The Ontogeny of Fear-Potentiated Startle: Effects of Earlier-Acquired
Carol S. L. Yap and Rick Richardson
University of New South Wales
Research has shown that learned fear emerges in a response-specific sequence. For example, freezing is
observed at a younger age than is potentiated startle (P. Hunt & B. A. Campbell, 1997). The present study
shows that the age at which a specific learned fear response emerges is influenced by the animal’s early
experiences. Specifically, fear potentiation of startle emerges earlier in development if the rat is given
prior fear conditioning to a different stimulus. Some constraints of this “facilitation” effect are deter-
mined in follow-up experiments. This facilitation effect may provide a novel way of testing the
development of the neural circuits underlying learned fear.
Keywords: fear-potentiated startle, odor, development, learned fear, early learning
In studies using rodents, fear memories elicited by a conditioned
stimulus (CS) such as a tone or odor that had been previously
paired with an aversive unconditioned stimulus (US; e.g., shock)
are typically indexed by behavioral responses such as avoidance,
freezing, and heart rate changes (e.g., LeDoux, 1993). Another
common measure of learned fear in the rat is fear-potentiated
startle (FPS, e.g., Brown, Kalish, & Farber, 1951; Falls & Davis,
1993; Hunt, 1999; Richardson, Vishney, & Lee, 1999). The startle
reflex consists of a rapid sequential contraction of muscles, pre-
dominantly around the face, neck, and shoulders, elicited by a
sudden noise. When an aversive CS precedes the startle-eliciting
noise, the amplitude of the elicited startle reflex is greater than it
is when the startle stimulus is presented alone (e.g., Davis, Falls,
Campeau, & Kim, 1993; Fendt & Fanselow, 1999). The condi-
tioned fear potentiation of startle has been demonstrated using a
variety of conditioned stimuli in rats (e.g., Campeau & Davis,
1995; Richardson et al., 1999) and has been found in humans as
well (Grillon, Pellowski, Merikangas, & Davis, 1997).
Past studies with adult rats suggest that various indices of
learned fear covary consistently across parametric variables and
experimental manipulations (Hunt & Campbell, 1997; Leaton &
Cranney, 1990; Stanton, 2000). However, developmental analyses
of classical conditioning suggest that learned fear emerges in a
sensory- and response-specific manner. In addition, the ontogeny
of modality-specific associative learning also mirrors the trajectory
for sensory development (see Alberts, 1984, for a review of
sensory detection across development). The ability to learn about
an olfactory stimulus precedes auditory learning, and the ability to
learn about a visual CS is the last to emerge. In addition to this
sensory-specific sequence, recent studies have also demonstrated a
response-specific sequence (Hunt & Campbell, 1997; Richardson
& Hunt, in press). A consistent finding in this work is that FPS
emerges later in development than does freezing or heart rate
changes. Studies on FPS indicate that this particular index of
conditioned fear emerges at Postnatal Day (PN) 22 or 23 for visual
(Hunt, 1999), auditory (Hunt, Richardson, & Campbell, 1994), and
olfactory stimuli (Richardson, Paxinos, & Lee, 2000). In contrast,
heart rate responses to an odor CS, for example, have been de-
tected as early as PN 12 (Hunt, Hess, & Campbell, 1998), and
freezing responses have been detected as early as PN 16 (Hunt,
1997). Further, avoidance of an odor previously paired with shock
occurs as early as PN 10 (Sullivan, Landers, Yeaman, & Wilson,
Given that FPS matures late in development, Richardson and
colleagues have used this index of learned fear to examine whether
the expression of early learning is appropriate to the animal’s age
at training or to their age at test (Richardson & Fan, 2002;
Richardson et al., 2000). For example, in Richardson et al. (2000),
rats were given odor–shock pairings at PN 16 and were tested for
avoidance and FPS at PN 23. Results showed that rats avoided the
odor CS, indicating that they retained the CS–US association
acquired 7 days earlier. However, Richardson et al. (2000) also
reported that FPS was absent for animals trained at 16 days and
tested at 23 days of age. In other words, although these subjects
possessed the ability to exhibit FPS at the age of test (?PN 23),
they did not express this particular conditioned response if it was
absent at the age of training (PN 16). In a recent extension of this
finding, Barnet and Hunt (2006) reported that PN 18 rats given
light–shock pairings did not exhibit FPS to the light CS when
tested at PN 25 even though they displayed light-elicited freezing
comparable with that seen in rats that were trained at PN 24. Taken
together, these studies suggest that the expression of learning is
appropriate to the animal’s age at training and not its age at test.
Yap, Stapinski, and Richardson (2005) recently reported, how-
ever, that early memories can be “updated” so that learned fear is
Carol S. L. Yap and Rick Richardson, School of Psychology, University
of New South Wales, Sydney, Australia.
This research was conducted as part of the Ph.D. thesis by Carol S. L.
Yap and was supported by grants from the Australian Research Council
and the Faculty of Science, University of New South Wales to Rick
Richardson and by an Australian Postgraduate Award to Carol S. L. Yap.
We thank Rose Bagot and Jee Kim for their assistance in data collection.
Correspondence concerning this article should be addressed to Carol
S. L. Yap, School of Psychology, University of New South Wales, Sydney
2052, Australia. E-mail: firstname.lastname@example.org
2007, Vol. 121, No. 5, 1053–1062
Copyright 2007 by the American Psychological Association
0735-7044/07/$12.00 DOI: 10.1037/0735-7044.121.5.1053
expressed in a manner appropriate to the rat’s age at test. The first
experiment by Yap et al. consisted of three groups. Two odor CSs
were used in that experiment, and for clarity of exposition, the
odor trained at PN 16 is referred to as “O1” and the odor trained
at PN 23 is referred to as “O2”. The two odors used by the
researchers were grape and eucalyptus and were counterbalanced
as O1 and O2. In Group 16 (O1), rats received O1–shock pairings
at PN 16 and were tested for freezing and FPS at PN 24 or 25 to
both odor CSs. Consistent with previous findings from our lab, rats
in this group exhibited freezing but not FPS at test to O1 and did
not exhibit any fear responses to O2, indicating that rats trained at
PN 16 did not generalize across these two different odors. Rats in
Group 23 (O2) were given O2–shock pairings at PN 23 and were
tested on the following 2 days for freezing and FPS to O1 and O2.
