In Pavlovian fear conditioning, a neutral stimulus (con-
ditional stimulus, or CS), usually a light or tone, is typi-
cally paired with an aversive stimulus (unconditional stim-
ulus, or UCS) such as an electric shock. Animals learn that
the CS predicts the UCS as the stimuli are repeatedly
paired, and this learning is manifested as changes in be-
havioral and autonomic responses (Rescorla, 1988). Pre-
sentation of the CS alone provokes the expression of a
conditional fear response once learning has occurred. The
thalamus, hippocampus, and sensory cortex each contribute
to conditional response (CR) acquisition, and the amyg-
dala is a principal site of CS–UCS convergence and criti-
cal synaptic plasticity (Bailey, Kim, Sun, Thompson, &
Helmstetter, 1999; Davis, 2000; LeDoux, 1995; Maren,
2001). In extinction, CR magnitude or probability tends to
decrease as the CS is repeatedly presented without the
UCS. Extinction of conditional fear probably represents
the learning of a new CS–no-UCS relationship rather than
the weakening of the initial CS–shock association and can
be prevented by blockade of NMDA receptors within the
amygdala (Baker & Azorlosa, 1996; Falls, Miserendino,
& Davis, 1992; Meyers & Davis, 2002; Santini, Muller, &
Quirk, 2001). Thus, the amygdala also appears to play a
317 Copyright 2004 Psychonomic Society, Inc.
This work was supported by NIMH Fellowship MH11722 to D.C.K.,
NIDA Grant DA09465 to E.A.S., NIMH Grant MH50864 to F.J.H., a
McDonnell Foundation grant to F.J.H., and General Clinical Research
Centers Grant M01 RR00058 to the Medical College of Wisconsin.Cor-
respondence should be addressed to F. J. Helmstetter, Department of
Psychology, University of Wisconsin, P.O. Box 413, Milwaukee, WI 53201
Amygdala and hippocampal activity during
acquisition and extinction of
human fear conditioning
DAVID C. KNIGHT
University of Wisconsin, Milwaukee, Wisconsin
and National Institute of Mental Health, Bethesda, Maryland
CHRISTINE N. SMITH
University of Wisconsin, Milwaukee, Wisconsin
and University of California, San Diego, La Jolla, California
DOMINIC T. CHENG
University of Wisconsin, Milwaukee, Wisconsin
ELLIOT A. STEIN
Medical College of Wisconsin, Milwaukee, Wisconsin
and National Institute on Drug Abuse/Intramural Research Program, Baltimore, Maryland
FRED J. HELMSTETTER
University of Wisconsin, Milwaukee, Wisconsin
and Medical College of Wisconsin, Milwaukee, Wisconsin
Previous functional magnetic resonance imaging (fMRI) studies have characterized brain systems in-
volved in conditional response acquisition during Pavlovian fear conditioning. However, the functional
neuroanatomy underlying the extinction of human conditional fear remains largely undetermined. The
present study used fMRI to examine brain activity during acquisition and extinction of fear condition-
ing. During the acquisition phase, participants were either exposed to light (CS) presentations that sig-
naled a brief electrical stimulation (paired group) or received light presentations that did not serve as
a warning signal (control group). During the extinction phase, half of the paired group subjects con-
tinued to receive the same treatment, whereas the remainder received light alone. Control subjects
also received light alone during the extinction phase. Changes in metabolic activity within the amyg-
dala and hippocampus support the involvement of these regions in each of the procedural phases of
fear conditioning. Hippocampal activity developed during acquisition of the fear response. Amygdala
activity increased whenever experimental contingencies were altered, suggesting that this region is in-
volved in processing changes in environmental relationships. The present data show learning-related
amygdala and hippocampal activity during human Pavlovian fear conditioning and suggest that the
amygdala is particularly important for forming new associations as relationships between stimuli
Cognitive, Affective, & Behavioral Neuroscience
2004, 4 (3), 317–325
318 KNIGHT, SMITH, CHENG, STEIN, AND HELMSTETTER
critical role in the extinction, as well as the acquisition, of
conditional fear. Recent studies indicate that CR extinc-
tion relies upon both short-term (within session) and long-
term (across session) memory processes (Santini et al.,
2001). Short-term extinction processes appear to be me-
diated through amygdala-dependent circuitry, and modu-
lation of this circuitry by the ventromedial prefrontal cor-
tex may be critical for long-term extinction (Milad &
Quirk, 2002; Quirk, Russo, Barron, & Lebron, 2000; San-
tini et al., 2001).
Imaging studies examining neural activity during ac-
quisition of conditional fear in humans have reported
learning-related changes within the amygdala and hippo-
campus, as well as in other brain regions (Büchel, Dolan,
Armony, & Friston, 1999; Büchel, Morris, Dolan, & Fris-
ton, 1998; Cheng, Knight, Smith, Stein, & Helmstetter,
2003; Knight, Cheng, Smith, Stein, & Helmstetter, 2004;
Knight, Smith, Stein, & Helmstetter, 1999; LaBar,
Gatenby, Gore, LeDoux, & Phelps, 1998). Differential re-
sponding to CS presentations rapidly decreased in the
amygdala as training continued in some of these studies
(Büchel etal., 1999; Büchel etal., 1998; LaBar etal., 1998).
