Psychopharmacology (2000) 152:376382

DOI 10.1007/s002130000557

O R I G I N A L I N V E S T I G AT I O N

Keith P. Corodimas Jennifer C. Pruitt

Jennifer M. Stieg

Acute exposure to caffeine selectively disrupts context conditioning

in rats
Received: 20 January 2000 / Accepted: 4 August 2000 / Published online: 12 October 2000
Springer-Verlag 2000

Abstract Rationale and objective: Acute caffeine administration has both beneficial and adverse effects on
learning and memory; however, the brain regions underlying these effects remain unclear. Several experiments
were conducted to examine the effects of acutely administered caffeine on the acquisition and expression of
hippocampal-dependent and hippocampal-independent
forms of conditioned fear. Methods: In the first experiment, caffeine (10, 20, or 30 mg/kg; IP) or vehicle was
administered to rats 15 min prior to classical fear conditioning, which consisted of ten tone-shock pairings.
Freezing to the conditioning context was measured 24 h
later, whereas tone-elicited fear was measured 48 h later.
A second experiment examined possible state-dependent
effects of caffeine by administering caffeine (30 mg/kg)
or vehicle 15 min before conditioning and before testing.
Results: Pretreatment of acute caffeine severely impaired
the acquisition of context conditioning, a hippocampaldependent task. Tone conditioning, a hippocampal-independent task, was only modestly and non-significantly
affected by caffeine (421% suppression compared with
controls). The disruption of context conditioning was
dose-dependent: 10 mg/kg had little effect on context or
tone conditioning, whereas doses of 20 and 30 mg/kg
caffeine severely disrupted context conditioning (7387%
suppression). In two subsequent experiments, it was
found that caffeines selective disruption of context conditioning could not be attributed to the fact that it is a
weaker form of learning than tone conditioning or to
state-dependent learning. Conclusions: Considered together, these results suggest that acute administration of
caffeine may preferentially disrupt the acquisition of hippocampal-dependent learning, including context conditioning.

K.P. Corodimas () J.C. Pruitt J.M. Stieg

School of Sciences, Psychology Program, Lynchburg College,
1501 Lakeside Drive, Lynchburg, VA 24501-3199, USA
e-mail: Corodimas_k@mail.lynchburg.edu
Tel.: +1-804-544-8666, Fax: +1-804-544-8499

Keywords Adenosine receptor Amygdala Caffeine

Context conditioning Fear Hippocampus

Introduction
The effects of caffeine on learning and memory in humans and laboratory animals have been well documented
(Nehlig et al. 1992; Buffalo et al. 1993; Angelucci et al.
1999; Kant 1993; Fisher and Guillet 1997). Despite this,
the specific cognitive processes affected by caffeine and
the brain regions underlying these effects are unclear.
Many of the behavioral effects of caffeine and other
methylxanthines are mediated by adenosine receptors
(Nehlig et al. 1992; Ledent 1997; Svenningsson et al.
1997), which have been divided into the A1 and A2
types. Both types of adenosine receptors are found in the
hippocampal formation (Fastbom et al. 1987a, 1987b;
Swanson et al. 1995; Johansson et al. 1997), a brain region that is essential for declarative or explicit memory
in humans (Parkin 1987; Squire 1987, 1993) and spatial/
contextual learning in many animals (Seldon et al. 1991;
Phillips and LeDoux 1992, 1994; Rudy 1993; Young et
al. 1994). Moreover, several studies have shown that
adenosine receptor antagonists alter hippocampal longterm potentiation (Arai et al. 1990; de Mendonca and
Ribeiro 1990; Sekino et al. 1991; Forghani and Krnjevic
1995; Kessey et al. 1997), an enduring form of synaptic
plasticity that is believed to underlie some forms of
learning and memory (Bliss and Collingridge 1993), including fear-related learning and memory (Maren 1999).
For example, direct application of specific adenosine receptor antagonists onto hippocampal neurons prevents
the induction of tetanus-induced long-term potentiation
(Sekino et al. 1991; Kessey et al. 1997). This raises the
possibility that caffeine, which acts as an antagonist at
all adenosine receptor types (Nehlig et al. 1992), may
impair hippocampal-dependent learning, possibly by altering long-term potentiation.
Several studies have demonstrated that pretraining lesions placed in the hippocampus impair the formation of