Rats in this group exhibited freezing and FPS to O2 but did not
display any conditioned fear when tested in the presence of O1.
This again demonstrated that rats were able to discriminate be-
tween these two odors at test. Particularly interesting was the
performance of Group 16–23. Rats in this group were given
O1–shock pairings at PN 16 and O2–shock pairings at PN 23. At
test, these rats exhibited freezing and FPS to O2, consistent with
results found in Group 23 (O2). However, in addition to condi-
tioned freezing, Group 16–23 also displayed FPS to an odor
trained at PN 16 (O1). That is, if rats were subsequently trained to
a second odor at PN 23 after receiving O1–shock pairings at PN
16, they then showed FPS as well as freezing to the first odor. It
seems that training to a second stimulus (O2) at an age when FPS
has emerged (PN 23) effectively “updates” a memory acquired at
PN 16 by activating the FPS response to O1. A similar result was
also found by Barnet and Hunt (2006) with visual–auditory CSs.
Yap et al. suggested that rats must be able to express the target FPS
response at the time of the second odor training phase in order to
trigger the activation of FPS to the first odor (trained at PN 16).
This hypothesis predicts that training to a second odor at an age
prior to the onset of FPS would not lead to an activation of FPS to
an odor CS trained at PN 16.
To this end, Experiment 1 of this study examined the effects of
shifting the second phase of training from PN 23 to an age when
FPS is absent (PN 20). There were three groups in the first
experiment. All rats were conditioned at PN 16 to O1. In Group
16–22, rats were given O2–shock pairings at PN 22. This group is
essentially the same as Group 16–23 in Yap et al.’s previous study
(2005, Experiment 1). Therefore, Group 16–22 was expected to
show FPS to O2 and, as a consequence, activate FPS to O1. Rats
in Group 16–20, however, received O2–shock pairings when they
were 20 days of age. On the basis of previous experiments that
have shown that rats do not exhibit FPS to an odor trained at PN
20 (e.g., Richardson et al., 2000), we did not expect Group 16–20
to exhibit FPS to O2. Consequently, no updating should occur in
this group, and these rats should also not display FPS to O1. The
final group received only O1–shock pairings at PN 16 and no
subsequent training prior to test (Group 16). As has been reported
in previous experiments, rats that are trained to an odor CS at PN
16 do not express FPS to that CS regardless of the age at test
(Richardson & Fan, 2002; Richardson et al., 2000). Therefore, rats
conditioned at PN 16 (Group 16) should not exhibit any FPS at
test. All rats were assessed for FPS to O1 and O2 at PN 23. As
learning about odor–shock associations has been demonstrated in
PN 16 rats in numerous studies across both 24-hr and 7-day
intervals between conditioning and test using both avoidance (Ri-
chardson et al., 2000) and freezing (Richardson, Tronson, Bailey,
& Parnas, 2002; Yap et al., 2005), these indices of fear were not
measured in this study.
Thirty-four male Sprague–Dawley rats, obtained from the
breeding colony maintained by the School of Psychology at the
University of New South Wales (Sydney, Australia), were used.
One rat from Group 16–22 was removed from the data set because
its performance was 3 SDs from the group mean. Therefore, there
were 11 subjects in Group 16–22, 12 subjects in Group 16–20, and
10 subjects in Group 16. Rats were 16 days of age at the com-
mencement of training for all groups. They were housed in groups
of 8 in plastic boxes (37.0 cm ? 24.5 cm ? 27.0 cm) with their
mother and were kept in a room with a 12-hr light–dark cycle
(lights on at 6 a.m.). No more than 1 subject per litter was used in
any experimental group. Food and water were continuously avail-
able for all rats. All animals were treated according to the princi-
ples of animal use outlined in the Australian Code of Practice for
the Care and Use of Animals for Scientific Purposes (Australian
Government Publishing Service, 2004), and the Animal Care and
Ethics Committee at the University of New South Wales approved
Conditioning and tests occurred in two identical chambers (13
cm ? 9 cm ? 9 cm). The front wall, rear wall, and ceiling of each
chamber were constructed of clear Plexiglas. The two sidewalls
were made of 3-mm stainless steel rods. The rods were vertically
positioned relative to the floor of the chamber. The floor was
composed of stainless steel rods spaced 13 mm apart, center-to-
center. Electric shock (0.6 mA, 1 s in duration) could be delivered
to the floor of each chamber via a custom-built shock generator.
The chamber was attached to a piece of Plexiglas onto which a
sheet of piezoelectric film had been laminated. Movement in the
chamber caused the piece of Plexiglas to flex, which consequently
produced voltage in the piezoelectric film. This voltage is propor-
tional to the amount of movement within the chamber, with larger
movements producing larger voltages. These voltages were ampli-
fied and digitized (at a 1-kHz rate) to measure startle amplitude.
The peak voltage (converted into arbitrary units ranging from
0–32,000; data rounded to 0–320 prior to analysis) in the 250-ms
period after stimulus onset was taken as the index of the startle
The acoustic startle stimulus was delivered through two high-
frequency speakers mounted 8 cm on either side of the chamber.
The startle stimulus was a 100-ms, 100-dB burst of white noise,
with a 1-ms rise–fall time. The intensity of the startle stimulus and
background noise was measured with a Bru ¨el and Kjaer precision
sound level meter (Type 2235) placed in the center of the startle
cage. A computer controlled the stimulus presentations and record-
ing of data. The software and hardware were custom developed at
the University of New South Wales.
YAP AND RICHARDSON
The chambers were housed in wood cabinets to reduce external
noise and visual stimulation. Ventilation fans located on a sidewall
of the wood cabinets provided a low-level background noise (?60
dB) at all times, and illumination was provided by a 15-W red light
on the front door of each cabinet. A removable tray that contained
animal bedding was located beneath each cage and was cleaned
and replaced with the removal of each rat.
