Interstimulus interval (ISI) discrimination of the conditioned eyeblink response in a rodent model of autism
Rats exposed to valproic acid (VPA) on gestational day 12 (GD12) have been advanced as a rodent model of autism [Arndt TL, Stodgell, Rodier PM. The teratology of autism. Int J Dev Neurosci 2005;23: 189-99.]. These rats show cerebellar anomalies and alterations in eyeblink conditioning that are associated with autism. Autistic humans and VPA-exposed rats show normal responses to conditioned and unconditioned stimuli, but they show marked differences from comparison groups in acquisition, magnitude, and timing of the conditioned response (CR). This study examined VPA-induced eyeblink CR timing differences by training rats on an interstimulus interval (ISI) discrimination task, in which two distinct conditioned stimuli (CS; tone and light) are paired with the same unconditioned stimulus (US; periocular shock) at two distinct CS-US intervals. Previous findings suggest that this task would produce abnormally large and prematurely timed CRs for VPA-exposed rats relative to controls. Adult male Long-Evans rats that were exposed to either VPA or saline on GD 12.5 were trained on an ISI discrimination task [Brown KL, Pagani JH, Stanton ME. The ontogeny of interstimulus interval (ISI) discrimination of the conditioned eyeblink response in rats. Behav Neurosci 2006;120: 1057-70.]. In support of earlier findings, we observed early acquisition and enhanced magnitude of the CR in VPA rats compared with controls on long CS trials. VPA rats also showed prematurely timed CRs to long- CS trials, but not to short- CS trials. The ISI discrimination procedure used in the current study reveals differential timed responses in this animal model of autism not previously seen.
This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
Author's personal copy
Behavioural Brain Research 196 (2009) 297–303
Contents lists available at ScienceDirect
Behavioural Brain Research
journal homepage: www.elsevier.com/locate/bbr
Interstimulus interval (ISI) discrimination of the conditioned
eyeblink response in a rodent model of autism
Nathen J. Murawski, Kevin L. Brown, Mark E. Stanton
Department of Psychology, University of Delaware, Newark, DE 19716, United States
Received 2 April 2008
Received in revised form
14 September 2008
Accepted 19 September 2008
Available online 2 October 2008
Rats exposed to valproic acid (VPA) on gestational day 12 (GD12) have been advanced as a rodent model
of autism [Arndt TL, Stodgell, Rodier PM. The teratology of autism. Int J Dev Neurosci 2005;23: 189–99.].
These rats show cerebellar anomalies and alterations in eyeblink conditioning that are associated with
autism. Autistic humans and VPA-exposed rats show normal responses to conditioned and unconditioned
stimuli, but they show marked differences from comparison groups in acquisition, magnitude, and timing
of the conditioned response (CR). This study examined VPA-induced eyeblink CR timing differences by
training rats on an interstimulus interval (ISI) discrimination task, in which two distinct conditioned
stimuli (CS; tone and light) are paired with the same unconditioned stimulus (US; periocular shock) at
two distinct CS–US intervals. Previous ﬁndings suggest that this task would produce abnormally large and
prematurely timed CRs for VPA-exposed rats relative to controls. Adult male Long-Evans rats that were
exposed to either VPA or saline on GD 12.5 were trained on an ISI discrimination task [Brown KL, Pagani
JH, Stanton ME. The ontogeny of interstimulus interval (ISI) discrimination of the conditioned eyeblink
response in rats. Behav Neurosci 2006;120: 1057–70.]. In support of earlier ﬁndings, we observed early
acquisition and enhanced magnitude of the CR in VPA rats compared with controls on long CS trials. VPA
rats also showed prematurely timed CRs to long- CS trials, but not to short- CS trials. The ISI discrimination
procedure used in the current study reveals differential timed responses in this animal model of autism
not previously seen.
© 2008 Elsevier B.V. All rights reserved.
Autism spectrum disorder (ASD) is characterized by social and
language impairments, and restricted or repetitive behaviors .
The incidence of ASDs is on the increase and recent reports
suggest an overall prevalence of 6.7 per 1000 children aged 8
years . Reports concerning an increased incidence of autism
in the offspring of mothers exposed to the speciﬁc teratogens
thalidomide [47,48] and valproic acid (VPA)  during the 1st
trimester supports the hypothesis that early gestational injury to
the brainstem plays a role in the etiology of autism . The
gestational day (GD) 12.5 VPA animal model of autism devel-
oped by Rodier et al.  also gives support to this hypothesis.
Rats exposed to VPA on GD 12.5 show neuroanatomical abnor-
malities which parallel those in human cases of autism .
Speciﬁcally, VPA-exposed rats have an overall reduction of cere-
bellar and posterior vermis volume, of Purkinje cells in both
Corresponding author at: Wolf 132, Department of Psychology. University of
Delaware, Newark, DE 19716, United States. Tel.: +1 302 831 0575;
fax: +1 302 831 3645.
E-mail address: Stanton@psych.udel.edu (M.E. Stanton).
the cerebellar hemispheres and posterior vermis , of neurons
in the interpositus nucleus  as well as neurons in brain-
stem motor nuclei . The advent of a rodent model of autism
based on common developmental etiology and neuroanatom-
ical phenotype creates an opportunity to examine behavioral
parallels between this animal model and the human disorder
As pointed out by Stanton et al. , eyeblink conditioning (EBC)
offers a number of advantages as a behavioral test of a rodent
model of autism. First, an identiﬁed brainstem-cerebellar circuit
is essential for the acquisition and maintenance of delay EBC (for
a review see ref. ). Second, the neuroanatomical abnormali-
ties present in autistic patients and in the GD 12.5 VPA animal
model overlap with the brainstem-cerebellar circuit mediating EBC
[3,8,15,18,19,39,40]. Third, the necessary and sufﬁcient circuitry
underlying EBC is conserved across mammalian species [51,52].
