INSIGHT REVIEW NATURE|Vol 444|16 November 2006|doi:10.

1038/nature05405

Smell images and the flavour system in the

human brain
Gordon M. Shepherd1

Flavour perception is one of the most complex of human behaviours. It involves almost all of the senses,
particularly the sense of smell, which is involved through odour images generated in the olfactory pathway.
In the human brain, the perceptual systems are closely linked to systems for learning, memory, emotion
and language, so distributed neural mechanisms contribute to food preference and food cravings. Greater
recognition of the role of the brain’s flavour system and its connection with eating behaviour is needed for a
deeper understanding of why people eat what they do, and to generate better recommendations about diet
and nutrition.

Many humans regard their sense of smell to be of minor importance.
The dominant role that smell has in sensing the flavours of the foods
we eat and influencing what we like to eat is largely unrecognized.
This is unfortunate, because our diet is causing a public health crisis in
many western countries. Knowledge of the importance of smell to the
perception of flavour and the formation of cognitive and emotional
responses to foods and beverages could help to improve health and
prevent chronic conditions such as obesity and diabetes.
The aim of this article is to review recent advances in the brain mecha-
nisms of smell perception and to put forward several hypotheses for
integrating these mechanisms into current theories of the neural basis of
flavour as a complex multisystem percept. My hope is that this will help
bridge the divide that impedes neuroscientists, behavioural psycholo-
gists, food scientists, nutritionists and public-policy makers from com-
municating effectively across disciplines about why we eat what we do
and how to achieve healthy diets. I focus on three critical aspects of the
brain mechanisms of smell: how odour molecules are represented in the
form of ‘odour images’ as the basis for smell perception; the importance
of retronasal smell for flavour perception; and the close relationships
between the brain mechanisms for flavour and the brain systems for
emotions and cravings.
Odour molecules are represented by odour images
Odour stimulation of olfactory sensory cells in the nose involves odour
molecules interacting with olfactory-receptor proteins. The discovery of
the genes that encode these proteins in 1991 by Richard Axel and Linda
Buck1 moved olfaction into the front rank among scientific endeav-
ours, not only opening the way to understanding olfactory (and taste)
transduction at the receptor level, but also providing new genetic and
molecular tools for pursuing the organization of the olfactory pathway
in the brain.
The genome projects show that the family of olfactory genes totals
more than 1,000 — the largest in the mammalian genome2. In the major-
ity of mammals, most of the genes are functional, but in primates the
number of functional genes decreases, in humans to about 350 (archived
in the Olfactory Receptor Database; see ref. 3). This has been taken to
confirm the belief that humans have degenerate senses of smell. How-
ever, olfactory genes do not map directly onto smell acuity; dogs, which
have superior tracking abilities, have only about 850 functional genes4.
1
Department of Neurobiology, Yale University School of Medicine, 333 Cedar Street, New Haven, Connecticut 06510, USA.
In fact, behavioural tests show that primates have surprisingly good
senses of smell5, and it has been argued that the decline in olfactory-gene
number is more than offset in humans by their much enlarged brains
with their increased capacities for analysis and complex processing of
smell to guide critical human behaviours6. As we shall see, a notable
share of this enlarged brain capacity is involved in flavour perception
and behavioural responses to flavours.
The axons of sensory neurons in the olfactory epithelium project
to the olfactory bulb, where they converge onto modules known as
glomeruli (Fig. 1a). Direct evidence of the brain mechanisms involved
in smell first emerged in the 1970s with the application of the earliest
brain-imaging methods to the olfactory bulb. This showed that odour
stimuli in the rat produce spatial patterns of activity in the glomerular
layer that are overlapping but that vary for different odours7. From this
was derived the fundamental principle, first suggested by Adrian8, that
different odour molecules are represented by different patterns of spatial
activity in the olfactory bulb. A homologous chemical series — such as
aldehydes — elicits patterns that are overlapping but different, as shown
by 2-deoxyglucose9 and high-resolution functional magnetic resonance
imaging (fMRI)10 (Fig. 1b). Behavioural studies11 have confirmed that
mice can readily discriminate such single-carbon differences between
molecules.
The olfactory system thus resembles other sensory pathways in that
it uses activity patterns in two-dimensional neural space to represent its
sensory modality. And such patterns, analogous to those of other sen-
sory systems, constitute ‘odour images’, or ‘odour maps’. Much research
has confirmed and extended this basic finding in the olfactory system
(for a review see ref. 12). Some recent studies have focused on charac-
terizing how these images evolve during a respiratory cycle13, whereas
others have analysed the contribution of temporal firing patterns to the
encoding process14. Work on tracing the connections in the olfactory
bulb is revealing the distributed circuits that process the images15. These
allow two-dimensional images (Fig. 1b) to encode the multidimensional
odour space inherent in the complex structures of odour molecules.
This principle of odour-molecule representation by glomerular activity
patterns applies across phyla16.
Despite this emerging consensus, the idea of odour images and odour
maps has not yet been incorporated into current theories on the neural
basis of smell and flavour perception (but see ref. 17), or in psychophysical

