The Association Between Sequence Learning on the Serial Reaction Time Task and Social Impairments in Autism
Journal of Autism and Developmental Disorders
The Association Between Sequence Learning on the Serial Reaction Time Task and Social Impairments in Autism
Fenny S. Zwart 0 1 2
Constance Th. W. M. Vissers 0 1 2
Joseph H. R. Maes 0 1 2
0 Royal Dutch Kentalis , Sint-Michielsgestel , The Netherlands
1 Behavioural Science Institute , Nijmegen , The Netherlands
2 Donders Institute for Brain Cognition and Behaviour, Radboud University , Montessorilaan 3, 6500 HE Nijmegen , The Netherlands
3 Fenny S. Zwart
It is assumed that learning on the Serial Reaction Time (SRT) task is related to learning involved in social skill development affected in autism, but this assumption has hardly been investigated. We have therefore examined associations between SRT task learning and social impairment measured by the Social Responsiveness Scale in 72 autistic and non-autistic adults. Results revealed a positive correlation between deterministic sequence learning, putatively involving explicit learning, and social impairment in autistic adults but not in non-autistic adults. No correlations with probabilistic learning were found. These results suggest that the type of learning that helps autistic adults during a deterministic SRT task hinders them during social development, and call for further investigating the ecological validity of the SRT task.
Implicit learning; SRT task; ASD; Social impairments; SRS-A
Social communication skills are believed to develop largely
through implicit, or automatic, learning mechanisms
. Learning what distance to keep or how
to make small talk seems to come natural for most of us,
without much explicit effort. This does not seem to be the
case for people with Autism Spectrum Disorder (ASD), a
neurodevelopmental disorder characterized by impairments
in social communication skills
. This has led to the hypothesis that altered
implicit learning mechanisms play a role in the
development of ASD-related symptoms. Although some studies
Electronic supplementary material The online version of this
article (https://doi.org/10.1007/s10803-018-3529-6) contains
supplementary material, which is available to authorized users.
The link to the data repository is active and is as follows: http://
have found (subtle) learning problems or reported altered
brain mechanisms during implicit learning in ASD
and Stark 2007; Mostofsky et al. 2000; Sharer et al. 2015,
2016; Travers et al. 2015; Zwart et al. 2017b)
majority of studies have found intact implicit learning in ASD
(for meta-analyses see: Foti et al. 2015; Obeid et al. 2016;
for a review see: Zwart et al. 2017a)
, hence challenging the
hypothesized association between implicit learning as
measured in scientific studies and social communication skills.
Most of these studies have used the Serial Reaction Time
(SRT task; Nissen and Bullemer 1987)
. In this task,
participants have to respond to a stimulus that appears on
one of four locations on the screen as fast as possible by
pressing a corresponding button. Unknown to the
participant, these locations follow a sequence. Implicit learning is
reflected by shorter reaction times (RTs) over time, without
any (verbal) knowledge about the sequence. As a general
reduction in RTs may reflect overall motor learning rather
than sequence-specific learning, it is common to include
(blocks of) random trials and investigate sequence
learning as the difference in RTs between random and sequenced
trials. The implicit nature of the task is confirmed by two
features: (i) there is no intention to learn (i.e., no
instruction), and (ii) there is limited awareness of the sequence
knowledge. The latter is usually confirmed by
post-experimental interviews, although other methods based on RTs
ble too (e.g., Wessel et al. 2012
). However, these
measures show that a substantial number of participants
do gain explicit knowledge in this task
(e.g., Haider and
. Such knowledge is believed to be prevented
when random trials are inserted in the sequence, making
the sequence probabilistic rather than deterministic
(Jiménez et al. 1996; in the literature a probabilistic SRT task is
often referred to as ‘Alternating SRT task’, see Howard and
It is assumed that learning on the SRT task relates to
the development of social communication skills
and, hence, also that any deficits in learning on this
task may be related to the social communication deficits in
(e.g., Mostofsky et al. 2000; Sharer et al. 2016)
However, there is not much direct empirical evidence supporting
these claims. One study found no correlation between ASD
symptoms measured as raw scores on the Social
Responsiveness Scale (SRS; Constantino and Gruber 2002) and implicit
learning on the SRT task
(Travers et al. 2010)
. In a later
fMRI-study, the same researchers found that symptoms of
repetitive behavior but not social communication deficits
as measured by the Autism Diagnostic Interview-Revised
(ADI-R; Lord et al. 1994)
negatively predicted brain
activation related to learning
(Travers et al. 2015)
A factor complicating the interpretation of SRT task
performance in ASD is the hypothesis that autistic participants
use more explicit learning strategies
(e.g., Klinger et al.
