Lateralized Response Timing Deficits in Autism
Anna-Maria D’Cruz, Matthew W. Mosconi, Shelly Steele, Leah H. Rubin, Beatriz Luna, Nancy Minshew,
and John A. Sweeney
Background: Procedural learning is an implicit process in which a behavioral response is refined through repeated performance. Neural
systems supporting this cognitive process include specific frontostriatal systems responsible for the preparation and timing of planned
motor responses. Evaluating performance on procedural learning tasks can provide unique information about neurodevelopmental
disorders in which frontostriatal disturbances have been reported, such as autism.
Methods: Fifty-two individuals with autism and 54 age-, IQ-, and gender-matched healthy individuals performed an oculomotor serial
reaction time task and a sensorimotor control task.
Results: Whereas the rate of procedural learning and the precision of planned motor responses were unimpaired in autism, a lateralized
alteration in the ability to time predictive responses was observed. Rightward saccadic responses were speeded in individuals with autism
relative to healthy control subjects.
Conclusions: Speeded rightward predictive saccades suggest atypical functioning of left hemisphere striatal chronometric systems in
cortex, procedural learning, striatum
itive patterns of interests and behavior. A broad pattern of
neuropsychological and neurological impairments have been
associated with autism (1).
One cognitive process that has received little attention in
autism research is procedural learning, an implicit process in
which behavioral responses are refined through repeated perfor-
mance. Well-characterized frontostriatal and frontocerebellar
loops subserve procedural learning (2,3), with the relative con-
tribution of these systems determined in part by the duration of
time intervals between component responses of behavioral
sequences. Learning and enacting precise motor sequences
depends on internal chronometric systems, neural time-keeping
systems that maintain and regulate timed motor responding.
Typically, the cerebellum controls the execution of rapid sequen-
tial motor responses with movement intervals of up to .5–1 sec,
whereas the timing of slower motor sequences relies more on
striatal chronometric systems (4,5). Deficits in striatum and
cerebellum have been observed in neuroimaging (6–11) and
histopathological studies (12,13) of autism, providing a rationale
for investigating components of procedural learning in this
disorder. Moreover, some but not all studies have found that
timing abilities are impaired in patients with autism (14,15).
A common approach for investigating procedural learning is
to study skill acquisition in serial reaction time tasks, in which the
speed of performance of a motor sequence improves with
practice (16). Mostofsky et al. (14) examined implicit learning of
utism is a pervasive neurodevelopmental disorder char-
acterized by disturbances in social interactions, verbal
and nonverbal communication, and restricted and repet-
a repeating 10-component sequence and reported that adoles-
cents with autism showed less decline than healthy control
subjects in response latencies with repeated presentation of the
The predictive saccade task provides a simple and rapid test
of procedural learning. It typically requires subjects to track
visual targets that alternate between two locations at a fixed time
interval, to which individuals quickly learn to anticipate the
sequence as evidenced by rapid speeding of reaction times
(17,18). Predictive saccades have latencies that are sufficiently
brief (?90 msec) to indicate that they are planned and initiated
in advance of target appearance.
The periodicity of the alternating target in the predictive
saccade task is usually within the seconds range. Therefore,
learning to accurately time predictive responses on this task
depends more on frontostriatal systems than cerebellum, consis-
tent with evidence from previous studies (19,20). Also, impaired
performance on the predictive saccade task has been observed in
disorders affecting the basal ganglia, including Parkinson’s (17)
and Huntington’s diseases (21).
Several types of information can be derived from an analysis
of predictive saccade task performance. The rate of learning over
trials can be evaluated by examining the reduction in reaction
times with practice. The ability to accurately time predictive
saccades serves as an index of the integrity of striatal response-
timing systems. Evaluating the accuracy of predictive saccades
provides information on the precision of voluntary motor re-
sponses initiated without sensory guidance. Because of the
strong lateralization of oculomotor systems, the presence of
lateralized deficits in each of these processes can be identified. In
this study, individuals with autism and matched healthy control
participants performed a predictive saccade task and a sensori-
motor control paradigm to assess the integrity of these three
processes in autism.
