3 5 2 | MAY 2001 |VOLUME 2
R E V I E W S
In recent years,there has been an explosion ofinterest
in the neuroscience of human emotion.Several stud-
ies have focused on a debate concerning whether the
neural representation of emotion involves individual
systems for separate emotions,or an integrated sys-
tem able to code all emotions.This relates to a more
general psychological debate concerning whether
emotions are best described in terms of category-
based frameworks or unifying dimensional accounts.
Exponents of the category-based approach posit that
a limited number ofemotions (that generally include
happiness, sadness, anger, fear and disgust) have a
‘basic’ status,and that signals of these basic emotions
are identified by activating discrete category represen-
tations (one for each emotion).The principal support
for this model comes from the finding that these basic
emotions are pan-cultural and that facial expressions
of each emotion are represented by the same distinct
facial musculatures across different cultures1,2.The
main alternative to the category-based account is the
view that all emotions can be represented in a single
unifying framework; for example, a limited set of
dimensions coding specified emotional constructs
such as pleasure,arousal,attention/rejection and so
In the past seven years,human neuropsychology
has begun to make a significant contribution to this
debate on two different fronts.First,the discovery that
certain types ofbrain injury and psychiatric disorder
can cause selective impairments in the recognition of
human signals offear and disgust.Second,functional
imaging research has revealed distinct neural corre-
lates for processing fear and disgust in healthy individ-
uals. In this review, we first address research on
humans showing that the AMYGDALA (FIG.1) has a central
role in processing signals of fear.We then discuss the
less extensive literature showing that a different neural
circuit that primarily involves the insula and the BASAL
GANGLIA serves recognition of disgust signals (FIG.1).
Last,we consider the implications ofthis research for
the way we think about human emotion.
Involvement of the amygdala in fear recognition
The amygdala and fear in non-humans. Animal
research has clearly shown that the amygdala is impor-
tant in emotion3. This was first indicated by the
reduced levels of aggression and fear, and the
increased tameness,observed in monkeys with bilater-
al lesions that included4, or were restricted to5, the
amygdalae.Until recently,however,the specificity of
these effects was unclear, as the ablations made in
these studies also destroyed fibres ofpassage coursing
through the amygdala.So,it is important that more
recent work using fibre-sparing excitotoxic lesions6,7
has confirmed that the amygdala has a significant role
in the processing ofemotions.
NEUROPSYCHOLOGY OF FEAR
Andrew J.Calder*,Andrew D.Lawrence* and Andrew W.Young‡
For over 60 years, ideas about emotion in neuroscience and psychology have been dominated
by a debate on whether emotion can be encompassed within a single, unifying model. In
neuroscience, this approach is epitomized by the limbic system theory and, in psychology, by
dimensional models of emotion. Comparative research has gradually eroded the limbic model,
and some scientists have proposed that certain individual emotions are represented separately in
the brain. Evidence from humans consistent with this approach has recently been obtained by
studies indicating that signals of fear and disgust are processed by distinct neural substrates. We
review this research and its implications for theories of emotion.
A small almond-shaped
structure,comprising 13 nuclei,
buried in the anterior medial
section ofeach temporal lobe.
A group ofinterconnected
subcortical nuclei in the
forebrain and midbrain that
includes the striatum (putamen
and caudate nucleus),globus
ventral tegmental area and
*MRC Cognition and Brain
Sciences Unit,15 Chaucer
Road,Cambridge CB2 2EF,
© 2001 Macmillan Magazines Ltd
NATURE REVIEWS | NEUROSCIENCE
VOLUME 2 |MAY 2001 | 3 5 3
R E V I E W S
various aetiologies13. This study found that, when
analysed as a group,fear processing was most affected in
these nine patients,but not all ofthem showed obvious
impairments for fear. In fact, an earlier study by
Hamann and colleagues42had reported that two of
these patients with complete bilateral damage to the
amygdala showed no significant impairment for any
emotion on Adolphs’s task.In the light ofdifferences
between the patients in this study and patient SM11,12,
Hamann et al.42suggested that a fear-recognition
Following on from these findings,a significant area
ofcomparative research on emotions has focused on the
role of the amygdala in a particular type of emotion
processing — FEAR CONDITIONING8,9.These studies have
shown that lesions ofthe amygdala and related areas
interfere with the acquisition and expression ofvarious
indices ofconditioned fear,including fear-potentiated
startle,freezing and disruption ofongoing behaviour
(that is,conditioned suppression).
Amygdala damage in humans.In line with observations
from comparative research,the most frequently docu-
mented consequence ofamygdala damage in humans is
a change in emotional behaviour,although the effects
are less pronounced than those found in non-human
primates10.In addition,cognitive impairments after
amygdala damage are remarkably limited.Deficits have
been found to affect some aspects offace perception11–18,
particularly recognition offacial expression11–15,17–20,
with less consistent evidence showing impaired learning
ofnew faces16,18,19.Additional studies have shown that
the amygdala is involved in memory for emotional
material21–23.Overall,these human data concur with
non-human primate research in showing that the amyg-
dala is involved in emotion3,5,24and face perception25–27.
These findings are further substantiated by functional
imaging and patient-based research that implicates the
human amygdala in facial-expression processing28–35
and in other aspects ofsocial signalling36,37.
Building on these findings and the role ofthe amyg-
dala in fear conditioning,Adolphs and colleagues11–13
addressed the more specific question of whether the
human amygdala is preferentially involved in processing
facial signals offear.Their first study included a patient
with complete and largely selective bilateral lesions of
the amygdala (patient SM),and six patients with unilat-
eral damage to the left or right amygdala12.Participants
rated examples of six facial expressions of emotion
(happiness,sadness,anger,fear,disgust and surprise)
plus neutral expressions on several emotional scales.In
comparison to controls with or without neurological
damage to areas other than the amygdala,SM showed
abnormal ratings offacial expressions offear and,to a
lesser extent,anger and surprise.By contrast,patients
with damage to the right amygdala showed no signifi-
cant impairments,whereas patients with lesions ofthe
left amygdala showed some evidence ofabnormal per-
formance,but not for fear.However,contrary to these
results obtained for the patients with unilateral amyg-
dala damage,Anderson et al.38have used an adapted
version ofthe task used by Adolphs and colleagues,and
found significant impairments in a larger group of
patients.These patients had suffered right unilateral
anteromedial temporal lobectomies that included the
amygdala,and showed abnormal processing offaces
expressing sadness,fear,disgust and happiness.Patients
with similar damage to the left hemisphere showed no
Adolphs’s rating task has also been used in several
other investigations13,20,39–42,including a larger study that
involved nine patients with amygdala damage due to
Tail of caudate
Head of caudate nucleus
Figure 1 | The human amygdala, basal ganglia and insula.
a | The location of the amygdala and selected nuclei in the
basal ganglia (putamen, caudate nucleus and globus pallidus)
in the human brain. The amygdala is a small almond-shaped
structure buried in the anterior section of the temporal lobe.
The putamen and caudate nucleus (striatum) lie beneath the
cerebral cortex with the putamen positioned adjacent to the
insula. The globus pallidus is positioned on the inner surface of
the putamen adjacent to the thalamus. b | A view of the right
hemisphere of the human brain with the superior temporal lobe
and sections of the frontal and parietal lobes removed to reveal
the insula. In an intact brain, the insula lies at the base of the
lateral (Sylvian) fissure.
A form ofPavlovian (classical)
conditioning in which the
animal learns that an innocuous
stimulus (for example,an
auditory tone — the
conditioned stimulus or CS),
comes to reliably predict the
occurrence ofa noxious
stimulus (for example,foot
shock — the unconditioned
stimulus or US) following their
repeated paired presentation.As
a result ofthis procedure,
presentation ofthe CS alone
elicits conditioned fear
responses previously associated
with the noxious stimulus only.
© 2001 Macmillan Magazines Ltd
3 5 4 | MAY 2001 |VOLUME 2
R E V I E W S
Ofthe two cases investigated by Calder et al.14,DR’s
lesions resulted from a series ofstereotactic operations
for intractable epilepsy,targeted first at the left amygdala
and then subsequently at both left and right amygdalae.
SE’s injury,by contrast,was caused by encephalitis;mag-
netic resonance imaging (MRI) of his brain shows
extensive destruction ofthe right temporal lobe (includ-
ing the amygdala) and a small region ofabnormality in
the left amygdala. Across the two facial-expression
recognition tasks,both DR and SE showed impaired
identification offear,and to a lesser extent anger.DR
also showed some difficulty in recognizing expressions
ofdisgust.These findings complement the findings of
Adolphs et al.12,13by showing that bilateral amygdala
damage affects the recognition offacial expressions of
fear and anger on different types oftask.Reports offur-
ther patients with bilateral amygdala damage have since
shown similar fear-recognition impairments on the
same forced-choice tasks15,19.
impairment might be contingent on acquiring amyg-
dala damage early in life.However,subsequent evi-
dence of impaired fear recognition from patients that
do not satisfy this criterion has made this explanation
unlikely14,15,19.The puzzle of the contradictory results
obtained by Hamann et al.42has recently been
resolved, however, by Schmolck and Squire41, who
found that a different method of analysing these
patient ratings revealed impaired performance.