The odors used in this study were 0.2 ml of grape flavor (Grape
No. 182380019 from Wild Flavours; Heidelberg, Germany) and
eucalyptus (Goanna Eucalyptus Oil; Herron Pharmaceuticals;
Queensland, Australia). The odors were squirted onto a piece of
paper towel inside a plastic specimen jar.
two-stage training procedure was used for odor conditioning. The
doors of the wood cabinets in which the experimental chambers
were located were kept slightly ajar throughout both stages of
training. In Stage 1 of training, rats were placed in the chamber and
were given a 5-min adaptation period followed by the first of 15
shocks. Seven to 10 s prior to the delivery of shock, a specimen jar
containing an odor was placed approximately 10 cm beneath the
chambers. The jar contained grape for half the rats and eucalyptus
for the remaining rats. For all groups, immediately after the shock,
the jar was removed and the lid replaced. The jar was then placed
on a bench approximately 2 m from the startle chamber. The
interval between the shock presentations was 1.5, 2.0, or 2.5 min
(average ? 2.0 min) and varied pseudorandomly. After the last
shock, rats were removed from the chambers and placed in their
home cages. In addition to producing an association between the
odor and shock, training also produces an association between the
context, (i.e., the chamber) and shock. To reduce the level of
contextual fear, we returned the rats to the chambers 50 to 80 min
later for the second phase of training. In Stage 2 of training, a
specimen jar containing water was placed beneath the chamber for
7–10 s after a 2-min adaptation, but no shock was delivered. There
were 15 water presentations in Stage 2 (i.e., the same number of
trials as Stage 1), and the interval between presentations was 1.5,
2.0, or 2.5 min (average ? 2.0 min) and varied pseudorandomly
(i.e., same as Stage 1). Both stages of training lasted approximately
30 min, and at the end of each stage, the doors and windows were
opened for at least 10 min to allow ventilation of the experimental
Rats trained to the second odor CS at 20 days of age (Group
16–20) or 22 days of age (Group 16–22) received five O2–shock
pairings at those ages. The procedures were identical to those used
when rats were 16 days of age, with the exception that training
consisted of 5 trials instead of 15 and the training session lasted
approximately 15 min instead of 30 min. Rats that received O2
training at PN 20 or PN 22 also experienced Stage 2 training, in
which they received contextual extinction. There were five water
presentations in Stage 2. The interval between water presentations
was the same as Stage 1 odor presentations (1.5, 2.0, and 2.5 min,
average ? 2.0 min). Rats in Group 16 did not receive any further
training prior to test at 23 days of age.
All rats received 15 O1–shock pairings at PN 16. A
Test for FPS.
had been trained and remained there for approximately 25 min.
After a 5-min adaptation period, 30 startle-eliciting noise bursts,
each separated by 30 s, were presented. No odor was present
during these initial 30 bursts. The average startle response on these
first 30 trials was taken as an estimate of the rat’s baseline startle
response. During a 60-s period following the 30th noise burst, a jar
containing either the grape or eucalyptus odor was placed beneath
the startle chamber floor and remained there for the rest of the test
session. After introduction of the odor, an additional 30 noise
bursts, each separated by 30 s, were presented. At the end of the
testing session, subjects were removed from the chamber and
returned to their home cage. The interval between test for O1 and
the test for O2 was approximately 1.5 hr.
Odor potentiation of startle (OPS) was measured by comparing
the average startle amplitude when the odor was present with the
average startle amplitude during baseline and was calculated using
the following equation: [(T?B) ? B] ? 100 ? OPS, where T ?
mean startle amplitude over the 30 test trials, B ? mean startle
amplitude over the 30 baseline trials, and OPS ? percentage
change in startle amplitude.
Rats were returned to the chamber in which they
All data presented in this study, with the exception of those in
Experiment 3, were analyzed using a post hoc analysis of variance
(Rodgers, 1967), and the decision-wise error was set at .05.
Results and Discussion
There were three main findings in Experiment 1. First, rats
given O1–shock pairings at PN 16 did not exhibit FPS to that odor
CS when tested at PN 23 [Group 16; see Figure 1]. This finding is
startle amplitude to Odor 1 (O1) and Odor 2 (O2). All rats received O1
conditioning at Postnatal Day (PN) 16. The rats received O2 conditioning
at either PN 22 (Group PN16–22) or PN 20 (Group PN16–20). Rats in
Group PN 16 did not receive any O2 conditioning prior to test.
Results of Experiment 1. Mean (? SEM) percentage change in
FACILITATION OF MEMORY
consistent with previous reports from our laboratory (Richardson
& Fan, 2002; Richardson et al., 2000; Yap et al., 2005). The
second finding is that training to another odor at PN 22 after initial
training to O1 at PN 16 activated FPS to O1 [Group 16–22]. This
finding is consistent with Yap et al.’s (2005) results. Finally, the
third (and unexpected) finding was that rats given O2–shock
pairings at PN 20 exhibited FPS to both odors at test. This latter
result illustrates two important points: (a) Previous studies have
shown that rats trained to an odor CS at PN 20 do not display FPS
to that CS, and therefore, this is the first demonstration of PN 20
rats exhibiting FPS, and (b) FPS to an odor trained at PN 20 in turn
activated FPS to O1, a CS that was trained at PN 16. These
descriptions of the data were confirmed by statistical analysis.
Differences between groups for baseline startle amplitude were
not significant, largest F(1, 30) ? 1.1.
Test for O1
Figure 1 illustrates the mean percentage change in startle am-
plitude in the presence of O1. Inspection of Figure 1 suggests that
Group 16 demonstrated less FPS to O1 than did Group 16–22 and
Group 16–20. That is, rats that had been trained to O2 at PN 20 or
PN 22 after initial training with O1 at PN 16 exhibited greater FPS
to O1 than did rats that experienced only O1–shock pairings at PN
16 and were tested at PN 23. Statistical analyses confirmed our
observation that the magnitude of FPS to O1 in Group 16–22 and
Group 16–20 was significantly different from that seen in Group
16, F(2, 30) ? 5.07, Fcritical? 5.04. Moreover, there were no
differences between Group 16–22 and Group 16–20 (F ? 1).
Baseline for O2
Analyses of baseline startle amplitudes showed that there were
no significant group differences (Fs ? 1).
Test for O2
Figure 1 shows that rats that had been given O2–shock pairings
at PN 20 or PN 22 (Group 16–20 and Group 16–22) exhibited
more FPS to O2 than did rats that had not been conditioned to this
odor (Group 16), F(2, 30) ? 5.48. In addition, FPS to O2 in Group
16–22 was not significantly different from that found in Group
16–20 (F ? 1).
The results obtained in Experiment 1 replicated past studies
demonstrating that rats trained at PN 16 do not exhibit FPS when
tested at PN 23 (e.g., Richardson, Fan, & Parnas, 2003). This
experiment also replicated Yap et al.’s (2005) updating effects.