377

an association between complex contextual stimuli such

as the conditioning chamber and footshock, but do not
impair the formation of an association between a tone
and footshock (Seldon et al. 1991; Kim and Fanselow
1992; Phillips and LeDoux 1992, 1994; Rudy 1993;
Young et al. 1994). Similarly, temporary inactivation of
the hippocampus 24 h after conditioning impairs the expression of context-elicited fear (Holt and Maren 1999).
Animals with pretraining or post-training hippocampal
lesions, therefore, display severe deficits in the amount
of conditioned fear they exhibit to the environment in
which conditioning occurred, but normal conditioned
fear to the tone.
To test the hypothesis that caffeine may preferentially
interfere with hippocampal-dependent learning, we conducted several experiments to examine the effect of caffeine on classical fear conditioning. In a typical fear conditioning paradigm, an emotionally neutral stimulus,
such as a tone, is paired with an affectively charged aversive stimulus, usually footshock. The tone (conditioned
stimulus), because of its temporal relationship with the
shock (unconditioned stimulus), acquires aversive properties and thus is capable, by itself, of eliciting responses
characteristically elicited by threatening stimuli (Bolles
and Fanselow 1980; Bouton and Bolles 1980). Conditioned emotional responses are also elicited by the static
contextual or background stimuli that are present in the
conditioning chamber during conditioning (Blanchard
and Blanchard 1972; Rescorla and Wagner 1972; Bolles
and Fanselow 1980; Bouton and Bolles 1980; Bouton
and King 1983).

Materials and methods

Animals
The subjects were male Sprague-Dawley rats (Harlan, Indianapolis, Ind., USA) housed in pairs in polycarbonate cages
(48.226.720.3 cm) on a 12-h light-dark cycle with free access to
food and water. Rats weighed 400450 g at the time of behavioral
testing. Before all of the experiments, animals were handled once
daily for approximately 7 days. For all experiments, the number of
animals/group is presented in each figure legend. Animals were
housed and maintained in accordance with the 1996 NIH Principles of laboratory animal care.

posed to 7 days of continuous caffeine before classical fear conditioning began. Although plasma levels of caffeine were not measured in the present study, the following has been reported for
each pellet: 25 mg pellet=2.02.5 g/ml serum, 10 mg pellet
=800 ng to 1.0 g/ml serum, and 5 mg pellet=400500 ng/ml (Innovative Research of America). For each pellet, plasma levels of
caffeine are sustained for approximately 21 days.
Fear conditioning
Rats were trained and tested in individual rodent conditioning
chambers (Coulbourn Instruments, Allentown, PA, E1-10) enclosed in sound attenuating chambers (Coulbourn Instruments,
E10-20). The conditioning chamber had two Plexiglas and two
brushed aluminum walls. The conditioned stimulus was a 10-kHz,
70-dB tone presented through a speaker located on the side panel
of the conditioning chamber. During training, a single trial consisted of a 20-s tone that co-terminated with a 500-ms, 1.0-mA footshock delivered through a grid floor. The inter-trial interval was
120 s. The presentation of the conditioned stimulus and unconditioned stimulus was controlled by an IBM-type computer running
experiment control and data acquisition software (WinLinc,
Coulbourn Instruments).
During behavioral testing, which began 24 h after training,
freezing, used as a marker of conditioned fear, was measured in all
experiments while animals were exposed to the conditioning context for 60 s (context conditioning). The following day, the conditioning context was altered before testing tone-elicited fear to minimize freezing to the conditioning context. The context was altered by placing opaque, mottled (black/white) Formica sheets on
the grid floor and the transparent rear wall of the conditioning
chamber. Freezing was then measured in the altered context while
animals were exposed to a 60-s tone. Shock was never presented
during behavioral testing. Scoring of freezing was conducted with
a stopwatch by an experimenter who was unaware of the specific
treatment (vehicle, placebo, caffeine) that animals received
(Corodimas et al. 1994; Corodimas and LeDoux 1995).
Freezing is a species-specific defensive response involving a
stereotyped crouching position and the absence of all movement
except for those involved in respiration (Blanchard and Blanchard
1969). Studies have demonstrated that freezing to both the conditioning context and tone-conditioned stimulus are conditioned responses. For instance, animals exposed to shock in one context
freeze when re-exposed to the same context but not when exposed
to a novel context (Fanselow 1980, 1990). Moreover, the magnitude of freezing to a tone-conditioned stimulus is dependent upon
its explicit pairing with footshock (Rescorla 1968; Phillips and
LeDoux 1994). Finally, freezing is highly correlated with other
measures of conditioned fear, including autonomic and endocrine
(Bolles and Fanselow 1980; Bouton and Bolles 1980; LeDoux et
al. 1984).
Acute caffeine exposure experiments