Methods and Materials
Fifty-two high-functioning individuals with autism and 54
healthy control participants (5 female subjects/group) were
matched on age [mean age (SD), range: 19.6 (11.3), 8–53 years,
and 20.3 (12.2), 8–56 years, respectively] and full-scale IQ [mean
(BL, NM), Pittsburgh, Pennsylvania.
Address reprint requests to John A. Sweeney, Ph.D., Center for Cognitive
Medicine, 912 S. Wood St., MC 913, University of Illinois at Chicago,
Chicago, IL 60612; E-mail: email@example.com.
Received August 29, 2008; revised January 7, 2009; accepted January 7,
BIOL PSYCHIATRY 2009;xx:xxx
© 2009 Society of Biological Psychiatry
ARTICLE IN PRESS
IQ (SD): 108.0 (16.8) and 110.5 (15.4), respectively]. All partici-
pants had a full-scale IQ ? 80 and far visual acuity of at least
20/40 (corrected or uncorrected).
Individuals with autism met DSM-IV criteria for autistic disor-
der on the basis of the Autism Diagnostic Interview–Revised (22)
and Autism Diagnostic Observation Schedule-Generic (23). Par-
ticipants diagnosed with a genetic or metabolic disorder known
to be associated with autism were excluded (e.g., Fragile-X,
tuberous sclerosis), as were those with a lifetime history of head
injury or seizure disorder. Participants were free of medications
known to affect cognitive or oculomotor abilities, including
antipsychotics, methylphenidate, amphetamine, and anticonvul-
Healthy participants had no first- or second-degree relatives
with a history of a neuropsychiatric disorder known to have a
genetic component, including autism. Written informed consent
or written assent from minors (in addition to written parental
consent) was obtained from all participants. Study procedures
were approved by the Institutional Review Board at the Univer-
sity of Pittsburgh.
Eye Movement Studies
Participants were seated in a darkened black room facing a
black arc of 1 m radius containing red light-emitting diodes
(LEDs) embedded in the horizontal plane at eye-level. The LEDs
subtended approximately .2° of visual angle and were not visible
unless illuminated. A chin and forehead rest with occipital
restraints and head strap was used to minimize head movement.
Eye movements were recorded with infra-red (IR) scleral-
reflection sensors mounted on spectacle frames (Model 210;
Applied Science Laboratories, Bedford, Massachusetts) or DC
electro-oculography (EOG) for those whose uncorrected far
visual acuity was ? 20/40, so that subjects could use corrective
lenses (Grass Neurodata 12, Astro-Med, West Warwick, Rhode
Island). Blinks were monitored with AC-coupled electrodes
placed above and below the left eye. Recordings were digitized
at 500 Hz (DI-210 14-bit A/D; DATAQ Instruments, Akron, Ohio),
stored for off-line analysis, and analyzed with custom software
developed in our laboratory.
Visually Guided Saccade Task
A visually guided saccade control task was first administered
to evaluate basic sensorimotor processes. Participants main-
tained central fixation for 1.5–2.5 sec at the start of each trial and
then looked toward a peripheral target presented pseudo-ran-
domly at one of six angular displacements in the horizontal plane
(? 10°, ? 20°, or ? 30°). Fifty-four trials were presented.
Peripheral targets appeared concurrently with the offset of the
central target. Latency (time from appearance of target to saccade
initiation) and gain (% of distance moved to the target location)
of primary saccades (the first saccade to the target) were
Predictive Saccade Procedural Learning Task
During the predictive saccade task, participants visually
tracked a target that stepped between two locations (? 7.5° from
center) every 1.5 sec. This sequence was repeated 10 times (i.e.,
20 target presentations) (Figure 1). The first saccade, made from
central fixation to one of the peripheral targets to start the task,
was not included in analyses. Latency and gain of primary
saccades were measured.
Primary saccades were classified qualitatively according to
their latency. Sensory-guided saccades (?160 msec) represent
reflexive responses to target appearance. Anticipatory saccades
(?90 msec) reflect saccades made in anticipation of target
appearance, before sensorimotor processes can respond to vi-
sual cues. We also considered an intermediate group of fast
saccades (?160 msec but ?90 msec). This classification of
saccades is consistent with neurophysiological evidence (24) and
previous oculomotor studies (20).
