including sniIIing and head
movement; only movement oI the animal`s Ilanks that accompanies respiration is
tolerated in the criteria we employ. Although these rats can suppress their movement
enough not to cross between sections oI an open Iield, the lesions may have rendered
them unable to generate a Ireezing response. Recent Iindings do not support this
characterization. Maren
12
demonstrated that with very extended overtraining,
amygdala-lesioned rats could Ireeze at normal levels, and he has IorceIully argued
FIGURE 3. Results Ior conditioning to the context are displayed. Erontotemporal
amygdala (ETA) lesions reduced Iear to a memory Iormed 1 day beIore the lesion (Recent)
and a year and a halI beIore the lesion (Remote) to a similar degree.
FIGURE 4. Number oI crossovers between sections oI an open Iield Ior each minute oI
an 8-minute session. The Iirst 4 minutes were in the dark and the last 4 were in bright light.
Erontotemporal amygdala (ETA) lesions did not aIIect this measure oI general activity.
130 ANNALS NEW YORK ACADEMY OF SCIENCES
that the deIicits in Ireezing shown by amygdala-damaged rats do not stem Irom an
inability to respond.

ThereIore, aIter the open Iield test, we gave the animals an over-
training session, similar to that used by Maren,
12
containing 75 unsignaled shocks.
As can be seen in EIGURE 5, overtraining resulted in high levels oI Ireezing in
lesioned animals when they were later tested in the overtrained context. The Ireezing
was associative, as these rats did not Ireeze when placed in a novel chamber. The lev-
el oI Ireezing obtained with this overtraining was comparable to that attained by the
unlesioned controls, whose data are presented in EIGURE 5 Ior comparison purposes.
We do not know what brain structures subsume the role oI the amygdala with this
overtraining protocol; those have yet to be determined. Here, we merely wish to
point out that these rats are capable oI executing a Ireezing response. It is important
to note that Maren
12
has shown that iI intact rats are Iirst overtrained and then given
a lesion, all Ireezing behavior is lost, and there is no savings when a second course
oI overtraining is begun. ThereIore, whatever neural structures compensate Ior loss
oI the amygdala, they are only engaged when substantial overtraining is given to an
amygdala-damaged subject. Thus, the amygdala is usually responsible Ior Iear learn-
ing whether normal or overtraining procedures are used.
This set oI Iindings makes it unlikely that the deIicit in Ireezing aIter conditioning
can be explained by an inability to react to signiIicant stimuli, hyperactivity, or an
inability to Ireeze. Rather, we believe the total pattern oI results points to a mnemon-
ic Iunction Ior the amygdala.
WHAT DOES THE AMYGDALA ENCODE?
Eear memories are complex; they contain many components. It is incorrect to
think that any one regions encodes the 'Iear memory. DiIIerent components oI the
Iear circuit extract and encode diIIerent aspects oI the Iear experience, and the mem-
ory is as much in the interconnections between these regions as it is in any particular
FIGURE 5. Rats with Irontotemporal amygdala (ETA) lesions shown in the previous
three Iigures were capable oI Ireezing when they were given substantial overtraining. Le-
sioned rats were overtrained and tested in both the overtrained context and a novel context.
Data on sham rats were obtained Irom the context tests oI EIGURE 1.
131 FANSELOW & GALE: AMYGDALA, FEAR, AND MEMORY
region. That being said, it is important to determine just what component oI the Iear
memory the ETA encodes. Given its general and permanent involvement in Iear
memory, that component must be essential. But what is the content oI that memory?
The question oI the content oI memory has long been a Iocus oI learning theo-
rists, particularly with respect to pavlovian conditioning. One early and rather intu-
itive hypothesis, Iormally proposed by Tolman,
13
is that the conditional stimulus
(CS) calls up a memory oI the unconditional stimulus (US). Eor the memory oI the
US to generate behavior, one component oI that US memory would have to be its
aIIective or hedonic value. According to this view, conditioning requires the encod-
ing oI an association between the CS and US and also a hedonic representation oI
the US. The association determines how well the CS can arouse the US representa-
tion, and behavior is a joint Iunction oI the associative strength and hedonic value oI
the US representation.
The most satisIying empirical attack on this view was accomplished in a series oI
experiments conducted by Rescorla.
14,15
He reasoned that we should be able to alter
conditional behavior not only by changing the associative strength between the CS
and US but also by independently manipulating the hedonic value oI the US. In one
experiment, he Iirst gave rats several pairing oI a tone and a mild shock.