Rats that were given O2–shock pairings at PN 22 displayed FPS to
an odor CS that had been trained at PN 16 (Group 16–22).
However, Experiment 1 also produced some unexpected results.
Specifically, this is the first time that FPS has been observed in rats
that have been trained before PN 23. In past studies, rats given
odor–shock pairings at 20 days of age did not display FPS to the
odor CS (Richardson et al., 2003, 2000). In Experiment 1, how-
ever, rats trained to O2 at PN 20 exhibited FPS to this odor as well
as FPS to an odor CS that had been trained at PN 16. The FPS to
O1 is likely to be a consequence of activation and is not surprising
given that subjects from Group 16–20 exhibited FPS to O2 at test.
This finding is consistent with Yap et al.’s hypothesis that if rats
exhibit FPS to a second different odor CS, they would also exhibit
FPS for an odor trained at PN 16. However, the observation of FPS
to an odor trained at PN 20 prevented an assessment of the original
hypothesis that this study was designed to examine. That is, rats
should not exhibit FPS to an odor trained at PN 16 if the second
odor was trained at an age when rats could not express FPS (? PN
Given the unexpected result in this experiment, subsequent
experiments explored the factors that produced FPS in PN 20 rats.
There are several possible explanations for the observed early
onset of FPS in PN 20 rats in this study. For example, in past
studies, rats were tested for FPS 24 hr after conditioning (e.g.,
Richardson et al., 2000, 2003). In the present study, however, rats
in Group 16–20 were conditioned to O2 at PN 20 and were tested
at PN 23. Thus, there was a 3-day delay between conditioning and
test. This 3-day interval between conditioning and test could
explain the occurrence of FPS in PN 20 rats. Past studies have
shown that the age of onset for a fear response sometimes differs
depending on the retention interval between conditioning and test.
For instance, conditioned changes in heart rate are observed at
younger ages when a delay is introduced between training and
testing (Campbell & Ampuero, 1985; Hunt et al., 1998). Specifi-
cally, conditioned bradycardia to a light CS during training does
not emerge until PN 28. However, if test occurs 2 hr after condi-
tioning, then conditioned bradycardia to the light CS is observed as
early as PN 14 (Hunt et al., 1998). A delay between conditioning
and test could therefore determine the age at which a specific fear
response emerges. For this reason, a 3-day delay between condi-
tioning and test for Group 16–20 in this experiment could explain
the accelerated age of onset for FPS. This possibility was exam-
ined in the next experiment.
To determine whether a 3-day lag between conditioning and test
explains the early emergence of FPS in Experiment 1, we manip-
ulated the interval between acquisition and test in Experiment 2.
Rats in Group PN 22 were given odor–shock pairings on PN 22
and were tested the next day. On the basis of prior results, these
rats should show FPS to the odor CS (e.g., Experiment 1 in this
study; Richardson et al., 2003). The rats in the other two groups
were trained on PN 20. Rats in one group were tested the next day
(Group 20 [24 hr]), and rats in the other group were tested 3 days
later (Group 20 [3 days]). On the basis of previous studies, we did
not expect Group 20 (24 hr) to exhibit FPS to the odor. The
interesting comparison involved those rats trained on PN 20 and
tested on PN 23. If a delay of 3 days between conditioning and test
results in the emergence of FPS in rats trained at PN 20, a
possibility suggested by Experiment 1, then Group PN 20 (3 days)
should exhibit FPS at test.
Twenty-four Sprague–Dawley rats were used in this experiment.
They were obtained from the same source as in the previous
YAP AND RICHARDSON
experiment. Rats in Group 22 were 22 days of age on the day of
training. Rats in the other two groups (Group 20 [24 hr] and Group
20 [3 days]) were 20 days of age on the first day of training (ns ?
The equipment used was the same as that used in the previous
Odor used was 0.2 ml of grape.
All training and test procedures were identical to those used for
O2 training in Experiment 1 (i.e., five odor–shock pairings) with
the exception that the interval between training and test was varied.
Rats in Group 22 were given odor–shock pairings at PN 22 and
were tested for FPS the following day. In the other two groups, rats
were given odor–shock pairings at PN 20 and were tested either
the following day (Group 20 [(24hr]) or 3 days later (Group 20 [3
Results and Discussion
The results of this experiment are depicted in Figure 2. As
expected, rats in Group 22 exhibited FPS, whereas rats in Group
20 (24 hr) did not. These results replicate past studies (Richardson
et al., 2000, 2003). Moreover, rats conditioned at PN 20 and tested
3 days later did not display FPS (Group 20 [3 days]). This finding
does not support the hypothesis that a lag of 3 days between
conditioning and test explains the occurrence of FPS in PN 20 rats
in Experiment 1. Statistical analysis confirmed these observations.
Groups were not significantly different at baseline, largest F(1,
21) ? 1.47.
Statistical analyses revealed that: (a) Group 22 exhibited signif-
icantly more FPS at test than did Group 20 (24 hr) and Group 20
(3 days), F(2, 21) ? 13.78, Fcritical? 5.24, and (b) Group 20 (24
hr) and Group 20 (3 days) did not display any FPS at test, and there
was no significant difference between these two groups (F ? 1).
The results are consistent with previous studies that have shown
that FPS emerges at approximately PN 23 but is absent in rats that
are 20 days of age at the time of training (Richardson et al., 2000).
Furthermore, this experiment indicates that an extended delay
between acquisition and test did not produce the FPS observed in
PN 20 rats in Experiment 1.
In the present experiment, we compared the performance of two
groups of rats trained with an odor CS at PN 20. Rats in Group 20
were conditioned to O2 and then tested the next day. These rats did
not receive any early training at PN 16. Rats in Group 16–20 were
given O1–shock pairings at PN 16 and O2–shock pairings at PN
20. Unlike Group 16–20 in Experiment 1, Group 16–20 in this
experiment was tested on PN 21. Therefore, both groups in this
experiment were tested at the same age (PN 21). Rats in Group 20
were not expected to show FPS at test, as this group is identical to
Group 20 (24 hr) in Experiment 2. However, if early training at PN
16 facilitates the emergence of FPS in PN 20 rats, then Group
16–20 should exhibit FPS to O2 when tested at PN 21.
Twenty-two Sprague–Dawley rats were used in this experiment.