And fourth, EBC conducted in human cases of autism reveal a
unique phenotype of conditioned response (CR) acquisition and
timing . EBC within the VPA model, therefore, can reveal how
early brainstem-cerebellar injury affects subsequent performance
on this behavioral task, which allows for inferences to be made con-
0166-4328/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
Author's personal copy
298 N.J. Murawski et al. / Behavioural Brain Research 196 (2009) 297–303
cerning brainstem-cerebellar function in both VPA treated rodents
and human autism cases.
During an EBC task, autistic patients show a unique pattern
of eyeblink conditioning with enhanced CR acquisition and early
onset CR timing . Stanton and co-workers [43,46] demonstrate
that VPA rats subjected to EBC show a similar phenotypic learning
curve, with dramatically enhanced CR acquisition in VPA rats rela-
tive to controls, giving further behavioral support to the VPA animal
model. The CR timing effects have been demonstrated in the VPA
model during some experiments , but not in others . This
may reﬂect insensitivity of the species or of the particular EBC task
The current experiment attempts to reconcile the apparent dis-
parity of CR timing effects in the VPA model through the use of an
interstimulus interval (ISI) discrimination task that emphasizes CR
timing functions of the cerebellum . During ISI discrimination,
distinctively different conditioning stimuli (CSs) are paired with
the same unconditione d stimulus (US) following different ISIs on
different trials. Given enough training, subjects learn to discrim-
inate the CSs in such a way that CRs are appropriately timed to
each ISI associated with its corresponding CS [6,17,21]. For the cur-
rent experiment, adult rats exposed to either VPA or saline (SAL)
on GD 12.5 were run in a tone/light ISI discrimination task. We pre-
dicted that VPA rats would show enhanced CR acquisition relative
to SAL controls. We also predicted that the CR timing demands of
the ISI discrimination task would reveal early onset CRs to both the
short and long CSs in VPA rats. Portions of this experiment were
previously reported in abstract form .
1.1. VPA dosing
Valproic acid(Sigma) was purchased as the sodium salt and was dissolved in 0.9%
saline for a concentration of 250 mg/ml, pH 7.3, as veriﬁed by enzyme-multiplied
immunoassay conducted at the University of Rochester. The dosing procedure was
based on that of Rodier et al. .
Time-mated females were bred in the animal facility at the University of
Delaware. They were housed with breeder males in suspended cages overnight and
if an ejaculatory plug was found the following morning, that day was designated as
GD1 [cf. 39]. They remained undisturbed in individual cages with ad lib food and
water until GD 7 when they were moved from the vivarium into the lab’s animal
housing unit. On GD 7 and GD 11 the females were weighed (at 10:00 am). On GD
12.5, females were weighed just prior to dosing, in order to determine the dosing
volume. Females were randomly assigned to receive VPA or SAL vehicle but in a
manner that matched the two groups as closely as possible to their pre-dosing body
weight. Weights were also taken daily for 3 days following dosing (GD 13–15) and
again a week after dosing (GD 19).
VPA-treated dams (n = 14) received a single i.p. injection of 600 mg/kg NaVP in
a volume of 2.4 ml/kg at about 10:00 am on GD 12, while control dams (n =11)were
treated with a similar volume of saline vehicle.
The subjects were 33 male Long-Evans rats derived from 25 litters. They were
the offspring of time d-pregnant females bred in the animal facility at the Univer-
sity of Delaware dosed with either VPA or SAL as described above. Illumination was
provided on a 12:12 h light–dark cycle, with lights on at 7:00 a.m. Age of pups was
determined by checking for births during the light cycle and designating the date
of birth as post-natal day (PND) 0 (all births occurred on GD 23). On PND3, litters
were culled to eight pups (usually four males and four females) and were given sub-
cutaneous injections of nontoxic black ink into one or more paws for identiﬁcation.
On PND21, pups were weaned from their mothers and housed in groups of same-
sex littermates in 45 cm × 24 cm × 16 cm plastic cages with standard bedding and
continuously supplied with rat chow and water. On PND30 the pups were further
separated and housed two per cage (45 cm × 24 cm × 21 cm) until surgery. A total
of 22 subjects contributed data to this study. Data from 11 animals were excluded
due to excessive noise in electromyography (EMG) records (1 subject), electrode
displacement and/or excessively low unconditioned response (UR) amplitudes (ﬁve
subjects), equipment failure (three subjects), and loss of head-stages during con-
ditioning (two subjects). The extensive amount of training in the current study (12
sessions) relative to previous reports using this freely moving rodent preparation
(usually six sessions) contributed to the loss of subjects. For the majority of behav-
ioral measures, subjects were assigned to Groups VPA (n = 13) and SAL (n = 9) such
that no more than one rat from a given litter was assigned to a given behavioral
condition (short- CS = tone/long- CS = light, or vice versa) and the treatment groups
were counterbalanced as closely as possible for CS modality. For CR latency and
topography analysis, an additional animal was dropped due to a missing data point
thus giving Groups VPA (n = 13) and SAL (n = 8) (see below). All animals were treated
within the guidelines set forth by the Institutional Animal Care and Use Committee
(IACUC) at the University of Delaware.