Analyses of Eye Movements
Eye position recordings obtained during fixation of targets in
each trial were used to convert voltage recordings to eye position
in degrees of visual angle. This “within-trial” calibration mini-
mizes artifacts resulting from drift in DC-EOG signals over the
course of testing. Performance was examined to identify primary
saccades, artifacts (e.g., blinks, signal clipping), and failures of
software algorithms to correctly identify saccades. Before analysis,
digitized eye movement signals were smoothed with a linear-phase,
finite-impulse response low-pass filter.
For the visually guided saccade task, repeated-measures
analyses of variance (ANOVAs) were used to examine effects of
target step amplitude (10°, 20°, 30°), direction (left, right), and
participant group (autism, control) on saccade latency and gain.
For the predictive saccade task, mixed-effects regression
models were used to accommodate repeated measurements
across trials that were correlated to different degrees, as is the
case when learning occurs. Latency and gain data were each
modeled as quadratic functions over trials, allowing for expected
nonlinear rates of learning. Initial models included group (au-
tism, control), linear and quadratic terms for change across trials,
response direction (left, right), and all two- and three-way
interactions. Terms were eliminated from the model with a
backwards elimination procedure to arrive at the most parsimo-
nious model. Mixed-effects models were analyzed with SAS
(v.8.02 for Windows; SAS, Cary, North Carolina).
Repeated-measures ANOVAs were used to compare the pro-
portion, gain, and latency of each saccade type as a function of
group and saccade direction. Correlations of age with all param-
eters were nonsignificant in both participant groups.
Figure 1. Schematic representation of the predictive saccade task. Partici-
pants were instructed to follow the dot with their eyes. Targets appear at
?/? 7.5° of visual angle every 1.5 sec.
2 BIOL PSYCHIATRY 2009;xx:xxx
A-M. D’Cruz et al.
ARTICLE IN PRESS
Visually Guided Saccade Task
There were no differences between individuals with autism and
healthy control subjects in visually guided saccade latencies
[F(1,104) ? .43, p ? .51] or gain [F(1,104) ? .66, p ? .42]. No group
interactions with target location or direction were significant.
Predictive Saccade Task
groups [F(1,1699) ? .61, p ? .43], and there were no group
differences in the gain of different saccade types (Table 1).
Latency. Whereas there was a significant overall reduction in
response latencies over trials [F(1,1664) ? 16.91, p ? .001], the rate
of learning (i.e., the overall latency reduction over trials) did not
differ across groups [F(1,1664) ? .03, p ? .90]. However, the
three-way interaction was significant [F(1,1664) ? 5.97, p ? .01],
driven by progressively faster rightward responses over trials in
individuals with autism, relative to control subjects. Follow-up
analyses confirmed that the interaction was significant for rightward
but not leftward saccades [F(1,779) ? 6.70, p ? .01, and F(1,677) ?
1.25, p ? .26, respectively] (Figures 2 and 3).
Consistent with these observations, individuals with autism
had a higher proportion of rightward than leftward anticipatory
saccades [F(1,104) ? 9.77, p ? .02] (Table 1), and the mean
latency of anticipatory saccades in the autism group was signif-
icantly faster than control subjects for rightward movements only
[F(1,320) ? 4.73, p ? .03].
Primary saccade gain did not differ between
We examined procedural learning with a predictive saccade
task known to engage frontostriatal systems in a relatively large
group of individuals with autism. We did not observe abnormal-
ities in the overall rate of procedural learning in autism (i.e., the
reduction in response latencies over trials). However, individuals
with autism displayed a speeding of rightward predictive/antic-
ipatory responses. Internal clocks, in the form of striatal temporal
oscillators (25), are the means by which precise interval timing is
perceived and used to initiate planned motor sequences (26).
The specific speeding of rightward saccades exhibited by indi-
viduals with autism on the predictive saccade task suggests a
lateralized acceleration of the striatal temporal coding system that
Table 1. Saccade Tasks Results
Predictive Saccade Task
Visually Guided Saccade Task
Percentage of each of the three saccade types (sensory-guided, fast, and anticipatory) among rightward and leftward responses, and the mean (M) and
each of the three target locations in the visually guided saccade task. Percentages of saccade types for the predictive saccade task are presented separately
as a proportion of all scorable leftward and rightward saccades for each subject group. Saccades were classified as follows: sensory-guided (?160 msec);
anticipatory (?90 msec); fast (?160 msec but ?90 msec).
aSignificant group differences, p ? .05.
