European Journal of Neuroscience, Vol. 22, pp.

1775–1783, 2005 ª Federation of European Neuroscience Societies

Fear learning induces persistent facilitation of amygdala

synaptic transmission

Bradley W. Schroeder and Patricia Shinnick-Gallagher

Department of Pharmacology and Toxicology, Neuroscience Graduate Program, The University of Texas Medical Branch,
301 University Blvd, Galveston, TX 77555–1031, USA

Keywords: anxiety, fear conditioning, long-term potentiation (LTP), memory and learning, rat, synaptic plasticity

Abstract
In the maintenance phase of fear memory, synaptic transmission is potentiated and the stimulus requirements and signalling
mechanisms are altered for long-term potentiation (LTP) in the cortico-lateral amygdala (LA) pathway. These findings link amygdala
synaptic plasticity to the coding of fear memories. Behavioural experiments suggest that the amygdala serves to store long-term fear
memories. Here we provide electrophysiological evidence showing that synaptic alterations in rats induced by fear conditioning are
evident in vitro 10 days after fear conditioning. We show that synaptic transmission was facilitated and that high-frequency
stimulation dependent LTP (HFS–LTP) of the cortico-lateral amygdala pathway remained attenuated 10 days following fear
conditioning. Additionally, we found that the low-frequency stimulation dependent LTP (LFS–LTP) measured 24 h after fear
conditioning was absent 10 days post-training. The persistent facilitation of synaptic transmission and occlusion of HFS–LTP
suggests that, unlike hippocampal coding of contextual fear memory, the cortico-lateral amygdala synapse is involved in the storage
of long-term fear memories. However, the absence of LFS–LTP 10 days following fear conditioning suggests that amygdala
physiology 1 day following fear learning may reflect a dynamic state during memory stabilization that is inactive during the long-term
storage of fear memory. Results from these experiments have significant implications regarding the locus of storage for maladaptive
fear memories and the synaptic alterations induced by these memories.

Introduction
The amygdala is critically involved in the formation of emotional
memories. The formation of fear memory is accompanied by increases
in synaptic strength (McKernan & Shinnick-Gallagher, 1997; Rogan
et al., 1997) and modifications in synaptic plasticity (Tsvetkov et al.,
2002; Schroeder & Shinnick-Gallagher, 2004) in afferent pathways to
the cortico-lateral amygdala (LA). Experimental manipulations that
block amygdala plasticity inhibit the formation and expression of fear
memory (Huang et al., 2000; Schafe et al., 2000; Fendt, 2001; Bauer
et al., 2002; Rodrigues et al., 2002) suggesting that these alterations in
synaptic physiology are necessary for acquisition of fear memories.
More specifically, cued fear memories occlude long-term potentiation
(LTP; Tsvetkov et al., 2002) and alter the stimulus requirements and
signalling mechanisms for LTP in the cortico-LA pathway (Schroeder
& Shinnick-Gallagher, 2004) suggesting fear memories encompass
overlapping and saturating mechanisms similar to those of underlying
amygdala synaptic plasticity.
The role of the amygdala in the storage of long-term memory is
somewhat controversial. Studies analysing inhibitory avoidance
suggest that the amygdala may play a time defined role in
consolidation of fear memory (Cahill et al., 1999; Vazdarjanova,
2000) whereas fear-conditioning studies showed little evidence for a
temporal gradient regarding the involvement of the amygdala in the
expression of fear memories (Chapman et al., 1990; Cahill et al.,
1999; Maren, 2000; Vazdarjanova, 2000). Recent experiments lasting
Correspondence: Professor Patricia Shinnick-Gallagher, as above.
Email: psgallag@utmb.edu
Received 24 February 2005, revised 25 July 2005, accepted 27 July 2005
the lifetime of an adult rat also indicated a role for the amygdala in the
permanent storage of fear memories (Gale et al., 2004). One
complicating factor in behavioural studies may be the effects of
experimental manipulations on the behavioural performance of the
animal (Maren, 1999; Vazdarjanova et al., 2001; Wallace & Rosen,
2001). Here we examined whether storage of fear memories in the
amygdala is accompanied by transient or long-lasting alterations in
synaptic mechanisms that are not dependent on recording of animal
performance.
To examine electrophysiological evidence of whether fear memories
are stored long-term in the amygdala, we examined the effect of fear
conditioning on amygdala synaptic efficacy and plasticity 10 days
following fear learning. We show that synaptic transmission and
efficacy was enhanced in the amygdala 10 days following fear
conditioning and that high-frequency stimulation dependent LTP
(HFS–LTP) of the cortico-LA pathway remained attenuated, while
low-frequency stimulation dependent LTP (LFS–LTP) recorded 24 h
after fear conditioning was absent 10 days post-training. The persist-
ent synaptic facilitation and occlusion of HFS–LTP suggested that the
cortico-LA synapse was involved in the long-term storage of fear
memory.
Materials and methods
Animals
Male albino Sprague–Dawley rats (virus free, Harlan, Houston, TX,
USA) 4–7-weeks-old were used. All rats were acclimated in an
isolated animal facility with food and water available ad libitum for a

doi:10.1111/j.1460-9568.2005.04343.x
1776 B. W. Schroeder and P. Shinnick-Gallagher