2007; Ullman and Pullman 2015)
. Such explicit strategies
may lead to similar behavioral performance as implicit
learning does, at least under certain conditions
et al. 2010; Zwart et al. 2017b)
. It is therefore questionable
whether we should interpret performance on the SRT task
in ASD in terms of the same underlying learning
mechanism (i.e., implicit) as we do for typical development (TD).
Following this line of reasoning, an association between
SRT task performance and social behavior impairments
in ASD might reflect a relation between explicit learning
and social functioning, whereas the same association in TD
would reflect a relation between implicit learning and social
The aim of the current study was to further investigate
the association between SRT task performance and social
impairments and other autistic symptoms as measured by
the questionnaire SRS for adults (SRS-A; Constantino and
). Given that social impairments measured on
the SRS are continuously distributed amongst the general
(Constantino and Todd 2003)
, we first assessed
this association in a group of 72 autistic and non-autistic
individuals. As implicit learning is thought to play an
important role in social skill development
predicted that overall, learning performance on a
probabilistic and a deterministic condition of the SRT task would
be negatively correlated to social impairments measured by
the SRS. We expected this effect to be stronger during the
(more implicit) probabilistic learning condition compared
to the deterministic condition. Based on the hypothesis that
autistic individuals may learn the task explicitly, whereas
non-autistic individuals rely more on implicit learning, we
also conducted within-group analyses to examine the
possibility that the association between SRT task performance
and social impairments may be different in autistic compared
to non-autistic individuals.
Data from 72 participants from two studies was analyzed
(see Table 1 for demographic details).
The first study included 19 autistic adults and 19
nonautistic adults from a previous EEG study
(Zwart et al.
. The second study included 16 young autistic adults
and 18 young non-autistic adults. There was no overlap in
participants between the two studies, i.e., none of the
participants took part in both studies. All participants were
free of major neurological disorders and all autistic
participants were diagnosed with ASD by a clinician. Participants
signed a written informed consent after being informed of
the details of the study. Both studies were approved by a
local ethical committee and in line with the guidelines of
the Declaration of Helsinki.
Procedure and SRT task
For study 1, the SRT task was administered while
Electroencephalogram (EEG) was recorded for other purposes.
The participant was instructed to respond to the direction
of an arrow by a corresponding button press as fast as
possible. Unknown to the participant, after 48 practice trials,
the arrows followed a sequence (Fig. 1). We designed an
SRT task that started with a probabilistic part, directly
followed by a deterministic part. The probabilistic part of the
task consisted of 72 repetitions of a probabilistic sequence
(2-1-3-4-3-2-4-1) in which one stimulus in every sequence
was replaced by a deviant (random) stimulus. In other words,
the stimuli were only predictable with a certain probability.
This part was directly followed by a deterministic part,
consisting of 72 repetitions of a deterministic (i.e., no deviant
stimuli) sequence (4-3-1-2-1-4-2-3). Both sequences were
second-order in nature, in which two stimuli predicted the
next stimulus (i.e., in the first sequence, ‘2-1’ predicted
‘3’). We ensured both sequences contained: (1) no
repeating elements; (2) only one “serial” triplet (e.g., 1-2-3); and
(3) only two “alternating” triplets (i.e., 1-2-1, 3-4-3). The
Fig. 1 Serial reaction time task:
the participant was asked to
respond to the direction of the
arrow, which—unknown to the
repeating 8-element sequence (e.g.,
2-1-3-4-3-2-4-1). Copyright by
INSAR/Wiley Periodicals, Inc.
(2017), adapted from Zwart
et al. (2017b)
deviant trials in the probabilistic sequence never repeated
the adjacent sequenced trials, and equally represented the
different stimuli. Response-to-stimulus interval was set at
500 ms. After the task, a short verbal interview was
administered to assess levels of awareness of the final deterministic
The SRT task used in study 2 was very similar to that
used in study 1. The only three differences were: (i) only
60 (instead of 72) repetitions per sequence were used; (ii)
deterministic sequence was slightly different (i.e.,
4-3-4-13-2-1-2); (iii) within group, each participant received a
different set of semi-randomized positions of the deviant trials.