One interpretation of these data is that the amygdala is im-
portant for initial acquisition of fear but contributes less
once the response has been learned. However, procedural
factors may also explain these data. Some of these studies
included CS-alone trials prior to the imaging session, re-
quiring subjects to learn a new contingency when training
with the UCS began (Büchel et al., 1999; LaBar et al.,
1998). Therefore, the amygdala activation observed may
be associated with factors related to the altered relation-
ships among stimuli. Our prior work on fear conditioning
and fMRI has not used preconditioning exposure to the
CS, and although we have shown significant activation
within a number of brain regions such as the anterior cin-
gulate, medial thalamus, and primary visual cortices, we
typically do not observe much CS-related amygdala ac-
tivity during acquisition (see, e.g., Cheng et al., 2003;
Knight et al., 2004; Knight et al., 1999). One possible ex-
planation for these divergent amygdala results is the
change in stimulus relationships associated with amyg-
dala activation. In fact, the amygdala has been implicated
in novelty detection (Fischer, et al., 2003; Montag-Sallaz,
Welzl, Kuhl, Montag, & Schachner, 1999) and is an im-
portant component of the attentional circuitry involved in
the formation of new associations during unexpected learn-
ing experiences (Holland & Gallagher, 1999; Holland,
Han, & Gallagher, 2000). Therefore, activation of this at-
tentional circuitry would be expected following a change
in stimulus relationships given that such changes produce
a novel experience through which new associations must
Several theories have been offered to explain the in-
volvement of the hippocampus during Pavlovian condi-
tioning procedures. Although the hippocampus is not nor-
mally included in the neural circuitry thought to be critical
for single-cue delay learning, multiple technical approaches
indicate that this region is differentially active during such
conditioning tasks. It is generally held that more complex
conditioning procedures, including trace conditioning,
discrimination reversal, and context conditioning, are de-
pendent upon the hippocampus, whereas single-cue delay
conditioning is not. Learning in these more complex pro-
cedural variants is generally disrupted by hippocampal le-
sions, whereas simple delay conditioning remains intact
(Berger & Orr, 1983; Moyer, Deyo, & Disterhoft, 1990;
Solomon, Vander Schaaf, Thompson, & Weisz, 1986). How-
ever, hippocampal unit recordings show learning-related
changes in neuronal activation during both simple and
complex conditioning tasks (Berger & Orr, 1983; Weiss,
Kronforst-Collins, & Disterhoft, 1996). These studies sug-
gest that although the hippocampus is only necessary for
performance in more complex conditioning procedures, it
may be metabolically active during simple conditioning
tasks as well. This neuronal activation may be measurable
through BOLD (blood oxygen level dependent) fMRI sig-
nal changes. In an attempt to better define the role of the
amygdala and hippocampus, we used fMRI during the ac-
quisition and extinction of Pavlovian fear conditioning in
healthy human subjects.
Thirty healthy, right-handed subjects (17 women and 13 men),
ranging in age from 18 to 39 (mean ? SEM ? 24.5 ? 5.17 years),
were recruited for this study. All participants provided informed
consent, and all procedures were approved by the Institutional Re-
view Boards of the University of Wisconsin-Milwaukee and the
Medical College of Wisconsin.
Functional MRI. Imaging was performed on a 1.5-Tesla GE
Signa scanner using a gradient-echo, echo-planar pulse sequence.
Contiguous 8-mm sagittal slices were collected (TR ? 3,000 msec,
TE?40msec, FOV?24cm, 64 ?64 matrix, 3.75mm ?3.75mm
in-plane resolution, flip angle ? 90º) in a series of 86 sequential
images during each 258-sec block of stimulus presentations. High-
resolution spoiled gradient recalled acquisition at steady state
(SPGR) anatomical images were acquired prior to echo–planar
imag-ing to allow subsequent anatomical localization of functional
activation. Foam padding was used to limit head movements within
a customized three-axis local gradient coil (Wong, Tan & Hyde,
Electricalstimulus. A customized AC (60-Hz) source was used
to provide transcutaneous electrical stimulation (the UCS; 0.5-sec
duration) through two aluminum surface electrodes (2-cm diameter)
placed over the right tibial nerve above the medial malleolus. The in-
tensity of electrical stimulation was individually set for each partic-
ipant. Participants determined the intensity of the electrical stimu-
lation by rating practice trials from 0 to 5 (0 ? no sensation, 5 ?
painful but tolerable). The intensity of the first practice trial was
equal to 0mA and was gradually increased until the participant rated
the stimulation as a 5, or the stimulus intensity reached the maximal
current of 7.5mA. Once set, the electrical stimulation was maintained
at a constant intensity throughout the experiment.