Impairments were also found on two additional tests
of facial-expression recognition,particularly for fear
and sadness. One of these additional tasks used a
forced-choice labelling procedure, a format used
widely in facial-expression research43.This particular
forced-choice task was one of two used originally by
Calder et al.14with two additional patients (DR and
SE) with bilateral amygdala lesions;the second task
involved morphed facial expressions and is illustrated
Box 1 | Unifying accounts of emotion
Dimensional models in psychology
Dimensional accounts ofemotion were borne out ofthe observation that human errors in recognizing facial expressions
are not random,but instead form consistent,replicable patterns that can be accommodated by a model in which facial
expressions are recognized by registering their positions in a continuous two-dimensional space124.This approach has
survived to the present day,its most recent variant being Russell’s circumplex model111,112,a two-dimensional system
coding pleasure–displeasure and arousal–sleepiness.
Several researchers have shown that this type oftwo-dimensional model can be applied to the recognition ofemotion
from multiple modalities111,125–127,and even to emotional experience112.The approach therefore reflects a multi-modal level
ofprocessing that can be accessed by all ofthe above modalities.
Dimensional models in neuroscience
A good example is Rolls’ theory ofemotion119,120.In the tradition ofanimal learning research,Rolls defines emotions in
terms ofstates elicited by positive and negative instrumental reinforcers.Positive reinforcers are rewarding,whereas
negative reinforcers are associated with punishment.The vertical axis ofthe graphic summarises the emotions elicited
by the presentation ofa reward (S+) or punishment (S–),whereas the horizontal axis shows emotions produced by the
termination (or omission) ofa reward (S+or S+!)1or punishment (S–or S–!)1.Moving away from the midpoint ofeach
axis,the intensity ofthe emotion increases.A distinction is made between the emotions elicited where an active or
passive behavioural response is possible.For example,on termination (or omission) ofa positive reinforcer,anger
might follow for an active response,but sadness/griefmight follow where only a passive response is possible.Rolls119,120
discusses experimental evidence that links the reward value (positive or negative) ofreinforcers to the orbitofrontal
cortex and amygdala.
The circumplex model — Russell (1980)
Theory of emotion — Rolls (1999)
S+ or S+!
S– or S–!
© 2001 Macmillan Magazines Ltd
NATURE REVIEWS | NEUROSCIENCE
VOLUME 2 |MAY 2001 | 3 5 5
R E V I E W S
Functional imaging studies of fear recognition. The
aforementioned patient-based studies clearly show
impaired recognition offacial signals ofemotion after
amygdala damage; the most consistent and severe
impairments are seen for fear.Given these observations,
it is of considerable interest that functional imaging
research has found increased activity in the amygdala for
tasks in which participants view facial expressions of
fear,relative to control conditions in which the faces con-
vey happiness28,29,33,34,44,disgust30,31,anger45or no emo-
tion (neutral)30–33,45,46.The location ofthe maximally
activated VOXELSin the left and right amygdala found in
these studies is summarized in FIG.3.This figure also
shows the maximally activated voxels from several func-
tional imaging studies investigating fear conditioning
(see also BOX 2). Functional imaging research has
revealed important insights into the nature ofthe amyg-
dala response to fear stimuli.Four key findings are high-
Time course ofamygdala response.There is debate in
the literature as to the temporal properties of amyg-
dala responses to fear-related stimuli47. So, it is of
interest that several studies,beginning with Breiter et
al.33, have shown that the response of the human
amygdala to fearful facial expressions diminishes with
repeated presentations (that is,the response habitu-
ates)33–35,44. Phillips et al.35and Wright et al.44have
addressed this observation in more detail by showing
that the habituation rate to facial expressions offear is
more rapid for the right than for the left amygdala.
This might explain the larger number of significant
left-hemisphere amygdala signals reported in studies
using fearful facial expressions (FIG.3).In addition,
Phillips et al.35found that the haemodynamic
response to neutral expressions increased over the
course ofthe experiment in the right,but not the left,
amygdala. The authors interpreted this finding as
reflecting conditioning of the neutral stimuli to the
aversive fear expressions in the right amygdala (see
BOX 2for a summary of fear-conditioning research
made in humans).
Non-conscious processing offacial expressions.Öhman
and colleagues have investigated non-conscious pro-
cessing of fear-conditioned facial expressions using a
method ofpresentation in which faces are shown very
briefly,followed by an obscuring (neutral facial expres-
sion) mask48. Although participants are unable to
identify the facial expression consciously,an emotional
response is elicited.Using the same method ofpresen-
tation,Morris et al.49have shown increased relative
regional cerebral blood flow (rCBF) in the amygdala to
fear-conditioned facial expressions ofanger (relative to
unconditioned angry expressions) for both masked
(non-conscious) and unmasked (conscious) presen-
tation formats.In related work,Whalen et al.34have
shown that briefly presented masked facial presentations
of fearful, but not happy, facial expressions increase
amygdala activation relative to a neutral fixation-cross
condition (see also REF. 50). Together with previous
It is important to note that despite their problems in
recognizing fear,the patients discussed above can provide
plausible situations in which a person might experience
fear.In other words,their deficits do not seem to arise
from impaired understanding ofthe concept offear.
Figure 2 | The Emotion Hexagon test of facial-expression recognition14. The six rows of this
illustration contain morphed (blended) continua ranging between the following six expression
pairs. From top to bottom, the continua shown in each row are happiness–surprise (top row),
surprise–fear (second row), fear–sadness (third row), sadness–disgust (fourth row), disgust–anger
(fifth row), anger–happiness (bottom row). Going from left to right, the columns show 90%, 70%,
50%, 30% and 10% morphs along each continuum. For example, from left to right, the top row of
images contain the following percentages of the happy and surprised expressions: 90%
happy–10% surprise, and then 70%–30%, 50%–50%, 30%–70% and 10%–90% of the same
two expressions. Data from neurologically intact participants show that stimuli that contain 90%
and 70% of an expression are consistently identified as the intended emotion14,82,138,139.
Volume element.The smallest
ofa three-dimensional space.
© 2001 Macmillan Magazines Ltd
3 5 6 | MAY 2001 |VOLUME 2
R E V I E W S
observed when subjects view pictures ofaversive visual
Functional imaging research has also highlighted a
neuromodulatory role for the amygdala in memory57,58.
For example,Hamann et al.57have shown that recogni-
tion memory for both aversive and pleasant emotional
scenes was positively correlated with rCBF in certain
areas of the amygdala,hippocampus and parahippo-
campal gyrus at encoding.Moreover,for both positive
and negative emotional scenes,rCBF in the identified
amygdala areas was positively correlated with rCBF in
the identified areas of hippocampus and parahippo-
campal gyrus, two areas known to be involved in
memory.Amygdala involvement in processing positive
affect is intriguing and we return to this issue at the
end ofthis section.
Opposite effects offear and happiness.Some studies have
found that,whereas fearful facial expressions produce
an increase in amygdala activity,significant decreases in
amygdala activity are produced by happy facial expres-
sions28,29,34(but see REF.33).The significance of these
findings is currently unclear,especially given the inherent
difficulties in interpreting relative deactivations in neu-
roimaging studies.However,it is nonetheless interesting
that levels offear induced by the local anaesthetic pro-
caine are positively correlated with left-amygdala rCBF,
whereas levels ofprocaine-induced euphoria are nega-
tively correlated with left-amygdala rCBF59.Note that the
emotional effect ofprocaine differs between individuals.
In relation to the neuromodulatory role ofthe left amyg-
dala discussed earlier,the procaine study also found that
levels offear were positively correlated with rCBF in the
medial occipital cortex, whereas levels of euphoria
showed a negative correlation with this brain region.
Together with the observations summarized in the
previous section,these findings indicate that the extras-
triate cortex is modulated by an amygdala function that
is influenced by emotions (that is,fear and happiness)
perceived in others,and by emotional states (that is,fear
and euphoria) experienced by the self.At a psychologi-
cal level,these data indicate that seeing someone who is
afraid (or happy) might have a similar effect on our level
ofvigilance towards potential threat ofexperiencing fear
(or happiness) oneself.Clearly,this parallel between per-
ception and experience merits further investigation.
Amygdala and non-facial signals ofemotion.In general,
functional imaging studies complement the observa-
tions ofthe patient-based research by showing dispro-
portionate amygdala involvement in processing facial
signals offear.In addition,they go substantially further
by identifying functions of the intact amygdala in
healthy individuals.An important issue,however,con-
cerns whether the role ofthe amygdala is restricted to
facial signals ofemotion (in particular,fear),or whether
it is involved in processing signals of emotion from
other sensory modalities.Anatomical studies ofnon-
human primates show that the amygdala receives inputs
from multiple sensory modalities60. So, one might
expect that the human amygdala would be involved in
research28–33,these findings highlight the involvement of
the amygdala in both conscious and non-conscious pro-
cessing offear-relevant information8,9,48.
A neuromodulatory role for the amygdala.Morris and
colleagues28,29reported that the emotional intensity of
fearful facial expressions was positively correlated with
changes in rCBF in the left amygdala,whereas the inten-
sity ofhappy expressions was negatively correlated with
rCBF in this area. In addition, regression analyses
showed that the rCBF in an area ofextrastriate cortex
showed a significant positive correlation with left-amyg-
dala rCBF for fear expressions, and a strong trend
towards a negative correlation with left-amygdala rCBF
for happy expressions.These findings were interpreted
as support for the idea that efferent connections from
the amygdala have a context-specific role in modulating
extrastriate cortical function51.A plausible psychological
interpretation ofthese data is that following the presen-
tation of a fearful facial expression, our vigilance is
enhanced by amygdala function,which increases the
sensitivity ofearly visual processing.This interpretation
concurs with the finding that phobic reactions induced
in people with phobias to spiders and snakes also cause
increased metabolic activity in the extrastriate cortex52,53.