No EEG was recorded.
Participants were asked to fill out the SRS-A. The SRS-A
consists of 65 items with a 4-point Likert-scale answer
format and measures social impairments and related
autistic symptoms (Constantino and Gru
). Outcome of
the SRS-A is a Total Score and four Subscale Scores: (1)
Social Awareness, (2) Social Communication, (3) Social
Motivation, and (4) Rigidity/Repetitive Behavior; with
higher scores indicating higher levels of social impairments.
It could be argued that the fourth subscale is not a direct
aspect of social behavior, but rather indirectly related.
Studies investigating the psychometric properties of the SRS
have shown that reliability and validity are satisfactory (e.g.,
Chan et al. 2017
Gau et al. 2013
). In addition,
the participants filled out the Autism Quotient
et al. 2001)
, a questionnaire regarding autistic traits with
50 items that are answered with ‘definitely agree’, ‘slightly
agree’, ‘slightly disagree’, ‘definitely disagree’, with total
scores ranging between 0 and 50
(more details can be found
in Baron-Cohen et al. 2001)
, and higher scores indicating
higher degrees of autistic traits. The reliability and
validity of the AQ have also been found to be satisfactory
Baron-Cohen et al. 2001; Hoekstra et al. 2008)
abbreviated version of the WAIS-IV (Wechsler 2008) was
administered to estimate IQ, including the subtests Block Design,
Similarities, Digit Span and Information. For 7 participants
from study 1, a full WAIS-III
administered within 12 months prior to participation, and
this IQ-score was used instead.
For all main analyses, the alpha level was set at 0.05.
Greenhouse Geisser correction was applied where the sphericity
assumption was violated, and corrected statistics including
adjusted degrees of freedom are reported. Effect sizes are
expressed as partial eta squared (ƞp2).
For study 1, the RT data was split into 12 Blocks of 6
sequences (48 trials) in each condition to assess learning
over time. For study 2, the RT data was split into 9 Blocks
of 6 sequences. In order to make the two studies comparable,
the last three blocks of study 1 were discarded from all
analyses. Extreme outliers were determined by the Interquartile
Range (IQR) criterion, i.e. values 1.5 × IQR ± the median RT
over each Block for the standard trials, and over two large
blocks for the deviant trials. On average, 24.2 (range 7–48;
out of 378 trials; 6.40%) standard and 1.25 (range 0–5; out
of 54 trials; 2.31%) deviant outlier trials were removed in
the probabilistic condition, and 32.5 (range 8–60; out of 432
trials; 7.52%) outlier trials were removed from the
deterministic condition. Trials with erroneous responses and the
subsequent trials, as well as trials directly after a deviant
trial, were removed.
ANOVAs of Probabilistic and Deterministic Learning
Although not the focus of the current paper, learning in
the probabilistic condition was analyzed with a Group
(ASD, TD) × Trial Type (Standard, Deviant) × Block (9)
ANOVA, and in the deterministic condition with a Group
(ASD,TD) × Block (9) ANOVA. Initial ANOVAs were
conducted with Study (1, 2) added as additional
betweensubjects factor. Non-significant effects involving the Study
factor confirmed comparable learning and justified pooling
data from the two studies (see Supplementary Materials 1
for details of these analyses).
Relation Between Sequence Learning and SRSA‑ Scores
An overall probabilistic learning score was computed by
subtracting the mean RT of all standard trials (Block 1–9)
from the mean RT of all deviant trials (Block 1–9). An
overall deterministic learning score was computed by
subtracting the mean RT of the final Block (9) from the first
Block (1). Pearson’s correlations between the probabilistic/
deterministic learning score and SRS-A were analyzed.
First, all participants (i.e., TD and ASD) were included to
investigate social impairments as a spectrum. Because of
the uncertainty regarding different learning mechanisms in
ASD, subsequent within-group correlations were analyzed.