316
©2006 Nature Publishing Group
NATURE|Vol 444|16 November 2006 INSIGHT REVIEW

a * Olfactory This is the route used to sense odours in the environment (Table 1). The
Olfactory bulb resultant odour images in the olfactory bulb are subject to processing by
epithelium
the olfactory cortex and are relayed to the primary olfactory cortex in
the orbitofrontal cortex (OFC), a part of the prefrontal lobe. This direct
input to the highest cognitive centres of the brain is a special property
Odour molecules
of smell that is critical to the experience of human flavour.
Retronasal stimulation occurs during food ingestion, when volatile
molecules released from the food in the mouth are pumped, by move-
Olfactory
ments of the mouth, from the back of the oral cavity up through the
sensory
neurons nasopharynx to the olfactory epithelium. Orthonasal olfaction does not
Glomeruli
contribute to this sensation. Retronasal olfactory stimulation can be
Odour molecules studied by introducing odour stimuli at the back of the mouth20 or by
their release from ingested foods21. Marcel Proust vividly described how
a tea-soaked madeleine brought back powerful childhood memories22,
b and this would have occurred primarily through the retronasal path-
way. It is activated only when breathing out through the nose between
mastications or swallowings23 (Fig. 2b). Because the molecules arise
from the food in the mouth, they are referred to the mouth; that is, they
are perceived as if they are sensed within the mouth24. This retronasal
food-molecule-laden air is judged to be part of the ‘taste’ of the food25
and has been shown to be necessary for flavour identification26. Thus,
* although a large part of flavour is due to smell, it is attributed to ‘taste’,
and is therefore almost entirely hidden from our conscious perception.
To avoid giving ‘taste’ a double meaning, we incorporate retronasal smell
as one of the modalities covered by the term ‘flavour’.

O O O Retronasal smell and the brain flavour system

The traditional view in the literature on eating behaviour in human
Figure 1 | Odour images in the olfactory glomerular layer. a, Diagram culture27 is that the flavour of prepared foods is humanity’s greatest
showing the relationship between the olfactory receptor cell sheet in the universal shared behaviour, experienced by individuals of all ages in
nose and the glomeruli of the olfactory bulb53. b, fMRI images of the the course of daily life. Flavour is also among the most complex and
different but overlapping activity patterns seen in the glomerular layer of
powerful of all human sensations. It engages almost all of the sensory
the olfactory bulb of a mouse exposed to members of the straight-chain
aldehyde series, varying from four to six carbon atoms. The lower part modalities. It also engages the complex facial, swallowing and respira-
of the image in the left panel corresponds to the image on the medial tory motor systems. Flavour is therefore an active sensation — we use
side of the olfactory glomerular layer as shown in a (see asterisk). (Image ‘active taste’ to palpate our food with our tongue as we use ‘active touch’
in a adapted, with permission, from ref. 53; image in b adapted, with to palpate an object we are examining with our fingers. Some of these
permission, from ref. 10.) systems are indicated in the diagram of Fig. 2b. Above these sensory and
motor systems are the cognitive systems for memory, emotion, abstract
thinking and language. The importance of retronasal smell images is
studies of the relationship between odour molecule structure and percep- illustrated by the massive extent to which they interact with these brain
tual response. This resistance seems to be due to at least two factors. systems compared with orthonasal smell images (compare Fig. 2a with
First, unlike the conscious images we perceive through the visual sys- Fig. 2b). Below, we briefly consider the main sensory systems.
tem, the odour patterns that arise in the olfactory bulb are unconscious. Taste is the most obvious component of food flavour. The receptors that
When we perceive a smell we do not think of it as a spatial image, even relate to sweet, salt, sour, bitter and umami taste are described in detail
though it arose from a spatial image in the olfactory bulb. This uncon- by Zuker in this issue (page 288). Here, I touch only on certain aspects of
scious nature may reflect the fact that the human olfactory pathway does taste relevant to the role of smell in flavour. In humans, the taste pathway
not include a significant relay in the thalamus (see below). ascends from the nucleus of the solitary tract in the brainstem to the
Second, odour patterns are complex and highly irregular (Fig. 1). hypothalamus and to the taste area of the somatosensory thalamus, and
Recognition of odour images is thus an example of pattern recognition.
We have suggested that this is analogous to recognizing complex visual
Table 1 | The dual olfactory system
patterns such as those of human faces6. Just as the identity of a face is
encoded in its visual image, so is the identity of an odour molecule or Operations Orthonasal olfaction Retronasal olfaction
complex odour object — such as the scent of a flower or the aroma of Stimulation Through the external From the back of the mouth through
coffee — encoded in an odour image. route nares the nasopharynx
Unconscious pattern recognition of odour images therefore seems Stimuli Floral scents Food volatiles
to lie at the heart of smell perception. This view is being incorporated Perfumes
by food scientists into their concepts of the neural basis of flavour18. Smoke
Food aromas
Neuroscientists and psychophysicists have the opportunity to follow Prey/predator smells
suit and devise experiments that test this hypothesis, which we will use Social odors
to describe the role of smell in flavour perception. Pheromones
MHC molecules
Retronasal smell is essential for human flavour perception Processed by Olfactory pathway Olfactory pathway combined with
As noted by Paul Rozin19, smell is unique in having a ‘dual nature’ — influenced by the pathways for taste, touch, sound
visual pathway and active sensing by proprioception
that is, it can sense signals originating outside (orthonasal) and inside form a ‘flavour system’
(retronasal) the body (Table 1). Note the interesting contrast, that orthonasal olfactory perception involves a wide range of types
Orthonasal stimulation refers to sniffing in through the external nares of odors processed through only the olfactory pathway, in comparison with retronasal olfactory
perception which involves only food volatiles but processed in combination with many brain pathways.
of the nose to activate the sensory cells in the olfactory epithelium (Fig. 2a).
317
©2006 Nature Publishing Group
INSIGHT REVIEW NATURE|Vol 444|16 November 2006