Similarly, amygdala and extrastriate activation are
Box 2 | The human amygdala and conditioned fear
There have now been numerous investigations offear conditioning in humans using
non-invasive imaging techniques such as positron-emission tomography (PET) and
functional magnetic resonance imaging (fMRI).Early PET studies showed a remarkable
lack ofamygdala activation during fear conditioning,possibly due to the poor temporal
resolution ofthe technique,resulting in an inability to disentangle fear conditioning and
expression from extinction processes.Notable exceptions were a study128showing a
positive correlation between conditioned fear,as indexed by heightened electrodermal
activity (EDA) and relative regional cerebral blood flow (rCBF) in the right amygdala,
and studies by Dolan and colleagues129showing amygdala activation using a covert fear-
learning paradigm.In fact,some blocked-presentation studies have shown decreased
relative rCBF in the amygdala129,130,although this might relate to the inability to
adequately separate extinction from acquisition and expression offear.
The development ofevent-related fMRI has allowed investigators to study evoked
haemodynamic activity to single stimulus categories,and to separate responses related to
acquisition,expression and extinction ofconditioned fear.Three ofthese studies131–133
have been successful in showing amygdala involvement in conditioned fear.Ofnote,these
studies showed enhanced amygdala activation during initial acquisition and extinction of
fear,although how such findings relate to current debates on amygdala involvement in
the acquisition and storage ofconditioned fear in other species47remains unclear.
Although functional imaging studies have shown that the amygdala is involved in fear
conditioning in humans,such studies cannot demonstrate that it is necessary.Such
evidence has been provided in studies ofpatients with lesions involving the amygdala134,
who show impaired fear conditioning,as indexed by changes in EDA (see also REF.135).Of
note,damage to the amygdala does not impair explicit memories ofthe fear-conditioning
protocol,only the affective consequences ofsuch conditioning.
The neuroimaging studies are particularly interesting in showing extra-amygdaloid
structures involved in fear conditioning.In particular,midbrain/hypothalamic regions
and the anterior cingulate cortex are consistently activated129,130,136.These regions have
been implicated in studies offear processing in several animal species137,but their roles in
human fear are not fully understood.One concept that might help in this regard is the
idea ofa hierarchical external defencesystem,in which progressively higher-level
components provide increasingly sophisticated and flexible solutions to the problem of
© 2001 Macmillan Magazines Ltd
NATURE REVIEWS | NEUROSCIENCE
VOLUME 2 |MAY 2001 | 3 5 7
R E V I E W S
ofvocal signals other than fear (that is,disgust and sur-
prise)61.Hence,at the very least,the results from the
analysis ofSP are consistent with amygdala involvement
in processing emotional signals from both visual and
So,two patients with bilateral amygdala damage
(SM and RH) show no obvious impairments in rec-
ognizing vocal signals of emotion,whereas two other
patients (DR and NM) with similar damage do.This
paradox requires explanation.In light of research on
the role of the basal ganglia in the interpretation of
prosody63,some researchers have suggested that the
impairment of DR in processing vocal cues might be
caused by her limited extra-amygdaloid damage in
this area40,61.Similarly,NM has additional damage to
the thalamus and internal capsule.However,given the
manner in which these patients’ emotional deficits
are mirrored across the visual and auditory process-
ing domains,it seems unlikely that this could arise
from damage to separate underlying systems. So,
although we agree that extra-amygdala damage might
be a necessary prerequisite of cross-modal impair-
ments,the current patient data nonetheless concur
with animal research showing that the amygdala is
involved in coding fear cues from different domains
and sensory modalities64.
Unfortunately,few functional imaging studies have
used fear cues other than facial expressions.One of
these studies has shown considerable overlap in the
neural correlates ofprocessing facial and vocal signals of
fear,which include the amygdala31.However,a second
study that compared vocal signals offear,sadness and
happiness to a neutral condition,found an interaction
between increased rCBF in the anterior insula and
decreased rCBF in the right amygdala that was specific
to vocal expressions offear65.The reason for the discrep-
ancy between these two studies is unclear,but it might
relate to the different control conditions — mild happi-
ness31and neutral ‘voiced nasals’65.A third relevant
study to the cross-modal debate used printed words
presented in the form ofan EMOTIONAL STROOP TASK66.The
results showed that threat-related words (for example,
assault,abuse,torture) resulted in increased rCBF in
both amygdalae,and longer naming times relative to a
neutral word condition (but see REF.67).Last,it is worth
noting that pictures ofaversive visual scenes (for exam-
ple,human violence,frightening animals and so on)
also engage the amygdala54,57.
Amygdala and the conscious experience offear.An
obvious question is whether the system involved in
recognizing fear in others also contributes to the
experience of this emotion. At present, there is a
paucity of empirical data on this issue;however,at
least three pieces of evidence indicate that the answer
to this question might be affirmative.First,patient
NM, whose case was discussed earlier, showed an
abnormal score on a self-assessment questionnaire
tapping the experience of fear15but normal scores for
similar questionnaires assessing his experience of
anger and disgust.Second,Halgren68has pointed out
coding fear cues regardless ofmodality.Nonetheless,
some have argued that the amygdala is not essential for
recognizing emotion from vocal signals,or more specif-
Strong support for the cross-modal hypothesis
comes from two bilateral amygdala patients — DR14,62
and NM15— that were tested with largely the same test
battery.DR showed impaired recognition offear and
anger from both facial and vocal cues14,62,whereas NM’s
deficits with facial,vocal and body posture cues were
predominantly restricted to fear15.In other words,both
patients’ impairments in the visual domain were mir-
rored by their impairments for processing auditory sig-
By contrast,Adolphs et al.40found that SM and a sec-
ond bilateral amygdala patient (RH) showed no signifi-
cant deficit on a vocal variant (emotional prosody) of
Adolphs’s facial-expression-rating task,despite both
showing impairments on the original facial-expression
task11,40.However,one method ofanalysing the vocal
data indicated that RH actually experienced limited dif-
ficulty with vocal cues offear and sadness40.In related
work,Anderson and colleagues20,61have reported intact
recognition ofvocal but not facial signals offear in an
additional patient with bilateral amygdala damage (SP).
However,the results obtained with SP need to be inter-
preted in relation to this patient’s impaired recognition
Figure 3 | Functional imaging studies of fear recognition. A summary of the maximally
activated voxels reported as involving the amygdala from functional imaging studies investigating
(i) fearful facial expression processing (green squares)29-31,33,34,44,46,89, and (ii) conditioned fear (red
circles)49,131–133,140,141. Different analysis conventions standardize to the brain in the Talairach atlas
or the Montreal Neurological Institute (MNI) template. Hence, for display purposes all Talairach
coordinates were converted to MNI space142and are displayed on a brain standardized to the
MNI template. The xand ycoordinates of the maximally significant voxel cluster of each contrast
of these studies are plotted on one of two axial slices (z= –14 and z= –24); these axial slices were
chosen because the zcoordinates from all studies were proximal to one or other slice (mean
deviation = 2.04 mm, standard deviation = 1.69 mm). Note that there is a tendency for facial
expressions to engage the left amygdala, whereas conditioned fear seems to produce more
bilateral activation. It also seems that fearful facial expressions are associated with more dorsal
amygdala activity, whereas conditioned fear shows ventral and dorsal involvement. In fact, a
Mann–Whitney comparison of the zcoordinates of all maxima (converted to the MNI template)
showed a significant difference between the conditioned fear and facial expression studies; U =
54.0, z = –2.23, p<0.025; no significant effects were found for the xand ycoordinates (p>0.2).
These images were prepared using MRIcro143.
EMOTIONAL STROOP TASK
A task in which participants are
asked to name the colour ofthe
font in which neutral and threat
words are printed.Results show
that colour-naming times for the
threat words are slower than for
neutral words.The generally
accepted interpretation is that
emotional words involuntarily
capture attention,distracting the
participants from naming.
© 2001 Macmillan Magazines Ltd
3 5 8 | MAY 2001 |VOLUME 2
R E V I E W S
authors13,14,34,61,62have suggested that these findings
might relate to the posited role ofthe amygdala in the
detection,evaluation and coordination ofresponse to
signals ofdanger in the environment8,9.At present,the
issue ofwhether the amygdala is also involved in coding
fear cues from other sensory modalities (for example,
vocal expressions) is less clear-cut,although we feel that,
on balance, the current evidence favours the cross-
modal hypothesis.However,there is clear need for more
research investigating the cross-modal issue and,given
the rarity ofpatients with selective amygdala lesions,
additional functional imaging studies would be a good
At the present time,it is worth pointing out some
cautionary notes.First,Rapcsak et al.69found that in a
group ofpatients with various focal lesions,the recog-
nition offearful facial expressions was disproportion-
ately impaired relative to other emotions,regardless of
whether the damage included the amygdala or not.
They attributed these findings to the fact that neuro-
logically intact controls find fear more difficult to rec-
ognize than other emotions,and that the marked fear
impairments were simply an effect of the difficulty
level.We agree that marked fear-recognition impair-
ments in the context ofa general emotion-recognition
impairment can arise from this kind ofeffect,and con-
sequently should be interpreted with caution.
However,it is important to note that the patients stud-
ied by Rapcsak et al.69differ from the bilateral amyg-
dala patients in several important aspects.First,none
ofthese patients was screened for basic visual defects.