Figure 2. Average saccade latency (msec) for rightward and leftward sac-
cades in participants with autism and healthy control subjects over trials in
the predictive saccade task. The percentage of scorable saccades included
in each curve (of the possible number of trials presented in each condition)
is: Control Subjects, Right: 95.6, Left: 96.1; Autism, Right: 86.0, Left: 89.0.
There was no significant group difference between the percentages of
saccades included in each condition.
A-M. D’Cruz et al.
BIOL PSYCHIATRY 2009;xx:xxx 3
ARTICLE IN PRESS
is used to time internally generated motor response sequences by
the contralateral (left) hemisphere.
Neuroimaging (27) and animal studies (28) have demon-
strated the importance of the basal ganglia in tasks requiring
precisely timed responses, including predictive saccades (20),
and in the implicit judgment of time intervals (29). Patients with
cerebellar lesions typically have a reduced ability to acquire
learned motor sequences during predictive saccade tasks rather
than specific deficits in response timing (30). The integrity of
visually guided saccades in this sample of individuals with autism
also argues against a cerebellar explanation for our results.
Alterations in response timing on our predictive saccade task are
within the response interval range regulated by striatal chrono-
metric systems. Thus, our findings suggest that a speeded
chronometric system in the left striatum is responsible for the
observed pattern of results.
The neurophysiology of the striatal clock is known to be
plastic, with regulation mediated by neurochemical systems. In
rodent studies, acceleration and deceleration of interval timing
has been demonstrated with striatal administration of dopamine
agonists and antagonists respectively (31) and with systemic
administration of haloperidol and clozapine (32). Thus, our
findings are both consistent with striatal abnormalities observed
in functional and structural neuroimaging studies of autism
(7,10,11) and might be related to neurochemical as well as
An alternative explanation for our findings is that they
might represent an inability to withhold planned motor re-
sponses until they are appropriate to execute. Deficits in
prefrontally mediated inhibitory control in autism have been
suggested by neuropsychological (1) and oculomotor studies
using the antisaccade task (33). However, deficits on antisac-
cade tasks reflect a reduced ability to suppress responses to
external stimuli rather than internally generated responses,
and lateralization of these inhibitory deficits has not been
observed. Furthermore, behavioral responses in some cogni-
tive paradigms, including some oculomotor tasks, are slowed
rather than accelerated in autism (33,34). Together, these
findings suggest that reduced prefrontal inhibition of planned
behavior is likely not the cause of speeded predictive saccades
It is, of course, noteworthy that speeding selectively affected
rightward anticipatory saccades, indicating a lateralized neurobi-
ological alteration in individuals with autism. Although autism
clearly affects functions localized in both hemispheres, our
observation of a left-lateralized alteration leading to speeded
rightward saccades is consistent with findings of greater left
hemisphere abnormalities in some studies of autism (35). Also,
language deficits are a core feature of autism, whereas spatial
and musical abilities are often less impaired in higher-functioning
patients (36,37). Neuroimaging studies have found abnormal
growth trajectories (38) and increased disorder and density of
white matter bundles in left frontal language regions and supe-
rior temporal gyrus in autism (39,40). Electroencephalography
studies of autism have reported left fronto-temporal abnormali-
ties (41) and altered connectivity of left frontal and temporal
cortex (42). Manual (43) and pursuit eye-movement performance
(44) and some neuroimaging findings (45) provide evidence for
left-lateralized disturbances of sensorimotor systems in autism.
Thus, our findings add to a growing body of literature suggesting
that some left hemisphere brain systems and the cognitive
abilities they support are more compromised in at least some
individuals with autism.