Surgeries occurred between PND 60–75 on male offspring of the dams
treated with VPA or saline. Rats were surgically implanted with head stages
containing stimulating and recording electrodes (see ref.  for full details)
under ketamine/xylazine anesthesia (i.p. injection of 87 mg/kg ketamine/13 mg/kg
xylazine in a 1.0 ml/kg injection volume). Differential EMG electrodes were
implanted in the left upper eyelid muscle, and a ground electrode was placed sub-
dermally at the back of the neck. A bipolar stimulating electrode for delivering the
US was implanted subdermally just caudal to the left eye. Electrode connectors were
secured to the skull with dental acrylic and via skull screws. Following surgery, sub-
jects were returned to individual cages that were temporarily heated (<45 min) on
one side by an electric heating pad (GE model #E12107) set to the lowest setting
to allow the rats to self-regulate their temperature during the recovery period after
anesthesia and surgery. The cages were supplied with ad lib food and water and
housed subjects throughout the experiments (except during test sessions). Adult
rats were handled 3 days prior to and 3 days after surgery (∼3 min. each) and given
a minimum of 5 days post surgery to recover before testing.
The “freely-moving rat” conditioning apparatus has been described previously
[31,44,45] (available from JSA Designs, Raleigh, NC). Brieﬂy, it consisted of 16 ani-
mal chambers (BRS/LVE, Laurel, MD) lined with sound-absorbing foam. Within each
chamber, rats were kept in stainless steel wire mesh cages during testing. Each cham-
ber was equipped with a fan that produced background noise of approximately
70 dB, a house light (15 W), and two speakers generating tones, one of which deliv-
ered the auditory CS. The present study used a 70 dB, 2.8 kHz tone presented for
either 380 or 980 ms as the auditory (tone) CS and activation of the house light
(against the dark background) for 380 or 980 ms as the visual (light) CS. A 70 dB
tone was chosen to more closely match conditioning rates to the tone and light .
The US was produced by a constant-current, 60 Hz square wave stimulator (World
Precision Instruments, Sarasota, FL) set to deliver a 2 mA, 100 ms periocular shock.
During conditioning sessions, subjects’ head stages were connected to wire leads
which passed through an opening in the chamber to a commutator which allowed
subjects to move freely about the chamber.
1.5. Design and procedures
On the 6th day post-surgery animals wereplaced in the apparatus and connected
to the recording equipment during a brief (1–2 min) adaptation session were the
quality of the EMG signal was assessed independent of a US through a puff-test (see
ref.  for further details).
On the 7th day following surgery, subjects were again placed into the apparatus
and connected to the recording equipment for two 45 min acclimation periods, once
in the morning and again 5 h (±30 min) later.
On the 8th day post-surgery, subjects began training in a tone-light discrimina-
tion task (described in detail in ref. ). Brieﬂy, each session consisted of 100 trials:
50 trials of a 380 ms CS and 50 trials of a 980 ms CS. Both the short and long CS pre-
ceded and coterminated with a 2 mA , 100 ms periocular-shock US, producing ISIs
of 280 and 880 ms, respectively. CS modality was counterbalanced such that light
was the short CS (and therefore tone was the long CS) for about half of the subjects,
and vice versa for the other half. Trials were presented in a pseudorandom order
such that there were no more than three consecutive trials with the same CS. The
ITI varied randomly around an average of 30 s (range = 18–42 s). Within a block of
10 trials, there were ﬁve presentations of each CS, and the CS+ was paired with the
US on 4 out of 5 trials (the 5th trial was a CS-alone test trial). Training consisted of 2
sessions per day spaced 5 h apart (±30 min) over 6 consecutive days (between PND
68 and PND 83 depending on age at surgery) for a total of 12 sessions.
1.6. Data analysis
EMG signals were sampled in 3.5 ms bins during the 1400 ms epoch of each trial
type. The raw signal was rectiﬁed and integrated for analysis.As described elsewhere
, each trial epoch was divided into (1) a 280-ms pre-CS baseline period, (2) an
80-ms alpha or startle period (ﬁrst 80 ms after CS onset), (3) a CS period (200-ms
for short CS period or 800-ms for long CS), and (4) a 140-ms US period (time from
offset of the US to the end of the trial – the recording was interrupted during the
100-ms presentation of US to avoid stimulation artifact). On CS-alone test trials,
the CR sampling periods extended to the end of the trial and therefore additionally
Author's personal copy
N.J. Murawski et al. / Behavioural Brain Research 196 (2009) 297–303 299
Mean (±S.E.) UR maximum amplitudes (UMA), and startle maximum amplitudes (SMA) during acquisition as a function of treatment group (SAL, Saline treated controls; and
VPA, 600 mg/kg VPA administered on GD12, see text for full explanation). The absence of treatment effects indicates that alterations in conditioning in this study were not
secondary to primary sensory or motor effects of VPA exposure.
Group UMA SMA
Short CS Long CS
Tone CS Light CS Tone CS Light CS
SAL (n = 9) 671.64 ± 60.14 8.14 ± 5.90 1.54 ± 1.54 22.00 ± 8.85 0.63 ± 0.63
VPA (n = 13) 704.36 ± 52.69 43.04 ± 34.53 2.74 ± 1.22 35.33 ± 8.85 5.05 ± 2.59
included the period designated as the “US period” on paired trials. The threshold for
registering an EMG response was set to 40 arbitrary units above the average baseline
amplitude during the pre-CS period (e.g., [42,46]). Following previous convention
with this preparation, CR amplitude measures counted trials in which CRs were not
present as “zero amplitude”. CR percentage and amplitude measures are reported
from CS+ trials, whereas CR latency measures are taken from CS-alone trials.