They were obtained from the same source as in the previous
experiment. Rats in Group 16–20 were 16 days of age when
training commenced; rats in Group 20–21 were 20 days of age
when training commenced. Two rats from Group 16–20 were
excluded from the data set because they were 3 SDs from the group
mean. There were 10 subjects in each group.
The equipment used was the same as that used in the previous
The odors used were 0.2 ml of grape and 0.2 ml of eucalyptus,
identical to those used in Experiment 1.
startle amplitude to Odor 2. All rats received Odor 1 conditioning at PN 16.
Odor 2 conditioning occurred at either Postnatal Day (PN) 22 (Group PN
22) or PN 20. Rats that were conditioned at PN 20 were tested either 3 days
or 24 hr after test.
Results of Experiment 2. Mean (?SEM) percentage change in
FACILITATION OF MEMORY
Rats in Group 16–20 were given 15 O1–shock pairings at PN 16
and five O2–shock pairings at PN 20. Rats in Group 20 received
five O2–shock pairings at PN 20. All rats were tested with O2
when they were 21 days of age. All parameters for training and test
were identical to those in previous experiments. Grape and euca-
lyptus odors were counterbalanced as O1 and O2.
Results and Discussion
Inspection of Figure 3 suggests that Group 16–20 exhibited FPS
to O2 but that Group 20 did not. This was confirmed by statistical
The two groups were not significantly different at baseline, F(1,
18) ? 1.29.
Group 16–20 exhibited significantly more FPS to O2 than did
Group 20, F(1, 18) ? 6.41, Fcritical? 4.41.
The results of this experiment indicate that rats trained to O1 at
PN 16 showed FPS to an odor trained at PN 20. This replicates the
surprising result from Experiment 1. In contrast, rats that were
trained only to O2 at PN 20 did not exhibit FPS when tested in the
presence of the odor CS. This replicates past findings from our lab
(e.g., Richardson et al., 2000). These data suggest that some aspect
of the training that occurs at PN 16 facilitates an earlier onset of
FPS so that it can be observed at PN 20.
As the facilitation effect appears to be a robust and reliable
finding given that it was observed in Experiment 1 and in the
present experiment, the next experiment assessed the various as-
pects of the training that occurred at PN 16 that could account for
the observation of FPS in PN 20 rats.
Several studies have shown that a variety of early experiences
not involving explicit learning about CS–US contingencies can
result in the early emergence of certain types of learning. For
example, Woodcock and Richardson (2000) have shown that rats
reared in enriched environments from PN 2 exhibited contextual
learning at an earlier age (PN 18) than had been previously
documented in past studies (e.g., PN 23 in Rudy, 1993). Other
researchers have shown that contextual learning can be enhanced
at PN 45 if rats are handled for the first 15 days after birth (Beane,
Cole, Spencer, & Rudy, 2002).
The current experiment examined whether the early experience
that led to the FPS observed in rats trained on PN 20 in Experi-
ments 1 and 3 is modality-specific and/or contingent on a paired
association of the odor with the shock at PN 16. Experiment 4
consisted of three groups. In the light–paired group, rats received
light–shock pairings at PN 16. In the odor–unpaired group, rats
received unpaired presentations of odor with a shock. The odor–
paired group received odor–shock pairings at PN 16. All rats then
received O2–shock presentations at PN 20 and were tested at PN
21 to O2. The odor–paired was identical to Group 16–20 in
Experiments 1 and 3 and was therefore expected to exhibit con-
ditioned FPS to O2 at test. The performance of the rats in the other
two groups would assess whether the facilitation effect was de-
pendent on rats learning about a discrete CS at PN 16 and whether
this learning had to be about a CS from the same sensory modality
as the CS trained at PN 20.
Twenty-six Sprague–Dawley rats were used in this experiment.
They were obtained from the same source as in the previous
experiments. All rats were 16 days (?1) of age at the beginning of
The equipment was the same as that used in the previous
experiment with the exception that, instead of an odor, an 18-W
white light was used as a CS at PN 16 for some rats.
The odors used were 0.2 ml of grape and 0.2 ml of eucalyptus,
the same as in Experiment 1.
same training and test procedures as Group 16–20 in Experiment
3. Rats in the light–paired group (n ? 10) received paired light–
shock training at PN 16. Rats in the odor–unpaired group (n ? 8)
received odor and shock presentations in an explicitly unpaired
Rats in the odor–paired group (n ? 8) received the
startle amplitude to an odor conditioned stimulus trained at Postnatal Day
(PN) 20. Rats in Group PN16–20 received Odor 1–shock pairings at PN
16, whereas Group PN 20 did not.
Results of Experiment 3. Mean (?SEM) percentage change in
YAP AND RICHARDSON
preparation at PN 16. At 20 days of age, all rats received O2–shock
Light training at PN 16 consisted of two phases for the light–
paired group. In the first phase of training, rats were placed in the
chambers with the cabinet doors closed and received a 5-min
adaptation period. Following this, they were exposed to an 18-W
white light for 8 s. A shock (0.6 mA for 1 s) was administered
during the last second of this light CS. There were 15 light–shock
pairings, with the intertrial interval varying pseudorandomly be-
tween 2–3 min. Following administration of the final shock, rats
were returned to their home cage. Rats in this group then received
a second phase of training in which they were placed in the startle
chamber for 35 min with no programmed stimuli (i.e., no lights or
shocks). This second phase of training occurred at least 2 hr after
the initial phase. Rats in the unpaired–odor group also received a
two-stage procedure at PN 16. In Stage 1, rats in the unpaired–
odor group received 15 water–shock presentations, and in Stage 2
they received 15 odor-only presentations. The interval between
presentations in both stages was 1.5, 2.0, or 2.5 min (average ? 2.0
min) and varied pseudorandomly.
Tests occurred when rats were 21 days of age. The
procedures were identical to those used in previous experiments.
Results and Discussion
Inspection of Figure 4 indicates that FPS was observed only in
the odor–paired group and not in the light–paired or odor–
unpaired groups. Statistical analyses confirmed these observations.
The groups were not significantly different at baseline, largest
F(2, 23) ? 2.09.
The odor–paired group exhibited FPS in the presence of O2 and
was significantly different from the light–paired and odor–
unpaired groups, F(2, 23) ? 9.41, Fcritical? 5.24. There were no
differences between the light–paired group and the odor–unpaired
group, as neither exhibited any FPS in the presence of O2 (F ? 1).