15
The tone
acquired the ability to produce a low level oI Iear, presumably because the represen-
tation oI the shock reIlected the low hedonic impact oI the mild US. On another day,
Rescorla gave the rats a series oI strong shocks in the absence oI the CS. Because
there was no CS during this treatment, such manipulation should not enhance asso-
ciative strength. However, receiving strong shocks should change how the subject re-
members shock, inIlating the hedonic value oI the memory. Later, Rescorla tested
the tone in the absence oI any shock. Rats that received the inIlation shocks showed
enhanced Iear, behaving as iI they were remembering the strong inIlation shock rath-
er than the mild shock that was actually paired with the tone. The inIlation eIIect de-
pended on the tone being paired with the weak shock; rats that received tone and
weak shock unpaired and the inIlation shocks did not show the eIIect. ThereIore, a
CS-US association was critical to perIormance. Because no US was present during
the test, the eIIect oI the CS had to occur via a memory oI the US. Inasmuch as the
strong shock was never directly paired with the CS, the inIlation manipulation was
not having its eIIect on the CS-US association. Thus, Rescorla demonstrated that the
hedonic value oI the memory oI the US could independently be changed with the in-
Ilation procedure.
Weiskrantz
16
suggested that the amygdala allows animals to recognize that stim-
uli are reinIorcing, and since then, many variants oI this idea have been put Iorth.
17
19
How do you recognize that a stimulus is reinIorcing? Perhaps, it can be done by
recalling its hedonic value. By combing these ideas with those oI Rescorla, we pro-
pose that what the ETA encodes is the hedonic value oI the US. A loss oI this com-
ponent oI memory would go a long way in explaining why ETA lesions have such a
devastating eIIect on pavlovian Iear conditioning. Rescorla`s inIlation procedure oI-
Iers a way to empirically test this conjecture. Studies examining both rodents
20
and
primates
21
with amygdala lesions perIorming instrumental behavior (e.g., lever
pressing Ior Iood) are consistent with this idea. Amygdala lesions do not disrupt in-
strumental perIormance per se, but they do prevent reenIorcer devaluation Irom
changing perIormance. However, the Iact that amygdala lesions produce such subtle
eIIects on positively reinIorced instrumental behaviors compared to the devastating
132 ANNALS NEW YORK ACADEMY OF SCIENCES
eIIects the same lesions have on pavlovian Iear conditioning suggests that the struc-
ture plays a diIIerent role in these two types oI behavior. Indeed, since a lesion oI the
ETA would eliminate perIormance oI pavlovian conditional Iear responses, it would
be impossible to determine, using lesion techniques, whether the ETA plays a role in
encoding the value oI the US. An alternative approach that can give us leverage on
this question is to temporarily inactivate the amygdala. II we prevent ETA activity
just during the inIlation procedure we should prevent the animal Irom encoding the
altered hedonic value oI the shock. This procedure should not aIIect the CS-US
association, as the ETA would be Iunctional during initial training with the tone and
mild shock. PerIormance oI the original Iear memory should not be compromised
during testing, because ETA activity will have returned to normal.
We conducted just such an experiment.
22
Rats with bilateral chronic indwelling
cannulas implanted into the ETA received pairings oI tone with a mild shock, but the
amygdala was not manipulated at this time. The next day the rats received a series
oI strong shocks, but no tone was presented. All rats received an inIusion oI the
GABA agonist muscimol to inactivate cells within the ETA. Eor halI the rats the in-
Iusion was perIormed just beIore the inIlation manipulation, so that the amygdala
was shut down at the time oI memory inIlation (INEL/BeIore). The other halI oI the
rats received the same muscimol inIusion, but immediately aIter the inIlation shock
(INEL/AIter). Eor this control group the amygdala would have been able to encode
the inIlation shocks, so they served as a control Ior any nonspeciIic eIIects oI mus-
cimol. Two other control groups (CNTRL/BeIore and CNTRL/AIter) were treated
like the two inIlation groups, but they did not receive the inIlation shocks; they were
merely placed in the inIlation context. The next day the rats were placed in a novel
context and presented with the tone. No inIusions and no shocks were given, and
Ireezing to the tone was assessed (EIG. 6).
FIGURE 6. Data Irom an inIlation experiment are summarized in which rats Iirst re-
ceived pairing oI tone and an 0.5-mA shock. On the next day, rats were presented with a se-
ries oI 3-mA shocks (INEL/) or simply placed in the conditioning chamber (CNTRL/).