Data were statistically analyzed with analyses of variance (ANOVAs) and post
hoc tests (Newman-Keuls) unless otherwise noted. Performance to the short- CS
and long- CS were analyzed separately. ANOVA involved the between-subjects fac-
tors of dosing condition (VPA vs. SAL) and CS modality (tone vs. light) as well
as the within-subject factor of sessions. The effects of CS modality were gener-
ally not statistically signiﬁcant or, when signiﬁcant, were too small numerically to
alter conclusions concerning the effects of the other factors and therefore are not
2.1. Growth and body weight
Prior to dosing on GD12.5, mean (±SE) body weights of SAL- and
VPA-treated pregnant females were 299 (±7.5) and 311 (±5.85) g,
respectively, and did not differ signiﬁcantly (t < 1.25). Following
dosing, body weights of SAL-treated females on GD13, 14, 15 and 19,
respectively, were 304 (±7.98), 306 (±7.80), 311 (±8.07), and 345
(±8.95) g; whereas the corresponding values for VPA females were
306 (±5.90), 310.18 (±5.76), 312.18 (±5.96), and 342.09 (±6.61) g.
Maternal body weights of the two groups never differed signiﬁ-
cantly across any of these days (all ps > 0.88). When litters were
culled on PND3, SAL- and VPA-treated offspring weighed 8.28
(±0.48) and 7.08 (±0.35) g, respectively, a non-signiﬁcant differ-
ence (p > 0.34; averages are combined across sex because body
weights of males and females did not differ signiﬁcantly). There
was also no treatment effect on litter size, VPA, 11.0 (±0.25) lit-
termates vs. Saline, 11 (±0.30) littermates (p > 0.35). As adults, the
mean body weights for VPA animals prior to the ﬁrst session of
eyeblink conditioning was 363.62 (±30.44) g and that of SAL con-
trols was 370.22 (±52.91) g. An independent groups t-test revealed
no signiﬁcant differences in mean body weights between groups
(p > 0.70).
2.2. Sensory processing of CS and US
Measures of sensory processing and motor performance that
are independent of learning were assessed. Startle response (SR)
maximum amplitude during Session 1 acquisition, a measure of CS
processing, and unconditioned response (UR) maximum amplitude
during the ﬁrst 10-trial acquisition block, a measure of US efﬁcacy
and motor performance, are shown in Table 1. VPA rats showed
considerable variability in SRs to the short tone CS, though no main
effects or interactions were present (ps > 0.32). SR response to the
long CS revealed a signiﬁcant modality effect (p < .01) with both
VPA and SAL rats showing signiﬁcantly higher SRs to the long tone
CS (mean VPA = 35.33, SAL = 22.00) compared to the long light CS
(mean VPA = 5.05, SAL 0.63), without group differences or inter-
actions of dosing condition × modality (ps > .29). The absence of
treatment effects on these measures indicates that alterations in
conditioning in this study were not secondary to primary sensory
or motor effects of VPA exposure.
For both dosing conditions, there was a signiﬁcant increase in
CR percentage (Fig. 1, top panel) and in CR amplitude (Fig. 1,bottom
panel) across sessions for both the short and long CSs. VPA treated
rats showed enhance d CR acquisition (CR% and CR amplitude) to
the long CS relative to saline controls (Fig. 1, middle panels), and
this also occurred in the adaptive CR measure (Fig. 1, right panels).
Differences in CR timing were present, with the onset of CRs for the
long CS in group VPA occurring signiﬁcantly earlier than the onset
of CRs for group SAL (Fig. 2, right panel).
2.4. Short- CS trials
In the CR percentage measure, both groups displayed robust
conditioning to the short CS with CRs averaging around 20% dur-
ing Session 1 and reaching asymptotic levels of 85% by Session 3
, top left panel). Statistically there were no signiﬁcant effects
(all Fs < 0.40) except for a main effect of sessions, F(11, 198) = 52.14,
p < 0.01. In the CR peak amplitude measure, both groups displayed
a statistically signiﬁcant increase in peak amplitude over sessions,
F(11, 19 8) = 33.25, p < 0.01, though no other statistically signiﬁcant
main effects (F < 1.94) or interactions (F < 0.78) were present (Fig. 1,
bottom left panel).
2.5. Long- CS trials
Analysis of CR percent to the long CS revealed a statistically
signiﬁcant main effect of group, F(1, 18) = 4.47, p < 0.05, and ses-
sions, F(11, 198) = 15.28, p < 0.01, but no signiﬁcant interaction of
Group × Sessions (Fig. 1, top middle panel). There was also a sig-
niﬁcant main effect of sessions in the CR peak amplitude measure,
F(11, 198) = 15.84, p < 0.01, and a trend towards a signiﬁcant main
effect of group (p < 0.063) (Fig. 1, bottom middle panel). A separate
analysis of CR peak amplitude consisting of only trials where CRs
occurred revealed a similar, though non-signiﬁcant enhancement
in VPA rats relative to SAL controls (data not shown). These mea-
sures indicate an overall enhancement to CR acquisition in group
VPA relative to SAL controls.
2.6. Long- CS trials (Adaptive CR Measure)
Measures of the well-timed CR to the long CS are conﬁned to
the last 200 ms of each trial for analysis . This measure empha-
sizes CRs that anticipate the time of US onset and are commonly
referred to as the “adaptive CR” in the literature . This measure
is also more directly comparable with measures of the CR to the
short CS because the sampling period (200 ms) is the same. During
long CS trials, there was a trend towards a group main effect in the
adaptive CR percentage measure, F(1, 18) = 4.05, p < 0.06 (Fig. 1,top
right panel). In the CR peak amplitude measure, analysis revealed a
Author's personal copy
300 N.J. Murawski et al. / Behavioural Brain Research 196 (2009) 297–303
Fig. 1. Mean (±S.E.) conditioned response (CR) percentage (top panels) and CR amplitude (bottom panels) for the short CS (left panels), long CS (middle panels), and long
CS (“adaptive” CR measure) for groups VPA and SAL as a function of 12, 100-trial sessions. VPA rats show enhanced CR acquisition in the long CS and adaptive CR measures
relative to SAL controls. VPA (n = 13); SAL (n = 9). “Adaptive” measure represents only the last 200 ms of the long CS trial .
statistically signiﬁcant main effect of group, F(1, 18) = 5.37, p < 0.04,
and Group x Sessions interaction, F(11, 198) = 2.86, p < 0.01 (Fig. 1,
bottom right panel). Newman-Keuls analyses of this latter interac-
tion revealed higher CR peak amplitudes for group VPA relative to
group SAL only over sessions 7–12 (all ps< 0.05).