Some (14) but not all (46) studies of procedural learning
provide evidence for deficits in autism. Laterality effects were not
investigated in these studies, because all responses were made
with one hand. Variations in task complexity might account for
these differences. Mostofsky et al. (14) used a much more
difficult serial reaction time task and reported reduced proce-
dural learning in autism. The simplicity of our oculomotor
paradigm might place less demand on prefrontally mediated
skills, such as maintaining a longer stimulus sequence in working
memory during the learning process. Also, Mostofsky et al.
studied adolescents, whereas the present study used a wider age
range, mostly adults. Future studies are needed to address the
importance of task complexity and developmental trends in
relation to response timing and motor learning deficits in autism,
as well as whether lateralized speeding in autism is seen in other
tasks requiring precisely timed responses. Nonetheless, our
findings are notable for demonstrating that the rate of procedural
learning, at least on simple tasks, might be a relatively intact
cognitive domain in autism in the context of widespread deficits
in cognitive function (1).
Our results identify functional abnormalities in an important
cognitive process, coding temporal information for anticipatory
behavior. Our findings implicate left striatal chronometric sys-
tems and might have clinical and developmental implications for
impaired higher-order functions such as praxis and imitation
This study was supported by HD35469, HD055751, the National
Illinois Graduate Student Fellowship. We also gratefully acknowl-
edge the participation of the individual subjects and their families.
The authors report no biomedical financial interests or po-
tential conflicts relevant to this manuscript.
2. Joseph RM, Tanaka J (2003): Holistic and part-based face recognition in
(2007): The dynamic network subserving the three phases of cognitive
Figure 3. Cumulative frequency of saccade latencies over all trials during
saccades for participants with autism and healthy control subjects.
4 BIOL PSYCHIATRY 2009;xx:xxx
A-M. D’Cruz et al.
ARTICLE IN PRESS
4. Ivry RB (1996): The representation of temporal information in percep- Download full-text
tion and motor control. Curr Opin Neurobiol 6:851–857.
5. Meck WH, Benson AM (2002): Dissecting the brain’s internal clock: how
frontal–striatal circuitry keeps time and shifts attention. Brain Cogn 48:
6. Courchesne E, Saitoh O, Townsend JP, Yeung-Courchesne R, Press GA,
Lincoln AJ, et al. (1994): Cerebellar hypoplasia and hyperplasia in infan-
tile autism. Lancet 343:63–64.
E, et al. (2005): Striatal volume on magnetic resonance imaging and
repetitive behaviors in autism. Biol Psychiatry 58:226–232.
8. Kaufmann WE, Cooper KL, Mostofsky SH, Capone GT, Kates WR, News-
autism: A quantitative magnetic resonance imaging study. J Child Neu-
9. Muller RA, Pierce K, Ambrose JB, Allen G, Courchesne E (2001): Atypical
patterns of cerebral motor activation in autism: A functional magnetic
resonance study. Biol Psychiatry 49:665–676.
10. Sears LL, Vest C, Mohamed S, Bailey J, Ranson BJ, Piven J (1999): An MRI
study of the basal ganglia in autism. Prog Neuropsychopharmacol Biol
11. Takarae Y, Minshew NJ, Luna B, Sweeney JA (2007): Atypical involve-
ment of frontostriatal systems during sensorimotor control in autism.
12. Bailey A, Luthert P, Dean A, Harding B, Janota I, Montgomery M, et al.
(1998): A clinicopathological study of autism. Brain 121:889–905.
13. Bauman M, Kemper TL (1985): Histoanatomic observations of the brain
in early infantile autism. Neurology 35:866–874.
Implications for cerebellar contribution. J Int Neuropsychol Soc 6:752–
15. Sears LL, Finn PR, Steinmetz JE (1994): Abnormal classical eye-blink
16. Robertson EM (2007): The serial reaction time task: Implicit motor skill
learning? J Neurosci 27:10073–10075.
17. Muslimovic D, Post B, Speelman JD, Schmand B (2007): Motor proce-
dural learning in Parkinson’s disease. Brain 130:2887–2897.
18. Harris MS, Wiseman CL, Reilly JL, Keshavan MS, Sweeney JA (2009):
Effects of risperidone on procedural learning in antipsychotic-naive
first-episode schizophrenia. Neuropsychopharmacology 4:468–476.