minimum of five days. Animals were routinely handled prior to use in
electrophysiological or behavioural experiments. All experiments
were performed with prior approval from the University of Texas
Medical Branch Institutional Animal Care and Use Committee.
Fear conditioning
Animals were fear-conditioned using a fear-potentiated startle para-
digm (Fig. 1A) adapted from (Campeau & Davis, 1995) as described
previously (Shinnick-Gallagher et al., 2003; Zinebi et al., 2003). Prior
to training, rats were acclimated to the experimental chambers and
subsequently trained and tested in enclosures attached to a stabilime-
ter ⁄ accelerometer device (San Diego Instruments, San Diego, CA,
USA). Animal movement displaced an accelerometer and the
magnitude of the displacement was converted to a voltage and stored
on a desktop computer. The acoustic startle response was elicited
using a white noise burst of 95 db. Startle amplitude was defined as
peak displacement within 200 ms after the onset of the startle
stimulus. A 72 db filtered, white noise tone of 2.7 s duration was used
as the conditioned stimulus (CS) and paired ten times a day for 2 days
with the unconditioned stimulus (US), a 0.5 mA co-terminating foot
shock of 0.5 s duration for the fear-conditioned group. Unpaired (UP)
controls received the same number of CS and US presentations in
pseudorandom fashion. On the third day, the rats were tested using ten
startle stimuli to habituate the animals followed by ten startle stimuli
alone randomly interspersed with ten startle stimuli paired with the
CS. Fear-potentiated startle was measured as a change in startle
magnitude elicited by CS-startle stimulus pairing compared to that
elicited by the startle stimulus alone. Testing sessions were performed
in separate chambers to control for effects of contextual conditioning.
Amygdala slices were prepared from fear-conditioned (FC1) and
unpaired (UP1) animals 24 h after the testing for use in the 1-day
electrophysiological experiments. Animals used in the 10-day elec-
trophysiological experiments (FC10 and UP10) were returned to the
animal housing facility following testing where they remained
undisturbed for 10 days.
Amygdala slice preparation
Rats were decapitated, and brains rapidly removed and placed in ice-
cold artificial cerebrospinal fluid (ACSF) solution bubbled continu-
ously with 95% O2 and 5% CO2 to maintain appropriate pH (7.4–
7.45). Anaesthetics were not used prior to this procedure because they
can mask changes in emotional learning. Brains were blocked,
mounted, and sliced into 500-lm serial coronal sections containing the
LA. Slices were placed in oxygenated ACSF and allowed to recover at
room temperature for at least two hours. Slices were submerged in a
0.8 mL tissue bath (Harvard Instruments, Holliston, MA, USA)
continuously perfused with oxygenated ACSF (2–3 mL ⁄ min) and
maintained at 33 ± 1
C. The composition of the ACSF was (in
millimolar): NaCl, 117; KCl, 3.0; CaCl2, 2.5; MgCl2, 1.2; NaHCO2,
25; NaH2PO4, 1.2; and glucose, 11.5.
Extracellular field recording
Extracellular field potentials were recorded using a tungsten electrode
(2–5MW) and amplified using a Microelectrode AC Amplifier (A-M
Systems, Carlsborg, WA, USA). Bipolar concentric stimulating

Fig. 1. Diagram of preparation and time frame for experimental paradigm and fear-potentiated startle in the different animal groups. (A) Diagram of slice
preparation taken from Paxinos & Watson (1998) with stimulation and recording sites indicated. (B) In vivo fear training and testing occurred on similar days while
in vitro recordings were compared 1 and 10 days after fear testing. (C) Startle responses were significantly altered in fear-conditioned animals subsequently analysed
electrophysiologically in vitro 1 and 10 days post-testing. The mean ± SEM of the startle stimulus alone (startle alone), CS-startle stimulus pairing (tone + startle),
and the difference of the two values are plotted for fear-conditioned (FC1) and unpaired (UP1) animal groups 1 day (n ¼ 13 per group) and fear-conditioned (FC10)
and unpaired (UP10) groups 10 days (n ¼ 14 per group) post-testing. *P < 0.05 relative to startle in FC; ^P < 0.05 relative to difference in UP.