2.7. CR latency
For both groups separate ANOVAs were used to analyze CR onset
and peak latency to the short and long CS on CS-alone test trials in
which a CR occurred. Due to a missing data point in one SAL animal,
latency and CR topography analyses (see below) are taken from
thirteen VPA and eight SAL animals. Measures to both the short and
long CS declined for both treatment groups, approaching asymptote
by Session 7. Therefore all latency measures are analyzed only over
Sessions 7–12 (see Fig. 2). CR peak latencies to either the short CS
or long CS did not differ signiﬁcantly between groups (Fs < 1.30). CR
onset latencies to the short CS also did not differ between groups
(F < 0.37). However, analysis revealed a signiﬁcant main effect of
groups for the long- CS CR- onset latency measure, F(1, 17) = 4.74,
p < 0.05 (see Fig. 2, right panel).
2.8. CR topography
Fig. 3 depicts trial tracings representing group averages to the
short CS (left panel) and long CS (right panel) for Session 12
CS-alone trials. Trial tracings reveal a distinctive enhancement of
double-peaks to the long CS for VPA rats relative to SAL controls.
Double-peak CRs were analyzed from Session 12 CS-alone trials fol-
lowing similar criteria to that outlined by Brown et al. . Brieﬂy,
CRs were deﬁned as having an initial increase of 100 arbitrary units
from baseline, having a minimum duration of 50 ms, and having
a decrease of 50 arbitrary units from peak amplitude. A second CR
(double-peak) was assessed by the same criteria with the additional
qualiﬁcation of the second peak being a minimum 200 ms apart
from the ﬁrst peak. VPA-treated rats showed an average of 64.46%
double-peaks, while SAL controls averaged 31.78% double-peaks
(see Fig. 4). A one-way ANOVA revealed a statistically signiﬁcant dif-
ference between these groups, F(1, 20) = 5.12, p < 0.04. Double-peak
CRs to the short CS did not reveal signiﬁcant differences between
groups (data not shown).
The current experiment tested VPA-treated and SAL- control rats
on an interstimulus interval discrimination eyeblink conditioning
task. VPA and SAL rats both acquired appropriately timed CRs, as
measured by CR peak latencies, to both the short and long CSs. The
two groups did not differ on CR acquisition measures to the short
CS. However, VPA rats did show signiﬁcant increases in CR percent
and CR amplitude and signiﬁcantly earlier CR onset latencies to the
long CS relative to SAL controls. The dif ferences between treatment
groups were not due to differences in CS or US efﬁcacy or motor
performance as no differences were seen in SR and UR measures.
The current ﬁndings of appropriately timed CR peak latencies
(in both VPA and SAL rats) and CR onset latencies (in SAL rats) to
both short and long CSs are similar to those found in other studies
using an ISI discrimination task [6,17]. Differences in onset laten-
cies to the long CS in VPA rats seems unique to this VPA animal
model (see below). A number of ﬁndings from the current experi-
ment are in accord with previous studies of this VPA rodent model
and human autism [3,41,46]. First, the current and previous stud-
ies report that primary sensory processing of CS, US, and motor
performance (unconditioned eyeblink reﬂex) are not altered in the
VPA rodent model or in human autism. Second, these studies also
show enhanced CR acquisition as measured by CR percent and CR
amplitude that was observed in the current study. We did not see
latency differences between groups to the short (380 ms) CS. This
Author's personal copy
N.J. Murawski et al. / Behavioural Brain Research 196 (2009) 297–303 301
Fig. 2. Mean (±S.E.) conditioned response (CR) onset latency (open or ﬁlled circles) and peak latency (open or ﬁlled downward triangles) from conditioned stimulus (CS)-alone
trials to the short (left) and long (right) CS for Groups VPA and SAL. Zero on the y-axis indicates CS onset. US onset to the short CS occurs at 280 ms while US onset to the long
CS occurs at 880 ms (dashed horizontal lines). VPA rats show early CR onset latencies to the long CS relative to SAL controls while there were no differences between groups
observed in CR onset to the short CS or peak latencies to both short and long CS. VPA (n = 13); SAL (n = 8). US, unconditioned stimulus.