Drugs
For the three acute caffeine experiments, anhydrous caffeine was
purchased from Sigma Chemical Co. (St Louis, Mo., USA) and
was dissolved in a sodium benzoate solution (37.5 mg/ml). Control animals received sodium benzoate vehicle. For the chronic
caffeine experiment, caffeine and placebo pellets were purchased
from Innovative Research of America (Sarasota, Fla., USA). Control animals received placebo pellets.
Surgery
Minor surgery was performed only on rats in experiment 4. Rats
were anesthetized with a mixture of ketamine (90 mg/kg, IP) and
xylazine (10 mg/kg, IP) and were subcutaneously implanted behind the neck with one of the following size pellets: 25 mg,
10 mg, 5 mg caffeine pellet or 25 mg placebo pellet. Rats were
placed back in their home cages following surgery and were ex-

The basic conditioning paradigm was as follows. On the day of

training, separate groups of rats were injected IP with sodium benzoate solution (vehicle) or caffeine, which was dissolved in sodium benzoate solution (see below for doses). Fifteen minutes later,
animals were placed into individual conditioning chambers where,
120 s later, they were presented with ten tone-shock pairings.
Twenty-four hours later, animals were placed in the same conditioning chambers and freezing to the conditioning context was
measured. The following day tone-elicited fear was measured as
described above.
Experiment 1: effect of increasing doses of caffeine
on the acquisition of context and tone conditioning
On the day of training, separate groups of animals (n=34) were injected IP with vehicle (37.5 mg/ml sodium benzoate solution) or
10, 20, 30 mg/kg caffeine. Fifteen minutes after injection, at a

378
time when plasma levels of caffeine are high (Wang and Lau
1998), animals were placed into individual conditioning chambers
and presented with ten tone-shock pairings. Conditioned fear responses were measured as previously described.
Experiment 2: role of state-dependent learning in the effect
of caffeine on conditioning
On the day of training, separate groups of animals (n=33) were injected with vehicle or 30 mg/kg caffeine then, 15 min later, presented with ten tone-shock pairings. Twenty-four hours later, half
of each group was injected with vehicle, whereas the other half
was injected with 30 mg/kg caffeine. This paradigm resulted in the
following groups: vehicle before training and before testing (vehicle-vehicle), vehicle before training and caffeine before testing
(vehicle-caffeine), caffeine before training and before testing (caffeine-caffeine, state-dependent learning control), and caffeine before training and vehicle before testing (caffeine-vehicle). Fifteen
minutes after the pre-testing injection of vehicle or caffeine, freezing responses to the original conditioning context were measured.
Twenty-four hours later, animals were once again given the same
pre-testing injection of caffeine or vehicle, then freezing responses
to the tone were measured.
Experiment 3: effect of caffeine on the acquisition of context
conditioning only
On the day of training, two groups of animals (n=12) were injected with vehicle or 30 mg/kg caffeine then, 15 min later, received
ten conditioning trials consisting of an unsignaled shock. The conditioning and testing parameters were identical to the previous two
experiments except no tone was presented during conditioning.

Fig. 1 Mean (SEM) duration of freezing to the conditioning context (open bars) and to the tone-conditioned stimulus (filled bars)
in animals injected with vehicle or increasing doses of caffeine
prior to classical fear conditioning. Freezing to the context was
significantly suppressed in animals receiving 20 (n=8) or 30 (n=8)
mg/kg caffeine compared with vehicle-injected controls (n=11).
Animals receiving 10 mg/kg (n=8) caffeine were not significantly
different from the remaining groups. *Significantly different from
freezing to the context in vehicle-injected animals
was a within-subjects factor. Significant drug-by-stimulus interactions were followed by separate one-way ANOVAs or t-tests performed on the context and tone data. Where appropriate, post hoc
comparisons were made using the Fisher protected least significant differences test (PLSD). All differences were considered significant when P<0.05.

Experiment 4: effect of continuous caffeine exposure

on the acquisition of context and tone conditioning

Results

Separate groups of rats were continuously exposed to caffeine

(three different size pellets) or placebo for 7 days (see above for
details). The following day, animals (n=40), with pellets still present, were placed into individual conditioning chambers where,
120 s later, they were presented with ten tone-shock pairings. Context- and tone-elicited fear were measured as previously described.

Experiment 1: effect of increasing doses of caffeine

on the acquisition of context and tone conditioning

Experiment 5: tone sensitization controls

After animals (n=24) were placed in individual conditioning
chambers, they received ten predictive conditioning trials or ten
non-predictive conditioning trials. A single predictive trial began
with a 120-s interval followed by a 20-s tone, which co-terminated
with a 500-ms shock. In contrast, a single non-predictive trial began with a 90-s interval, which co-terminated with a 500-ms
shock. The shock was followed by a second interval of 30 s then a
20-s tone. No conditioned or unconditioned stimuli were presented
during the aforementioned intervals (30 s, 90 s, 120 s). For both
predictive and non-predictive conditioning trials, the duration and
intensity of the tone and the shock were identical, as was the overall duration of each conditioning trial.
Experiment 6: habituation controls
After being placed in conditioning chambers, animals (n=8) received ten conditioning trials, each of which began with a 120-s
inter-trial interval and ended with a 20-s tone. Animals did not
receive any presentations of the shock.
Statistical analyses
Behavioral data were analyzed using a two-way analysis of variance (ANOVA), where drug (vehicle or placebo versus caffeine)
was a between-subjects factor, and stimulus (context versus tone)