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 22, 1775–1783

electrodes (50–75KW) were placed on the external capsule, and
orthodromic synaptic potentials were elicited with an Isolated Pulse
Stimulator (A-M Systems, Carlsborg, WA, USA). While the external
capsule also includes fibers originating but not terminating in the LA,
the conventional term ‘cortico-LA pathway’ was used here to describe
dorsolateral LA field potentials elicited by external capsule stimula-
tion. Input–output (IO) relationships were assessed by stimulating
afferents with increasing levels of intensity (100-ls biphasic pulses,
typically ranging from 3 to 11 V) while recording the evoked field
excitatory postsynaptic potentials (fEPSPs) in the LA. Field potentials
were included for study if the maximum amplitude was above 0.8 mV
(without spikes). For analysis of synaptic plasticity, the stimulus
intensity was adjusted to 33–50% of maximum fEPSP amplitude.
Experimenters were blinded to the animal groups and slices from
experimental and control animals were recorded simultaneously in the
same chamber to control for day-to-day variation. Stimulation
protocols were generated and off-line analyses of fEPSP recordings
performed using pClamp 9.0 (Axon Instruments, Foster City, CA,
USA). Averaged traces of six successive sweeps were used for
analysis of fEPSP amplitude and 10–90% slope. No significant
differences were found between the two measurements.
Statistical analysis
One slice per treatment per animal was used for analyses such that n
represents the number of animals. Data were presented as mean ± -
SEM. A nonlinear regression analysis was used to compare the IO
relationships for control and fear-conditioned animals. Differences
between groups and treatment conditions were assessed using one-
way and two-way anovas. When the overall F ratio was significant,
Tukey posthoc tests were used to compare individual means.
Statistical significance was defined as P < 0.05.
Results
Fear potentiates startle
Animals were fear-conditioned using a fear-potentiated startle para-
digm adapted from Campeau & Davis (1995). The fear-conditioned
group was exposed to ten unconditioned stimuli (US, 0.5 mA, 0.5 ms
foot shock) paired with ten conditioned stimuli (CS, 2.7 s tone) per
day for 2 days and learned that CS presentation predicts subsequent
US presentation. The unpaired control groups received ten CS and ten
US exposures per day for 2 days presented in pseudorandom fashion
and, consequently, the unpaired animals did not learn the CS–US
association. Fear potentiation of startle was measured on the third day
by recording startle amplitude in response to the startle stimulus in the
presence and absence of the CS (Fig. 1). As a result of fear learning,
CS presentation in the absence of the foot shock elicited a potentiated
startle response. Startle values for the startle stimulus alone, the startle
stimulus in the presence of the CS, and the subtracted value of those
measurements were significantly different in fear-conditioned and
unpaired groups in both 1 day and 10 day animal populations
(Fig. 1B). Startle amplitudes for the three measures, startle stimulus
alone, startle stimulus plus CS, and the difference of the two values
(142.2 ± 9.8, 237.8 ± 18.7, and 95.6 ± 11.5, respectively) for the
1 day fear-conditioned group (FC1) showed a significant effect of CS
presentation on startle magnitude (67% potentiation of startle,
P < 0.05 between startle alone and startle plus CS). In 1 day unpaired
group (UP1), startle values for the three measures (147.6 ± 21.6,
151.2 ± 24.3, and 3.7 ± 5.9, respectively) showed no significant effect
of CS presentation (P > 0.05 between startle alone and startle plus
ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 22, 1775–1783
Fear learning and persistent synaptic facilitation 1777
CS). Training also did not significantly affect startle response to the
startle stimulus alone in the 1 day populations (P > 0.05, between
FC1 and UP1 groups), but startle in the presence of the CS showed a
significant group effect (P < 0.05, for subtracted values between FC1
and UP1 groups).
Fear-potentiated startle amplitudes were determined in the animal
groups whose brains were subsequently analysed in vitro 10 days after
testing (Fig. 1B). Responses to the startle stimulus in the presence and
absence of the CS, and the difference in these two values were
118.3 ± 14.3, 191.7 ± 19, and 73.4 ± 7, respectively, for the fear-
conditioned group (FC10) and 123.6 ± 15.6, 131.2 ± 16.5, and
7.6 ± 4.8, respectively, for the unpaired group (UP10). Startle
amplitudes were potentiated 62% above baseline values (P < 0.05)
in the FC10 group. There was no significant effect of the CS on startle
magnitude in the UP10 population (P > 0.05) and no significant
difference in baseline response to startle alone for the FC10 and UP10
groups (P > 0.05).
Analysis of startle values revealed no significant difference between
FC1 and FC10 groups for startle stimulus alone, conditioned tone plus
startle stimulus, and the difference of the two values (P > 0.05).
Additionally, there was no significant difference in startle values
between UP1 and UP10 groups. These data indicated that fear-
potentiated startle responses were not different in the 1 and 10 days
post-training groups, which suggested that the degree of fear learning
did not differ in the animal populations analysed electrophysiologi-
cally at the two time points.
Fear memory formation facilitates cortico-LA synaptic
transmission
Synaptic strength was increased at the thalamo-LA pathway (McKernan
& Shinnick-Gallagher, 1997) and LTP of the cortico-lateral amygdala
synapse was attenuated 24 h after fear conditioning (McKernan &
Shinnick-Gallagher, 1997; Tsvetkov et al., 2002; Schroeder & Shinnick-
Gallagher, 2004). These findings suggested that emotional learning
facilitated synaptic transmission at the cortico-LA synapse through LTP-
like mechanisms, but the attenuation of LTP may be independent of a
change in synaptic strength. To analyse synaptic efficacy of cortico-LA
synapses, we evaluated IO relationships by recording fEPSP slope as a
function of increasing stimulus intensity in slices from naı̈ve, UP1, and
FC1 animals (Fig. 2A). Non-linear regression analysis of IO relation-
ships revealed no significant differences between the naı̈ve and UP1
groups (P > 0.05) whereas the IO relationship for FC1 animals was
significantly different than that in the control groups (Fig. 2A,
P < 0.0001). These data suggested that fear memory strengthens
synaptic transmission at the cortico-LA synapse.
Persistent synaptic facilitation in the storage of fear memories
To determine the persistence of the increase in synaptic strength, we
examined the IO relationship at the cortico-LA synapse 10 days after
testing. Synaptic transmission was recorded in response to afferent
stimulation at increasing stimulus intensities for naı̈ve, UP10, and
FC10 groups (Fig. 2B). Non-linear regression analysis of the IO
relationships for naı̈ve and UP10 animals revealed no significant
differences in synaptic strength (Fig. 2B, P > 0.05). However, the IO
relationship for the FC10 group showed a significant difference when
compared to naı̈ve and UP10 groups (P < 0.0001). These data
indicated that synaptic strength remained facilitated 10 days after fear
learning, suggesting that enhanced cortico-LA synaptic efficacy may
play a role in the long-term storage of fear memory.
1778 B. W. Schroeder and P. Shinnick-Gallagher

Fig. 2. Fear learning facilitated synaptic transmission in the cortico-LA pathway 1 day post-testing and this potentiation persisted 10 days after fear conditioning.
(A) Traces of fEPSPs were recorded at increasing stimulus intensities in slices from naı̈ve control rats and from unpaired and fear-conditioned animals one day post-
testing. (B) fEPSPs were measured in slices from naı̈ve control and from unpaired and fear-conditioned animals 10 days post-testing. (C) Synaptic strength (fEPSP
slope) is plotted as a function of increasing afferent stimulation intensity in slices from naı̈ve control (d), unpaired (UP1, s) control, and fear-conditioned (FC1, .)
animals. The input–output (IO) relationship for synaptic transmission was shifted upward in FC1 animals, P < 0.0001 vs. control, nonlinear regression analysis. (D)
The upward shift in the IO relationship for synaptic transmission remained in slices from fear-conditioned animals analysed 10 days post-testing. Synaptic strength
(fEPSP slope) is plotted as a function of increasing afferent stimulation intensity for naı̈ve (d), unpaired (UP10, s), and fear-conditioned (FC10, .) animals,
P < 0.0001 vs. control, linear regression analysis.