a Orthonasal olfactory b Retronasal olfactory

perceptual system flavour system
Motor
cortex SOM
DI faces
Insula SOM SOM
taste taste tongue
AVI PPC
ACC

MOFC LH
MOFC VPM VC
LOFC AM LOFC AM
OC OC
OB OB
OE OE

NST

VII
Odour in Odour in
inspired expired Taste buds IX X
air air
Tongue muscles
Salivary glands Smell
Masticating muscles Taste
Texture
Vision
Motor

Figure 2 | The dual olfactory system. a, Brain systems involved in smell perception during orthonasal olfaction (sniffing in). b, Brain systems involved
in smell perception during retronasal olfaction (breathing out), with food in the oral cavity. Air flows indicated by dashed and dotted lines; dotted lines
indicate air carrying odour molecules. ACC, accumbens; AM, amygdala; AVI, anterior ventral insular cortex; DI, dorsal insular cortex; LH, lateral
hypothalamus; LOFC, lateral orbitofrontal cortex; MOFC, medial orbitofrontal cortex; NST, nucleus of the solitary tract; OB, olfactory bulb; OC, olfactory
cortex; OE, olfactory epithelium; PPC, posterior parietal cortex; SOM, somatosensory cortex; V, VII, IX, X, cranial nerves; VC, primary visual cortex;
VPM, ventral posteromedial thalamic nucleus.