Visualstimulus. A red 25-W light bulb was flashed (0.5 Hz) for
15 sec as the CS. The light bulb was housed in a black plastic light
fixture, positioned 10º to the right side of a fixation point (a small
red light-emitting diode [LED]), mounted 5.5 m from the scanner.
The CS was viewed with a pair of prism glasses.
fMRI AND FEAR CONDITIONING 319
Skinconductanceresponse(SCR). A J&J thermal monitoring
system (Model T-68) with GSR Preamp (Model IG-3) was used to
monitor each participant’s skin conductance throughout the experi-
ment. Skin conductance was monitored with a pair of surface gel
cup electrodes (Ag/AgCl, 1-cm diameter) attached 2cm apart on the
sole of the participant’s left foot. Although SCR is traditionally mon-
itored from the palm of the hand, compatibility issues between the
SCR and MRI equipment required measurement from this alterna-
tive location. All skin conductance data were range corrected by di-
viding each participant’s raw skin conductance by his or her mean
skin conductance level across each training block. SCR was calcu-
lated as the peak response that occurred during the 15-sec CS pre-
sentation and was expressed as a percentage of the evoked SCR to
the first conditioning trial. SCR data were averaged for each train-
ing block of the acquisition and extinction phases of the study.
Participants were randomly assigned to one of two groups (paired,
n ? 20; control, n ? 10). During the acquisition phase, paired sub-
jects received CS (15sec) and UCS (0.5sec; intertrial interval [ITI]?
30 sec) presentations that coterminated, whereas the control group
received explicitly unpaired CS (15 sec) and UCS (0.5 sec; ITI ?
30 sec, ISI ? 15 sec) presentations. Following four blocks of acqui-
sition trials (5 trials per block), half of the paired group subjects were
randomly assigned to a new extinction group. During the extinction
phase, the paired group (n ? 10) continued to receive the same
CS–UCS pairings, but the extinction (n ? 10) and control (n ? 10)
groups were presented light alone (see Figure1). Subjective pain rat-
ings of the UCS were obtained after each block of trials.
Functional MRI Analysis
Functional image analyses were performed using AFNI Ver-
sion 2.2 (Cox & Hyde, 1997). In-plane motion correction and edge-
detection algorithms were applied to the functional data. Activations
within the amygdala and hippocampus were identified through
cross-correlation analysis with a reference waveform representing
the temporal pattern of visual stimulation. The reference waveform
was shifted by one image to account for the delay in the hemody-
namic response. Given the anticipated hemodynamic delay and rise
time, BOLD signal changes elicited by UCS presentation were ex-
cluded from this analysis. Functional and anatomical images for
each subject were transformed into stereotactic coordinate space rel-
ative to the line between the anterior and posterior commisures (Ta-
lairach & Tournoux, 1988). Functional maps were spatially blurred
using a 3.0-mm RMS Gaussian filter. A two-factor analysis of vari-
ance (ANOVA) was performed on a voxel-wise basis, resulting in
an F statistic (F ? 3.07) that was used to restrict additional analyses
to those voxels that were significantly activated (p ? .01). Changes
in the pattern of activation across the conditioning session were eval-
uated by t test comparisons (t ? 2.89, p ? .01).
Figure 1. Illustration of the experimental design. Participants were randomly
assigned to one of two groups (paired group, n ? 20, or control group, n ? 10)
that received four blocks of stimulus presentations (5 trials per block) during
the acquisition phase. Acquisition phase: Paired subjects received coterminat-
ing CS and UCS presentations, whereas control subjects received explicitly un-
paired CS and UCS presentations. During the extinction phase, the paired
group was divided, and half of the subjects (paired group, n ? 10) continued
to receive the same CS–UCS pairings, whereas the other half (extinction group,
n ? 10) and the control group (n ? 10) were presented with the light alone.
320 KNIGHT, SMITH, CHENG, STEIN, AND HELMSTETTER
All groups received UCS presentations of similar in-
tensity (mean ? SEM ? paired, 3.78 ? 0.76 mA; extinc-
tion, 4.21 ? 0.63 mA; control, 3.72 ? 0.6 mA; F ? 1).
All groups reported similar pain ratings during the ac-
quisition phase of the experiment (mean?SEM?paired,
3.95?0.11; control, 3.68?0.18; F?1). Paired subjects
provided comparable pain ratings during both the acqui-
sition and extinction phases, whereas extinction and con-
trol subjects reported essentially no pain during the ex-
tinction phase when no UCS was presented (mean ?
SEM?paired, 3.82?0.18; extinction, 0.07 ?0.07; con-
trol, 0.12 ? 0.07; F ? 84.46, p ? .05).