To ensure that none of these correlations were due to
outlier participants, data points with Cook’s distances > 1.00
were removed from the analysis
(Cook and Weisberg 1982)
Cook’s distances were determined by using a simple linear
regression analysis with SRS-A score as independent
variable, and learning score as dependent variable.
Differential Analyses on Age and IQ
In order to confirm that any potential correlational findings
were not driven by the factors age and IQ, the same
correlational analyses were conducted controlling for these factors.
Block Analyses of Learning
Figure 2 shows probabilistic and deterministic learning in
both studies and suggests that these learning effects were
similar. This suggestion was indeed statistically confirmed
(see Supplementary Materials 1).
For the probabilistic condition, ANOVA revealed a
main Trial Type effect, F(1,70) = 170, p < .001, ƞp2=.71,
with larger RTs for deviant (M = 571 ms) than standard
trials (M = 510 ms), confirming sequence-specific
learning. Furthermore, significant Block, F(4.7,328) = 4.8,
(N = 38)
(N = 34)
p < .001, ƞp2=.064, and Trial Type × Block interaction,
F(4.3,299) = 3.8, p = .004, ƞp2 = .052, reflecting a linear
trend, F(1,70) = 17.5, p < .001, ƞp2 = .20, effects suggested
that learning increased over time. No effects involving the
Group factor were found, p’s ≥ .19, suggesting similar motor
speed and learning in ASD and TD.
For the deterministic condition, ANOVA revealed a main
Block effect, F(4.4,311) = 13.1, p < .001, ƞp2 = .16, reflecting
a linear trend, F(1,70) = 28.0, p < .001, ƞp2 = .29, suggesting
learning. No main Group, p = .27, or Group × Block
interaction, p = .097, effect was found, suggesting similar speed and
learning in ASD and TD.
Relation Between Sequence Learning and Social
Probabilistic Learning and Social Impairment
Figure 3 suggests no clear association between the
probabilistic learning score and the SRS-A total score, and
one outlier ASD participant. After excluding this
participant (Cook’s distance = 1.035), indeed no correlation
across groups (i.e., TD and ASD collapsed) was present,
r(69) = .077, p = .53. No correlations were found within
groups either, TD: r(35) = .11, p = .53; ASD: r(32) = .057,
p = .75, after excluding the same outlier participant (Cook’s
distance = 1.01).
Deterministic Learning and Social Impairment
Figure 4 suggests a positive correlation between the
deterministic learning score and the SRS-A total score, which was
statistically confirmed, r(70) = .27, p = .023 (across groups).
Within groups, a significant correlation was found for ASD,
r(33) = .35, p = .041, but not for TD, r(35) = −.078, p = .65.
Controlling for age and IQ did not change the general pattern
of findings described above. That is, no correlations were
found between probabilistic learning and the SRS-A total
score, neither across groups or within groups. A positive
correlation between deterministic learning and the SRS-A
total score was found across group, and within-group
confirmed for ASD but not for TD (see Supplementary
The aim of the current paper was to investigate the
association between sequence learning on the SRT task and levels
of social impairment related to ASD in autistic and
nonautistic individuals. We used an SRT task with a
probabilistic condition, designed to evoke implicit learning, and a
SRS-A Total T-Score
deterministic learning condition that would allow for
successful performance using explicit strategies. Overall
performance on both conditions of the SRT task was similar in
autistic and non-autistic individuals, in line with conclusions
of previous meta-analyses and reviews on the topic
et al. 2015; Obeid et al. 2016; Zwart et al. 2017a)
indepth discussion of the current SRT task findings in ASD
and TD falls beyond the scope of the current manuscript, but
can be found in Zwart et al. 2017b.
The current findings suggest no association between
probabilistic learning and social impairments as measured
with the SRS-A. This was true for both the ASD and the TD
group. This finding, especially in TD, does not corroborate
the general idea that the type of implicit learning needed
to successfully complete an SRT task is also involved in
the development of social communication skills. A
possible explanation is that although social communication
skills depends upon the detection of temporal and spatial
sequences of facial, gestural and vocal cues
, these sequences differ from the sequences used
in an SRT task in terms of complexity and probability. For
example, successful social communication skills includes
understanding the other person, which requires
probabilistically associating the facial, gestural and vocal sequences
to internal emotional states. These sequences can be much
longer and can occur with a much lower probability than
the eight element sequences used in the current SRT task.