from there extends to the primary taste cortex (Fig. 2b). A key fact about
taste stimuli is that they elicit the most basic human emotions of pleasure
(sweet) and disgust (bitter), which are not learned; they are hard-wired in
the brainstem from birth28,29. In contrast to taste, the affective responses
to odour images seem to be mostly learned, which presumably accounts
for the enormous diversity of flavours in the world’s cuisines.
Using brain imaging, researchers have studied how smell and taste
pathways function, separately and together, in humans30. At the neocor-
tical level, taste stimuli alone activated the primary taste centres in the
tongue area and insula, whereas odour stimuli by themselves activated the
primary olfactory receiving area in the orbitofrontal region of prefrontal
cortex (Fig. 2b), as expected. The investigators then tested for simultane-
ous presentation of the taste and odour stimuli. This produced enhanced
activity in many of the regions activated by the independent stimulations
as well as additional activity in contiguous areas around the primary
receiving areas (Fig. 2b). This indicates that a wider range of integrative
activity underlies the more complex perception of flavour than would
occur through simple addition of the taste and smell pathways.
By contrast, neocortical integration in rodents is minimal. The impor-
tant principle is that in humans the increase in processing at this higher
cognitive level is independent of the smaller size of the receptor reper-
toire. A possible analogy is with language: the development of human
language was not dependent on more hair cells or a more acute sense
of hearing, but on the elaboration of neocortical areas for the complex
processing of low- to medium-frequency auditory signals. Similarly, it
can be proposed that the increased processing and elaboration of fla-
vour through these larger cognitive regions is unique to humans (Box
1). Processing in this case encompasses odour images together with the
other senses that contribute to flavour. But how does this occur?
The orbitofrontal sensory packaging centre
The prefrontal cortex has long been recognized as a special cortical area
in the human brain that is crucial for planning, abstract thought and
318
©2006 Nature Publishing Group
working memory. As indicated in Fig. 2b, it is also particularly impor-
tant in flavour perception. Anatomical studies31 in monkeys show that
in addition to the primary neocortical olfactory area there are various
subregions differentiated to receive connections from other sensory neo-
cortical areas, including taste, hearing, touch and vision. There are also
Box 1 | Smell images, flavours and language
It is commonly assumed that there is a remote relationship between
language and smell; for example, it is difficult to find words to describe
the taste of a wine. It has been postulated that odours are difficult to
name because odour and language processing rely on the same neural
substrates56. Olfactory perceptions are also heavily conditioned by
verbal cues, in the same way that they are influenced by other senses57.
However, the naming of smells recognized through odour images may
be no more difficult than naming of visual images such as faces; both
involve the inherent difficulty of describing complex spatial patterns
in words. We are very good at recognizing human faces but have great
difficulty describing them, and it is similarly difficult for us to describe
a smell.
The relationship between flavour and language is even more
interesting when retronasal odour images are considered to be a main
component of food flavours. Cooking had a central role in human
evolution58; the shared common meal has long been regarded as the
defining social activity of early humans27. Within this social context,
language arose not only in relation to the activities of hunting and
gathering, but also as a necessary means to communicate about
the preparation of the meal and its assessment in terms of desirable
flavours. It can be proposed that the adaptive advantage of language
allowed humans to attach verbal labels to the odour images associated
with new flavours produced by cooking, facilitating the process of
extracting more nutritious food from otherwise unpalatable foodstuffs
such as roots, nuts or toxic plants58, or for the cultural advantage it
conferred in the competition between rival human groups.
NATURE|Vol 444|16 November 2006 INSIGHT REVIEW

subregions with reciprocal connections with limbic subcortical regions
such as the amygdala and the hypothalamus. An extensive series of phys-
iological studies32 in monkeys with recordings from single cells in these
areas has shown that responses to taste and both orthonasal and retro-
nasal stimuli are heavily dependent on context, especially in connection
with learning and reward. Similar results have been shown by fMRI
studies in humans33,34. These suggest that taste and, in particular, smell
perceptions are more heavily conditioned by multi-sensory inputs than
by perceptions in other systems such as those for vision and hearing.
Integral components of our experience of eating also arise from all
submodalities of the somatosensory system: fine touch, creaminess,
deep pressure (for example, crunchy), temperature and pain (such as
the burn of chillies). Astringency, such as that of a dry wine, also gives
a sense of body.
These systems and their submodalities are represented in Fig. 3, and
their pathways from the sensors to the receiving cortical regions are
indicated in Fig. 2b. Flavour is also influenced by the sight of the food to
be consumed35. A recent study36 tested both naive and experienced sub-
jects on red and white wines. Both groups classified the wines accord-
ing to the contrasting tastes of red and white wines. However, the reds
included whites that had been coloured red, showing that the colour
had an overriding effect on the perceived flavour. The visual cortex is
therefore included in Fig. 2b as a part of the greater flavour system (see
also Table 1).
In summary, the multisystem nature of the flavour centre of the OFC
(Fig. 2b) suggests that this can be regarded as a sort of ‘sensory packag-
ing centre’ in the brain, one of the areas that makes us susceptible to the
pitch of advertisers to buy food in different coloured boxes or with more
crunchy texture. This centre has a critical and asymmetric relationship
with human language areas (Fig. 3). On the one hand, there is the dif-
ficulty humans have describing in words the spatiotemporal patterns
of odour images as part of flavour perception. On the other hand, there
is the strong influence of words on our flavour perceptions and prefer-
ences (Box 1).
From images of smell to images of desire
The perception of flavour thus involves many sensory and motor sys-
tems. But in order for a flavour to play a part in food preference and
consumption, it must also connect with other systems that are involved
in the complex control of feeding. Recent research has begun to show
how flavour perception might drive these systems as well.
The olfactory cortex and its interactions with other limbic areas form a
central hub in this process (Figs 2b, 3). In the olfactory cortex, the odour
images are reformatted into content-addressable memory form37–40. (The
olfactory cortex also contains a sensor for essential amino acids41, which
is indicated by the asterisk in Fig. 3.) A number of studies have identi-
fied the sub-areas that are activated by food flavours. These include the
OFC, parahippocampal gyrus, anterior fusiform gyrus, insula, striatum
and cingulate. These are mostly cortical areas, or centres closely associ-
ated with them. Moreover, the OFC and amygdala have been found to
have associations with the desirability of foods, with pleasurable foods
activating the medial OFC and unpleasant foods activating the lateral
OFC42. It is therefore possible to begin to identify the different feelings
and states of motivation about feeding in relation to levels of activation
of the extended amygdalar system, from a level of liking or disliking a
flavour to feeling strongly attracted or repelled by it (Fig. 3).
The motivation to eat is also involved in cravings, overeating and
conditions such as obesity, in which energy intake chronically exceeds