ﬁnding replicates those of Stanton et al.  who reported no onset
latencies differences between VPA and SAL rats during a 380 ms
tone/light CS discrimination reversal task. However our latency
results to the short CS differ, from those reported in an additional
experiment by Stanton et al.  in which early onset latencies
were detected in VPA rats relative to SAL controls during short
delay (380 ms) conditioning. Our results also differ from human
autism studies that show prematurely timed CRs during eyeblink
conditioning [3,41]. As suggested by Stanton et al. , the fail-
ure to show prematurely timed CRs to the short CS in the current
study may reﬂect the nature of the current task. Those experiments
reporting prematurely timed CRs used single-cue eyeblink condi-
tioning, while Stanton et al.  and the present study employed
discriminative conditioning. Why differences in CR timing are seen
in single-cue vs. discrimination learning requires further study. The
early onset CR to the long CS in VPA rats in the current study how-
ever, suggest that CR timing is affected in this animal model of
autism, especially under long-ISI training conditions in which CR
timing is arguably more difﬁcult. It would be of interest to examine
CR timing during single-cue long-delay conditioning in this VPA
rodent model. Taken together, the current ISI discrimination task
used in this study has revealed a unique within-subject compari-
son of CR timing differences in the VPA animal model of autism,
whereby VPA rats show appropriately timed peak CR latencies to
both short and long CSs with early onset latencies only to the long
VPA rats in the current study also demonstrated double-peak
CRs to the long CS, a unique CR topography that is less pronounced
in SAL controls. Both the short and long CSs in the current study are
of different modalities and the presence of double-peak CRs sug-
gests CS generalization from the short CS to the long CS. This ﬁnding
of double-peak CRs in addition to the enhanced CR amplitude to
Fig. 3. Tracings of integrated electromyography activity from Session 12 CS-alone trials for groups VPA and SAL. Average tracing across all rats for the short (left) and long
(right) CS-alone trials. The solid vertical line represents the onset time of the CS, and the two dashed vertical lines represent the onset times of the short and long unconditioned
stimulus, respectively. VPA (n = 13); SAL (n = 8).
Author's personal copy
302 N.J. Murawski et al. / Behavioural Brain Research 196 (2009) 297–303
Fig. 4. Mean (±S.E.) percent double-peak conditioned responses (CR) from long
conditioned stimulus (CS)-alone trials for groups VPA and SAL during Session 12.
The presence of double-peak CRs, especially in VPA rats, suggests poorer temporal
discrimination or CR timing to the long CS. VPA (n = 13); SAL (n = 8).
the long CS in adult VPA rats relative to SAL controls is particularly
interesting in light of ﬁndings by Brown et al. , which show that
post-weanling (PND 23) and juvenile (PND 30) rats run in this ISI
discrimination procedure demonstrated enhanced CR amplitudes
and double-peak CRs to the long CS relative to adults (PND 80–85).
Additional ﬁndings also suggest that juvenile rats show early onset
CR latencies to the long CS (Brown et al., unpublished observations).
CR timing has been shown to improve over development (see ref.
 for a discussion of potential underlying mechanisms) and it is
possible that VPA exposure delayed this developmental improve-
ment in the current study. It is possible that poor discrimination
of the tone and light CSs contributed to the performance of VPA
rats in this study, although this is unlikely to be a sensory effect
because these rats show enhanced discrimination of these cues in
a CS+/CS-paradigm . The mechanisms responsible for adult VPA
rats showing CR timing characteristics of weanling and juvenile rat
warrant further study.
Eyeblink conditioning is a powerful tool in behavioral neu-
roscience because the underlying neural pathways mediating
conditioning have been elucidated by lesions, reversible inacti-
vation, and electrophysiological recordings [11,13,49]. CS and US
afferents travel, respectively, from pontine nuclei via mossy ﬁbers
and the inferior olive via climbing ﬁbers to converge at both
the interpositus nucleus and the cerebellar cortex. The interposi-
tus nucleus is an essential convergent site mediating associative
eyeblink learning [26,49] while the cerebellar cortex has been
hypothesized to modulate appropriate CR timing [25,26,32]. Purk-
inje cells form the sole output from the cerebellar cortex and
produce an inhibitory tone on the interpositus, the main motor
output of the cerebellum. Recordings from lobule HVI (see ref. ,
references therein) and the anterior lobe  reveal Purkinje cell
activation in the early to middle of the CS–US interval that decreases
later in the CS–US interval allowing for appropriately timed CR
expression. Lesions of the cerebellar cortex, or pharmacological dis-
ruption of cerebellar cortical projections to the deep nuclei, produce
early onset and peak latencies and a reduction of CR amplitude (cal-
culated only from CR trials with CRs present) [16,29,32]. The current
study reveals early onset latencies to the long CS in VPA animals.
However, the appropriately timed peak latencies and enhancement
in CR amplitude in VPA animals is inconsistent with lesion or inac-
tivation studies of the cerebellar cortex.
In addition to the cerebellar cortex, the hippocampus and
the amygdala also contribute to CR timing. In some studies,
hippocampal lesions are shown to produce early onset CRs
with CR acquisition enhancement [22,33] (but also see ref.
[30,34]). In contrast, lesions and reversible inactivation of the
amygdala produce delayed onset latencies with reduced CR acqui-
sition [5,10,22] (see also ref.  for reduced CR acquisition
following amygdala lesions). It is possible that disruptions or alter-
ations at any or all of these levels (e.g., the cerebellar cortex,
the hippocampus, and the amygdala) could directly affect tim-
ing properties of the CR during eyeblink conditioning in VPA
An important question still to be answered is why VPA rats
and human autism subjects show facilitated eyeblink condition-
ing. Impairments in eyeblink conditioning have consistently been
shown in subjects with cerebellar injuries [51,52]. As addressed
by Stanton et al. , the early 1st trimester insult sustained by
VPA rats may allow for compensatory changes in brain devel-
opment both within or outside the cerebellum to occur. Though
the cerebella of VPA animals and human autistic subjects show
reductions in Purkinje cell number and cerebellar volume [14,19],
those reductions do not necessarily correlate with reduced neu-
ral activity of the cerebellum. For example, in a recent fMRI study
by Allen et al. , autistic and control subjects showed activa-
tion of the anterior cerebellum and lobule HVI during a simple
motor task. However, in autistic subjects the spatial extent and
magnitude of activation was greater, even though the regions of
interest in the autistic cerebellum were 12% smaller than con-
trols . A neurophysiological study of the VPA rat cerebellum has
yet to be undertaken, but the facilitation of eyeblink conditioning
seen in VPA rats may be due to an underlying enhanced activa-
tion of the eyeblink circuit. Markram and co-workers [35,36] have
shown neocortical enhancements in VPA rats: in vitro electrophys-
iological recordings of pyramidal neurons of the somatosensory
cortex of VPA rats show enhancement of local recurrent connec-
tivity ; an additional study also revealed overexpression of
NR2A and NR2B subunits of n-methyl-d-aspartate (NMDA) recep-
tors, increased NMDA receptor-mediated synaptic currents, and
increased postsynaptic long-term potentiation in the neocortex of
VPA rats . Such an enhancement within the cerebellum in this
model of autism could help explain our ﬁndings. The facilitatory
mechanisms may also have their origin outside of the cerebellum.