To examine the effect of caffeine on hippocampal-dependent (context) versus hippocampal-independent (tone)

learning, animals were injected with vehicle or caffeine
before the presentation of ten tone-shock pairings and
then tested for conditioned fear to the context 24 h later.
The following day, tone-elicited fear was measured.
While pretraining administration of the lowest dose of
caffeine had little effect on either context or tone conditioning, pretraining administration of higher doses led to
marked reductions in context conditioning (Fig. 1).
A two-way ANOVA revealed significant effects of
stimulus [F(1,31)=326.33, P<0.0001], no significant effects of drug [F(3,31)=1.24, P>0.05], and a significant
stimulus by drug interaction [F(3,31)=3.64, P<0.02], reflecting the fact that caffeine had a greater effect on context conditioning than on tone conditioning. Post hoc
analysis of the context data demonstrated that animals
receiving 20 mg/kg and 30 mg/kg caffeine froze significantly less than vehicle treated animals (P<0.01). Animals receiving 10 mg/kg caffeine were not significantly
different from vehicle treated animals or animals receiving the 20 or 30 mg/kg doses of caffeine (P>0.05). Post
hoc analysis of the tone conditioning data did not reveal
any differences between groups.
To examine the effects of caffeine on context and tone
conditioning more directly, we calculated the percentage

379

change in freezing to the context and the tone in caffeine

treated animals relative to vehicle treated animals. Percentage change in freezing was calculated as follows:
mean of the vehicle control group minus freezing of individual caffeine treated animals divided by the mean of
the vehicle control group. The lowest dose of caffeine
had only a slight effect on context or tone conditioning
(producing a 27.92% suppression of context conditioning and a 9.01% suppression of tone conditioning). Pretraining administration of 20 or 30 mg/kg caffeine, on
the other hand, severely and selectively suppressed context conditioning by 73.16% and 74.70%, respectively.
In contrast, freezing responses to the tone in the
20 mg/kg animals were slightly suppressed (6.46%),
and in the 30 mg/kg animals they were slightly enhanced
(+6.01%). Importantly, paired t-tests revealed that for
each of these two doses, caffeine suppressed context
conditioning significantly greater than it suppressed tone
conditioning (P<0.0001 and P<0.001, respectively).
Experiment 2: role of state-dependent learning
in the effect of caffeine on conditioning
In the previous experiment, animals were administered
caffeine prior to training but not prior to testing. To test
whether caffeines suppression of context conditioning
was secondary to this change in state between training
and testing, and to examine the effects of caffeine on the
expression of conditioned fear, animals were administered 30 mg/kg caffeine or vehicle 15 min before training, as described in experiment 1. Twenty-four hours later, half of each group was administered the same dose
of caffeine (30 mg/kg) and the other half was given vehicle 15 min before testing context-elicited fear. The following day, animals received the same injection 15 min
before tone-elicited fear was measured. As expected, animals that received a pretraining dose of 30 mg/kg caffeine showed severe and selective deficits in context
conditioning regardless of whether animals were tested
in the same (C-C group) or opposite state (C-V group)
(Fig. 2). In contrast, tone conditioning was not adversely
effected by pretraining administration of 30 mg/kg caffeine. These results are consistent with experiment 1.
A two-way ANOVA revealed significant effects of
stimulus [F(1,29)=61.77, P<0.0001], significant effects
of drug [F(3,29)=5.56, P<0.004], and a significant stimulus by drug interaction [F(3,29)=4.31, P<0.01], reflecting, once again, the fact that caffeine had a greater effect
on context conditioning than on tone conditioning. Post
hoc analysis of the context data demonstrated that animals in the caffeine/vehicle group froze significantly less
than animals in either the vehicle/vehicle (P<0.0001) or
the vehicle/caffeine (P<0.007) groups, but not significantly less than animals in the caffeine/caffeine group
(P>0.05). Animals in the caffeine/caffeine group froze
significantly less to the context compared with animals
in the vehicle/vehicle (P<0.0001) and vehicle/caffeine
(P<0.001) groups. Finally, animals in the vehicle/caf-