Skin Conductance Response
Comparisons of SCR (see Figure 2) during the acquisi-
tion phase demonstrated that CRs were greater for sub-
jects who received paired CS–UCS presentations as op-
posed to controls (t ? 2.19, p ? .05). These differences
confirm that the CS provoked a conditional fear response
in paired but not in control subjects. During the extinction
phase, subjects within the extinction and control groups
received presentations of the light alone, whereas the
paired group continued to receive the same CS–UCS pair-
ings. Large conditioned SCRs were maintained in paired
group subjects, as opposed to controls (t?2.30, p?.05),
during this phase of the study, whereas CR amplitude for
extinction group subjects fell to an intermediate level that
did not differ from that of the paired (t ? 1.10) or control
(t ? 1.15) groups. CR timing (mean ? SEM ? 7.68 ?
0.57 sec) was consistent with second interval responding,
which is often used as an index of learning CS–UCS asso-
ciations (Prokasy & Raskin, 1973; Wolter & Lachnit, 1993).
Functional MRI Results
An ANOVA revealed the development of significant
stimulus-evoked activity within the bilateral amygdala
(Talairach coordinate system: right, 19, ?11, ?19; left,
?22, ?6, ?19) and left hippocampus (Talairach coordi-
nate system: ?29, ?15, ?12) over the course of training.
Amygdala activation was similar for all groups during the
acquisition phase of the study (Figure 3A). However, as
can be seen in Figure3B, right amygdala activity increased
in subjects receiving CS alone trials (extinction and con-
trol groups) during the extinction phase of the experiment,
whereas left amygdala activity decreased in these groups.
Paired subjects showed no change in amygdala activation.
As is illustrated in Figure 3C, differences in right amyg-
dala activation quickly returned to pre-extinction phase
levels, whereas differences in left amygdala activation
gradually decreased across the extinction phase.
Differences in left hippocampal activation developed
over the course of acquisition and were maintained into
the extinction phase of the experiment. As can be seen in
Figure 4A, hippocampal activation was similar for all
groups during the first training block. However, subjects
receiving paired CS–UCS presentations showed greater
left hippocampal activity than did controls by the last block
of acquisition (see Figure 4A). Figure 4B demonstrates
that left hippocampal activity decreased in extinction sub-
jects as they began the extinction phase of the study, but
that no change in hippocampal activity was revealed for
Figure 2. SCR for paired, extinction, and control groups during the acquisi-
tion and extinction phases. During the acquisition phase, the paired group (n?
20) produced larger CRs than did controls (n ? 10). During the extinction
phase, the paired group (n ? 10) continued to produce large CRs, whereas the
extinction group’s (n?10) CRs fell to an intermediate level in comparison with
the small responses elicited by controls (n ? 10).
fMRI AND FEAR CONDITIONING321
the paired or the control groups. Figure 4C shows differ-
ences in hippocampal activity for paired versus extinction
and control groups for each block of the extinction phase
of the experiment. Left hippocampal activity was greater
in paired subjects initially, but it decreased over the course
of the extinction phase.
Figure 3. Amygdala activity. (A) Paired (Prd) compared with control (Cnt) group activation on
the first and last blocks of the acquisition phase. No differences in amygdala activity were demon-
strated between groups during this phase of the study. (B)Comparison of amygdala activity during
the last stimulus block of the acquisition phase and first block of the extinction phase for each group.
Right amygdala activity increased in extinction (Ext) and Cnt subjects, and left amygdala activity
decreased for the Ext group during the extinction phase of the study. Subjects who continued to re-
ceive the same CS–UCS pairings (Prd group) throughout the study showed no change in amygdala
activation. (C) Comparison of group differences (Cnt & Ext vs. Prd) during the extinction phase of
the study. A brief increase in right amygdala activity coincided with the initiation of extinction tri-
als for Cnt and Ext subjects and then quickly returned to pre-extinction phase levels. In contrast,
left amygdala activity, during extinction, was greater for the paired group than for subjects receiv-
ing the CS alone (Cnt & Ext), but gradually diminished across the extinction phase.
322 KNIGHT, SMITH, CHENG, STEIN, AND HELMSTETTER
The amygdala is a critical component of the neural cir-
cuit thought to mediate acquisition of conditional fear. Af-
ferent projections carrying information about the CS and
UCS converge within subnuclei of the amygdala, and pro-
jections from the amygdala to brainstem targets appear to
be important for the expression of critical autonomic and
behavioral responses (Davis, 2000; Helmstetter, Tershner,
Poore, & Bellgowan, 1998; LeDoux, 1995; Price & Ama-
ral, 1981). Although most of the research exploring the
amygdala’s contribution to conditional fear has been con-
ducted with laboratory animals, functional brain imaging
Figure 4. Left hippocampal activity. (A) Paired (Prd) compared with control (Cnt) group activa-
tion on the first and last blocks of the acquisition phase. Although left hippocampal activity was sim-
ilar for both groups at the beginning of the session, significant differences (Prd ? Cnt) developed
with repeated training. (B) Comparison of left hippocampal activity during the last acquisition
block and first extinction block for each group. Activity decreased during the extinction phase for
the extinction (Ext) group, whereas no change was observed for Cnt and Prd subjects. (C) Com-
parison of group differences (Cnt & Ext vs. Prd) during the extinction phase of the study. Left hippo-
campal activity remained greater for Prd subjects initially, but habituated across the extinction
phase of the study.