Presynaptic involvement in cortico-LA synaptic facilitation
Paired-pulse ratio (PPR) is a measure of short-term plasticity widely
used to probe for changes in presynaptic function. In this paradigm,
pairs of fEPSPs are elicited by two temporally close stimuli. The effect
of the first stimulus on the second can be used to infer changes in
probability of transmitter release from the presynaptic terminal. To
examine whether the increased synaptic strength recorded 1 day after
fear conditioning is mediated by a presynaptic component, we
analysed PPRs for naı̈ve, UP1, and FC1 groups (Fig. 3). Ratios of
the slope of the second fEPSP to the slope of the first fEPSP were
examined at varying interpulse intervals (Fig. 3A) in slices from
control and fear-conditioned animals. No significant differences were
measured between naı̈ve and UP1 controls, however, analysis of PPRs
revealed a significant depression in the FC1 group at 25, 35, and
50 ms intervals (naı̈ve, n ¼ 8; UP1, n ¼ 9 and FC1, n ¼ 12
P < 0.05). The PPR depression measured in the FC1 group suggested
that the synaptic facilitation generated by fear learning is mediated at
least in part by an increase in presynaptic release probability.
Enhanced probability of transmitter release in fear memory
To examine whether the change in PPR recorded 1 day after fear
conditioning is persistent, we recorded cortico-LA synapses 10 days
after fear conditioning (Fig. 3). Analysis of PPRs revealed no
significant differences between naı̈ve and UP10 control groups, while
PPRs from the FC10 group were significantly depressed at 25, 35, and
50 ms interpulse intervals (naı̈ve, n ¼ 8; UP10, n ¼ 8 and FC10,
n ¼ 11, P < 0.05, Fig. 3B). These data suggested that the modulation
of presynaptic function induced by fear memory formation persisted in
the long-term storage of fear memory.
HFS-LTP remains attenuated 10 days after fear conditioning
At cortico-LA synapses HFS-induced LTP is attenuated 1 day after
fear conditioning (Fig. 4A; Schroeder & Shinnick-Gallagher, 2004).
To investigate whether fear-induced alterations in synaptic plasticity at
the cortico-LA synapse were enduring, we examined the effects of

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 22, 1775–1783
 Fear learning and persistent synaptic facilitation 1779

Fig. 3. Alterations in PPR in fear-conditioned animals that were recorded 1 day post-testing persisted 10 days post-testing. (A) Pairs of fEPSPs were recorded at
25 ms intervals in slices from naive control and from unpaired and fear-conditioned groups 1 day after fear testing. (B) fEPSP pairs were measured at 25 ms
intervals in naı̈ve, unpaired and fear-conditioned groups 10 days post-testing. (C) PPRs (fEPSP2 ⁄ fEPSP1) are plotted as a function of varying interpulse intervals in
naı̈ve (d), unpaired (UP1, s), and fear-conditioned (FC1, .) animals (n ¼ 8 per group for naı̈ve and FC1, n ¼ 7 for UP1). *P < 0.05 for FC1 relative to control
groups. (D) PPRs (fEPSP2 ⁄ fEPSP1) are plotted as a function of varying interpulse intervals in naı̈ve (d), unpaired (UP10, s), and fear-conditioned (FC10, .)
animals (n ¼ 8 per group for naı̈ve and FC10, n ¼ 7 for UP10). *P < 0.05 for FC10 compared to control groups.