The results suggest that the early emergence of FPS to an odor
CS at PN 20 is contingent on two factors occurring during training
at PN 16: (a) The experience must involve learning about an
association between two events (i.e., O1–shock), and (b) the CS
modality used at PN 16 must be the same as that used at PN 20. If
a light CS was used at PN 16, then FPS was not observed in the
presence of an odor that was trained at PN 20. Thus, arousal via
odors and shock in an unpaired preparation or light–shock pairings
at PN 16 does not facilitate the early emergence of FPS to an odor
CS trained at PN 20. It is also unlikely that handling produces the
facilitation effect. Specifically, the light–paired and odor–unpaired
groups received the same amount of handling as the odor–paired
group at PN 16, and neither group exhibited FPS to O2 at PN 20.
There were several findings in this series of experiments. The
first experiment replicated Yap et al.’s (2005) original updating
effects: Rats exhibited FPS to an odor CS that had been paired with
shock at PN 16 if they also received pairings of a second odor CS
with shock at PN 22. In other words, rats given the second training
episode at PN 22 were able to express their fear of an odor CS
trained at PN 16 by a response system that matured during the
retention interval. The first experiment in this study was also
designed to test a hypothesis put forward by Yap et al. Specifi-
cally, Yap et al. suggested that one interpretation of their findings
is that rats must exhibit FPS to the second odor to activate FPS for
an odor trained at PN 16. From this perspective, if O2–shock
pairings occur at an age when FPS is absent (i.e., PN 20), then no
updating would occur. Surprisingly, the results of the first exper-
iment in this study showed that training a second odor CS at PN 20
led to the updating of an odor CS trained at PN 16 (i.e., rats now
expressed their fear of that first odor CS via the late-maturing FPS
response system). However, even more surprising was the finding
that these rats exhibited FPS to the odor CS trained at PN 20. This
result is inconsistent with the well-documented finding that FPS to
an odor CS is first observed in rats trained at PN 22 (Richardson
et al., 2000, 2003) and does not allow us to test our original
hypothesis (i.e., that the updating effect would not occur if rats
were trained at an age when they did not acquire the FPS response
to a fear-eliciting CS).
Although we were not able to test our original hypothesis that
the activation, or updating, effect is contingent on FPS being
expressed to the second odor CS in the present study, there are
alternative approaches available to examine this issue. For in-
stance, future experiments could arrange for the FPS to O2 to be
abolished or reduced prior to testing O1. As one example, the
startle amplitude to an odor conditioned stimulus trained at Postnatal Day
(PN) 20. Rats received one of the following at PN 16: (a) light–shock
pairings (light–paired group), (b) Odor 1 and shock in an unpaired manner
(unpaired group), or (c) odor–shock pairings (odor–paired group).
Results of Experiment 4. Mean (?SEM) percentage change in
FACILITATION OF MEMORY
learned fear response to O2 could be extinguished prior to testing
O1. In this preparation, rats would be conditioned to O1 at PN 16
and to O2 at PN 22. Some rats would have their learned fear of O2
extinguished prior to test, whereas other rats would not. The rats in
the nonextinguished condition should replicate the findings re-
ported here: FPS to both O2 and O1. Rats in the extinguished
condition should not exhibit FPS to O2; the question of interest is
whether they still exhibit the activated OPS to O1. The results of
this experiment would pinpoint whether it is only necessary for
rats to be trained at an age when they can express FPS to the CS
or whether they need to actually express FPS to that second CS.
Another approach would be to give O2–shock pairings even earlier
than PN 20, which should reduce the likelihood of observing OPS
to O2. The results of this experiment would show whether O1–
shock pairings at PN 16 facilitates the emergence of FPS to O2
even if training to the second odor occurs as early as PN 17.
Despite being unable to explore our original hypothesis, we
were able to further explore the surprising finding that rats trained
at PN 20 exhibited FPS—an effect that we had never observed in
rats trained prior to PN 23. These subsequent experiments reliably
demonstrated that conditioning at PN 16 accelerated the age of
onset for FPS. Further, Experiments 2 and 4 demonstrated that the
facilitation effect was not due to a lag of 3 days between condi-
tioning at PN 20 and test or an unspecified arousal or enrichment
effect, given that unpaired presentations of the O1 and shock at PN
16 did not lead to FPS in PN 20 rats. Finally, the facilitation effect
also appeared to be modality specific. That is, light–shock training
at PN 16 did not produce a facilitation effect to an odor CS that
was trained at PN 20.
Previous studies have shown that preweanling rats exposed to
CS–US pairings exhibited enhanced learning at later ages. For
example, Rudy and Hyson (1982) reported that conditioned
mouthing to a tone CS in an appetitive preparation did not emerge
until PN 16. In a follow-up study, Rudy, Vogt, and Hyson (1984)
demonstrated that rats given tone–sucrose pairings from 14 days of
age performed better at 16 days of age than did rats that began
training at PN 16. Thus, it appears that early training using the
same tone CS can facilitate the level of learning observed at a later
age in an appetitive preparation.
Although Rudy et al.’s (1984) study shares some similarities to
the current series of experiments in that both explored the facili-
tation of learning, or at least the expression of that learning, via
Pavlovian conditioning (i.e., tone–sucrose and O1–shock) in de-
veloping rats, direct comparisons between Rudy et al.’s study and
the current study are somewhat difficult. For example, Rudy et al.
used the same tone CS in their experiment, whereas the current
study examined whether training to one odor CS facilitated the age
of onset for FPS to another odor CS. The current study could be
thought of as being similar to Rudy et al.’s if our rats perceived O1
and O2 as being the same. There is, however, evidence that
preweanling rats can discriminate between the two odors used in
these experiments (Yap et al., 2005). Another difference between
the present study and that of Rudy et al. is that they examined
whether training at PN 14 facilitated the rate of learning to a tone
CS at PN 16, whereas we examined whether training to one odor
CS at PN 16 led to an earlier age of onset for FPS when rats were
trained to a second odor CS at PN 20. Nonetheless, both studies do
show that it is possible for particular early experiences to facilitate
learning in the developing rat.