Fig. 2 Mean (SEM) duration of freezing to the conditioning context (open bars) and to the tone-conditioned stimulus (filled bars)
in animals injected with vehicle or 30 mg/kg caffeine prior to
training and tested 24 and 48 h later following a pre-testing injection of vehicle or 30 mg/kg caffeine. Pretraining administration of
caffeine severely suppressed freezing to the context in animals
given caffeine or vehicle before testing (C-C/n=8, C-V/n=8) compared with animals receiving vehicle before training and either vehicle or caffeine before testing (V-V/n=8, V-C/n=9). Animals receiving vehicle before training and caffeine before testing (V-C)
froze less to the conditioning context than animals receiving vehicle before training and before testing (V-V). **Significantly different from contextual freezing in V-V and V-C animals. *Significantly different from contextual freezing in V-V animals

feine group froze significantly less than animals in the

vehicle/vehicle group (P<0.03). In contrast, there were
no significant differences in freezing to the tone. Importantly, since pretraining administration of caffeine severely suppressed context conditioning regardless of the
animals drug state before training and testing, it is very
unlikely that caffeine-induced deficits in context conditioning are secondary to state-dependent learning.
As expected, the percentage suppression of freezing
to the context in caffeine/caffeine and caffeine/vehicle
treated animals (relative to vehicle/vehicle animals) was
severe, producing an 86.78% and 76.43% suppression,
respectively. In contrast, tone conditioning was relatively
unaffected by caffeine administration in the C-C and
C-V groups (21.16% and 4.09% suppression, respectively). Importantly, paired t-tests revealed that for each of
these two groups, caffeine suppressed context conditioning significantly more that it suppressed tone conditioning (P<0.001 and P<0.0001, respectively). These results
are consistent with those in experiment 1. Finally, the
suppression of freezing to the context in both the caffeine/caffeine (86.78%) and caffeine/vehicle (76.43%)
groups was significantly greater (P<0.006) than the suppression of freezing in the vehicle/caffeine group
(33.15%).
Experiment 3: effect of caffeine on the acquisition
of context conditioning only
In the first two experiments, conditioning to the context
was weaker than conditioning to the tone, which is consistent with other studies, and is thought to be due to the

380

Fig. 3 Mean (SEM) duration of freezing to the conditioning context in animals injected with vehicle (n=6) or 30 mg/kg (n=6) caffeine prior to ten presentations of the unconditioned stimulus only.
*Significantly different from freezing to the context in vehicleinjected animals

fact that the tone is more salient (discrete/unisensory)

and a better predictor of the shock because it is phasically presented only immediately prior to the shocks presentation (Rescorla 1968). This raises the possibility that
caffeine disrupted context conditioning because it is a
weaker form of learning. To test this alternative hypothesis, animals were injected with vehicle or 30 mg/kg caffeine 15 min before the presentation of ten unsignaled
shocks. The conditioning and testing parameters were
identical to our previous experiments except no tone was
presented during conditioning. Our objective was to
bring the conditioning context to the foreground so that
freezing to the context would be comparable to freezing
to the tone in our previous two experiments. Indeed, as
depicted in Fig. 3, the magnitude of context conditioning
in vehicle treated controls receiving ten unsignaled
shocks was very similar to the magnitude of tone conditioning in our first two experiments (35 s versus 40 s, respectively). Despite this, an unpaired t-test revealed that
caffeine pretreatment severely impaired the acquisition
of context conditioning (nearly five-fold decrement versus controls) [t(10)=3.741; P<0.0001]. As expected, the
percentage suppression of freezing to the context in caffeine treated animals relative to controls was similar to
our first two experiments (79.6%). The fact that context
conditioning was markedly suppressed by 30 mg/kg caffeine even though it was made comparably strong to tone
conditioning in our previous two experiments, suggests
that caffeine did not preferentially disrupt context conditioning because it is a weaker form of learning.
Experiment 4: effect of continuous caffeine exposure
on the acquisition of context and tone conditioning
The first three experiments examined the effects of an
acute, single injection of caffeine on fear conditioning.
To examine the effect of chronic caffeine exposure on

Fig. 4 Mean (SEM) duration of freezing to the conditioning context (open bars) and to the tone-conditioned stimulus (filled bars)
in animals (n=10/group) exposed to placebo or caffeine (25 mg
pellet, 10 mg pellet, or 5 mg pellet) for 7 continuous days. No significant differences were found in the freezing responses of animals to the tone or the context