HFS (100 Hz for 1 s, repeated five times at 3 min intervals) on
synaptic transmission 10 days following training (Fig. 4B). HFS in
slices from naı̈ve and UP10 animals elicited LTP (naive
135.5% ± 3.6%; UP 130.7% ± 4.6%; P > 0.05 between groups)
while HFS–LTP in slices from FC10 animals remained attenuated
(109.3% ± 2.8%, P < 0.001 between fear-conditioned and controls;
Fig. 4A). These data suggested that the attenuation of HFS–LTP was
associated with the long-term storage of fear memory.
LFS–LTP is absent 10 days after training
In addition to attenuation of HFS–LTP, we previously described a low-
frequency stimulation (LFS) dependent form of LTP (LFS–LTP) that
is measured at the cortico-LA synapse 24 h after fear testing (Fig. 5A;
Schroeder & Shinnick-Gallagher, 2004). We examined whether this
form of plasticity is present 10 days after fear conditioning (Fig. 4B).
LFS 1 Hz for 15 min elicited no significant change in fEPSPs in slices
from naı̈ve (n ¼ 8), UP10 (n ¼ 7), and FC10 (n ¼ 10) animals
(98.8 ± 4.7, 103.3 ± 4.9, and 103.8 ± 3.2, respectively, P > 0.05,
Fig. 4B). These data suggest that the potentiation of synaptic strength
in response to LFS in the period shortly after fear learning does not
persist in long-term fear memory.
Discussion
The major findings of this study are: (i) fear learning induced synaptic
facilitation at the cortico-LA synapse that was expressed in part
through an enhanced presynaptic function; (ii) the increased synaptic
weight and presynaptic enhancement recorded 1 day after fear
learning persisted for 10 days and (iii) HFS–LTP remained attenuated
10 days after fear conditioning, but LFS–LTP measured in slices from
fear-conditioned animals 1 day after training was not present 10 days
following fear learning.
Fear learning facilitates synaptic transmission
Previous studies have demonstrated synaptic facilitation at thalamic
sensory input pathways to the LA following fear conditioning
(McKernan & Shinnick-Gallagher, 1997; Rogan et al., 1997). Other
studies in the cortico-LA pathway, showed that HFS-induced and
pairing-induced LTP are occluded 1 day after fear conditioning
(Tsvetkov et al., 2002; Schroeder & Shinnick-Gallagher, 2004). One
hypothesis is that fear learning induces a long-lasting synaptic
facilitation at the cortico-LA pathway and that the occlusion of
HFS–LTP recorded after fear conditioning is due to saturation of
signalling mechanisms necessary to support LTP (Tsvetkov et al.,
2002). Although definitive proof showing that synaptic facilitation
underlies the attenuated HFS–LTP in fear conditioning requires
extensive occlusion studies (Lledo et al., 1995; Schulz & Fitzgibbons,
1997), the data presented here provide evidence that cortico-LA
synapses are facilitated in fear memory and could contribute to
attenuation of HFS–LTP.
Paired-pulse facilitation is a form of short-term plasticity that is
widely believed to rely on enhanced presynaptic function (Schulz
et al., 1994; Zinebi et al., 2001; Zucker & Regehr, 2002). In this
paradigm, facilitation is thought to be due to a residual Ca++ in the

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 22, 1775–1783
1780 B. W. Schroeder and P. Shinnick-Gallagher

Fig. 4. Long-term storage of fear memory was associated with occlusion of HFS–LTP. (A) Plot of previous data (Schroeder & Shinnick-Gallagher, 2004) showing
occlusion of HFS–LTP 1 day after fear training. Per cent potentiation of fEPSPs following HFS in slices is graphed with respect to time from naı̈ve (d), UP (s), and
FC (.) animals. (B, upper) fEPSPs before (baseline) and after HFS–LTP stimulation in slices from naı̈ve, unpaired (UP10) and fear-conditioned (FC10) animals. (B,
lower) Graph of percentage potentiation of fEPSPs following HFS is plotted with respect to time in slices from naı̈ve (d), UP10 (s), and FC10 (.) animals (n ¼ 9
animals per group). *1-day data as in (Schroeder & Shinnick-Gallagher, 2004).

presynaptic terminal that has not been cleared following the first action
potential and as a result second action potential is augmented,
resulting in a greater probability of transmitter release and facilitated
synaptic transmission. Conversely, if presynaptic function is already
enhanced the ratio of the second to the first event could be reduced
either through depletion of the readily releasable pool of vesicles,
activation of presynaptic autoreceptors, or decreased by desensitiza-
tion of postsynaptic receptors initiated by the first synaptic event
(Zucker & Regehr, 2002). The data presented here demonstrated PPRs
were depressed in the fear conditioning group 1 day after fear
conditioning, suggesting that the increase in synaptic strength
recorded 1 day following fear testing is due, at least in part, to
enhanced presynaptic function. The end result of an enhanced
presynaptic function is an increase in transmitter release for a given
stimulus input, which is reflected as an increase in the slope of the
input–output relationship of the fEPSP. These data support previous
findings showing that enhanced presynaptic function is associated
with the maintenance phase of fear memory (Tsvetkov et al., 2002)
and the expression of LTP in the cortico-LA pathway (Huang &
Kandel, 1998) as well as thalamo-LA pathway (McKernan &
Shinnick-Gallagher, 1997). These findings provide an additional
mechanistic link between the synaptic changes induced by fear
learning and those underlying LTP.
Enduring synaptic facilitation in long-term fear memory
The increased synaptic weight and enhanced presynaptic function
recorded 1 day after fear conditioning persisted 10 days after testing.
Similar synaptic potentiation has been described at the Schaffer
collateral-CA1 synapse in the hippocampus following contextual fear
conditioning but hippocampal synaptic facilitation lasts only 7 days
following training (Sacchetti et al., 2001). These electrophysiological
changes are thought to reflect a dynamic process by which memory is
initially formed and coded by changes in hippocampal synaptic
strength and then conveyed to remote cortical areas for long-term
storage (Kim & Fanselow, 1992; Squire & Alvarez, 1995; Anagnost-
aras et al., 2001; Nader, 2003). Conversely, synaptic facilitation in the
amygdala is present for at least 11 days after fear conditioning
(10 days post-testing). It is possible that this enduring facilitation of
synaptic transmission provides a synaptic trace for the long-term
storage of emotional memories in the amygdala.
HFS-LTP remains occluded
In addition to alterations in synaptic efficacy and short-term
plasticity, HFS–LTP was attenuated 10 days after fear conditioning
further suggesting that the cortico-LA synapse is involved in the
long-term storage of fear memory. Similar occlusion of LTP has
been described at the Schaffer–CA1 synapse in animals that have
been subjected to hippocampal-dependent learning tasks (Sacchetti
et al., 2002). In that study, HFS–LTP was attenuated immediately
following learning; however, complete reinstatement of synaptic
plasticity occurred within 7 days following training. The persistent
occlusion of amygdala HFS–LTP 11 days after cued fear condi-
tioning suggests that the cued fear engram in the amygdala may not
undergo a similar form of systems consolidation as in the
hippocampus (Debiec et al., 2002) where the memory trace is
exported to other brain areas for permanent storage. Rather, this