fMRI AND FEAR CONDITIONING323
studies support the amygdala’s importance for this type of
learning in humans as well (Büchel et al., 1999; Büchel
et al., 1998; Cheng et al., 2003; LaBar et al., 1998). In the
present study, differential task-related amygdala activation
was only observed following a change in stimulus rela-
tionships, which is consistent with the view that this re-
gion is an important component of the attentional circuitry
supporting the formation of new associations when rela-
tionships between stimuli change (Holland & Gallagher,
1999; Holland etal., 2000). These data provide further ev-
idence that the amygdala is critically involved in human
Pavlovian fear conditioning and suggest that such differ-
ential amygdala activity may be related to specific fea-
tures of the training procedure that require the learning of
new relationships between stimuli.
Previous fMRI fear conditioning research has typically
observed unilateral activation of the right amygdala (Büchel
etal., 1999; Büchel etal., 1998; Cheng etal., 2003; LaBar
et al., 1998). The right amygdala activation observed in
the present study is largely consistent with this prior work.
LaBar et al. observed amygdala activity during both early
acquisition and extinction. In the present study, however,
differential amygdala activity was observed only during
extinction. One possible explanation for these differences
is the use of CS habituation trials prior to training in the
previous study. The differential activation observed dur-
ing early acquisition trials may have been driven by the
change in contingency for the CS?, but not CS?, during
the transition from habituation to acquisition rather than
by acquisition of the CR itself. In the present study, sub-
jects had no prior exposure to target stimuli. Therefore, all
CS presentations were initially novel and appear to have
produced similar levels of right amygdala activation. Other
fMRI fear learning research has observed that the peak
amygdala response follows stimulus onset and then atten-
uates after several seconds of long-duration (e.g., 18 sec)
stimuli (Phelps et al., 2001). The CS duration in the pres-
ent study approximated the length of time during which
amygdala responses were maintained in prior fMRI work.
However, our analysis may be insensitive to more rapid
within-trial attenuation of amygdala activity. Future stud-
ies using long-duration CS periods should explore dynamic
within-trials changes of the amygdala response.
Research with laboratory animals indicates that the
amygdala is also important for extinction of conditional fear
(Baker & Azorlosa, 1996; Falls etal., 1992). Our data lend
support to this concept. During the first block of the extinc-
tion phase, right amygdala activity increased when ex-
tinction group subjects were presented the light alone. In
contrast, paired group subjects who continued to receive
the same CS–UCS pairings throughout both phases showed
no change in amygdala activation. Further, our data sug-
gest that this amygdala activity may have been related to
changes in stimulus relationships, as is seen when subjects
are switched from acquisition to extinction trials. Control
subjects who received unpaired training and then had UCS
presentations withheld in the final phase of the experiment
also showed an increase in right amygdala activity that
was similar to the response observed in extinction subjects.
By design, control group subjects should have no CR to
extinguish since the treatment they received does not sup-
port excitatory conditioning. Therefore, the transient amyg-
dala activation observed in control subjects cannot be gen-
erated by a process exclusively related to extinction of the
CR. An alternative interpretation is that the right amyg-
dala activity for both the control and extinction groups
was driven by the change in stimulus relationships that oc-
curred during the transition from the acquisition phase to
the extinction phase. This interpretation is consistent with
suggestions that the amygdala is involved in orienting and
attentional processes during Pavlovian conditioning and
may be involved in the establishment of new associations
following the violation of current expectancies (Fischer
et al., 2003; Holland & Gallagher, 1999; Holland et al.,
2000; Montag-Sallaz etal., 1999). Each of these processes
is likely to be present during a conditioning session and may
be important for forming CS–UCS associations. However,
it will be important for future studies to determine which
of these processes drive amygdala activity, so that we may
gain a better understanding of this region’s relative contri-
butions to the various processes involved in Pavlovian fear
A distinct pattern of left amygdala activity was also ob-
served in the present study. Beginning with onset of the
extinction phase, subjects who continued to receive paired
CS–UCS presentations showed greater left amygdala ac-
tivation than did subjects who received the CS alone (i.e.,
the extinction and control groups). This activation of the
left amygdala is consistent with previous fear learning
studies that have ascribed higher level cognitive processes
to this region. Activation of the left amygdala has been as-
sociated with the cognitive representation of fear acquired
through verbal instructions (Phelps etal., 2001), and other
work has suggested that left amygdala activation is asso-
ciated with awareness of the CS–UCS relationship (Morris,
Öhman, & Dolan, 1998). Interestingly, the pattern of left
amygdala activity obtained in those studies appeared to
mirror the pattern of activation within the left hippocampus
in the present study. Although left amygdala activity was
delayed (by one block) relative to the activation pattern
observed in the left hippocampus, this similarity in activa-
tion patterns may reflect the interaction of the amygdala
and hippocampus during the encoding of fear memories
(Richardson, Strange, & Dolan, 2004).