The amygdala was inactivated with muscimol either immediately beIore (/BeIore) or aIter
(/AIter) this shock inIlation treatment. On the day aIter inIlation, tone Iear was tested in a
novel chamber. There was no inIusion and no shock on this test day.
133 FANSELOW & GALE: AMYGDALA, FEAR, AND MEMORY
A comparison oI the INEL/AIter and the CNTRL/AIter groups oI EIGURE 6 shows
a replication oI Rescorla`s inIlation eIIect. However, Irom a comparison oI the INEL/
AIter and the INEL/BeIore groups, it is obvious that inIusing muscimol immediately
beIore inIlation treatment completely abolished the enhancement oI Iear. Thus, the
ETA is essential Ior the inIlation oI Iear. Because all animals received muscimol in-
Iusions, the eIIects cannot be attributed to some nonspeciIic or long-term conse-
quence oI muscimol inIusion. These data provide strong evidence that the amygdala
encodes the hedonic memory oI the US. The eIIects cannot stem Irom disruption oI
some postinIlation consolidation processes, because animals that received inIusions
immediately aIter inIlation showed enhanced Iear.
CONCLUSIONS
The amygdala pays a crucial role in pavlovian Iear conditioning. The series oI ex-
periments summarized in this chapter go a long way towards pinpointing what this
crucial role is Ior the Irontotemporal amygdala. Taken together, the data suggest that
the ETA permanently encodes the hedonic value oI aversive stimuli and makes that
memory available to other structures that encode other aspects oI the Iear-condition-
ing experience. Eor example, the hippocampus can connect that aIIective inIorma-
tion to particular contexts and episodes.
23,24
The medial geniculate connects it to
speciIic auditory stimuli.
25
Projections to the striatum may provide aIIective tone to
instrumental behavior.
26
The aIIective inIormation stored by the ETA may be able to
inIluence the storage oI declarative memories mediated by other structures.
27
Cer-
tainly the ETA is not the only part oI the amygdala important Ior Iear. Eor example,
the ability Ior aIIective memories to activate the central nucleus via its connections
with the ETA is critical Ior initiating the constellation oI behaviors that allows a spe-
cies to eIIectively deIend against threats that jeopardize its survival.
2,4,5
A greater
appreciation oI the interrelations between these regions and other components oI the
Iear circuitry will enhance our understanding oI Iear processes and will be critical
Ior advancements in the treatment oI Iear-related disorders.
ACKNOWLEDGMENT
This research was supported by National Science Eoundation Grant IBN-0091487
(awarded to M.S.E.). During the preparation oI this paper, M.S.E. was a Iellow at the
Hanse-WissenschaItskolleg, Lehmkuhlenbusch 4, 27753 Delmenhorst, Germany.
REFERENCES
1. EANSELOW, M.S. & J.E. LEDOUX. 1999. Why we think plasticity underlying pavlovian
Iear conditioning occurs in the basolateral amygdala. Neuron 23: 229232.
2. EANSELOW, M.S. 1994. Neural organization oI the deIensive behavior system responsi-
ble Ior Iear. Psychon. Bull. Rev. 1: 429438.
3. EENDT, M. & M.S. EANSELOW. 1999. The neuroanatomical and neurochemical basis oI
conditioned Iear. Neurosci. Biobehav. Rev. 23: 743760.
4. SWANSON, L.W. & G.D. PETROVICH. 1998. What is the amygdala? Trends Neurosci. 8:
323331.
134 ANNALS NEW YORK ACADEMY OF SCIENCES
5. DAVIS, M. 1992. The role oI the amygdala in conditioned Iear. In The Amygdala: Neu-
robiological Aspects oI Emotion, Memory, and Mental DysIunction. John P. Aggle-
ton, Ed. :255306. Wiley-Liss. New York.
6. LEE, Y., D. WALKER & M. DAVIS. 1996. Lack oI a temporal gradient oI retrograde
amnesia Iollowing NMDA-induced lesions oI the basolateral amygdala assessed with
the Iear-potentiated startle paradigm. Behav. Neurosci. 110: 836839.
7. MAREN, S., G. AHARONOV & M.S. EANSELOW. 1996. Retrograde abolition oI condi-
tional Iear aIter excitotoxic lesions oI the basolateral amygdala oI rats: absence oI a
temporal gradient. Behav. Neurosci. 110: 718726.