Second,some ofthe cases that we have discussed show
highly circumscribed deficits affecting just fear15,or
fear and only one,or two,other emotions12,14,41,where-
as participants studied by Rapcsak et al.showed gener-
alized deficits in which fear was the worst affected
emotion. Third, the cases we have discussed show
replicable fear impairments that are evident on differ-
ent tests offacial-expression recognition11–15,18–20,41,and
in some cases,vocal-expression recognition15,62.The
presence of a consistent fear deficit across different
tests and different modalities makes it unlikely that an
artefact related to task difficulty can explain the overall
Second, although facial-expression research has
indicated that the amygdala is particularly involved in
coding facial signals offear,functional imaging studies
have demonstrated amygdala activation in response to
positive stimuli such as pleasant pictures and pleasant
tastes57,70.Similarly,there is currently a debate within
the comparative literature concerning whether,in addi-
tion to aversive (fear) conditioning,the amygdala is also
involved in appetitive conditioning71–74.A discussion of
this literature goes well beyond the remit ofthis article,
and at present,there are few human data to address this
debate.Nonetheless,these findings are intriguing and
might provide important clues concerning the nature
ofthe relationship between the amygdala and fear.One
possibility suggested in the literature is that different
subnuclei ofthe amygdala might have different roles in
processing positive and negative emotions,for example.
that electrical stimulation of the amygdala in human
subjects undergoing surgery for epilepsy can induce
various reactions,but when an emotion is reported it
is invariably fear.Last,anecdotal reports indicate that
bilateral amygdala lesions can disrupt the normal
response to fear-provoking situations.However,this is
not always expressed as a reduction,or even abolition
of the response16,19,implying that the amygdala is not
simply a fear generator.For example,the husband ofa
woman with bilateral amygdala damage described
how his wife seemed to misinterpret the actions of
youths who tried to mug him as ‘larking around’,
whereas she became terrified on another occasion by a
mildly aggressive exchange between characters in a
television drama19. The same patient also showed
inappropriate fear reactions to her carer and to close
members of her family (see also REF.16).These find-
ings might relate to research in monkeys,which posits
dissociable types of unconditioned fear responses for
the amygdala (acute fear responses such as reaction to
a snake) and orbitofrontal cortex (trait-like
anxiety/fear responses such as reactions to humans7).
A potentially promising line of enquiry concerns the
extent to which the human amygdala is involved in
the appraisal ofpotential threat.
Summary.The research that we have reviewed so far
provides strong support for the disproportionate
involvement ofthe amygdala in processing facial signals
offear and in fear conditioning.Consequently,several
Figure 4 | Functional imaging studies of disgust recognition. A summary of the maximally
activated voxels reported as involving the insula or basal ganglia from functional imaging studies
investigating the processing of facial expressions of disgust30,31,89. Different analysis conventions
standardize to the brain in the Talairach atlas or the Montreal Neurological Institute (MNI) template.
Hence, for display purposes all Talairach coordinates were converted to MNI space141and are
displayed on a brain standardized to the MNI template. The xand ycoordinates of the maximally
significant voxel cluster of each contrast of these studies are plotted on one of two axial slices
(z = 0 and z = 10); these axial slices were chosen because the z coordinates from all studies
were proximal to one or other slice (mean deviation = 2.75 mm, standard deviation = 2.08 mm).
Blue squares indicate maxima reported as insula, red squares indicate maxima reported as
basal ganglia. Open squares represent disgust minus neutral contrasts, filled circles represent
disgust minus fear contrasts. The figure shows that the basal ganglia signals are largely confined
to the right hemisphere, whereas the insula signals are more evenly distributed between the two
hemispheres. These images were prepared using MRIcro143.
© 2001 Macmillan Magazines Ltd
NATURE REVIEWS | NEUROSCIENCE
VOLUME 2 |MAY 2001 | 3 5 9
R E V I E W S
made significantly more errors in recognizing facial
expressions of disgust than either control group.
Furthermore,this finding held for participants from the
AR+group who showed no overt symptoms ofthe dis-
ease.So, the HD mutation was in itself sufficient to
cause a disgust-recognition impairment in the absence
Other disorders that affect disgust recognition.The ini-
tial stages ofHD principally affect the basal ganglia.It is
therefore particularly informative that two psychiatric
disorders associated with abnormal metabolic activity
in this brain system85–87— OBSESSIVE-COMPULSIVE DISORDER
(OCD) and TOURETTE’S SYNDROME — have also been
shown to produce marked impairments in recognizing
facial expressions of disgust, and to a lesser extent
anger88.It is important to stress,however,that people
with Tourette’s syndrome that participated in this study
only showed impaired recognition of disgust if they
showed co-morbid obsessive-compulsive symptoms.
So,the presence ofOCD seemed to be a defining feature
ofthe disgust deficit.
We discussed earlier that the most effective means of
showing that impaired recognition offear is not simply
attributable to effects of the level of difficulty is to
demonstrate that damage to a separate neural region
causes a disproportionate impairment in recognizing a
second emotion as compared with fear.Clearly,the fear
and disgust impairments we have described have
important complementary functions in strengthening
the case that each constitutes a genuine emotion-specif-
ic impairment.It is important to remember,however,
that HD,OCD and Tourette’s syndrome are not charac-
terized by focal neuropathology. So, although the
patient-based studies are consistent with the idea that
disgust and fear recognition are served by separate
neural structures,the actual brain regions involved for
disgust are unclear.In this regard,functional imaging
research has been particularly informative.
Insula involvement in recognizing disgust.In contrast
to functional imaging studies using fearful facial
expressions,the amygdala has rarely been reported as
a neural correlate for the processing offacial signals of
disgust.Instead,all of the functional imaging studies
that have addressed the neural correlates of viewing
facial expressions of disgust30–32,89have pinpointed
two areas — the insula and basal ganglia nuclei (most
consistently,the putamen/pallidum) (FIG.4).Insula
involvement is particularly interesting given its identi-
fied role in gustatory function90,91.This role is illus-
trated by the observation made by Penfield and
Faulk92that electrical stimulation ofthe insula ofcon-
scious human patients undergoing surgery produced
sensations ofnausea,unpleasant tastes and sensations
in the stomach92.It is also of interest that lesions of
the insula or pallidum in rats have been shown to
interfere with CONDITIONED TASTE AVERSION93,94.Together,
these findings concur with the proposal of Rozin and
colleagues79,95that disgust has developed from a more
primitive system involved in distaste.
However,this seems unlikely given that cell recording
in non-human primate amygdala has shown that the
same cells can respond to both positive and negative
stimuli75.Clearly,this is an important topic for future
study.Finally,there is also evidence from functional
imaging research to indicate that the amygdala might
be involved in coding facial expressions of sadness76.
However,this effect has not been observed in two addi-
As noted,some authors have preferred to interpret
fear impairments as level-of-difficulty effects rather
than emotion-specific impairments.Clearly,the most
convincing way to refute this argument is to show that
other types ofbrain injury can affect a different emotion
while leaving fear intact.It is ofparticular interest,then,
that recent studies have demonstrated disproportionate
impairments in the recognition offacial expressions of
disgust relative to fear.
The neural substrate of disgust
Rozin and his colleagues79,80have proposed that disgust
evolved from the phylogenetically more primitive sen-
sation ofdistaste.As such,non-human studies ofcon-
ditioned taste aversion provide an interesting animal
model of some aspects of this emotion81. However,
Rozin and colleagues also suggest that the full concept
ofdisgust is essentially unique to the humans.It can be
a morally as well as a physically based emotion.For
example, many people would not like the idea of
putting on Adolf Hitler’s jacket.In this example,the
distaste involves moral rather than physical repug-
nance.From this perspective,human studies ofdisgust
are not only informative about the neural representa-
tion ofthis emotion,they can also offer a perspective
unattainable in animal studies.
Disgust recognition in Huntington’s disease.Evidence
that disgust might be associated with a particular neural
substrate came first from an investigation ofpeople with
manifest Huntington’s disease (HD)82,an autosomal-
dominant neurogenetic disorder that in its early stages
particularly affects a region ofthe basal ganglia known
as the striatum (FIG.1).Participants in this study were
shown the same facial-expression identification tests
that were used with several of the bilateral amygdala
patients14,15,19(FIG. 2). The patients with HD showed
problems in recognizing several expressions,but a dis-
proportionately severe impairment was found for facial
Further evidence that HD particularly affects the
recognition offacial signals ofdisgust has been shown
in two additional studies83,84.The first provided detailed
case studies oftwo patients.The second was an investi-
gation offace processing (including facial-expression
recognition) in people at risk ofcarrying the mutation
associated to HD84.Participants that were subsequently
identified as gene carriers (AR+) were compared with
each oftwo control groups — participants that did not
carry the gene (AR–) and neurologically intact controls.
A comparison ofthe scores revealed just one significant
difference for recognition ofemotion — the AR+group
A psychological disorder in
which the person is burdened by
recurrent,persistent thoughts or
ideas,and/or feels compelled to
carry out a repetitive,ritualized
behaviour.Anxiety is increased
by attempts to resist the
compulsion and is relieved by
giving way to it.
A rare genetic disorder,
characterized by facial and vocal
tics,and less frequently by verbal
A form of memory in which a
taste is associated with digestive
malaise,leading to avoidance of
the taste in subsequent
presentations.This form of
memory depends on the
integrity of the insula.