19. Broerse A, Crawford TJ, den Boer JA (2001): Parsing cognition in schizo-
20. Simo LS, Krisky CM, Sweeney JA (2005): Functional neuroanatomy of
anticipatory behavior: Dissociation between sensory-driven and mem-
ory-driven systems. Cereb Cortex 15:1982–1991.
21. Lasker AG, Zee DS (1997): Ocular motor abnormalities in Huntington’s
disease. Vision Res 37:3639–3645.
22. Lord C, Rutter M, Le Couteur A (1994): Autism Diagnostic Interview-
Revised: A revised version of a diagnostic interview for caregivers of
individuals with possible pervasive developmental disorders. J Autism
23. Lord C, Risi S, Lambrecht L, Cook EH Jr, Leventhal BL, DiLavore PC, et al.
(2000): The Autism Diagnostic Observation Schedule-Generic: A stan-
24. Smit AC, Van Gisbergen JA (1989): A short-latency transition in saccade
dynamics during square-wave tracking and its significance for the dif-
ferentiation of visually-guided and predictive saccades. Exp Brain Res
25. Schoner G (2002): Timing, clocks, and dynamical systems. Brain Cogn
26. Hinton SC, Meck WH (1997): The ‘internal clocks’ of circadian and inter-
val timing [erratum]. Endeavour 21:82–87.
(2003): Basal ganglia and supplementary motor area subtend duration
perception: An fMRI study. NeuroImage 19:1532–1544.
busch HWM (2004): Premature responding following bilateral stimula-
tion of the rat subthalamic nucleus is amplitude and frequency depen-
dent. Brain Res 1008:198–204.
29. Hinton SC, Meck WH (2004): Frontal-striatal circuitry activated by hu-
man peak-interval timing in the supra-seconds range. Brain Res Cogn
30. Molinari M, Leggio M, Solida A, Ciorra R, Misciagna S, Silveri M, et al.
(1997): Cerebellum and procedural learning: Evidence from focal cere-
bellar lesions. Brain 120:1753–1762.
31. Meck WH (1996): Neuropharmacology of timing and time perception.
32. MacDonald CJ, Meck WH (2005): Differential effects of clozapine and
haloperidol on interval timing in the supraseconds range. Psychophar-
33. Luna B, Doll SK, Hegedus SJ, Minshew NJ, Sweeney JA (2007): Matura-
tion of executive function in autism. Biol Psychiatry 61:474–481.
34. Schmitz N, Rubia K, Daly E, Smith A, Williams S, Murphy DGM (2006):
Neural correlates of executive function in autistic spectrum disorders.
individuals with high-functioning autism and Asperger’s disorder: A
36. Heaton P (2003): Pitch memory, labeling and disembedding in autism.
37. Jolliffe T, Baron-Cohen S (1997): Are people with autism and Asperger
syndrome faster than normal on the Embedded Figures Test? J Child
(2007): Superior temporal gyrus, language function, and autism. Dev
39. Ben Bashat D, Kronfeld-Duenias V, Zachor DA, Ekstein PM, Hendler T,
Tarrasch R, et al. (2007): Accelerated maturation of white matter in
young children with autism: A high b value DWI study. NeuroImage
(2006): Structural correlates of intellectual impairment and autistic fea-
tures in adolescents. NeuroImage 33:1136–1144.
sitivity to acoustic change in children with autism: Electrophysiological
evidence of left frontal cortex dysfunctioning. Psychophysiology
42. Murias M, Webb SJ, Greenson J, Dawson G (2007): Resting state cortical
connectivity reflected in EEG coherence in individuals with autism. Biol
43. Dawson G, Warrenburg S, Fuller P (1983): Hemisphere functioning and
motor imitation in autistic persons. Brain Cogn 2:346–354.
44. Takarae Y, Minshew NJ, Luna B, Krisky CM, Sweeney JA (2004): Pursuit
eye movement deficits in autism. Brain 127:2584–2594.
45. Mostofsky SH, Burgess MP, Gidley Larson JC (2007): Increased motor
46. Barnes KA, Howard JH, Jr, Howard DV, Gilotty L, Kenworthy L, Gaillard
sequences in childhood autism spectrum disorder. Neuropsychology
A-M. D’Cruz et al.
BIOL PSYCHIATRY 2009;xx:xxx 5
ARTICLE IN PRESS