Perhaps even more surprising is the positive correlation
found for deterministic learning and social impairments.
Analyses of the ASD and TD groups separately revealed that
this association was only significantly present in ASD, and
did not seem to be driven by age or IQ. It seems that
autistic individuals who are better at deterministic learning also
experience more difficulties in social situations. Although
this may sound counterintuitive, it could be explained by
the idea of an overactive or compensatory explicit learning
system in ASD
(Klinger et al. 2007; Ullman and Pullman
. It has been found that explicit learning can be helpful
in one situation, but detrimental in more complex situations
(e.g., Howard and Howard 2001). It may be that autistic
individuals who have a stronger developed explicit learning
system benefit from this in a deterministic SRT task, but are
hindered during learning from complex social situations. For
example, learning the art of small talk in an explicit fashion
would be extremely difficult and load heavily on cognitive
resources, as in any such interaction the possible number of
verbal and non-verbal cue sequences would be practically
impossible to consciously predict and infer.
The finding of a positive correlation between
deterministic learning and social impairments in autistic individuals is
not in line with an exploratory analysis of a previous study
reporting no such correlation in ASD
(Travers et al. 2010)
this study, learning was measured as the difference between
blocks of a deterministic sequence and one random block of
trials near the end of the experiment. Important differences
between the two studies that could explain the inconsistent
findings, are the different SRS and learning measures.
Travers et al. (2010) used raw scores from the parental version
of the SRS for their group of 14- to 25-year-olds, whereas
we used T-scores from the self-report version for adults. The
parental version of the SRS (including norms to calculate
T-scores) used by Travers et al. (2010) is developed for
children and adolescents up to 18 years old, and was therefore
not suitable for their older adult participants. Hence it could
be argued that our measure of social impairments in adults
is more accurate. However, our deterministic learning score
included an RT difference over time, and could therefore
reflect a general improvement in motor speed rather than
sequence-specific learning (for references on motor
(learning) impairments in ASD, see
Dziuk et al. 2007
et al. 2010
). Travers et al. (2010) did control for general
motor speed by comparing sequenced trials with a block
of random trials, and thereby measured sequence-specific
learning more accurately than we did. Given a lack of such
control in our study, one cannot exclude the possibility that
the association with the SRS-A score was entirely driven by
this general motor learning. However, this possibility is quite
implausible given the direction of the observed association:
the larger the RT decrease the more social impairments.
Still, future research should include a random block at the
end of the experiment and incorporate the performance on
this block in the measure of sequence learning.
Taken together, it seems that social impairments in autism
are related to the tendency to use explicit strategies during
sequence learning (as presumably evoked by the
deterministic condition) rather than to impairments in implicit
learning (as presumably evoked by the probabilistic condition).
However, the lack of correlations with probabilistic learning
on the SRT task in both groups could be interpreted as a
failure of the task to measure the implicit learning abilities we
use to extract the complex statistical properties of our daily
life environment, perhaps due to the task’s simplicity. For
describes how non-verbal
communication requires a complex, and probabilistic
sequencing of cues, such as hand gestures and facial expressions.
The ecological validity of the SRT task could potentially
be improved by increasing task complexity, for example, by
decreasing the probability of the stimuli and by increasing
Previous studies on the association between learning on
the SRT task and everyday skills are mixed. Some studies
have reported that learning on the SRT task predicts
(Misyak et al. 2010; Lum et al. 2012)
one large study reported no association between learning
on an SRT task and reading ability (Waber et al. 2003).
The use of individual learning scores has been criticized
by researchers using a different, statistical learning
paradigm, mainly because quite a few individuals do not show
learning on this task
(Siegelman et al. 2017)
. This may also
be a concern to learning on the SRT task, which is closely
related to learning on the statistical learning paradigms
Perruchet and Pacton 2006)
. Indeed, some of the
deterministic learning scores in the current study were negative,
i.e., the participants became slower over time, perhaps due
to fatigue. It is important to develop a good derivative for
individual learning on the SRT task, perhaps in which RT
gains over all blocks are included, rather than the
difference between the first and last block, as we decided a-priori.
Several suggestions to develop a proper task and measure
for individual sequence learning abilities has been made by
Siegelman et al. (2017), including the use of trials with
different probabilities, i.e., varying the difficulty level, which
would increase the sensitivity to individual learning abilities.