Sensory modalities

Vision Primate neocortex

Colour Conscious Human neocortex
Shape Flavour perception Language circuits
Circuits

Sound
Frequency

Somatosensory
Temperature
Deep touch
Astringency Amygdala
Light touch systems
Creaminess
Pain Emotion
circuits
Taste Hippocampus, olfactory* and
Sweet limbic systems)
Motivation
Umami
Limbic subconscious circuits
Salt
memory systems
Sour
Bitter Craving
circuits

Smell
Pattern

Gut
Autonomic and Hypothalamus
metabolic * Feeding circuits
properties

Figure 3 | The human brain flavour systems that evaluate and regulate food intake. The diagram shows the areas involved in the perceptual, emotional,
memory-related, motivational and linguistic aspects of food evaluation mediated by flavour inputs32,41,54,55. Left, different sensory modalities and
submodalities that contribute to flavour perception. Middle and right, brain flavour system that evaluates and regulates food intake. Red regions mediate
conscious sensory perception; thicker outlines indicate their greater importance in humans and other primates. Green regions mediate subconscious
feeding regulation. Deficiencies in essential amino acids are sensed by the anterior olfactory cortex (asterisk).
319
©2006 Nature Publishing Group
INSIGHT REVIEW
energy expenditure. Various animal studies have shown that the
regions involved in cravings include the amygdala, anterior cingulate,
OFC, insula, hippocampus, caudate and dorsolateral prefrontal cortex.
Functional imaging studies in humans have tested for activity in these
areas43 by maintaining subjects on a monotonous diet and asking them
to imagine the taste, smell and texture (in other words, flavour) of their
favourite foods. Craving-related increases in the fMRI blood-oxygen-
level dependent (BOLD) signal were seen in the hippocampus, insula
and caudate nucleus. The authors characterized these activity sites as
‘images of desire’.
The regions activated in humans by food cravings are similar to those
activated during drug and alcohol addiction. The results are consistent
with the common substrates hypothesis — the idea, arising from many
converging lines of work, that the brain areas underlying food cravings
share cellular and systems mechanisms with those involved in alcohol
and drug abuse44,45.
Thus, the images of smells in the olfactory pathway lead to images of
desire for acquiring and ingesting the object supplying that smell image,
and so, it can be proposed, the cycle repeats itself. The key point is that
understanding overeating and obesity involves not only understanding
the hypothalamic feeding centres and how they respond to fats, carbo-
hydrates and proteins (Fig. 3), but how those centres are driven by the
brain mechanisms underlying the flavours of those foods and the desire
to consume them.
Food science and beyond
Behavioural testing of the effects of food flavours and consumer pref-
erence for them is carried out in food science, a field almost entirely
unknown to neuroscientists on the one hand and nutritionists on the
other. Using combinations of gas chromatographic methods linked to
odour testing, these studies are providing essential links between the
physical and chemical properties of foods and the flavour perceptions
they produce. The level of detail can be exquisite; one such study, for
example, analysed the type of grass consumed by cows hourly during
the day and in different seasons, and how this related to the chemical
composition and flavour qualities of cheese produced from the cows’
milk46. Another study investigated the change in retronasal volatiles
released during successive stages of ingestion, mastication and swallow-
ing47. Such research is contributing to the effort to enhance traditional
methods of producing foods with naturally rich flavours48, as well as to
the effort to maintain or simulate these qualities in produce for mass
markets. Recent studies have also assessed the decline that occurs in
flavour perception in humans as they age, when retronasal olfaction
seems more vulnerable to loss than orthonasal olfaction49.
Human flavour systems in the brain are also of interest in relation to