© 2001 Macmillan Magazines Ltd
3 6 0 | MAY 2001 |VOLUME 2
R E V I E W S
the basal ganglia dysfunction discussed earlier86,87,meta-
bolic activity in the insula is correlated with scores in the
Yale–Brown obsessive-compulsive scale (Y–BOCS)105,a
recognized measure ofOCD severity106.In addition,the
first two factors extracted from a factor analysis ofscores
on this scale were found to be correlated with rCBF in
the striatum and parietal cortex (factor 1:checking com-
pulsions,and religious,aggressive and sexual obsessions),
and the striatum and temporoinsular cortex (factor 2:
symmetry and ordering symptoms)105.It is also ofinter-
est that OCD symptoms represented by these factors are
frequently seen in Tourette’s syndrome with co-morbid
OCD107–109,particularly a liking for symmetry and order.
This might be important because patients with Tourette’s
syndrome in the study mentioned above88only showed a
disgust impairment ifthey showed OCD symptoms.In
this regard,it is also relevant that OCD is a recognized
psychiatric correlate ofHD.
Summary.Overall,studies on patients and functional
imaging research show a link between the recognition of
facial expressions ofdisgust and the insula–basal ganglia
regions.Whether these same brain areas underlie the
coding ofdisgust signals from multiple sensory modali-
ties is only beginning to be addressed.However,the cur-
rent data seem largely consistent with a cross-modal
hypothesis.The contribution of these regions to the
experience ofdisgust is also underinvestigated at pre-
sent,but the limited evidence indicates that there might
be a significant link82,96,98.As with the fear impairments
we have discussed, it is important to point out that
patients with disgust-recognition impairments are able
to provide examples ofplausible situations in which a
person might feel disgusted and do not show impaired
knowledge ofthe concept ofdisgust.
Implications for theories of emotion
We chose to focus on the neuropsychology ofboth fear
and disgust recognition for two reasons.First,remark-
ably similar pictures seem to be emerging for both emo-
tions.Second,fear and disgust provide a striking double
dissociation that substantially strengthens the theoreti-
cal impact ofeither area alone.As discussed,patient-
based research has shown that damage to the amygdala
causes a disproportionate impairment in recognizing
facial signals offear11–15,19,20,41,whereas abnormalities of
the insula–basal ganglia regions primarily affect the
recognition of facial signals of disgust82–84,88,96.
Functional imaging has substantiated this dissocia-
tion28–34,44–46,89and,in the case offear research in particu-
lar,it has identified important insights into the underly-
ing neural mechanisms. However, although there is
general agreement that human amygdala is involved in
recognizing fear from facial signals,its role in recogniz-
ing vocal signals of this emotion is still being debat-
ed15,40,61,62.On the basis ofthe current evidence,however,
we feel that the data for both fear and disgust favour the
cross-modal hypothesis;a position that concurs with
the limited human data indicating that the amygdala
and insula–basal ganglia might also contribute to the
experience offear and disgust,respectively.
Is there a cross-modal system for disgust?As is the case
with impairments in recognizing fear, an interesting
question is whether impaired recognition of facial
expressions ofdisgust reflects a more general deficit that
affects recognition ofthis emotion in multiple modalities
(for example,vocal expressions).A test ofvocal-expres-
sion recognition included in the study ofpatients with
HD82discussed earlier showed some evidence ofa cross-
modal impairment for disgust.However,more solid evi-
dence of this phenomenon has come recently from
patient NK,a person with a focal lesion ofthe brain areas
that have identified by functional imaging as the neural
correlates ofprocessing facial expressions ofdisgust —
the insula and basal ganglia96.The damage present in
NK is lateralized to the left hemisphere and includes the
insula,putamen,internal capsule,globus pallidus and,to
a lesser extent,the head ofthe caudate.On tests offacial-
expression recognition and vocal-expression recognition,
NK showed a highly selective deficit for disgust in the
context ofpredominantly preserved recognition ofother
emotions.NK also showed abnormal performance on a
questionnaire tapping his experience ofdisgust,whereas
his scores for comparable questionnaires assessing his
experience offear and anger were normal.These results
are consistent with damage to a system involved in recog-
nizing signals of disgust from both facial and vocal
modalities.In addition,they indicate that this system
might be significant in the experience ofthis emotion.It
is worth emphasizing that the performance ofNK shows
a marked double dissociation with the performance of
patients with amygdala damage, particularly patient
NM15,who was tested with a similar test battery.
To our knowledge,very few functional imaging stud-
ies have addressed whether a single system might under-
lie the recognition ofdisgust in different domains or
sensory modalities.In the most relevant study,Phillips et
al.31compared activation after viewing facial signals of
disgust to activity after listening to vocal signals ofthe
same emotions.As in previous work,facial signals of
disgust engaged areas of the insula and striatum.By
contrast,vocal cues produced significant signals in sev-
eral areas,including the superior temporal regions,thal-
amus,and rostral and dorsolateral prefrontal areas.A
second relevant study did not directly compare signals
ofdisgust from different modalities.However,it showed
that pictures ofdisgusting scenes (for example,decaying
food,cockroaches and so on) produce a similar pattern
ofinsula activity to that observed for facial expressions
ofdisgust97.A third imaging study showed that recall of
events associated with the experience ofguilt increased
insula activity relative to the recall ofneutral events98.
This is interesting,considering that guilt has been char-
acterized as disgust directed towards the self99.
Insula–basal ganglia in HD,OCD and Tourette’s.Studies
ofneuroimaging and case NK provide direct support for
the involvement ofthe insula and basal ganglia in disgust
recognition.It is worth emphasizing that these structures
are highly interconnected100,and that both regions have
been implicated in HD101–104.With regard to OCD,func-
tional imaging research has shown that,in addition to
© 2001 Macmillan Magazines Ltd
NATURE REVIEWS | NEUROSCIENCE
VOLUME 2 |MAY 2001 | 3 6 1
R E V I E W S
preferred states or ‘attractors’in a state-space defined by
N(where N> 2) higher-order dimensions115,116could
potentially account for some ofthe deficits reported
here.It is unclear,however,whether the nature ofsuch
abstract dimensions could be discovered on the basis of
purely descriptive analyses of,for example,self-report
data.Indeed,the thrust ofan attractor-state account is
to bridge the theoretical gap betweenthe between cate-
gory-based and multidimensional frameworks — such
a system will show properties ofboth.
We suggest that,in the future,it will be more useful
for emotion research to focus on causal mechanisms,
rather than descriptive taxonomies.This approach is
characteristic of‘basic’ emotion accounts117,118,which
posit a variable number ofdiscrete emotions that differ
from one another in several fundamental ways.
However,the same approach is also evident in other
emotional theories,such as those provided by Rolls119,120
(BOX 1),Gray121and Davidson122,123,which relate broad
classes of emotions to underlying reinforcement or
motivational systems.However,it is unclear whether
these models would predict the patterns of deficit
We would argue that an approach seeking to dis-
cover evidence for dissociable emotion systems after
brain injury might provide a useful model in which to
understand the structure ofemotion,both in terms of
proximate mechanisms (for example,neural struc-
tures) and ultimate causation (for example,adaptive
significance).So far,this research has described dis-
tinct neural pathways that underlie the processing of
signals of fear and disgust in humans.This dissocia-
tion can be related to the adaptive significance of
these emotions as responses to critical forms ofthreat
that are associated with external (fear) and internal
(disgust) defence systems81.
In a recent article,LeDoux9highlighted some ofthe
attributes ofthe comparative neuroscience approach
that,he believes,have facilitated this area’s substantial
contribution to our understanding ofemotion in recent
years.These included “… focusing on a psychologically
well-defined aspect ofemotion,… using an experimen-
tal approach to emotion that simplified the problem in
such a way as to make it tractable,… circumventing
vague and poorly defined aspects ofemotion,and …
removing subjective experience as a roadblock to exper-
imentation”.Similarly,the human neuropsychological
studies that we have reviewed here share several ofthese
attributes by focusing on what some maintain to be the
bedrock of human affect — the basic emotions. By
adopting this general approach,the neuropsychology of
emotion can be as useful in dissecting and understand-
ing the emotion system as cognitive neuropsychology
has been in understanding the cognitive system.
It is important to clarify that we are not ofthe opin-
ion that fear and disgust are represented by entirely dis-
tinct neural circuits.Nor do we wish to suggest that the
amygdala and insula–basal ganglia regions are simply
fear and disgust generators.Our conclusions seem war-
ranted by several lines of evidence.First, functional
imaging research shows that facial expressions offear,
disgust and anger all engage similar areas ofinferior pre-
frontal cortex32,76,78. Second, research on patients110
shows that damage to this and the surrounding areas
impairs recognition ofthese emotions.Third,work in
both humans and monkeys shows that bilateral amyg-
dala lesions do not abolish all fear reactions7,19.However,
we think that the studies included in this review do indi-
cate that the neural mechanisms underlying fear and
disgust might be separate in part,and this clearly has
implications for the representation ofemotion.
So,to return to our original question,in what way
have these studies ofthe neuroscience and neuropsy-
chology ofdisgust and fear contributed to our under-
standing of human emotion? As we discussed in the
introduction,contemporary theories ofemotion are
essentially of two generic forms — category-based
accounts and dimensional models (BOX 1).The idea that
individual emotions are represented as distinct psycho-
logical categories is clearly consistent with observations
that fear and disgust have different neural correlates and
can be selectively impaired. The issue of whether
dimensional models can account for these observations
is less straightforward.