It is possible that our finding that training rats with one odor CS
at PN 16 facilitates the emergence of FPS to a different odor CS
trained at PN 20 is a consequence of cumulative learning. More
specifically, multiple training sessions to O1 and O2 may have
resulted in what has been termed the learning-to-learn effect (e.g.,
Slotnick, 2001; Slotnick, Hanford, & Hodos, 2000). In other
words, training rats to O1 at PN 16 allowed them to learn the “rules
of the game” (that an odor predicts shock). At PN 20, when they
are reexposed to a second odor, this rule is applied in a general way
(i.e., any odor predicts shock) and potentially facilitated learning
about O2–shock associations. Specifically, PN 20 rats may have
learned more about O2–shock associations if they were given prior
training at PN 16 with O1–shock pairings. Stronger O2–shock
associations at PN 20 for these rats may then account for the
observation of FPS at test. McNish, Gewirtz, and Davis (1997)
have shown that the threshold for FPS usually requires a stronger
CS–US association (i.e., more pairings during conditioning) than
that required for freezing. Thus, if rats that had been conditioned
at PN 16 were learning more about O2–shock associations at PN
20 than were naive PN 20 rats, this may account for the observa-
tion of FPS in PN 20 rats that had also been trained earlier to O1.
This “learning-to-learn” account of the facilitation effect might
also explain the modality-specific effects found in Experiment
4—learning the rule that light predicts shock at PN 16 may not
facilitate odor conditioning at PN 20. There is an alternative
explanation, however, for the modality-specific effects observed in
this study. That is, PN 20 rats could have forgotten about the light
CS encountered at PN 16. As conditioned freezing to the light was
not assessed, there was no direct evidence that PN 20 rats remem-
bered the light CS that was trained 4 days earlier. Future studies
will have to assess this possibility by including freezing as an
additional measure for the light CS trained at PN 16.
The facilitation effect observed in this study could prove useful
in exploring the neural bases of learned fear. The current general
consensus on the neural bases of learned fear is that the encoding
and storage of fear occurs in the basolateral complex of the
amygdala in adult rats (Campeau & Davis, 1995; Cousens & Otto,
1998). The basolateral cortex projects to the central nucleus of the
amygdala (CeA), which has both direct and indirect projections to
various hypothalamic and brainstem structures that mediate spe-
cific behavioral expressions of fear. For instance, direct and indi-
rect projections from the CeA to the caudal pontine reticular
nucleus (PnC) mediate the potentiated acoustic startle response
(Davis et al., 1993; Fendt, Koch, & Schnitzler, 1996; Rosen,
Hitchcock, Sananes, Miserendino, & Davis, 1991; Walker &
Davis, 1997; Zhao & Davis, 2004). From the perspective of this
model, a potential explanation for the delayed development of FPS
is that the pathway between the CeA and PnC may not be func-
tional prior to PN 23 (Hunt & Campbell, 1997; Weber & Rich-
ardson, 2001). However, the current study indicates that early
training at PN 16 may accelerate the maturation of this specific
pathway between the CeA and PnC, thus enabling rats to exhibit
FPS at PN 20. This possibility, of course, will need to be empir-
In summary, this study has contributed in a number of ways to
our understanding of FPS. It is important to note that we initially
replicated the updating effect found in Yap et al. (2005). That is,
training at an older age activates FPS to an odor trained at an
earlier age. However, the results failed to support our main hy-
YAP AND RICHARDSON
pothesis that FPS would not be activated if the second training
phase occurred at an age when FPS is absent (i.e., PN 20) because
of the surprising finding that PN 20 rats expressed FPS in our
experimental arrangement. This unexpected finding was explored
in follow-up experiments, and it appears that initial training at PN
16 led to an early onset of FPS at PN 20. Further experiments
revealed that learning accrued at an early age to a same sensory
modality CS (i.e., odor in this study) facilitated the early emer-
gence of FPS at PN 20. That is, we found that the facilitation of
FPS onset was modality specific and was not due to unspecified
arousal at PN 16. Our novel findings in this study suggest that
conditioning to a similar CS at a young age may accelerate and/or
support the neural and learning processes for FPS at an older age.
Future experiments are required to more fully explore the condi-
tions necessary for producing the accelerated ontogeny of FPS and
for determining how this facilitation effect occurs.
Alberts, J. R. (1984). Sensory-perceptual development in the Norway rat:
A view towards comparative studies. In R. Kail & N. E. Spear (Eds.),
Comparative perspective on the development of memory (pp. 65–101).
Hillsdale, NJ: Erlbaum.
Australian Government Publishing Service. (2004). The Australian code of
practice for the care and use of animals for scientific purposes (7th ed.).
Canberra, Australia: Author.
Barnet, R. C., & Hunt, P. S. (2006). The expression of fear-potentiated
startle during development: Integration of learning and response sys-
tems. Behavioral Neuroscience, 120, 861–872.
Beane, M. L., Cole, M. A., Spencer, R. L., & Rudy, J. W. (2002). Neonatal
handling enhances contextual fear conditioning and alters corticosterone
stress responses in young rats. Hormones and Behavior, 41, 33–40.
Brown, J. S., Kalish, H. I., & Farber, I. E. (1951). Conditioned fear as
revealed by magnitude of startle response to an auditory stimulus.
Journal of Experimental Psychology, 41, 317–328.
Campbell, B. A., & Ampuero, M. X. (1985). Dissociation of autonomic
and behavioral components of conditioned fear during development in
the rat. Behavioral Neuroscience, 99, 1089–1102.
Campeau, S., & Davis, M. (1995). Involvement of central nucleus and the
basolateral complex of the amygdala in fear conditioning measured with
fear-potentiated startle in rats trained concurrently with auditory and
visual conditioned stimuli. Journal of Neuroscience, 15, 2301–2311.
Cousens, G., & Otto, T. (1998). Both pre- and posttraining excitotoxic
lesions of the basolateral amygdala abolish the expression of olfactory
and contextual fear conditioning. Behavioral Neuroscience, 112, 1092–
Davis, M., Falls, W. A., Campeau, S., & Kim, S. (1993). Fear-potentiated
startle: A neural and pharmacological analysis. Behavioural Brain Re-
search, 58, 175–198.
Falls, W. A., & Davis, M. (1993). Visual cortex ablations do not prevent
extinction of fear-potentiated startle using visual conditioned stimulus.
Behavioral & Neural Biology, 60, 259–270.
Fendt, M., & Fanselow, M. S. (1999). The neuroanatomical and neuro-
chemical basis of conditioned fear. Neuroscience and Biobehavioral
Reviews, 23, 743–760.