the popular movement, which began in the 1980s, to put the cooking of
food on a scientific basis. Explaining the chemistry and physics of food
preparation, and the biochemical basis of food flavours, is a lively area of
popular communication of food science that is associated with Shirley
Corriher50 and Harold McGee51, and termed ‘molecular gastronomy’ by
others52. It is possible to imagine a future ‘neurogastronomy’, combin-
ing the biochemistry of food preparation, the molecular biology of the
olfactory receptors, and the knowledge of odour images and the brain
flavour system reviewed here. Such a synthesis could have the potential
to improve our understanding of the human experience of eating and
ways to shape it towards more flavourful and healthier diets.
Challenges
The increasing evidence that odour molecules are encoded as odour
images provides testable hypotheses for further research. Building on
current animal studies, applications of genetic and molecular tools to
the large receptor gene family will provide the molecular basis of the
organization of the olfactory pathway from the sensory cells to the olfac-
tory cortex, and perhaps, ultimately, to the neocortical level as well. This
will merge with progress in the physiological and brain imaging studies
mentioned here to give a comprehensive understanding of the neural
basis of odour images during orthonasal olfaction.
320
©2006 Nature Publishing Group
NATURE|Vol 444|16 November 2006
Extending these studies to retronasal olfaction and the basis of flavour
will require more concerted efforts by neuroscientists to integrate olfac-
tion with taste and the other sensory systems, and with psychophysical
and ingestive behaviour studies. In this multidisciplinary approach, the
expertise of food scientists and nutritionists will also be needed. In addi-
tion, anthropologists need to provide a better understanding of the role
of smell in the development of cuisines during human evolution.
This integrated approach should make it possible for researchers to
attack the difficult questions about human flavour perception. What is
the relationship between flavour and nutrition? This would have been
critical for human evolution, and is at the core of the current concern
about healthy diets. Why do fast foods and certain flavours attract
humans to the point of craving? If the smell of the flavour is an image,
the work reviewed here suggests that the smell image plays a large part
in that attraction. Is a fast-food flavour an attractive ‘caricature’ of the
odour image of the natural food, in the same way that Mickey Mouse is
an attractive caricature of a real mouse?
The traditional view is that we are attracted to foods by their biochem-
ical composition, especially by those constituents that make them sweet
or fat; recently, attention has shifted to the biochemical actions of food
constituents on addiction centres in the brain. Here we have reviewed
evidence that flavour is a major factor. If we recognize the power that
conscious images have on our daily lives, the power of the unconscious
odour images that dominate flavours seems worth investigating. ■
1. Buck, L. & Axel, R. A novel multigene family may encode odorant receptors: a molecular
basis for odor recognition. Cell 65, 175–187 (1991).
2. Zhang, X. & Firestein, S. The olfactory receptor gene superfamily of the mouse. Nature
Neurosci. 5, 124–133 (2002).
3. Crasto, C., Marenco, L., Miller, P. & Shepherd, G. Olfactory receptor database: a metadata-
driven automated population from sources of gene and protein sequences. Nucleic Acids
Res. 30, 354–360 (2002).
4. Quignon, P. et al. The dog and rat olfactory receptor repertoires. Genome Biol. 6, R83
(2005).
5. Laska, M., Seibt, A. & Weber, A. ‘Microsmatic’ primates revisited: olfactory sensitivity in
the squirrel monkey. Chem. Senses 25, 47–53 (2000).
6. Shepherd, G. M. The human sense of smell: are we better than we think?PLoS Biol. 2, e146
(2004).
7. Stewart, W. B., Kauer, J. S. & Shepherd, G. M. Functional organization of rat olfactory bulb
analysed by the 2-deoxyglucose method. J. Comp. Neurol. 185, 715–734 (1979).