Limitations of the current study include a relatively
low sample size for correlational analyses, particularly the
within-group analyses, and potentially the fixed order of task
conditions in which the probabilistic condition was always
followed by the deterministic condition. Because
deterministic sequence learning is more likely to lead to awareness
than probabilistic learning
(e.g., Cleeremans and Jiménez
1998; Norman et al. 2007)
, starting with the deterministic
condition could have triggered an active search for prompts
in the following probabilistic condition, harming its implicit
nature. Although the current design suits the current study
aims best, we cannot rule out that learning the first,
probabilistic part, has influenced (e.g., enhanced) learning on the
second, deterministic part. The effect of task order could be
investigated by counterbalancing the conditions in a larger
study. Additionally, future studies could monitor learning
over time rather than investigating only the current moment,
as several clinical studies suggest deficits in consolidation of
learning rather than in initial learning
(e.g., Hedenius et al.
2011; Nemeth et al. 2013)
. And in everyday life, a skill is
only useful if it can be used at a later point in time.
In conclusion, the current study suggests that better
performance on a deterministic SRT task is associated
with higher levels of social impairments in autistic
participants as measured by the SRS-A. Probabilistic sequence
learning does not seem to be related to social impairments.
These findings suggest that caution should be taken in
translating findings from traditional SRT studies to learning in
everyday life, and call for further investigating the ecological
validity of the SRT task. Furthermore, it would be
interesting to replicate these findings using other measures of social
Acknowledgments The authors would like to thank ITVitae and Fontys
Hogeschool and their students for their participation and enthusiasm.
Author Contributions FZ conceived of the study, participated in its
design, in the coordination of the study and performed the
measurement, participated in the interpretation of the data and performed the
statistical analysis, and drafted the manuscript; CV participated in the
design of the study and edited the manuscript; JM participated in the
design of the study, the interpretation of the data and the statistical
analyses, and helped to draft the manuscript.
Compliance with Ethical Standards
Conflict of interest All authors declare no conflicts of interest.
Ethical Approval All procedures performed were approved by the local
ethical committee (Radboud University, Nijmegen,
ECSW2014-1003207) and in accordance with the 1964 Helsinki declaration and its later
amendments or comparable ethical standards.
Open Access This article is distributed under the terms of the
Creative Commons Attribution 4.0 International License (http://creativeco
mmons.org/licenses/by/4.0/), which permits unrestricted use,
distribution, and reproduction in any medium, provided you give appropriate
credit to the original author(s) and the source, provide a link to the
Creative Commons license, and indicate if changes were made.
American Psychiatric Association. ( 2013 ). Diagnostic and statistical manual of mental disorders (5th edn .). Washington, DC: American Psychiatric Association.
Baron-Cohen , S. , Wheelwright , S. , Skinner , R. , Martin , J. , & Clubley , E. ( 2001 ). The autism-spectrum quotient (AQ): Evidence from asperger syndrome/high-functioning autism, males and females, scientists and mathematicians . Journal of Autism and Developmental Disorders , 31 , 5 - 17 .
Bölte , S. ( 2012 ). Brief report: The social responsiveness scale for adults (SRS-A): Initial results in a German cohort . Journal of autism and developmental disorders , 42 , 1998 - 1999 .
Brown , J., Aczel , B. , Jiménez , L. , Kaufman , S. B., & Grant , K. P. ( 2010 ). Intact implicit learning in autism spectrum conditions . The Quarterly Journal of Experimental Psychology , 63 , 1789 - 1812 .
Chan , W. , Smith , L. E. , Hong , J. , Greenberg , J. S. , & Mailick , M. R. ( 2017 ). Validating the social responsiveness scale for adults with autism . Autism Research , 10 , 1663 - 1671 .
Cleeremans , A. , & Jiménez , L. ( 1998 ). Implicit sequence learning: The truth is in the details . In Handbook of implicit learning (pp. 323 - 364 ), Thousand Oaks, CA: Sage Publications
Constantino , J. , & Gruber , C. ( 2002 ). The social responsiveness scale . Los Angeles: Western Psychological Services.
Constantino , J. N. , & Gruber , C. P. ( 2012 ). Social responsiveness scale (SRS) . Torrance, CA: Western Psychological Services.