context and tone conditioning, animals were subcutaneously implanted with placebo or caffeine pellets 7 days
before the presentation of ten tone-shock pairings. Context and tone conditioning were measured over the next
48 h, as described above. The caffeine or placebo pellets
remained implanted during conditioning and testing. Unlike the previous three experiments that found a severe
and selective disruption of context conditioning following acute caffeine, 7 days of continuous caffeine exposure produced only a modest, non-significant effect on
context conditioning (Fig. 4).
Experiment 5: tone sensitization controls
To test whether sensitization was occurring to the tone
the following groups were compared: animals that received ten conditioning trials in which the tone predicted
the shock (temporally paired) versus animals that received ten conditioning trials in which the tone did not
predict the shock (temporally unpaired).
Unpaired t-tests revealed that animals receiving ten
unpaired tone-shock trials froze significantly less to the
tone compared with animals receiving ten paired conditioning trials [t(14)=5.078; P<0.0001]. Context conditioning was not significantly different between the two
groups [t(14)=0.341; P>0.05; Table 1].
Experiment 6: habituation controls
The goal of this experiment was to determine if habituation was occurring in the present study. Habituation is a
reduced responsiveness in animals to stimuli that have
been repeatedly presented. To test whether habituation
was occurring we compared animals that received ten
tone-shock pairings with animals that received ten conditioning trials of the tone only.
As shown in Table 1, unpaired t-tests revealed that
animals receiving ten tone only trials froze significantly
less to the tone [t(14)=4.951; P<0.0001] and to the conditioning context [t(14)=7.054; P<0.0001] compared
with animals receiving ten tone-shock trials.

381
Table 1 Mean (SEM) duration of freezing to the conditioning
context and to the tone in animals (n=8/group) given ten paired or
unpaired tone-shock trials, or ten tone only trials. Animals that received ten unpaired conditioning trials froze significantly less to
the tone than animals that received paired trials. Animals that received ten tone only trials froze significantly less to the conditioning context and tone than animals that received paired CS-US
trials
Group

10 paired trials
10 unpaired trials
10 tone only trials

Context

Tone

Mean secondsSEM

Mean secondsSEM

32.634.42
34.382.59
1.230.64*

34.133.15
13.132.68*
7.254.42*

*Significantly different from the duration of freezing of animals

receiving paired conditioning trials

Discussion
The first experiment demonstrated that acute administration of caffeine severely disrupted, in a dose-dependent
fashion, the acquisition of context conditioning, a hippocampal-dependent task (Seldon et al. 1991; Kim and
Fanselow 1992; Phillips and LeDoux 1992, 1994; Rudy
1993; Young et al. 1994). This finding, which was replicated in the second and third experiments, is consistent
with studies demonstrating adverse effects of caffeine on
other hippocampal-dependent behaviors including avoidance learning (Castellano et al. 1973; Sansone et al.
1994; Fisher and Guillet 1997; Angelucci et al. 1999),
performance on place conditioning (Brockwell et al.
1991), and performance on the Morris water maze (Kant
1993). While our results also indicate that the expression
of context conditioning was impaired in animals receiving caffeine before behavioral testing, the acquisition of
context conditioning was much more severely impaired.
In contrast to its severe effects on context conditioning,
caffeine produced only a modest, non-significant suppression of tone conditioning in the same animals. Our
results also suggest that sensitization and habituation
were not occurring to the tone-conditioned stimulus.
However, it is possible that non-associate processes were
induced by caffeine in the present study.
In all of the present experiments, freezing to the context was measured first, then 24 h later, freezing to the
tone was measured in the same conditioning chamber.
Although the context was altered prior to the measurement of tone conditioning, it is possible that the presence
of contextual stimuli influenced freezing to the tone.
Since we were expecting to see a selective suppression
of caffeine on context conditioning, any effect of contextual stimuli on freezing responses to the tone works
against rather than for the hypothesis being tested. Moreover, the order in which context and tone conditioning
was measured also works against our hypothesis. However, because we did not counterbalance the order in
which context and tone conditioning were tested, we
cannot completely rule out the possible influence of the
order of behavioral testing.

It is difficult to attribute caffeines selective disruption of context conditioning to changes in motivational