Differential hippocampal activity was also observed in
the present study. Paired group subjects showed greater
left hippocampal activation than did their controls by the
end of the acquisition phase. These differences were main-
tained through the first two blocks of the extinction phase.
In contrast, the participants who switched from paired
CS–UCS presentations to CS-alone trials (i.e., the extinc-
tion group) showed a significant decrease in hippocampal
activity during early trials of the extinction phase. Al-
though the hippocampus is not generally considered crit-
324 KNIGHT, SMITH, CHENG, STEIN, AND HELMSTETTER
ical for single-cue delay conditioning, the present results
support previous observations of hippocampal activity
during delay eyeblink and fear conditioning (Berger, Ri-
naldi, Weisz, & Thompson, 1983; Blaxton, Zeiffiro, &
Gabrieli, 1996; Knight etal., 2004; Ramnani, Toni, Josephs,
Ashburner, & Passingham, 2000; Weiss etal., 1996). Fur-
thermore, decreased hippocampal responsiveness, similar
to that seen in the extinction group in the present study,
has been previously demonstrated with behavioral extinc-
tion in animal models (Berger & Thompson, 1982; Segal,
The hippocampus has prominent reciprocal connec-
tions with the amygdala as well as cingulate, prefrontal,
and temporal cortical areas known to be important for cer-
tain forms of memory (Amaral & Cowan, 1980; Amaral &
Insausti, 1992). The hippocampus and medial temporal
cortex have been specifically implicated in explicit or de-
clarative memory processes (awareness or conscious rec-
ollection of facts and events), which are often contrasted
with implicit or nondeclarative memory (a skill, habit, or
other behavior that is unconsciously performed; see, e.g.,
Clark, Manns, & Squire, 2002; Milner, Squire, & Kandel,
1998; Squire, 1992). Left lateralization of hippocampal ac-
tivity, similar to that demonstrated in the present study, has
been previously observed in fMRI conditioning research
(Knight et al., 2004; Ramnani et al., 2000). Left hippo-
campal activation has been associated with declarative
memory processes that involve matching a stimulus to a
mnemonic representation (Iidaka et al., 2003). Further-
more, concurrent activation of the left hippocampus and
amygdala may reflect associative processing of emotional
material in declarative memory (Killgore, Casasanto,
Yurgelun-Todd, Maldjian, & Detre, 2000). Although suc-
cessful performance in single-cue Pavlovian conditioning
does not require conscious awareness of programmed con-
tingencies (e.g., Bechara etal., 1995; Clark & Squire, 1998),
declarative memory processes are clearly engaged in nor-
mal subjects during standard conditioning tasks (Knight
et al., 2004; Knight, Nguyen, & Bandettini, 2003). There-
fore, it is possible that the hippocampal activity seen in
the present study may reflect the development of declara-
tive knowledge of the CS–UCS relationship.
In sum, the present study explored amygdala and hippo-
campal activation during the acquisition and extinction of
human Pavlovian fear conditioning. These results high-
light the amygdala’s role in the short-term extinction of
conditional fear and suggest that region is particularly sen-
sitive to changes in environmental contingencies requiring
new learning. Our study focused explicitly on short-term
extinction processes that occur within a single session.
Recent studies suggest that long-term retention (?24h) of
extinction memory requires processes that occur within
the prefrontal cortex between training sessions (Milad &
Quirk, 2002; Quirk et al., 2000; Santini et al., 2001). Ad-
ditional imaging work is currently exploring the neural
mechanisms that support long-term extinction of condi-
tional fear (e.g., Richards, Cheng,Thomas, Smith, & Helm-
Amaral, D. G., & Cowan, W. M. (1980). Subcortical afferents to the
hippocampal formation in the monkey. Journal of Computational
Neurology, 189, 573-591.
Amaral, D. G., & Insausti, R.(1992). Retrograde transport of D-[3H]-
aspartate injected into the monkey amygdaloid complex. Experimen-
tal Brain Research, 88, 375-388.
Bailey, D. J., Kim, J. J., Sun, W., Thompson, R. F., & Helmstetter,
F. J. (1999). Acquisition of fear conditioning in rats requires the syn-
thesis of mRNA in the amygdala. BehavioralNeuroscience, 113, 276-
Baker, J. D., & Azorlosa, J. L. (1996). The NMDA antagonist MK-
801 blocks the extinction of Pavlovian fear conditioning. Behavioral
Neuroscience, 110, 618-620.
Bechara, A., Tranel, D., Damasio, H., Adolphs, R., Rockland, C.,
& Damasio, A. (1995). Double dissociation of conditioning and de-
clarative knowledge relative to the amygdala and hippocampus in hu-
mans. Science, 269, 1115-1118.