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 22, 1775–1783
 Fear learning and persistent synaptic facilitation 1781

Fig. 5. LFS–LTP was expressed at 1 but not 10 days after fear conditioning. (A) Plot of previous data (Schroeder & Shinnick-Gallagher, 2004) showing occlusion
of LFS–LTP 1 day after fear training. Per cent potentiation of fEPSPs following LFS in slices is plotted with respect to time from naı̈ve (d), UP (s), and FC (.)
animals. (B, upper) fEPSPs before (baseline) and after LFS–LTP stimulation in slices from naı̈ve, unpaired (UP10) and fear-conditioned (FC10) animals. (B, lower)
Graph of percentage potentiation of fEPSPs following LFS is plotted with respect to time in slices from naı̈ve (d), UP10 (s), and FC10 (.) animals (n ¼ 9 animals
per group). *1-day data as in (Schroeder & Shinnick-Gallagher, 2004).

suggests that the cortico-LA synapse may play a role in storage of
long-term fear memory.
LFS–LTP is absent 10 days following fear conditioning
The LFS–LTP recorded in amygdala slices from fear-conditioned
animals 1 day after testing was not present 10 days following fear
conditioning. Current theory suggests that reactivated memories
undergo reconsolidation during which they become labile and
susceptible to protein synthesis inhibitors (Nader et al., 2000). It is
possible that the difference in response to LFS 1 and 10 days after
fear conditioning may be related to the temporal proximity of the
recall of the fear memory and electrophysiological examination of
amygdala function. The LFS–LTP may represent a dynamic state
associated with reconsolidation 1 day post-testing but not in long-
term storage of a stabilized fear memory. Testing this hypothesis
requires definitive experimentation but the present data showing an
absence of LFS–LTP 10 days following fear conditioning suggest
that amygdala physiology in the period shortly after fear testing
may differ from that in the long-term storage of fear memory. It is
also possible that testing itself may induce extinction training and
that LFS may in fact be related to extinction rather than a
consolidation mechanism. However, previous studies have shown
that extinction of fear-potentiated startle requires approximately 60
presentations of the CS alone (Davis & Astrachan, 1978; Falls
et al., 1992) suggesting that ten presentations of tone during testing
may exert a relatively small influence on our responses.
LFS is analogous to slow, delta wave sleep of 1–4 Hz that plays a
key role in the consolidation and maintenance of multiple forms of
memory (Kenny et al., 2003; Higgins et al., 2004). In rats, lateral
amygdala projection neurons increase firing during delta wave sleep
(Pare & Gaudreau, 1996), and amygdala inactivation during sleep
periods following emotional learning leads to impaired memory
retention (Rasmussen et al., 2004). Although the mechanisms
underlying the transient nature of this event and its relationship to
memory consolidation ⁄ reconsolidation requires further analyses, these
data raise the possibility that LFS–LTP recorded 1 day but not 10 days
after fear conditioning may reflect an active consolidation ⁄ reconsol-
idation process of fear memory within the amygdala during sleep,
which in rats coincides with our recording time frame.
Taken together, the present data suggest that altered amygdala
synaptic physiology is not only associated with the formation of fear
memories but also likely plays an important role in the maintenance
and long-term storage of fear memories. Brain systems outside the
amygdala such as the thalamus and cortex also show neural changes
with fear conditioning (Weinberger, 1997; Maren et al., 2001), but
nearly all changes depend on the amygdala (Quirk et al., 1997;
Armony et al., 1998; Maren et al., 2001; Pare et al., 2002). A temporal
gradient for auditory fear responses is not evident after post-training
inhibition or lesioning of the amygdala suggesting auditory fear
memories are stored in the amygdala (Lee et al., 1996; Maren et al.,
1996). Additionally, a recent behavioural data showed that the
amygdala was essential for conditioned freezing behaviour over the
lifetime of the animal whether lesioned 1 day or over 1 years after