state (i.e., shock sensitivity) or to state-dependent learning. The fact that caffeine severely suppressed context
conditioning, but had only modest and non-significant
effects on tone conditioning in the same animals, suggests that shock sensitivity was not significantly altered
by caffeine pretreatment. Moreover, because caffeine severely suppressed context conditioning, but only had
modest effects on tone conditioning in the same animals
argues against state-dependent learning. Despite this, we
examined the issue of state-dependency directly by administering vehicle or caffeine before training and then
tested animals in the same or the opposite drug state. We
did not observe any state-dependent effects of caffeine
on fear conditioning to the context or tone.
Caffeines preferential effect of context conditioning
is also unlikely, due to the fact that context conditioning
is weaker than tone conditioning. If this is true, then caffeine should not disrupt context conditioning if it is
made comparably strong to tone conditioning. In our
third experiment, context conditioning was strengthened
by presenting animals with ten unsignaled unconditioned
stimuli. Importantly, despite the fact that context conditioning was strong and made comparable to tone conditioning in our first two experiments, caffeine, once
again, severely disrupted context conditioning.
Although acute administration of caffeine severely
impaired context conditioning, 7 days of continuous caffeine exposure resulted only in a modest and non-significant impairment of context conditioning. This observed
difference is consistent with studies showing that acute
and chronic administration of caffeine often have very
different effects on the same behavior (Nelhig et al.
1992). It is possible that the differences observed in the
present study could be due to the well-documented
changes in the number of adenosine receptors during
chronic caffeine exposure. Indeed, chronic caffeine treatment is correlated with significant changes in the density
of adensosine receptors in a variety of brain regions, including the hippocampus and lateral nucleus of the
amygdala (Guillet and Kellogg 1991; Johansson et al.
1997; Svenningsson et al. 1999), two areas that play an
essential role in the acquisition and the expression of
fear-related learning (LeDoux 1998). We also cannot
rule out the possibilty that a longer exposure to caffeine
would have disrupted context-related learning.
Our results are consistent with the hypothesis that caffeine impairs context conditioning, possibly by acting on
adenosine receptors in the hippocampus. Future studies,
however, will be needed to address this issue more directly by applying adenosine receptor agonistic and antagonistic drugs into the dorsal hippocampal formation
before and after classical fear conditioning.
Acknowledgements We thank Dr. Thomas A. Looney (Lynchburg College) and Dr. Elinore C. Lau (Rutgers University) for
their helpful suggestions and comments. This research was supported, in part, by a research grant from The Whitehall Foundation
(K.P.C.).

382

References
Angelucci ME, Vital MA, Cesario C, Zadusky CR, Rosalen PL,
Da Cunha C (1999) The effect of caffeine in animal models of
learning and memory. Eur J Pharmacol 373:135140
Arai A, Kessler M, Lynch G (1990) The effects of adenosine on
the development of long-term potentiation. Neurosci Lett 119:
4144
Blanchard DC, Blanchard RJ (1969) Crouching as an index of
fear. J Comp Physiol Psychol 67:370375
Blanchard DC, Blanchard RJ (1972) Innate and conditioned reactions to threat in rats with amygdaloid lesions. J Comp Physiol
Psychol 81:281290
Bliss TV, Collingridge GL (1993) A synaptic model of memory:
long-term potentiation in the hippocampus. Nature 361:3139
Bolles R, Fanselow M (1980) A perceptual-defensive-recuperative
model of fear and pain. Behav Brain Sci 3:291323
Bouton M, Bolles R (1980) Conditioned fear assessed by freezing
and by the suppression of three different baselines. Anim
Learn Behav 8:429434
Bouton ME, King DA (1983) Contextual control of the extinction
of conditioned fear: tests for the associative value of the context. J Exp Psychol Anim Behav Process 9:248265
Brockwell NT, Eikelboom R, Beninger RJ (1991) Caffeine-induced
place and taste conditioning: production of dose-dependent preference and aversion. Pharmacol Biochem Behav 38:513517
Buffalo EA, Gillam MP, Allen RR, Paule MG (1993) Acute effects of caffeine on several operant behaviors in rhesus monkeys. Pharmacol Biochem Behav 46:733737
Castellano C, Sansone M, Renzi P, Annecker L (1973) Central
stimulant drugs on avoidance behavior in hamsters. Pharmacol
Res Commun 5:287293
Corodimas KP, LeDoux JE (1995) Disruptive effects of posttraining perirhinal cortex lesions on conditioned fear: contributions
of contextual cues. Behav Neurosci 109:613619
Corodimas KP, LeDoux JE, Gold PW, Schulkin J (1994) Corticosterone potentiation of conditioned fear in rats. Ann NY Acad
Sci 746:392393
de Mendonca A, Ribeiro JA (1990) 2-Chloroadenosine decreases
long-term potentiation in the hippocampal CA1 area of the rat.
Neurosci Lett 118:107111
Fanselow MS (1980) Conditioned and unconditional components
of post-shock freezing. Pavlov J Biol Sci 15:177182
Fanselow M (1990) Factors governing one-trial contextual conditioning. Anim Learn Behav 15:177182
Fastbom J, Pazos A, Palacios JM (1987a) The distribution of
adenosine A1 receptors and 5-nucleotidase in the brain of
some commonly used experimental animals. Neuroscience 22:
813826
Fastbom J, Pazos A, Probst A, Palacios JM (1987b) Adenosine A1
receptors in the human brain:a quantitative autoradiographic
study. Neuroscience 22:827839
Fisher S, Guillet R (1997) Neonatal caffeine alters passive avoidance retention in rats in an age- and gender-related manner.
Brain Res Dev Brain Res 98:145149
Forghani R, Krnjevic K (1995) Adenosine antagonists have differential effects on induction of long-term potentiation in hippocampal slices. Hippocampus 5:7177
Guillet R, Kellogg C (1991) Neonatal exposure to therapeutic caffeine alters the ontogeny of adenosine A1 receptors in brain of
rats. Neuropharmacology 30:489496
Holt W, Maren S (1999) Muscimol inactivation of the dorsal hippocampus impairs contextual retrieval of fear memory. J Neurosci 19:90549062
Johansson B, Georgiev V, Lindstrom K, Fredholm BB (1997) A1
and A2A adenosine receptors and A1 mRNA in mouse brain:
effect of long-term caffeine treatment. Brain Res 762:153164
Kant G (1993) Effects of psychoactive drugs or stress on learning,
memory, and performance as assessed using a novel water
maze task. Pharmacol Biochem Behav 44:287295
Kessey K, Trommer BL, Overstreet LS, Ji T, Mogul DJ (1997) A
role for adenosine A2 receptors in the induction of long-term