Constantino , J. N. , & Todd , R. D. ( 2003 ). Autistic traits in the general population: A twin study . Archives of General Psychiatry , 60 , 524 - 530 .
Cook , R. D. , & Weisberg , S. ( 1982 ). Residuals and influence in regression . New York: Chapman and Hall.
Dziuk , M. A. , Larson , J. C. , Apostu , A. , Mahone , E. M. , Denckla , M. B. , & Mostofsky , S. H. ( 2007 ). Dyspraxia in autism: Association with motor, social, and communicative deficits . Developmental Medicine & Child Neurology , 49 , 734 - 739 .
Foti , F. , De Crescenzo , F. , Vivanti , G. , Menghini , D. , & Vicari , S. ( 2015 ). Implicit learning in individuals with autism spectrum disorders: A meta-analysis . Psychological Medicine , 45 , 897 - 910 .
Fournier , K. A. , Hass , C. J. , Naik , S. K. , Lodha , N. , & Cauraugh , J. H. ( 2010 ). Motor coordination in autism spectrum disorders: A synthesis and meta-analysis . Journal of Autism and Developmental Disorders , 40 , 1227 - 1240 .
Gau , S. S. F. , Liu , L. T. , Wu , Y. Y. , Chiu , Y. N. , & Tsai , W. C. ( 2013 ). Psychometric properties of the Chinese version of the social responsiveness scale . Research in Autism Spectrum Disorders , 7 , 349 - 360 .
Gordon , B. , & Stark , S. ( 2007 ). Procedural learning of a visual sequence in individuals with autism . Focus on Autism and Other Developmental Disabilities , 22 , 14 - 22 .
Haider , H. , & Rose , M. ( 2007 ). How to investigate insight: A proposal . Methods , 42 , 49 - 57 .
Hedenius , M. , Persson , J. , Tremblay , A. , Adi-Japha , E. , Veríssimo , J. , Dye , C. D. , … Ullman , M. T. ( 2011 ). Grammar predicts procedural learning and consolidation deficits in children with specific language impairment . Research in Developmental Disabilities, 32 , 2362 - 2375 .
Hoekstra , R. A. , Bartels , M. , Cath , D. C. , & Boomsma , D. I. ( 2008 ). Factor structure, reliability and criterion validity of the AutismSpectrum Quotient (AQ): A study in Dutch population and patient groups . Journal of Autism and Developmental Disorders , 38 ( 8 ), 1555 - 1566 .
Howard , D. V. , & Howard , J. H. ( 2001 ). When it does hurt to try: Adult age differences in the effects of instructions on implicit pattern learning . Psychonomic Bulletin & Review , 8 , 798 - 805 .
Howard , J. H. Jr. , & Howard , D. V. ( 1997 ). Age differences in implicit learning of higher order dependencies in serial patterns . Psychology and Aging , 12 , 634 .
Jiménez , L. , Méndez , C. , & Cleeremans , A. ( 1996 ). Comparing direct and indirect measures of sequence learning . Journal of Experimental Psychology: Learning, Memory, and Cognition , 22 , 948 .
Klinger , L. G. , Klinger , M. R. , & Pohlig , R. L. ( 2007 ). Implicit learning impairments in autism spectrum disorders. In New developments in autism: The future is today (pp. 76 - 103 ). London: Jessica Kingsley.
Lieberman , M. D. ( 2000 ). Intuition: A social cognitive neuroscience approach . Psychological Bulletin , 126 , 109 .
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 . Journal of Autism and Developmental Disorders , 24 , 659 - 685 .
Lum , J. A. , Conti-Ramsden , G. , Page , D. , & Ullman , M. T. ( 2012 ). Working, declarative and procedural memory in specific language impairment . Cortex , 48 , 1138 - 1154 .
Misyak , J. B. , Christiansen , M. H. , & Tomblin , J. B. ( 2010 ). On-line individual differences in statistical learning predict language processing . Frontiers in Psychology , 1 , 31 .
Mostofsky , S. H. , Goldberg , M. C. , Landa , R. J. , & Denckla , M. B. ( 2000 ). Evidence for a deficit in procedural learning in children and adolescents with autism: Implications for cerebellar contribution . Journal of the International Neuropsychological Society , 6 , 752 - 759 .