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 22, 1775–1783
1782 B. W. Schroeder and P. Shinnick-Gallagher
training (Gale et al., 2004). Our data provide direct electrophysiolog-
ical evidence to support a role for the amygdala in the long-term
storage and expression of fear memory.
These findings have significant implications as the locus of storage
for fear memories and the synaptic alterations induced by these
memories may guide the development of pharmacological and
behavioural treatment modalities for anxiety disorders. It has been
recently suggested that patients with learned anxiety disorders, such as
post-traumatic stress disorder, suffer from an inability to form effective
extinction engrams (Rothbaum & Davis, 2003). This study identifies
modifications of amygdala synaptic physiology that may serve as
potential therapeutic targets for intervention in the treatment of learned
anxiety disorders in patients with established disease.
Acknowledgements
We would like to thank Drs Joel Gallagher, Geoff Swanson, and Sebastian
Pollandt and Mr Michael Scott and Ms Julie Ross for their helpful reviews.
Supported by MH58327 and F30 MH067458.
Abbreviations
CS, conditioned stimulus; FC, fear-conditioned; fEPSPs, field excitatory
postsynaptic potentials; HFS–LTP, high-frequency stimulation dependent LTP;
IO, input–output; LA, lateral amygdala; LTP, long-term potentiation; LFS–LTP,
low-frequency stimulation dependent LTP; PPR, paired-pulse ratio; US,
unconditioned stimulus; UP, unpaired.
References
Anagnostaras, S.G., Gale, G.D. & Fanselow, M.S. (2001) Hippocampus and
contextual fear conditioning: recent controversies and advances. Hippocam-
pus, 11, 8–17.
Armony, J.L., Quirk, G.J. & Ledoux, J.E. (1998) Differential effects of
amygdala lesions on early and late plastic components of auditory cortex
spike trains during fear conditioning. J. Neurosci., 18, 2592–2601.
Bauer, E.P., Schafe, G.E. & Ledoux, J.E. (2002) NMDA receptors and L-type
voltage-gated calcium channels contribute to long-term potentiation and
different components of fear memory formation in the lateral amygdala.
J. Neurosci., 22, 5239–5249.
Cahill, L., Weinberger, N.M., Roozendaal, B. & Mcgaugh, J.L. (1999) Is the
amygdala a locus of ‘conditioned fear’? Some questions and caveats.
Neuron, 23, 227–228.
Campeau, S. & Davis, M. (1995) Involvement of subcortical and cortical
afferents to the lateral nucleus of the amygdala in fear conditioning measured
with fear-potentiated startle in rats trained concurrently with auditory and
visual conditioned stimuli. J. Neurosci., 15, 2312–2327.
Chapman, P.F., Kairiss, E.W., Keenan, C.L. & Brown, T.H. (1990) Long-term
synaptic potentiation in the amygdala. Synapse, 6, 271–278.
Davis, M. & Astrachan, D.I. (1978) Conditioned fear and startle magnitude:
effects of different footshock or backshock intensities used in training.
J. Exp. Psychol. Anim. Behav. Process, 4, 95–103.
Debiec, J., Ledoux, J.E. & Nader, K. (2002) Cellular and systems
reconsolidation in the hippocampus. Neuron, 36, 527–538.
Falls, W.A., Miserendino, M.J. & Davis, M. (1992) Extinction of fear-
potentiated startle: blockade by infusion of an NMDA antagonist into the
amygdala. J. Neurosci., 12, 854–863.
Fendt, M. (2001) Injections of the NMDA receptor antagonist aminophospho-
nopentanoic acid into the lateral nucleus of the amygdala block the
expression of fear-potentiated startle and freezing. J. Neurosci., 21, 4111–
4115.
Gale, G.D., Anagnostaras, S.G., Godsil, B.P., Mitchell, S., Nozawa, T., Sage,
J.R., Wiltgen, B. & Fanselow, M.S. (2004) Role of the basolateral amygdala
in the storage of fear memories across the adult lifetime of rats. J. Neurosci.,
24, 3810–3815.
Higgins, G.A., Ballard, T.M., Kew, J.N., Richards, J.G., Kemp, J.A., Adam, G.,
Woltering, T., Nakanishi, S. & Mutel, V. (2004) Pharmacological manipula-
tion of mGlu2 receptors influences cognitive performance in the rodent.
Neuropharmacology, 46, 907–917.
ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 22, 1775–1783
Huang, Y.Y. & Kandel, E.R. (1998) Postsynaptic induction and
PKA-dependent expression of LTP in the lateral amygdala. Neuron, 21,
169–178.
Huang, Y.Y., Martin, K.C. & Kandel, E.R. (2000) Both protein kinase A and
mitogen-activated protein kinase are required in the amygdala for the
macromolecular synthesis-dependent late phase of long-term potentiation.
J. Neurosci., 20, 6317–6325.
Kenny, P.J., Gasparini, F. & Markou, A. (2003) Group II metabotropic and
alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionate (AMPA) ⁄ kainate
glutamate receptors regulate the deficit in brain reward function associated
with nicotine withdrawal in rats. J. Pharmacol. Exp. Ther., 306, 1068–
1076.
Kim, J.J. & Fanselow, M.S. (1992) Modality-specific retrograde amnesia of
fear. Science, 256, 675–677.
Lee, Y., Walker, D. & Davis, M. (1996) Lack of a temporal gradient of
retrograde amnesia following NMDA-induced lesions of the basolateral
amygdala assessed with the fear-potentiated startle paradigm. Behav.
Neurosci., 110, 836–839.
Lledo, P.M., Hjelmstad, G.O., Mukherji, S., Soderling, T.R., Malenka, R.C. &
Nicoll, R.A. (1995) Calcium ⁄ calmodulin-dependent kinase II and long-term
potentiation enhance synaptic transmission by the same mechanism. Proc.
Natl Acad. Sci. USA, 92, 11175–11179.
Maren, S. (1999) Neurotoxic basolateral amygdala lesions impair learning and
memory but not the performance of conditional fear in rats. J. Neurosci., 19,
8696–8703.
Maren, S. (2000) Reply to Vazdarjanova. TINS, 23, 345–346.
Maren, S., Aharonov, G. & Fanselow, M.S. (1996) Retrograde abolition of
conditional fear after excitotoxic lesions in the basolateral amygdala of rats:
absence of a temporal gradient. Behav. Neurosci., 110, 718–726.
Maren, S., Yap, S.A. & Goosens, K.A. (2001) The amygdala is essential for the
development of neuronal plasticity in the medial geniculate nucleus during
auditory fear conditioning in rats. J. Neurosci., 21, RC135.
Mckernan, M.G. & Shinnick-Gallagher, P. (1997) Fear conditioning induces a
lasting potentiation of synaptic currents in vitro. Nature, 390, 607–611.
Nader, K. (2003) Memory traces unbound. TINS, 26, 65–72.
Nader, K., Schafe, G.E. & Ledoux, J.E. (2000) The labile nature of
consolidation theory. Nature Rev. Neurosci., 1, 216–219.
Pare, D., Collins, D.R. & Pelletier, J.G. (2002) Amygdala oscillations and the
consolidation of emotional memories. Trends Cogn. Sci., 6, 306–314.
Pare, D. & Gaudreau, H. (1996) Projection cells and interneurons of the
lateral and basolateral amygdala: distinct firing patterns and differential
relation to theta and delta rhythms in conscious cats. J. Neurosci., 16,
3334–3350.
Paxinos, G. & Watson, C. (1998) The Rat Brain in Stereotaxic Coordinates.
Academic Press, San Diego, CA.
Quirk, G.J., Armony, J.L. & Ledoux, J.E. (1997) Fear conditioning enhances
different temporal components of tone-evoked spike trains in auditory cortex
and lateral amygdala. Neuron, 19, 613–624.
Rasmussen, K., Hsu, M.A. & Vandergriff, J. (2004) The selective mGlu2 ⁄ 3
receptor antagonist LY341495 exacerbates behavioral signs of morphine
withdrawal and morphine-withdrawal-induced activation of locus coeruleus
neurons. Neuropharmacology, 46, 620–628.
Rodrigues, S.M., Bauer, E.P., Farb, C.R., Schafe, G.E. & Ledoux, J.E. (2002)
The group I metabotropic glutamate receptor mGluR5 is required for fear
memory formation and long-term potentiation in the lateral amygdala.
J. Neurosci., 22, 5219–5229.
Rogan, M.T., Staubli, U.V. & Ledoux, J.E. (1997) Fear conditioning induces
associative long-term potentiation in the amygdala. Nature, 390, 604–
607.
Rothbaum, B.O. & Davis, M. (2003) Applying learning principles to the
treatment of post-trauma reactions. Ann. N.Y. Acad. Sci., 1008, 112–121.
Sacchetti, B., Lorenzini, C.A., Baldi, E., Bucherelli, C., Roberto, M., Tassoni,
G. & Brunelli, M. (2001) Long-lasting hippocampal potentiation and
contextual memory consolidation. Eur. J. Neurosci., 13, 2291–2298.
Sacchetti, B., Lorenzini, C.A., Baldi, E., Bucherelli, C., Roberto, M., Tassoni,
G. & Brunelli, M. (2002) Time-dependent inhibition of hippocampal LTP in
vitro following contextual fear conditioning in the rat. Eur. J. Neurosci., 15,
143–150.
Schafe, G.E., Atkins, C.M., Swank, M.W., Bauer, E.P., Sweatt, J.D. & Ledoux,
J.E. (2000) Activation of ERK ⁄ MAP kinase in the amygdala is required for
memory consolidation of pavlovian fear conditioning. J. Neurosci., 20,
8177–8187.
Schroeder, B.W. & Shinnick-Gallagher, P. (2004) Fear memories induce a
switch in stimulus response and signaling mechanisms for long-term
potentiation in the lateral amygdala. Eur. J. Neurosci., 20, 549–556.