A recent study by Markram et al.  demonstrated enhanced fear
conditioning in VPA rats, whose amygdala were hyper-reactive to
electrical stimulation, showed enhanced synaptic plasticity, and
deﬁcits in inhibition. It is interesting to note that direct stimulation
of the central nucleus of the amygdala or prior fear condition-
ing enhances CR amplitude during eyeblink conditioning [7,28,50].
Taken together, an “extra-cerebellar” hypothesis of facilitated eye-
blink conditioning in VPA rats and human autism might suggest
that a hyperactive amygdala within these subjects facilitates eye-
blink conditioning. Electrophysiological and reversible inactivation
studies of the amygdala in VPA rats might help to elucidate this
Overall, the current experiment has replicated the enhanced
CR acquisition effect in eyeblink conditioning within the GD 12.5
animal model of autism. The ISI discrimination task revealed an
interesting proﬁle of CR timing effects where both VPA and SAL
rats showed appropriate peak latencies to both short and long CSs.
VPA subjects, however, showed early onset latencies to the long
CS, but not to the short CS. As suggested by Mauk and Ruiz ,
the ISI discrimination procedure used in the current study allows
for within-animal comparisons of CR timing to temporally distinct
CSs. The unique pattern of CR timing observed in VPA rats in the cur-
rent study suggests anatomical or physiological differences in brain
regions mediating appropriate CR timing in this ISI discrimination
Author's personal copy
N.J. Murawski et al. / Behavioural Brain Research 196 (2009) 297–303 303
The authors would like to thank Kathy K. Chadman, Deborah J.
Watson, and Huan Bao Duong for technical assistance. This research
was supported by the University of Delaware, and by NIH grant
 Diagnostic and Statistical Manual of Mental Disorders, American Psychiatric
 Allen G, Muller RA, Courchesne E. Cerebellar function in autism: functional
magnetic resonance image activation during a simple motor task. Biol Psychi-
 Arndt TL, Chadman KC, Watson DJ, Tsang V, Peloso ED, Rodier PM, Stanton ME.
Eyeblink conditioning in autism and a rodent model. II. Long-delay eyeblink
conditioning in autism. Dev Psychobiol [in press].
 Arndt TL, Stodgell CJ, Rodier PM. The teratology of autism. Int J Dev Neurosci
 Blankenship MR, Huckfeldt R, Steinmetz JJ, Steinmetz JE. The effects of amyg-
dala lesions on hippocampal activity and classical eyeblink conditioning in rats.
Brain Res 2005;1035:120–30.
 Brown KL, Pagani JH, Stanton ME. The ontogeny of interstimulus interval (ISI)
discrimination of the conditioned eyeblink response in rats. Behav Neurosci
 Canli T, Brown TH. Amygdala stimulation enhances the rat eyeblink reﬂex
through a short-latency mechanism. Behav Neurosci 1996;110:51–9.
 Carper RA, Courchesne E. Inverse correlation between frontal lobe and cerebel-
lum sizes in children with autism. Brain 2000;123:836–44.
 CDC. Prevalence of autism spectrum disorders – autism and developmen-
tal disabilities monitoring network, six Sites, United States. MMWR 2007;56:
 Choi JS, Lindquist DH, Brown TH. Amygdala lesions block conditioned enhance-
ment of the early component of the rat eyeblink reﬂex. Behav Neurosci
 Christian KM, Thompson RF. Neural substrates of eyeblink conditioning: acqui-
sition and retention. Learn Mem 2003;10:427–55.
 Christianson AL, Chesler N, Kromberg JG. Fetal valproate syndrome: clinical
and neuro-developmental features in two sibling pairs. Dev Med Child Neurol
 Fanselow MS, Poulos AM. The neuroscience of mammalian associative learning.
Annu Rev Psychol 2005;56:207–34.
 Fatemi SH, Halt AR, Realmuto G, Earle J, Kist DA, Thuras P, et al. Purkinje
cell size is reduced in cerebellum of patients with autism. Cell Mol Neurobiol
 Gaffney GR, Kuperman S, Tsai LY, Minchin S. Morphological evidence for brain-
stem involvement in infantile autism. Biol Psychiatry 1988;24:578–86.
 Garcia KS, Mauk MD. Pharmacological analysis of cerebellar contributions to
the timing and expression of conditioned eyelid responses. Neuropharmacol-
 Green JT, Steinmetz JE. Purkinje cell activity in the cerebellar anterior lobe after
rabbit eyeblink conditioning. Learn Mem 2005;12:260–9.
 Hashimoto T, Tayama M, Murakawa K, Yoshimoto T, Miyazaki M, Harada M, et
al. Development of the brainstem and cerebellum in autistic patients. J Autism
Dev Disord 1995;25:1–18.
 Ingram JL, Peckham SM, Tisdale B, Rodier PM. Prenatal exposure of rats to
valproic acid reproduces the cerebellar anomalies associated with autism. Neu-
rotoxicol Teratol 2000;22:319–24.