Berger, T. W., & Orr, W. B. (1983). Hippocampectomy selectively dis-
rupts discrimination reversal conditioning of the rabbit nictitating
membrane response. Behavioural Brain Research, 8, 49-68.
Berger, T. W., Rinaldi, P. C., Weisz, D. J., & Thompson, R. F. (1983).
Single-unit analysis of different hippocampal cell types during classi-
cal conditioning of rabbit nictitating membrane response. Journal of
Neurophysiology, 50, 1197-1219.
Berger, T. W., & Thompson, R. F. (1982). Hippocampal cellular plas-
ticity during extinction of classically conditioned nictitating mem-
brane behavior. Behavioural Brain Research, 4, 63-67.
Blaxton, T. A., Zeiffiro, T. A., & Gabrieli, J. D. E. (1996). Func-
tional mapping of human learning: A positron emission tomography
activation study of eyeblink conditioning. Journal of Neuroscience,
Büchel, C., Dolan, R. J., Armony, J. L., & Friston, K. J. (1999).
Amygdala–hippocampal involvement in human aversive trace condi-
tioning revealed through event-related functional magnetic resonance
imaging. Journal of Neuroscience, 19, 10869-10876.
Büchel, C., Morris, J., Dolan, R. J., & Friston, K. J. (1998). Brain
systems mediating aversive conditioning: An event-related fMRI study.
Neuron, 20, 947-957.
Cheng, D. T., Knight, D. C., Smith, C. N., Stein, E. A., & Helmstet-
ter, F. J. (2003). Functional MRI of human amygdala activity during
Pavlovian fear conditioning: Stimulus processing versus response ex-
pression. Behavioral Neuroscience, 117, 3-10.
Clark, R. E., Manns, J. R., & Squire, L. R. (2002). Classical condi-
tioning, awareness, and brain systems. Trends in Cognitive Sciences,
Clark, R. E., & Squire, L. R. (1998). Classical conditioning and brain
systems: The role of awareness. Science, 280, 77-81.
Cox, R. W., & Hyde, J. S.(1997). Software tools for the analysis and vi-
sualization of fMRI data. NMR in Biomedicine, 10, 171-178.
Davis, M. (2000). The role of the amygdala in conditioned and uncon-
ditioned fear and anxiety. In J. P. Aggleton (Ed.), The amygdala: A
functional analysis (2nd ed., pp. 213-287). New York: Oxford Univer-
Falls, W. A., Miserendino, M. J., & Davis, M. (1992). Extinction of
fear-potentiated startle: Blockade by infusion of an NMDA antago-
nist into the amygdala. Journal of Neuroscience, 12, 854-863.
Fischer, H., Wright, C. I., Whalen, P. J., McInerney, S. C., Shin,
L. M., & Rauch, S. L. (2003). Brain habituation during repeated ex-
posure to fearful and neutral faces: A functional MRI study. BrainRe-
search Bulletin, 59, 387-392.
Helmstetter, F. J., Tershner, S. A., Poore, L. H., & Bellgowan,
P. S. (1998). Antinociception following opioid stimulation of the ba-
solateral amygdala is expressed through the periaqueductal gray and
rostral ventromedial medulla. Brain Research, 779, 104-118.
Holland, P. C., & Gallagher, M. (1999). Amygdala circuitry in at-
tentional and representational processes. Trends in Cognitive Sci-
ences, 3, 65-73.
Holland, P. C., Han, J., & Gallagher, M. (2000). Lesions of the
fMRI AND FEAR CONDITIONING325 Download full-text
amygdala central nucleus alter performance on a selective attention
task. Journal of Neuroscience, 20, 6701-6706.
Iidaka, T., Terashima, S., Yamashita, K., Okada, T., Sadato, N., &
Yonekura, Y. (2003). Dissociable neural responses in the hippo-
campus to the retrieval of facial identity and emotion: An event-
related fMRI study. Hippocampus, 13, 429-436.
Killgore, W. D., Casasanto, D. J., Yurgelun-Todd, D. A., Mald-
jian, J. A., & Detre, J. A. (2000). Functional activation of the left
amygdala and hippocampus during associative encoding. Neuro-
Report, 11, 2259-2263.
Knight, D. C., Cheng, D. T., Smith, C. N., Stein, E. A., & Helm-
stetter, F. J. (2004). Neural substrates mediating human delay and
trace fear conditioning. Journal of Neuroscience, 24, 218-228.
Knight, D. C., Nguyen, H. T., & Bandettini, P. A. (2003). Expression
of conditional fear with and without awareness. Proceedings of the
National Academy of Sciences, 100, 15280-15283.
Knight, D.C., Smith, C. N., Stein, E. A., & Helmstetter, F. J.(1999).
Functional MRI of human Pavlovian fear conditioning: Patterns of ac-
tivation as a function of learning. NeuroReport, 10, 3665-3670.