potentiation in the CA1 region of rat hippocampus. Brain Res

756:184190
Kim JJ, Fanselow MS (1992) Modality-specific retrograde amnesia of fear. Science 256:675677
Ledent C VJ, Schiffmann SN, Pedrazzini T, Vassart G, Parmentier
M (1997) Aggressiveness, hypoalgesia and high blood pressure in mice lacking the adensosine A2A receptor. Nature
388:674678
LeDoux JE (1996) The emotional brain. Simon and Schuster, New
York
LeDoux JE, Sakaguchi A, Reis DJ (1984) Subcortical efferent
projections of the medial geniculate nucleus mediate emotional responses conditioned to acoustic stimuli. J Neurosci 4:683
698
Maren S (1999) Long-term potentiation in the amygdala: a mechanism for emotional learning and memory. Trends Neurosci
22:561567
Nehlig A, Daval JL, Debry G (1992) Caffeine and the central nervous system: mechanisms of action, biochemical, metabolic
and psychostimulant effects. Brain Res Brain Res Rev 17:
139170
Parkin A (1987) Memory and amnesia. Blackwell, Oxford
Phillips RG, LeDoux JE (1992) Differential contribution of amygdala and hippocampus to cued and contextual fear conditioning. Behav Neurosci 106:274285
Phillips RG, LeDoux JE (1994) Lesions of the dorsal hippocampal
formation interfere with background but not foreground contextual fear conditioning. Learn Mem 1:3444
Rescorla RA (1968) Probability of shock in the presence and absence of CS in fear conditioning. J Comp Physiol Psychol 66:
15
Rescorla R, Wagner A (1972) A theory of Pavlovian conditioning:
variations in the effectiveness of reinforcement and nonreinforcement. Appleton-Century-Crofts, New York
Rudy JW (1993) Contextual conditioning and auditory cue conditioning dissociate during development. Behav Neurosci 107:
887891
Sansone M, Battaglia M, Castellano C (1994) Effect of caffeine
and nicotine on avoidance learning in mice: lack of interaction. J Pharm Pharmacol 46:765767
Sekino Y, Ito K, Miyakawa H, Kato H, Kuroda Y (1991) Adenosine (A2) antagonist inhibits induction of long-term potentiation of evoked synaptic potentials but not of the population
spike in hippocampal CA1 neurons. Biochem Biophys Res
Commun 181:10101014
Seldon N, Everitt B, Jarrard L, Robbins T (1991) Complimentary
roles for the amygdala and hippocampus in aversive conditioning to explicit and contextual cues. Neuroscience 42:
335350
Squire L (1987) Memory and brain. Oxford University Press, New
York
Squire L (1993) The structure and organization of memory. Annu
Rev Psychol 44:453495
Svenningsson P, Nomikos GG, Ongini E, Fredholm BB (1997)
Antagonism of adenosine A2A receptors underlies the behavioural activating effect of caffeine and is associated with reduced expression of messenger RNA for NGFI-A and NGFI-B
in caudate-putamen and nucleus accumbens. Neuroscience
79:753764
Svenningsson P, Nomikos GG, Fredholm BB (1999) The stimulatory action and the development of tolerance to caffeine is associated with alterations in gene expression in specific brain
regions. J Neurosci 19:40114022
Swanson TH, Drazba JA, Rivkees SA (1995) Adenosine A1 receptors are located predominantly on axons in the rat hippocampal formation. J Comp Neurol 363:517531
Wang Y, Lau CE (1998) Caffeine has similar pharmacokinetics
and behavioral effects via the i.p. and p.o. routes of administration. Pharmacol Biochem Behav 60:271278
Young SL, Bohenek DL, Fanselow MS (1994) NMDA processes
mediate anterograde amnesia of contextual fear conditioning
induced by hippocampal damage:immunization against amnesia by context preexposure. Behav Neurosci 108:1929