There are several two-dimensional models that have
been proposed to account for the structure ofemotion,
including,for example,Russell’s circumplex model111,112
(BOX 1) and Watson and Tellegen’s113positive-affect/neg-
ative-affect model.Most models are based on purely
descriptive taxonomies,and they are reasonably success-
ful in describing the measures ofself-reported emotion
and the relative confusabilities ofdifferent facial expres-
sions (and other emotional cues).However,it is unclear
that such models can account for the specific deficits in
processing fear and disgust for at least two reasons.First,
it has been suggested that these dimensions are the
product or the expression ofseveral discrete systems
that code more basic emotional constructs.Ifthis is cor-
rect,then these models do not provide an appropriate
psychological level to account for the selective impair-
ments that we have described.Alternatively, if these
dimensions are indeed the foundations ofhuman affect,
then damage to just one ofthese broad dimensions (for
example, pleasure or arousal in the Russell model)
should produce effects on the wide range ofemotions.
But this does not seem to be the case (but see REF.114for
an attempt at such an account).
For the latter interpretation,however,a crucial con-
sideration might be the number of dimensions pro-
posed in such a unifying system.The fear versus disgust
dissociations show that a system with low (N = 2)
dimensionality runs into serious difficulties.However,
there are ways in which dimensional models could be
revised to account for the present data.For example,a
hybrid model,in which discrete emotions are seen as
ENCYCLOPEDIA OF LIFE SCIENCES Limbic system
MIT ENCYCLOPEDIA OF COGNITIVE SCIENCE Emotion
and the human brain | Emotion and the animal brain |
Amygdala | Limbic system | Basal ganglia
© 2001 Macmillan Magazines Ltd
3 6 2 | MAY 2001 |VOLUME 2
R E V I E W S
1. Ekman, P. Strong evidence for universals in facial
expression: a reply to Russell’s mistaken critique. Psychol.
Bull. 115, 268–287 (1994).
Ekman, P. An argument for basic emotions. Cognition
Emotion6, 169–200 (1992).
Aggleton, J . P. The Amygdala: A Functional Analysis (Oxford
Univ. Press, Oxford, 2000).
Book that includes several detailed reviews on the
role of the amygdala in emotion.
Kluver, H. & Bucy, P. C. Preliminary analysis of functions of
the temporal lobes in monkeys. Arch. Neurol. Psychiatry42,
Weiskrantz, L. Behavioral changes associated with ablation
of the amygdaloid complex in monkeys. J . Comp. Physiol.
Psychol. 49, 381–391 (1956).
Meunier, M., Bachevalier, J ., Murray, E. A., Malkova, L. &
Mishkin, M. Effects of aspiration versus neurotoxic lesions of
the amygdala on emotional responses in monkeys. Eur. J .
Neurosci. 11, 4403–4418 (1999).
Kalin, N. H., Shelton, S. E., Davidson, R. J . & Kelley, A. E.
The primate amygdala mediates acute fear but not the
behavioral and physiological components of anxious
temperament. J . Neurosci. 21, 2067–2074 (2001).
A study showing that monkeys with bilateral (fibre-
sparing) excitotoxin lesions of the amygdala show
reduced acute fear responses (for example, reactions
to snakes), whereas indices of a trait-like anxious
temperament (for example, reactions to human
intruders) were unaffected.
LeDoux, J . E. Emotion: clues from the Brain. Annu. Rev.
Psychol. 1995, 209–235 (1995).
LeDoux, J . E. Emotion circuits in the brain. Annu. Rev.
Neurosci. 23, 155–184 (2000).
A recent review of the role of the amygdala in fear
10. Aggleton, J . P. in The Amygdala (ed. Aggleton, J . P.)
485–504 (Wiley-Liss, New York, 1992).
11. Adolphs, R., Tranel, D., Damasio, H. & Damasio, A. Impaired
recognition of emotion in facial expressions following
bilateral damage to the human amygdala. Nature372,
Evidence of impaired recognition of fearful facial
expressions in a patient with selective bilateral
amygdala lesions, but not in patients with unilateral
12. Adolphs, R., Tranel, D., Damasio, H. & Damasio, A. R. Fear
and the human amygdala. J . Neurosci. 15, 5879–5891
13. Adolphs, R. et al. Recognition of facial emotion in nine
individuals with bilateral amygdala damage.
Neuropsychologia 37, 1111–1117 (1999).
14. Calder, A. J . et al. Facial emotion recognition after bilateral
amygdala damage: differentially severe impairment of fear.
Cogn. Neuropsychol. 13, 699–745 (1996).
15. Sprengelmeyer, R. et al. Knowing no fear. Proc. R. Soc.
Lond. B 266, 2451–2456 (1999).
16. J acobson, R. Disorders of facial recognition, social
behaviour and affect after combined bilateral amygdalotomy
and subcaudate tractotomy — a clinical and experimental
study. Psychol. Med. 16, 439–450 (1986).
17. Young, A. W., Hellawell, D. J ., van de Wal, C. & J ohnson, M.
Facial expression processing after amygdalectomy.
Neuropsychologia 34, 31–39 (1996).
18. Young, A. W. et al. Face processing impairments after
amygdalotomy. Brain118, 15–24 (1995).
19. Broks, P. et al. Face processing impairmnents after
encephalitis: amygdala damage and recognition of fear.
Neuropsychologia 36, 59–70 (1998).
20. Anderson, A. K. & Phelps, E. A. Expression without
recognition: contributions of the human amygdala to
emotional communication. Psychol. Sci. 11, 106–111
21. Cahill, L. et al. Amygdala activity at encoding correlated with
long-term, free recall of emotional information. Proc. Natl.
Acad. Sci. USA 93, 8016–8021 (1996).
22. Cahill, L., Babinsky, R., Markowitsch, H. J . & McGaugh, J . L.
The amygdala and emotional memory. Nature377,
23. LaBar, K. S. & Phelps, E. A. Arousal-mediated memory
consolidation: role of the medial temporal lobe in humans.
Psychol. Sci. 9, 490–493 (1998).
24. Aggleton, J . P. The Amygdala (Wiley-Liss, New York, 1992).
25. Rolls, E. T. Neurons in the cortex of the temporal lobe and in
the amygdala of the monkey with responses selective for
faces. Hum. Neurobiol. 3, 209–222 (1984).
26. Leonard, C. M., Rolls, E. T., Wilson, F. A. & Baylis, G. C.
Neurons in the amygdala of the monkey with responses
selective for faces. Behav. Brain Res. 15, 156–176 (1985).
27. Brothers, L., Ring, B. & Kling, A. Response of neurons in the
macaque amygdala to complex social-stimuli. Behav. Brain
Res. 41, 199–213 (1990).
28. Morris, J . S. et al. A neuromodulatory role for the human
amygdala in processing emotional facial expressions. Brain
121, 47–57 (1998).
29. Morris, J . S. et al. A differential neural response in the human
amygdala to fearful and happy facial expressions. Nature
383, 812–815 (1996).
References 28 and 29 are two functional imaging
studies showing, first, that rCBF in the amygdala is
positively related to the intensity of afraid facial
expressions, and second, that rCBF in the extrastriate
cortex is modulated by amygdala activity.
30. Phillips, M. L. et al. A specific neural substrate for perceiving
facial expressions of disgust. Nature389, 495–498 (1997).
Functional MRI study showing different neural
correlates to facial expressions of fear and disgust.
31. Phillips, M. L. et al. Neural responses to facial and vocal
expressions of fear and disgust. Proc. R. Soc. Lond. B 265,
32. Sprengelmeyer, R., Rausch, M., Eysel, U. T. & Przuntek, H.
Neural structures associated with recognition of facial
expressions of basic emotions. Proc. R. Soc. Lond. B 265,
33. Breiter, H. C. et al. Response and habituation of the human
amygdala during visual processing of facial expression.
Neuron17, 875–887 (1996).
34. Whalen, P. J . et al. Masked presentations of emotional facial
expressions modulate amygdala activity without explicit
knowledge. J . Neurosci. 18, 411–418 (1998).
35. Phillips, M. L. et al. Time courses of left and right amygdalar
responses to fearful facial expressions. Hum. Brain Mapping
12, 193–202 (1998).
36. Kawashima, R. et al. The human amygdala plays an
important role in gaze monitoring- A PET study. Brain122,
37. Adolphs, R., Tranel, D. & Damasio, A. R. The human
amygdala in social judgment. Nature393, 470–474 (1998).
38. Anderson, A. K., Spencer, D. D., Fulbright, R. K. & Phelps,
E. A. Contribution of the anteromedial temporal lobes to the
evaluation of facial emotion. Neuropsychology14, 526–536
39. Adolphs, R., Damasio, H., Tranel, D., Cooper, G. & Damasio,
A. R. A role for somatosensory cortices in the visual
recognition of emotion as revealed by three-dimensional
lesion mapping. J . Neurosci. 20, 2683–2690 (2000).
40. Adolphs, R. & Tranel, D. Intact recognition of emotional
prosody following amygdala damage. Neuropsychologia 37,
41. Schmolck, H. & Squire, L. Impaired perception of facial
emotions following bilateral damage to the anterior temporal
lobe. Neuropsychology15, 30–38 (2001).
42. Hamann, S. B. et al. Recognizing facial emotion. Nature
379, 497–497 (1996).
43. Ekman, P., Friesen, W. V. & Ellsworth, P. Emotion and the
Human Face: Guidelines for Research and an Integration of
Findings (Pergamon, New York, 1972).
44. Wright, C. I. et al. Differential prefrontal cortex and amygdala
habituation to repeatedly presented emotional stimuli.
Neuroreport12, 379–383 (2001).
45. Whalen, P. J ., Shin, L. M., McInerney, S. C. & Håkan, F. A
functional MRI study of human amygdala responses to facial
expressions of fear vs. anger. Emotion(in the press).