LaBar, K. S., Gatenby, J. C., Gore, J. C., LeDoux, J. E., & Phelps,
E. A. (1998). Human amygdala activation during conditioned fear ac-
quisition and extinction: A mixed-trial study. Neuron, 20, 937-945.
LeDoux, J. (1995). Emotion: Clues from the brain. Annual Review of
Psychology, 46, 209-235.
Maren, S.(2001). Neurobiology of Pavlovian fear conditioning. Annual
Review of Neuroscience, 24, 897-931.
Meyers, K. M., & Davis, M. (2002). Behavioral and neural analysis of
extinction. Neuron, 26, 567-584.
Milad, M. R., & Quirk, J. (2002). Neurons in the medial prefrontal cor-
tex signal memory for fear extinction. Nature, 420, 70-74.
Milner, B., Squire, L. R., & Kandel, E. R. (1998). Cognitive neuro-
science and the study of memory. Neuron, 20, 445-468.
Montag-Sallaz, M., Welzl, H., Kuhl, D., Montag, D., & Schach-
ner, M. (1999). Novelty-induced increased expression of immediate–
early genes c-fos and arg 3.1 in the mouse brain. Journal of Neuro-
biology, 38, 234-246.
Morris, J. S., Öhman, A., & Dolan, R. J. (1998). Conscious and un-
conscious emotional learning in the human amygdala. Nature, 393,
Moyer, J. R., Deyo, R. A., & Disterhoft, J. F. (1990). Hippocampec-
tomy disrupts trace eye-blink conditioning in rabbits. Behavioral
Neuroscience, 104, 243-252.
Phelps, E. A., O’Connor, K. J., Gatenby, J. C., Gore, J. C., Gril-
lion, C., & Davis, M. (2001). Activation of the left amygdala to a
cognitive representation of fear. Nature Neuroscience, 4, 437-441.
Price, J. L., & Amaral, D. G. (1981). An autoradiographic study of the
projections of the central nucleus of the monkey amygdala. Journalof
Neuroscience, 1, 1242-1259.
Prokasy, W. F., & Raskin, D. C. (1973). Electrodermal activity in psy-
chological research. New York: Academic Press.
Quirk, G. J., Russo, G. K., Barron, J. L., & Lebron, K. (2000). The
role of ventromedial prefrontal cortex in the recovery of extinguished
fear. Journal of Neuroscience, 20, 6225-6231.
Ramnani, N., Toni, I., Josephs, O., Ashburner, J., & Passingham,
R.E.(2000). Learning- and expectation-related changes in the human
brain during motor learning. Journal of Neurophysiology, 84, 3026-
Rescorla, R.(1988). Behavioral studies of Pavlovian conditioning. An-
nual Review of Neuroscience, 11, 329-352.
Richards, J. A., Cheng, D. T., Thomas, T., Smith, C. N., & Helm-
stetter, F. J. (2003, November). Frontal lobe activity during acqui-
sition, extinction, and retrieval of fear conditioning. SocietyforNeuro-
science Abstracts. Available at http://apu.sfn.org/
Richardson, M. P., Strange, B. A., & Dolan, R. J. (2004). Encoding
of emotional memories depends on amygdala and hippocampus and
their interactions. Nature Neuroscience, 7, 278-285.
Santini, E., Muller, R. U., & Quirk, G. J. (2001). Consolidation of
extinction learning involves transfer from NMDA-independent to
NMDA-dependent memory. JournalofNeuroscience, 21, 9009-9017.
Segal, M.(1973). Flow of conditioned responses in limbic telencephalic
system of the rat. Journal of Neurophysiology, 36, 840-854.
Solomon, P. R., Vander Schaaf, E. R., Thompson, R. F., & Weisz,
D.J.(1986). Hippocampus and trace conditioning of the rabbit’s clas-
sically conditioned nictitating membrane response. BehavioralNeuro-
science, 100, 729-744.
Squire, L. R. (1992). Memory and the hippocampus: A synthesis from
findings with rats, monkeys, and humans. Psychological Review, 99,
Talairach, J., & Tournoux, P. (1988). Co-planar stereotaxic atlas of
the human brain: 3-dimensional proportional system. An approach to
cerebral imaging. Stuttgart: Thieme.
Weiss, C., Kronforst-Collins, M. A., & Disterhoft, J. F. (1996).
Activity of hippocampal pyramidal neurons during trace eyeblink con-
ditioning. Hippocampus, 6, 192-209.
Wolter, J., & Lachnit, H. (1993). Are anticipatory first and second
interval skin conductance responses indicators of predicted aversive-
ness? Integrative Physiological & Behavioral Science, 28, 163-166.
Wong, E. C., Tan, G., & Hyde, J. S. (1993). A quadrature transmit–
receive endcapped birdcage coil for imaging of the human head at
125MHz. Proceedingsofthe InternationalSocietyforMagneticRes-
onance in Medicine (12th Annual Meeting), p. 1344.
(Manuscript received January 6, 2004;
revision accepted for publication April 4, 2004.)