Nemeth , D. , Janacsek , K. , Király , K. , Londe , Z. , Németh , K. , Fazekas , K. , … Csányi , A. ( 2013 ). Probabilistic sequence learning in mild cognitive impairment . Frontiers in Human Neuroscience , 7 , 318 .
Nissen , M. J. , & Bullemer , P. ( 1987 ). Attentional requirements of learning: Evidence from performance measures . Cognitive Psychology , 19 , 1 - 32 .
Norman , E. , Price , M. C. , Duff , S. C. , & Mentzoni , R. A. ( 2007 ). Gradations of awareness in a modified sequence learning task . Consciousness and Cognition , 16 , 809 - 837 .
Obeid , R. , Brooks , P. J. , Powers , K. L. , Gillespie-Lynch , K. , & Lum , J. A. ( 2016 ). Statistical learning in specific language impairment and autism spectrum disorder: A meta-analysis . Frontiers in Psychology , 7 , 1245 .
Perruchet , P. , & Pacton , S. ( 2006 ). Implicit learning and statistical learning: One phenomenon, two approaches . Trends in Cognitive Sciences , 10 , 233 - 238 .
Sharer , E. , Crocetti , D. , Muschelli , J. , Barber , A. D. , Nebel, M. B., Caffo , B. S. , … Mostofsky , S. H. ( 2015 ). Neural correlates of visuomotor learning in autism . Journal of Child Neurology , 30 , 1877 - 1886 .
Sharer , E. A. , Mostofsky , S. H. , Pascual-Leone , A. , & Oberman , L. M. ( 2016 ). Isolating visual and proprioceptive components of motor sequence learning in ASD . Autism Research , 9 , 563 - 569 .
Siegelman , N. , Bogaerts , L. , & Frost , R. ( 2017 ). Measuring individual differences in statistical learning: Current pitfalls and possible solutions . Behavior Research Methods , 49 , 418 - 432 .
Travers , B. G. , Kana , R. K. , Klinger , L. G. , Klein , C. L. , & Klinger , M. R. ( 2015 ). Motor learning in individuals with autism spectrum disorder: Activation in superior parietal lobule related to learning and repetitive behaviors . Autism Research , 8 , 38 - 51 .
Travers , B. G. , Klinger , M. R. , Mussey , J. L. , & Klinger , L. G. ( 2010 ). Motor-linked implicit learning in persons with autism spectrum disorders . Autism Research , 3 , 68 - 77 .
Ullman , M. T. , & Pullman , M. Y. ( 2015 ). A compensatory role for declarative memory in neurodevelopmental disorders . Neuroscience & Biobehavioral Reviews , 51 , 205 - 222 .
Waber , D. P. , Marcus , D. J. , Forbes , P. W. , Bellinger , D. C. , Weiler , M. D. , Sorensen , L. G. , & Curran , T. ( 2003 ). Motor sequence learning and reading ability: Is poor reading associated with sequencing deficits? Journal of Experimental Child Psychology , 84 , 338 - 354 .
Wechsler , D. ( 1997 ). WAIS-III, Wechsler adult intelligence scale: Administration and scoring manual . San Antonio, TX: Psychological Corporation.
Wechsler , D. ( 2008 ). Wechsler adult intelligence scale-Fourth Edition (WAIS-IV) . San Antonio, TX: NCS Pearson, (Vol. 22 , p. 498 ).
Wessel , J. R. , Haider , H. , & Rose , M. ( 2012 ). The transition from implicit to explicit representations in incidental learning situations: More evidence from high-frequency EEG coupling . Experimental Brain Research , 217 , 153 - 162 .
Zwart , F. S. , Vissers , C. Th. W. M. , Kessels , R. P. C. , & Maes , J. H. R. ( 2017a ). Procedural learning across the lifespan: A systematic review with implications for atypical development . Journal of Neuropsychology . https://doi.org/10.1111/jnp.12139.
Zwart , F. S. , Vissers , C. Th . W. M., van der Meij , R., Kessels , R. P. C. , & Maes , J. H. R. ( 2017b ). Autism: Too eager to learn? Event related potential findings of increased dependency on intentional learning in a serial reaction time task . Autism Research , 10 , 1533 - 1543 .