46. Thomas, K. M. et al. Amygdala response to facial
expressions in children and adults. Biol. Psychiatry49,
47. Maren, S. Auditory fear conditioning increases CS-elicited
spike firing in lateral amygdala neurons even after extensive
overtraining. Eur. J . Neurosci. 12, 4047–4054 (2000).
48. Esteves, F., Dimberg, U. & Ohman, A. Automatically elicited
fear — conditioned skin-conductance responses to masked
facial expressions. Cognition Emotion8, 393–413 (1994).
49. Morris, J . S., Ohman, A. & Dolan, R. J . Conscious and
unconscious emotional learning in the human amygdala.
Nature393, 467–470 (1998).
50. Rauch, S. L. et al. Exaggerated amygdala response to
masked facial stimuli in posttraumatic stress disorder: a
functional MRI study. Biol. Psychiatry47, 769–776 (2000).
51. Rolls, E. T. in The Amygdala (ed. Aggleton, J . P.) (Wiley-Liss,
New York, 1992).
52. Fredrikson, M. et al. Regional cerebral blood-flow during
experimental phobic fear. Psychophysiology30, 126–130
53. Fredrikson, M., Wik, G., Annas, P., Ericson, K. &
Stoneelander, S. Functional neuroanatomy of visually elicited
simple phobic fear — additional data and theoretical
analysis. Psychophysiology32, 43–48 (1995).
54. Lane, R. D. et al. Neuroanatomical correlates of pleasant
and unpleasant emotion. Neuropsychologia 35, 1437–1444
55. Lang, P. J . et al. Emotional arousal and activation of the
visual cortex: an fMRI analysis. Psychophysiology35,
56. Taylor, S. F., Liberzon, I. & Koeppe, R. A. The effect of
graded aversive stimuli on limbic and visual activation.
Neuropsychologia 38, 1415–1425 (2000).
57. Hamann, S. B., Ely, T. D., Grafton, S. T. & Kilts, C. D.
Amygdala activity related to enhanced memory for pleasant
and aversive stimuli. Nature Neurosci. 2, 289–293 (1999).
Functional imaging study showing that the amygdala
plays a neuromodulatory role in recognition memory
for pictures of both aversive and pleasant emotional
58. Canli, T., Zhao, Z., Brewer, J ., Gabrieli, J . D. E. & Cahill, L.
Event-related activation in the human amygdala associates
with later memory for individual emotional experience.
J . Neurosci. 20, RC99: 1–5 (2000).
59. Ketter, T. A. et al. Anterior paralimbic mediation of procaine-
induced emotional and psychosensory experiences. Arch.
Gen. Psychiatry53, 59–69 (1996).
60. Amaral, D. G., Price, D. L., Pitkanen, A. & Carmichael, S. T.
in The Amygdala (ed. Aggleton, J . P.) 1–66 (Wiley-Liss, New
Review of the research addressing the anatomical
organization of the primate amygdaloid complex.
61. Anderson, A. K. & Phelps, E. A. Intact recognition of vocal
expressions of fear following bilateral lesions of the human
amygdala. Neuroreport9, 3607–3613 (1998).
62. Scott, S. K. et al. Impaired auditory recognition of fear and
anger following bilateral amygdala lesions. Nature385,
Case study of a patient with bilateral amygdala
damage showing impaired recognition of auditory
signals of fear and anger. A previous study had shown
that this patient was impaired at recognizing facial
cues of these same emotions (see reference 14).
63. Starkstein, S. E., Federoff, J . P., Price, T. R., Leiguarda, R. C.
& Robinson, R. G. Neuropsychological and neuroradiologic
correlates of emotional prosody comprehension. Neurology
44, 515–522 (1994).
64. Davis, M. in The Amygdala (ed. Aggleton, J . P.) 255–306
(Wiley-Liss, New York, 1992).
65. Morris, J . S., Scott, S. K. & Dolan, R. J . Saying it with feeling:
neural responses to emotional vocalizations.
Neuropsychologia 37, 1155–1163 (1999).
66. Isenberg, N. et al. Linguistic threat activates the human
amygdala. Proc. Natl. Acad. Sci. USA 96, 10456–10459
67. Whalen, P. J . et al. The emotional counting Stroop
paradigm: a functional magnetic resonance imaging probe
of the anterior cingulate affective division. Biol. Psychiatry
44, 1219–1228 (1998).
68. Halgren, E. in The Amygdala (ed. Aggleton, J . P.) 191–228
(Wiley-Liss, New York/Chichester, 1992).
69. Rapcsak, S. Z. et al. Fear recognition deficits after focal brain
damage — a cautionary note. Neurology54, 575–581
70. O’Doherty, J ., Rolls, E. T., Francis, S., Bowtell, R. &
McGlone, F. Representation of pleasant and aversive taste in
the human brain. J . Neurophysiol. 85, 1315–1321 (2001).
71. Parkinson, J . A., Robbins, T. W. & Everitt, B. J . Dissociable
roles of the centraland basolateral amygdala in appetitive
emotional learning. Eur. J . Neurosci. 12, 405–13 (2000).
72. Baxter, M. G., Parker, A., Lindner, C. C. C., Izquierdo, A. D. &
Murray, E. A. Control of response selection by reinforcer
value requires interaction of amygdala and orbital prefrontal
cortex. J . Neurosci. 20, 4311–4319 (2000).
73. Hatfield, T., Han, J . S., Conley, M., Gallagher, M. & Holland,
P. Neurotoxic lesions of basolateral, but not central,
amygdala interfere with pavlovian second-order conditioning
and reinforcer devaluation effects. J . Neurosci. 16,
74. Malkova, L., Gaffan, D. & Murray, E. A. Excitotoxic lesions of
the amygdala fail to produce impairment in visual learning for
auditory secondary reinforcement but interfere with
reinforcer devaluation effects in rhesus monkeys. J .
Neurosci. 17, 6011–6020 (1997).
75. Ono, T. & Nishijo, H. in The Amygdala (ed. Aggleton, J . P.)
167–190 (Wiley-Liss, New York, 1992).
76. Blair, R. J . R., Morris, J . S., Frith, C. D., Perrett, D. I. & Dolan,
R. J . Dissociable neural responses to facial expressions of
sadness and anger. Brain122, 883–893 (1999).
77. Phillips, M. L. et al. Investigation of facial recognition memory
and happy and sad facial expression perception: an fMRI
study. Psychiatry Res. Neuroimaging 83, 127–138 (1998).
78. Kessler/West, M. L. et al. Neural substrates of facial emotion
processing using fMRI. Cogn. Brain Res. 11, 213–226
79. Rozin, P. & Fallon, A. E. A perspective on disgust. Psychol.
Rev. 94, 23–41 (1987).
Review of studies addressing several aspects of the
emotion disgust, including contamination, the
phylogeny and ontogeny of disgust, and the function
of this emotion in human society.
© 2001 Macmillan Magazines Ltd
NATURE REVIEWS | NEUROSCIENCE
VOLUME 2 |MAY 2001 | 3 6 3
R E V I E W S
80. Rozin, P., Haidt, J . & McCauley, C. R. in Handbook of
Emotions (eds Lewis, M. & Haviland, J .) 575–594 (Guilford,
New York, 1993).
81. Garcia, J ., Forthman Quick, D. & White, B. in Primary Neural
Substrates of Learning and Behavioral Change(eds Alkon,
D. L. & Farley, J .) 47–61 (Cambridge Univ. Press,
82. Sprengelmeyer, R. et al. Loss of disgust — perception of
faces and emotions in Huntington’s disease. Brain119,
A group study of facial and vocal expression
processing in patients with Huntington’s disease.
Patients showed a disproportionately severe deficit in
recognizing facial signals of disgust.
83. Sprengelmeyer, R. et al. Recognition of facial expressions of
basic emotions in Huntington’s disease. Cogn.
Neuropsychol. 14, 839–879 (1997).
84. Gray, J . M., Young, A. W., Barker, W. A., Curtis, A. & Gibson,
D. Impaired recognition of disgust in Huntington’s disease
gene carriers. Brain120, 2029–2038 (1997).
85. Braun, A. R. et al. The functional neuroanatomy of Tourettes
syndrome — an FDG-PET study. 2. Relationships between
regional cerebral metabolism and associated behavioral and
cognitive features of the illness. Neuropsychopharmacology
13, 151–168 (1995).
86. Rapoport, J . L. The neurobiology of obsessive-compulsive
disorder. J . Am. Med. Assoc. 260, 2888–2890 (1989).
87. Rapoport, J . L. & Fiske, A. The new biology of obsessive-
compulsive disorder: implications for evolutionary
psychology. Perspect. Biol. Med. 41, 159–175 (1998).
88. Sprengelmeyer, R. et al. Disgust implicated in obsessive-
compulsive disorder. Proc. R. Soc. Lond. B 264,
89. Phillips, M. L. et al. A differential neural response to
threatening and non-threatening negative facial expressions
in paranoid and non-paranoid schizophrenics. Psychiatry
Res. Neuroimaging 92, 11–31 (1999).
90. Augustine, J . R. Circuitry and functional aspects of the insula
lobe in primates including humans. Brain Res. Rev. 22,
91. Small, D. M. et al. Human cortical gustatory areas: a review
of functional neuroimaging. Neuroreport10, 7–14 (1999).
92. Penfield, W. & Faulk, M. E. The insula: further observations
of its function. Brain78, 445–470 (1955).
93. Hernadi, I., Zaradi, K., Faludi, B. & Lenard, L. Disturbances
of neophobia and taste-aversion learning after bilateral
kainate microlesions in the rat pallidum. Behavioral
Neuroscience111, 137–146 (1997).