8. Adrian, E. D. Sensory messages and sensation; the response of the olfactory organ to
different smells. Acta Physiol. Scand. 29, 5–14 (1953)
9. Johnson, B. A., Farahbod, H., Saber, S. & Leon, M. Effects of functional group position on
spatial representations of aliphatic odorants in the rat olfactory bulb. J. Comp. Neurol. 483,
192–204 (2005).
10. Xu, F. Q. et al. Odor maps of aldehydes and esters revealed by fMRI in the glomerular layer
of the mouse olfactory bulb. Proc. Natl Acad. Sci. USA 100, 11029–11034 (2003).
11. Laska, M., Joshi, D. & Shepherd, G. M. Olfactory sensitivity for aliphatic aldehydes in CD-1
mice. Behav. Brain Res. 167, 349–354 (2006).
12. Xu, F., Greer, C. A. & Shepherd, G. M. Odor maps in the olfactory bulb. J. Comp. Neurol. 422,
489–495 (2000).
13. Spors, H., Wachowiak, M., Cohen, L. B. & Friedrich, R. W. Temporal dynamics and latency
patterns of receptor neuron input to the olfactory bulb. J. Neurosci. 26, 1247–1259 (2006).
14. Laurent, G. Olfactory network dynamics and the coding of multidimensional signals.
Nature Rev. Neurosci. 3, 884–895 (2002).
15. Willhite, D. C. et al. Viral tracing identified distributed columnar organization in the
olfactory bulb. Proc. Natl Acad. Sci. USA 103, 12592–12597 (2006).
16. Hildebrand, J. G. & Shepherd, G. M. Molecular mechanisms of olfactory discrimination:
converging evidence for common principles across phyla. Annu. Rev. Neurosci. 20, 595–631
(1997).
17. Shepherd, G. M. Outline of a theory of olfactory processing and its relevance to humans.
Chem. Senses 30 (Suppl. 1), i3–i5 (2005).
18. Acree, T. E., Deibler, K. D. & Kittel, K. M. in Handbook of Flavor Characterization: Sensory
Analysis, Chemistry, and Physiology (eds Deibler, K. D. & Delwiche, J.) 83–91 (Marcel Dekker,
New York, 2004).
19. Rozin, P. ‘Taste–smell confusions’ and the duality of the olfactory sense. Percept.
Psychophys. 31, 397–401 (1982).
20. Sun, B. C. & Halpern, B. P. Identification of air phase retronasal and orthonasal odorant
pairs. Chem. Senses 30, 693–706 (2005).
21. Duffy, V. B., Backstrand, J. & Ferris, A. M. Olfactory dysfunction and related nutritional risk
in free-living, elderly women. J. Am. Diet. Assoc. 95, 879–884 (1995).
22. Shepherd-Barr, K. & Shepherd, G. M. Madeleines and neuromodernism: reassessing
mechanisms of autobiographical memory in Proust. Auto/Biography Studies 13, 40–60
(1998).
23. Taylor, A. J., Linforth, R. S. T., Harvey, B. A. & Blake, A. Atmospheric pressure chemical
ionisation for monitoring of volatile flavour release in vivo. Food Chem. 71, 327–338 (2000).
24. Bartoshuk, L. M. et al. From psychophysics to the clinic: missteps and advances. Food Qual.
Preference 15, 617–632 (2004).
25. Murphy, C., Cain, W. S. & Bartoshuk, L. M. Mutual action of taste and olfaction. Sens.
Process. 1, 204–211 (1977).
NATURE|Vol 444|16 November 2006
26. Mozell, M., Smith, B., Smith, P., Sullivan, L. & Swender, P. Nasal chemoreception in flavor
identification. Arch. Otolaryngol. 90, 367–373 (1969).
27. Farb, P. & Armelagos, G. Consuming Passions: The Anthropology of Eating (Houghton Mifflin,
Boston, 1980).
28. Scalera, G., Grigson, P. S. & Norgren, R. Gustatory functions, sodium appetite, and
conditioned taste aversion survive excitotoxic lesions of the thalamic taste area. Behav.
Neurosci. 111, 633–645 (1997).
29. Steiner, J. E. Discussion paper: innate, discriminative human facial expressions to taste and
smell stimulation. Ann. NY Acad. Sci. 237, 229–233 (1974).
30. Small, D. M., Gerber, J. C., Mak, Y. E. & Hummel, T. Differential neural responses evoked by
orthonasal versus retronasal odorant perception in humans. Neuron 47, 593–605 (2005).
31. Ongur, D., Ferry, A. T. & Price, J. L. Architectonic subdivision of the human orbital and
medial prefrontal cortex. J. Comp. Neurol. 460, 425–449 (2003).
32. Rolls, E. T. Taste, olfactory, and food texture processing in the brain, and the control of food
intake. Physiol. Behav. 85, 45–56 (2005).
33. Small, D. M. et al. Experience-dependent neural integration of taste and smell in the human
brain. J. Neurophysiol. 92, 1892–1903 (2004).
34. Gottfried, J. A. & Zald, D. H. On the scent of human olfactory orbitofrontal cortex: meta-