Schulz, P.E., Cook, E.P. & Johnston, D. (1994) Changes in paired-pulse
facilitation suggest presynaptic involvement in long-term potentiation. J.
Neurosci., 14, 5325–5337.
Schulz, P.E. & Fitzgibbons, J.C. (1997) Differing mechanisms of expression for
short- and long-term potentiation. J. Neurophysiol., 78, 321–334.
Shinnick-Gallagher, P., Mckernan, M.G., Xie, J. & Zinebi, F. (2003)
L-type voltage-gated calcium channels are involved in the in vivo and
in vitro expression of fear conditioning. Ann. N.Y. Acad. Sci., 985, 135–
149.
Squire, L.R. & Alvarez, P. (1995) Retrograde amnesia and memory con-
solidation: a neurobiological perspective. Curr. Opin. Neurobiol., 5,
169–177.
Tsvetkov, E., Carlezon, W.A., Benes, F.M., Kandel, E.R. & Bolshakov, V.Y.
(2002) Fear conditioning occludes LTP-induced presynaptic enhancement of
synaptic transmission in the cortical pathway to the lateral amygdala.
Neuron, 34, 289–300.
Vazdarjanova, A. (2000) Does the basolateral amygdala store memories for
emotional events? TINS, 23, 345–346.
ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 22, 1775–1783
Fear learning and persistent synaptic facilitation 1783
Vazdarjanova, A., Cahill, L. & Mcgaugh, J.L. (2001) Disrupting basolateral
amygdala function impairs unconditioned freezing and avoidance in rats.
Eur. J. Neurosci., 14, 709–718.
Wallace, K.J. & Rosen, J.B. (2001) Neurotoxic lesions of the lateral nucleus of
the amygdala decrease conditioned fear but not unconditioned fear of a
predator odor: comparison with electrolytic lesions. J. Neurosci., 21, 3619–
3627.
Weinberger, N. (1997) Learning-induced receptive field plasticity in the
primary auditory cortex. Sem. Neurosci., 9, 59–67.
Zinebi, F., Russell, R.T., Mckernan, M. & Shinnick-Gallagher, P. (2001)
Comparison of paired-pulse facilitation of AMPA and NMDA synaptic
currents in the lateral amygdala. Synapse, 42, 115–127.
Zinebi, F., Xie, J., Liu, J., Russell, R.T., Gallagher, J.P., Mckernan, M.G. &
Shinnick-Gallagher, P. (2003) NMDA currents and receptor protein are
downregulated in the amygdala during maintenance of fear memory.
J. Neurosci., 23, 10283–10291.
Zucker, R.S. & Regehr, W.G. (2002) Short-term synaptic plasticity. Annu. Rev.
Physiol., 64, 355–405.