8. KIM, J.J. & M.S. EANSELOW. 1992. Modality speciIic retrograde amnesia oI Iear Iol-
lowing hippocampal lesions. Science 256: 675677.
9. GALE, G.D., S.G. ANAGNOSTARAS, B.P. GODSIL, et al. 1999. The basolateral amygdala
and storage oI Iear memories spanning the adult liIetime oI rats. Society Ior Neuro-
science, Miami.
10. MAREN, S., G. AHARONOV & M.S. EANSELOW. 1997. Neurotoxic lesions oI the dorsal hip-
pocampus and Pavlovian Iear conditioning in rats. Behav. Brain Res. 88: 261274.
11. ANAGNOSTARAS, S.G., S. MAREN & M.S. EANSELOW. 1999. Temporally graded retro-
grade amnesia oI contextual Iear aIter hippocampal damage in rats: within-subjects
examination. J. Neurosci. 19: 11061114.
12. MAREN, S. 2001. Is there savings Ior pavlovian Iear conditioning aIter neurotoxic
basolateral amygdala lesions in rats? Neurobiol. Learn. Mem. 76: 268283.
13. TOLMAN, E.C. 1932. Purposive Behavior in Animals and Men. Appleton. New York.
14. RESCORLA, R.A. 1973. EIIects oI US habituation Iollowing conditioning. J. Compar.
Physiol. Psychol. 82: 137143.
15. RESCORLA, R.A. 1974. EIIect oI inIlation oI the unconditioned stimulus value Iollow-
ing conditioning. J. Compar. Physiol. Psychol. 86: 101106.
16. WEISKRANTZ, L. 1956. Behavioral changes associated with ablation oI the amygdaloid
complex in monkeys. J. Compar. Physiol. Psychol. 49: 381391.
17. CARDINAL, R.N., J.A. PARKINSON, J. HALL & B.J. EVERITT. 2002. Emotion and motiva-
tion: the role oI the amygdala, ventral striatum, and preIrontal cortex. Neurosci.
Biobehav. Rev. 26: 321352.
18. MAREN, S. & M.S. EANSELOW. 1996. The amygdala and Iear conditioning: has the nut
been cracked? Neuron 16: 237240.
19. SETLOW, B., M. GALLAGHER & P.C. HOLLAND. 2002. The basolateral complex oI the
amygdala is necessary Ior acquisition but not expression oI CS motivational value in
appetitive pavlovian second-order conditioning. Eur. J. Neurosci. 15: 18411853.
20. HATEIELD, T., J.S. HAN, M. CONLEY, et al. 1996. Neurotoxic lesions oI basolateral, but
not central, amygdala interIere with Pavlovian second-order conditioning and rein-
Iorcer devaluation eIIects. J Neurosci. 16: 52565265.
21. MALKOVA, L., D. GAEEAN & E.A. MURRAY. 1997. Excitotoxic lesions oI the amygdala
Iail to produce impairment in visual learning Ior auditory secondary reinIorcement
but interIere with reinIorcer devaluation eIIects in rhesus monkeys. J. Neurosci. 17:
60116020.
22. AKBARI, Y., R. MONGEAU, S. MAREN & M.S. EANSELOW. 1997. Reversible inactivation
oI the basolateral amygdala prevents inIlation oI Iear conditioning in rats. Society Ior
Neuroscience, New Orleans.
23. EANSELOW, M.S. 2000. Contextual Iear, gestalt memories, and the hippocampus.
Behav. Brain Res. 110: 7381.
24. MAREN, S. & M.S. EANSELOW. 1995. Synaptic plasticity in the basolateral amygdala
induced by hippocampal Iormation stimulation in vivo. J. Neurosci. 15: 75487564.
25. LEDOUX, J.E., A. SAKAGUCHI & D.J. REIS. 1984. Subcortical eIIerent projections oI the
medial geniculate nucleus mediate emotional responses conditioned to acoustic stim-
uli. J Neurosci. 4: 683698.
26. EVERITT, B.J., J.A. PARKINSON, M.C. OLMSTEad, et al. 1999. Associative processes in
addiction and reward. The role oI amygdala-ventral striatal subsystems. Ann. N.Y.
Acad. Sci. 877: 412438.
27. CAHILL L. & J.L. MCGAUGH. 1998. Mechanisms oI emotional arousal and lasting
declarative memory. Trends Neurosci. 7: 294299.

You might also like