Fendt, M., Koch, M., & Schnitzler, H.-U. (1996). Lesions of the central
gray block conditioned fear as measured with the potentiated startle
paradigm. Behavioral Brain Research, 74, 127–134.
Grillon, C., Pellowski, M., Merikangas, K. R., & Davis, M. (1997).
Darkness facilitates the acoustic startle reflex in humans. Biological
Psychiatry, 42, 453–460.
Hunt, P. (1997). Retention of conditioned autonomic and behavioural
responses in preweanling rats: Forgetting and reinstatement. Animal
Learning & Behavior, 25, 301–311.
Hunt, P. (1999). A further investigation of the developmental emergence of
fear-potentiated startle in rats. Developmental Psychobiology, 34, 281–
Hunt, P., & Campbell, B. A. (1997). Developmental dissociation of the
components of conditioned fear. In M. E. Bouton & M. S. Fanselow
(Eds.), Learning, motivation and cognition: The functional behaviorism
of Robert C. Bolles (pp. 53–74). Washington, DC: American Psycho-
Hunt, P., Hess, M. F., & Campbell, B. A. (1998). Inhibition of the
expression of conditioned cardiac responses in the developing rat. De-
velopmental Psychobiology, 33, 221–233.
Hunt, P., Richardson, R., & Campbell, B. A. (1994). Delayed development
of fear-potentiated startle in rats. Behavioral Neuroscience, 108, 69–80.
Leaton, R. N., & Cranney, J. (1990). Potentiation of acoustic startle
response by a conditioned stimulus paired with acoustic startle stimulus
in rats. Journal of Experimental Psychology: Animal Behavior Pro-
cesses, 16, 279–287.
LeDoux, J. E. (1993). Emotional memory systems in the brain. Behavioral
Brain Research, 58, 69–79.
McNish, K. A., Gewirtz, J. C., & Davis, M. (1997). Evidence of contextual
fear after lesions of the hippocampus: Disruption of freezing but not
fear-potentiated startle. Journal of Neuroscience, 17, 9353–9360.
Richardson, R., & Fan, M. (2002). Behavioral expression of learned fear in
rats is appropriate to their age at training, not their age at testing. Animal
Learning & Behavior, 30, 394–404.
Richardson, R., Fan, M., & Parnas, S. (2003). Latent inhibition of condi-
tioned odor potentiation of startle: A developmental analysis. Develop-
mental Psychobiology, 42, 261–268.
Richardson, R. & Hunt, P. S. (in press). The ontogeny of fear conditioning.
In M. S. Blumberg. J. H. Freeman, Jr., & S. R. Robinson (Eds.),
Developmental and comparative neuroscience: Epigenetics, evolution,
& behavior. Oxford, United Kingdom: Oxford University Press.
Richardson, R., Paxinos, G., & Lee, J. (2000). The ontogeny of conditioned
odor potentiation of startle. Behavioral Neuroscience, 114, 1167–1173.
Richardson, R., Tronson, N., Bailey, G., & Parnas, A. S. (2002). Extinction
of conditioned potentiation of startle. Neurobiology of Learning and
Memory, 78, 426–440.
Richardson, R., Vishney, A., & Lee, J. (1999). Conditioned odor potenti-
ation of startle in rats. Behavioral Neuroscience, 113, 787–794.
Rodgers, R. S. (1967). Type II errors and their decision basis. British
Journal of Mathematical and Statistical Psychology, 20, 187–204.
Rosen, J. B., Hitchcock, J. M., Sananes, C. B., Miserendino, M. J., &
Davis, M. (1991). A direct projection from the central nucleus of the
amygdala to the acoustic startle pathway: Anterograde and retrograde
tracing studies. Behavioral Neuroscience, 105, 817–825.
Rudy, J. W. (1993). Contextual conditioning and auditory cue conditioning
dissociate during development. Behavioural Neuroscience, 107, 887–
Rudy, J. W., & Hyson, R. L. (1982). Consummatory response conditioning
to an auditory stimulus in neonatal rats. Behavioral and Neural Biology,
Rudy, J. W., Vogt, M. B., & Hyson, R. L. (1984). A developmental
analysis of the rat’s learned reactions to gustatory and auditory stimu-
lation. In R. Kail & N. E. Spear (Eds.), Comparative perspectives on the
development of memory (pp. 181–208). Hillsdale, NJ: Erlbaum.
Slotnick, B. (2001). Animal cognition and the rat olfactory system. Trends
in Cognitive Sciences, 5, 216–222.
Slotnick, B., Hanford, L., & Hodos, W. (2000). Can rats acquire an
olfactory learning set? Journal of Experimental Psychology: Animal
Behavior Processes, 26, 399–415.
Stanton, M. E. (2000). Multiple memory systems, development and con-
ditioning. Behavioural Brain Research, 110, 25–37.
FACILITATION OF MEMORY
Sullivan, R. M., Landers, M., Yeaman, B., & Wilson, D. A. (2000,
September 7). Good memories of bad events in childhood. Nature, 407,
Walker, D. L., & Davis, M. (1997). Involvement of the dorsal periaque-
ductal gray in the loss of fear-potentiated startle accompanying high
footshock training. Behavioral Neuroscience, 111, 692–702.
Weber, M., & Richardson, R. (2001). Centrally administered corticotropin-
releasing hormone and peripheral injections of strychnine hydrochloride
potentiate the acoustic response in preweanling rats. Behavioral Neuro-
science, 115, 1273–1282.
Woodcock, E. A., & Richardson, R. (2000). Effects of multisensory envi-
ronmental stimulation on contextual conditioning in the developing rat.
Neurobiology of Learning and Memory, 74, 89–104.
Yap, C. S. L., Stapinski, L., & Richardson, R. (2005). Behavioral expres-
sion of learned fear: Updating of early memories. Behavioral Neuro-
science, 119, 1467–1476.
Zhao, Z., & Davis, M. (2004). Fear-potentiated startle in rats is mediated
by neurons in the deep layer of the superior colliculus/deep mesence-
phalic nucleus of the rostral midbrain through glutamate non-NMDA
receptors. Journal of Neuroscience, 24, 10326–10334.
Received March 26, 2007
Revision received May 14, 2007
Accepted May 15, 2007 ?
YAP AND RICHARDSON