94. Dunn, L. T. & Everitt, B. J . Double dissociations of the effects
of amygdala and insular cortex lesions on conditioned taste
aversion, passive-avoidance, and neophobia in the rat using
the excitotoxin ibotenic acid. Behav. Neurosci. 102, 3–23
95. Rozin, P., Lowery, L. & Ebert, R. Varieties of disgust faces
and the structure of disgust. J . Personality Social Psychol.
66, 870–881 (1994).
96. Calder, A. J ., Keane, J ., Manes, F., Antoun, N. & Young, A.
W. Impaired recognition and experience of disgust following
brain injury. Nature Neurosci. 3, 1077–1078 (2000).
Case study of a patient with left-hemisphere damage
affecting the insula and basal ganglia. The patient
showed a marked selective impairment in recognizing
both facial and vocal signals of disgust, and impaired
experience of disgust.
97. Phillips, M. L. et al. A differential neural response in
obsessive-compulsive disorder patients with washing
compared with checking symptoms to disgust. Psychol.
Med. 30, 1037–1050 (2000).
98. Shin, L. M. et al. Activation of anterior paralimbic structures
during guilt-related script-driven imagery. Biol. Psychiatry
48, 43–50 (2000).
99. Power, M. & Dalgleish, T. Cognition and Emotion: From
Order to Disorder(Psychology Press, Hove, 1997).
100. Chikama, M., McFarland, N. R., Amaral, D. G. & Haber, S.
N. Insular cortical projections to functional regions of the
striatum correlate with cortical cytoarchitectonic
organization in the primate. J . Neurosci. 15, 9686–9705
101. Weeks, R. A. et al. Cortical control of movement in
Huntington’s disease. Brain120, 1569–1578 (1997).
102. Boecker, H. et al. Sensory processing in Parkinson’s and
Huntington’s disease: investigations with 3D H2
Brain122, 1651–1665 (1999).
103. Lange, H. W. Quatitative changes of the telencephalon,
diencephalon, and mesencephalon in Huntington’s chorea,
postcencephalitic, and idiopahthic Parkinsonism. Verh.
Anat. Ges. (J ena)75, 923–925 (1981).
104. Fennema-Notestine, C. et al. Global pattern of
neuroanatomical changes in Huntington’s disease with
morphometric analyses of structural magnetic resonance
imaging. Am. Acad. Neurol. A152–A153 (2000).
105. Goodman, W. K. et al. The Yale-Brown obsessive
compulsive scale. 1. Development, use, and reliability. Arch.
Gen. Psychiatry46, 1006–1011 (1989).
106. Rauch, S. L. et al. Neural correlates of factor-analysed OCD
symptom dimensions: a PET study. CNS Spectrums 3,
107. Baer, L. Factor-analysis of symptom subtypes of obsessive-
compulsive disorder and their relation to personality and tic
disorders. J . Clin. Psychiatry55, 18–23 (1994).
108. Degroot, C. M., Bornstein, R. A., J anus, M. D. &
Mavissakalian, M. R. Patterns of obsessive-compulsive
symptoms in Tourette subjects are independent of severity.
Anxiety1, 268–274 (1994).
109. Petter, T., Richter, M. A. & Sandor, P. Clinical features
distinguishing patients with Tourette’s syndrome and
obsessive-compulsive disorder from patients with
obsessive-compulsive disorder without tics. J . Clin.
Psychiatry59, 456–459 (1998).
110. Hornak, J ., Rolls, E. T. & Wade, D. Face and voice
expression identification in patients with emotion and
behavioural changes following ventral frontal lobe damage.
Neuropsychologia 34, 247–261 (1996).
111. Russell, J . A. & Bullock, M. Multidimensional scaling of
emotional facial expressions: similarity from preschoolers to
adults. J . Personality Social Psychol. 48, 1290–1298 (1985).
112. Russell, J . A. A circumplex model of affect. J . Personality
Social Psychol. 39, 1161–1178 (1980).
113. Watson, D. & Tellegen, A. Toward a consensual structure of
mood. Psychol. Bull. 98, 219–235 (1985).
114. Adolphs, R., Russell, J . A. & Tranel, D. A role for the human
amygdala in recognising emotional arousal from unpleasant
stimuli. Psychol. Sci. 10, 167–171 (1999).
115. J ohnsen, B. H., Thayer, J . F. & Hugdahl, K. Affective
judgment of the Ekman faces — a dimensional approach. J .
Psychophysiol. 9, 193–202 (1995).
116. Calder, A. J ., Burton, A. M., Miller, P., Young, A. W. &
Akamatsu, S. A principal component analsysis of facial
expressions. Vision Res. 41, 1179–1208 (2001).
117. Ekman, P. & Davidson, R. J . The Nature of EmotionVol. 1,
7–47 (Oxford Univ. Press, New York, 1994).
118. Panksepp, J . Affective Neuroscience: The Foundations of
Human and Animal Emotions (Oxford Univ. Press, New
119. Rolls, E. T. The Brain and Emotion(Oxford Univ. Press,
120. Rolls, E. T. A theory of emotion, and its application to
understanding the neural basis of emotion. Cognition
Emotion4, 161–190 (1990).
121. Gray, J . A. The Psychology of Fear and Stress (Cambridge
Univ. Press, Cambridge, 1987).
122. Davidson, R. J . Prolegomenon to the structure of ermotion
— gleanings from neuropsychology. Cognition Emotion6,
123. Davidson, R. J ., J ackson, D. C. & Kalin, N. H. Emotion,
plasticity, context, and regulation: perspectives from
affective neuroscience. Psychol. Bull. 126, 890–909
124. Woodworth, R. S. & Schlosberg, H. Experimental
Psychology (Henry Holt, New York, 1954).
125. Schlosberg, H. The description of facial expressions in terms
of two dimensions. J . Exp. Psychol. 44, 229–237 (1952).
126. Green, R. S. & Cliff, N. Multidimensional comparisons of
structures of vocally and facially expressed emotions.
Perception Psychophysics 17, 429–438 (1975).
127. Bush, L. E. U. Individual differences in multidimensional
scaling of adjectives denoting feelings. J . Personality Social
Psychol. 25, 50–57 (1973).
128. Furmark, T., Fischer, H., Wik, G., Larsson, M. & Fredrikson,
M. The amygdala and individual differences in human fear
conditioning. Neuroreport8, 3957–3960 (1997).
129. Buchel, C. & Dolan, R. J . Classical fear conditioning in
functional neuroimaging. Curr. Opin. Neurobiol. 10, 219–223
130. Fischer, H., Andersson, J . L. R., Furmark, T. & Fredrikson,
M. Fear conditioning and brain activity: a positron emission
tomography study in humans. Behav. Neurosci. 114,
131. Buchel, C., Morris, J ., Dolan, R. J . & Friston, K. J . Brain
systems mediating aversive conditioning: an event-related
fMRI study. Neuron20, 947–957 (1998).
Demonstration of amygdala involvement in
conditioned fear in humans using event-related
imaging techniques (see also reference 133).
132. Buchel, C., Dolan, R. J ., Armony, J . L. & Friston, K. J .
Amygdala-hippocampal involvement in human aversive
trace conditioning revealed through event-related functional
magnetic resonance imaging. J . Neurosci. 19,
133. LaBar, K. S., Gatenby, J . C., Gore, J . C., LeDoux, J . E. &
Phelps, E. A. Human amygdala activation during
conditioned fear acquisition and extinction: a mixed-trial
fMRI study. Neuron20, 937–945 (1998).
134. Bechara, A. et al. Double dissociation of conditioning and
declarative knowledge relative to the amygdala and
hippocampus in humans. Science269, 1115–1118 (
135. Labar, K. S., Ledoux, J . E., Spencer, D. D. & Phelps, E. A.
Impaired fear conditioning following unilateral temporal
lobectomy in humans. J . Neurosci. 15, 6846–6855 (1995).
136. Knight, D. C., Smith, C. N., Stein, E. A. & Helmstetter, F. J .
Functional MRI of human Pavlovian fear conditioning:
patterns of activation as a function of learning. Neuroreport
10, 3665–3670 (1999).
137. Fendt, M. & Fanselow, M. S. The neuroanatomical and
neurochemical basis of conditioned fear. Neuroscience
Biobehav. Rev. 23, 743–760 (1999).
A review of the literature on conditioned fear .
138. Calder, A. J ., Young, A. W., Perrett, D. I., Etcoff, N. L. &
Rowland, D. Categorical perception of morphed facial
expressions. Visual Cogn. 3, 81–117 (1996).
139. Young, A. W. et al. Megamixing facial expressions. Cognition
63, 271–313 (1997).
140. Morris, J . S., Friston, K. J . & Dolan, R. J . Neural responses
to salient visual stimuli. Proc. R. Soc. Lond. B 264, 769–775
141. Morris, J . S., Friston, K. J . & Dolan, R. J . Experience-
dependent modulation of tonotopic neural responses in
human auditory cortex. Proc. R. Soc. Lond. B 265,
142. Brett, M., Christoff, K., Cusack, R. & Lancaster, J . Using the
Talairach atlas with the MNI template. NeuroImage(in the
143. Rorden, C. & Brett, M. Stereotaxic display of brain lesions.
Behav. Neurol. (in the press).
We would like to thank J ill Keane, Brian Cox and Matthew Brett for
their assistance in preparing this article, and Professor Paul Ekman
for giving us permission to reproduce examples of the Ekman and
Friesen (1976) faces.
© 2001 Macmillan Magazines Ltd