analysis and comparison to non-human primates. Brain Res. Brain Res. Rev. 50, 287–304
(2005).
35. Koza B. J., Cilimi, A., Dolese, M. & Zellner, D. A. Color enhances orthonasal olfactory
intensity and reduces retronasal intensity. Chem. Senses 30, 643–649 (2005).
36. Morrot, G., Brochet, F. & Dubourdieu, D. The color of odors. Brain Lang. 79, 309–320
(2001).
37. Haberly, L. B. Parallel-distributed processing in olfactory cortex: new insights from
morphological and physiological analysis of neuronal circuitry. Chem. Senses 26, 551–576
(2001).
38. Wilson, D. A. & Stevenson, R. J. Olfactory perceptual learning: the critical role of memory in
odor discrimination. Neurosci. Biobehav. 27, 307–328 (2003).
39. Ross, R. S., McGaughy, J. & Eichenbaum, H. Acetylcholine in the orbitofrontal cortex is
necessary for the acquisition of a socially transmitted food preference. Learn. Mem. 12,
302–306 (2005).
40. Linster, C., Maloney, M., Patil, M. & Hasselmo, M. E. Enhanced cholinergic suppression of
previously strengthened synapses enables the formation of self-organized representations
in olfactory cortex. Neurobiol. Learn. Mem. 80, 302–314 (2003).
41. Gietzen, D. W. & Rogers, Q. R. Nutritional homeostasis and indispensable amino acid
sensing: a new solution to an old puzzle. Trends Neurosci. 29, 91–99 (2006).
42. Rolls, E. T., Kringelbach, M. L. & de Araujo, I. E. Different representations of pleasant and
unpleasant odours in the human brain. Eur. J. Neurosci. 18, 695–703 (2003).
43. Pelchat, M. L., Johnson, A., Chan, R., Valdez, J. & Ragland, J. D. Images of desire: food-
craving activation during fMRI. Neuroimage 23, 1486–1493 (2004).
©2006 Nature Publishing Group
INSIGHT REVIEW
44. Volkow, N. D. & Wise, R. A. How can drug addiction help us understand obesity? Nature
Neurosci. 8, 555–560 (2005).
45. Pelchat, M. L. Of human bondage: food craving, obsession, compulsion and addiction.
Physiol. Behav. 76, 347–352 (2002).
46. Carpino, S. et al. Composition and aroma compounds of Ragusano cheese: native pasture
and total mixed rations. J. Dairy Sci. 87, 816–830 (2004).
47. Deibler, K. D. & Delwiche, J. (eds) Handbook of Flavor Characterization: Sensory Analysis,
Chemistry, and Physiology (Marcel Dekker, New York, 2003).
48. Licitra, G. (ed.) Cheese Art 2004. Ragusa: Consorzio Ricerca Filiera Lattiero Casearia. (Ragusa,
Italy, 5 Jun 2004).
49. Duffy, V. B., Cain, W. S. & Ferris, A. M. Measurement of sensitivity to olfactory flavor:
application in a study of aging and dentures. Chem. Senses 24, 671–677 (1999).
50. Corriher, S.O. CookWise (William Morrow, New York, 1997).
51. McGee, H. On Food and Cooking: The Science and Lore of the Kitchen (Scribner, New York,
2004).
52. This, H. Molecular gastronomy. Nature Mater. 4, 5–7 (2005).
53. Mori, K., Nagao, H. & Yoshihara, Y. The olfactory bulb: coding and processing of odor
molecule information. Science 286, 711–715 (1999).
54. Smith, D. V. & Frank, M. E. in Mechanisms of Taste Transduction (eds Simon, S. A. & Roper,
S. D.) 295–338 (CRC Press, Baton Rouge, 1993).
55. Rolls, E. T. The representation of information about faces in the temporal and frontal lobes.
Neuropsychologia [Epub ahead of print] 22 Jun 2006 (doi:10.1016/j.neuropsychologia.200
6.04.019).
56. Lorig, T. S. On the similarity of odor and language perception. Neurosci. Biobehav. Rev. 23,
391–398 (1999).
57. Herz, R. S. The effect of verbal context on olfactory perception. J. Exp. Psychol. Gen. 132,
595–606 (2003).
58. Wrangham, R. & Conklin-Brittain, N. Cooking as a biological trait. Comp. Biochem. Physiol. A
Mol. Integr. Physiol. 136, 35–46 (2003).
Acknowledgements Research in my laboratory is supported by the National
Institutes of Health National Institute for Deafness and Other Communicative
Disorders, and by the Human Brain Project. I thank M. L. Pelchat, E. T. Rolls, D. Small,
L. Bartoshuk, T. Acree, F. Q. Xu and V. Duffy for valuable advice.
Author Information Reprints and permissions information is available at
npg.nature.com/reprintsandpermissions. The author declares no competing
financial interests. Correspondence should be addressed to the author
(gordon.shepherd@yale.edu).
321