 Ivkovich D, Paczkowski CM, Stanton ME. Ontogeny of delay versus trace eye-
blink conditioning in the rat. Dev Psychobiol 20 00;36:148–60.
 Kehoe EJ, Horne PS, Horne AJ. Discrimination learning using different CS-
US intervals in classical conditioning of the rabbit’s nictitating membrane
response. Psychobiology 1993;21:277–85.
 Lee T, Kim JJ. Differential effects of cerebellar, amygdalar, and hippocam-
pal lesions on classical eyeblink conditioning in rats. J Neurosci 2004;24:
 Lindquist DH, Brown TH. Amygdalar NMDA receptors control the expression
of associative reﬂex facilitation and three other conditional responses. Behav
 Markram K, Rinaldi T, Mendola DL, Sandi C, Markram H. Abnormal Fear
conditioning and amygdala processing in an animal model of autism. Neuropsy-
 Mauk MD, Ruiz BP. Learning-dependent timing of Pavlovian eyelid responses:
differential conditioning using multiple interstimulus intervals. Behav Neu-
 McCormick DA, Thompson RF. Cerebellum: essential involvement in the clas-
sically conditioned eyelid response. Science 1984;223:296–9.
 Murawski NJ, Brown KL, Chadman KK, Stanton ME. Interstimulus interval dis-
crimination conditioning of the eyeblink reﬂex in a rodent model of autism.
Paper presented at the Neurobehavioral Teratology Society, Pittsburgh, Penn-
sylvania, June 23–27; 2007.
 Neufeld M, Mintz M. Involvement of the amygdala in classical conditioning of
eyeblink response in the rat. Brain Res 2001;889:112–7.
 Ohyama T, Mauk M. Latent acquisition of timed responses in cerebellar cortex.
J Neurosci 2001;21:682–90.
 Orr WB, Berger TW. Hippocampectomy disrupts the topography of conditioned
nictitating membrane responses during reversal learning. Behav Neurosci
 Paczkowski C, Ivkovich D, Stanton ME. Ontogeny of eyeblink conditioning using
a visual conditional stimulus. Dev Psychobiol 1999;35:253–63.
 Perrett SP, Ruiz BP, Mauk MD. Cerebellar cortex lesions disrupt
learning-dependent timing of conditioned eyelid responses. J Neurosci
 Port RL, Mikhail AA, Patterson MM. Differential effects of hippocampectomy
on classically conditioned rabbit nictitating membrane response related to
interstimulus interval. Behav Neurosci 1985;99:200–8.
 Port RL, Romano AG, Steinmetz JE, Mikhail AA, Patterson MM. Retention and
acquisition of classical trace conditioned responses by rabbits with hippocam-
pal lesions. Behav Neurosci 1986;100:745–52.
 Rinaldi T, Kulangara K, Antoniello K, Markram H. Elevated NMDA receptor lev-
els and enhanced postsynaptic long-term potentiation induced by prenatal
exposure to valproic acid. Proc Natl Acad SciUSA2007;104:13501–6.
 Rinaldi T, Silberberg G, Markram H. Hyperconnectivity of local neocortical
microcircuitry induced by prenatal exposure to valproic acid. Cereb Cortex
 Rodier PM. The early origins of autism. Sci Am 2000;282:56–63.
 Rodier PM. Converging evidence for brain stem injury in autism. Dev Psy-
 Rodier PM, Ingram JL, Tisdale B, Nelson S, Romano J. Embryological origin for
autism: developmental anomalies of the cranial nerve motor nuclei. J Comp
 Saitoh O, Courchesne E. Magnetic resonance imaging study of the brain in
autism. Psychiatry Clin Neurosci 1998;52 Suppl:S219–22.
 Sears LL, Finn PR, Steinmetz JE. Abnormal classical eye-blink conditioning in
autism. J Autism Dev Disord 1994;24:737–51.
 Skelton RW. Bilateral cerebellar lesions disrupt conditioned eyelid responses
in unrestrained rats. Behav Neurosci 1988;102:586–90.
 Stanton ME, Chadman KK, Watson DJ, Rodier PM. Eyeblink conditioning in
autism and a rodent model. II. Conditioned response acquisition and timing in
weanling and adult rats following early gestational exposure to valproic acid.
Dev Psychobiol [in press].
 Stanton ME, Freeman JH. Eyeblink conditioning in the infant rat: an animal
model of learning in developmental neurotoxicology. Environ Health Perspect
 Stanton ME, Freeman JH, Skelton RW. Eyeblink conditioning in the developing
rat. Behav Neurosci 1992;106:657–65.
 Stanton ME, Peloso E, Brown KL, Rodier PM. Discrimination learning and rever-
sal of the conditioned eyeblink reﬂex in a rodent model of autism. Behav Brain
 Stromland K, Miller MT. Thalidomide embryopathy: revisited 27 years later.
Acta Ophthalmol (Copenh) 1993;71:238–45.
 Stromland K, Nordin V, Miller M, Akerstrom B, Gillberg C. Autism in thalidomide
embryopathy: a population study. Dev Med Child Neurol 1994;36:351–6.
 Thompson RF. In search of memory traces. Annu Rev Psychol 2005;56:1–23.
 Whalen PJ, Kapp BS. Contributions of the amygdaloid central nucleus to the
modulation of the nictitating membrane reﬂex in the rabbit. Behav Neurosci
 Woodruff-Pak DS, Steinmetz JE. Eyeblink classical conditioning. Applications
in humans, vol. 1. Kluwer Academic Publishers; 20 00.
 Woodruff-Pak DS, Steinmetz JE. Eyeblink classical conditioning. Applications
models, vol. 2. Kluwer Academic Publishers; 2000.