Journal of Autism and Developmental Disorders

https://doi.org/10.1007/s10803-020-04780-1

ORIGINAL PAPER

Reduced Sensory Habituation in Autism and Its Correlation

with Behavioral Measures
Wasifa Jamal1 · Annie Cardinaux1 · Amanda J. Haskins1,2 · Margaret Kjelgaard1,3,4 · Pawan Sinha1

Accepted: 3 November 2020

© Springer Science+Business Media, LLC, part of Springer Nature 2020

Abstract
Autism is strongly associated with sensory processing difficulties. We investigate sensory habituation, given its relevance
for understanding important phenotypic traits like hyper- and hypo-sensitivities. We collected electroencephalography data
from 22 neuro-typical(NT) and 13 autistic(ASD) children during the presentation of visual and auditory sequences of
repeated stimuli. Our data show that the ASD children have significantly reduced habituation relative to the NT children for
both auditory and visual stimuli. These results point to impaired habituation as a modality-general phenomenon in ASD.
Additionally, the rates of habituation are correlated with several clinical scores associated with competence along diverse
phenotypic dimensions. These data suggest that the sensory difficulties in autism are likely to be associated with reduced
habituation and are related to clinical symptomology.

Keywords Sensory habituation · Autism · Hyper and hypo-sensitivity · Audition · Vision · Electroencephalography

Introduction Despite the prevalence of sensory issues in autism, a

mechanistic understanding of their causal basis has been elu-
Autism is a neurodevelopmental disorder characterized by sive. One possibility is that autism affects basic sensory sen-
impaired social communication skills and repetitive behav- sitivities, leading to a perceived enhancement or diminish-
iors (American Psychiatric Association 2013), estimated to ment of normal intensity stimuli (Heaton et al. 1998; Gomot
affect over 1% of all children (Christensen et al. 2016). A key et al. 2002; Bonnel et al. 2003; Ferri et al. 2003; Khalfa et al.
correlate of autism is unusual sensory perception. Observa- 2004; Ashwin et al. 2009). However, this account has been
tions of sensory differences date to the original reports by questioned by several follow-up studies that failed to find
Kanner (1943). Over 90% of individuals with autism report significant sensory acuity differences between NT and ASD
sensory difficulties in several modalities (Leekam et al. participant groups in visual, auditory and somatosensory
2007; Marco et al. 2011). The strong association with ASD domains (Bölte et al. 2012; DePape et al. 2012; Marco et al.
of hyper- and hypo-sensitivity to sensory stimuli has led to 2011; Dabbous 2012; Cascio et al. 2008; O’Riordan and
the DSM-5 (American Psychiatric Association 2013) now Passetti 2006).
including them as diagnostic features of ASD. A more recent alternative account suggests that abnormal
sensitivities may arise not because of instantaneous differ-
ences in sensory thresholds, but rather because of how stim-
* Wasifa Jamal
wjamal@mit.edu uli are processed over time. Specifically, it is hypothesized
that autism may be associated with a reduction in sensory
1
Department of Brain and Cognitive Sciences, Massachusetts habituation. On the one hand, this reduction may contrib-
Institute of Technology, Cambridge, MA, USA ute to an inability to detect novel stimuli in the presence
2
Present Address: Psychological and Brain Sciences, of ongoing ones, leading to a hypo-sensitivity to changes
Dartmouth College, Hanover, NH, USA in the sensory environment (Guiraud et al. 2011), while on
3
Present Address: Bridgewater State University, Bridgewater, the other hand, the compromised ability to habituate would
MA, USA reduce stimulus suppression and may lead to hyper-sensitiv-
4
Department of Communication Sciences and Disorders, ity due to immersion in unrelentingly salient stimuli (Sinha
Massachusetts General Hospital Institute of Health et al. 2014).
Professions, Boston, MA, USA

13
Vol.:(0123456789)
 Journal of Autism and Developmental Disorders

Habituation is broadly defined as response decrement due and slower N1 attenuation in response to oddball stimuli
to stimulus repetition (Montagu and Coles 1966; Rankin (Hudac et al. 2018).
et al. 2009). With behavioral assays, there is evidence to Notwithstanding the aforementioned evidence, the rela-
suggest that children with ASD do not show progressive tionship between autism features and reduced habituation
dampening of response to repeated standard tones or clicks is not yet firmly established. Several studies have not found
(Guiraud et al. 2011; Martineau et al. 1992). Perry et al. differences in habituation rates between ASD and NT par-
found that ASD participants required more trials to reach ticipants. For instance, work by Miller et al. (2001), showed
maximal habituation and suggested that ASD participants a lack of evidence for reduced habituation in autism with
have a generalized slowing of habituation to repeated stim- an electrodermal activity (EDA) assay. They suggest that
uli (Perry et al. 2007). In adults, Lawson et al. presented a differences in habituation could be dependent on factors
30 dB, 1 KHz continuous tone for 190 s to the right ear and such as stimulus intensity and modality that are specific
collected periodic subjective ratings of loudness (Lawson to the study. An earlier study by Bernal and Miller (1970)
et al. 2015). ASD loudness ratings were significantly higher showed hypo-responsiveness in the ASD group as indicated
than NTs as the stimulus progressed, consistent with reduced by reduced responses to the initial three visual stimuli in the
habituation. Barry and James (1988) found that the autistic ASD group, but no other differences in sensory reactivity.
group differed from the control groups in its failure to show Given these mixed findings regarding sensory habituation
response habituation to simple visual and auditory stimuli. in autism, we believe that this domain merits further inves-
Additionally, slower behavioral habituation to faces has tigation, ideally employing objective electrophysiological
been shown to be positively correlated with higher symptom measures and careful control for factors of cognitive abil-
severity in ASD (Webb et al. 2010). ity and stimulus delivery. With this motivation, the current
Reports of behavioral evidence showing reduced habitu- study has three goals. First, we investigate whether neural
ation in autism, mentioned above, can be strengthened by habituation is reduced in ASD even with simple sensory
neural assays that potentially sidestep response subjectiv- stimuli that do not have any social connotations. Second,
ity. Neural habituation would entail reduction in cortical we examine whether changes in habituation are specific to
responses over time with successive repetition of a stimulus a sensory modality or apply across modalities, suggesting a
(Plichta et al. 2014; Thompson and Spencer 1966). Although more general underlying mechanism. Third, to understand
data on neural habituation in ASD are still fairly sparse, the the practical implications of any observed habituation reduc-
findings thus far suggest that this is a potentially promising tion, we explore the relationship between neural habitua-
avenue of inquiry. An fMRI study showed reduced neural tion and clinical characteristics of ASD, as measured by
habituation to repeated exposures to faces in ASD wherein clinical assessments of sensory, cognitive and behavioral
individuals with highest symptom severity demonstrated the performance.
least amount of neural habituation to repetition (Kleinhans
et al. 2009). In a two-back working memory task flanked
by faces with different emotions as task-irrelevant distrac- Materials and Methods
tors, Tam et al. reported a lack of amygdala habituation in
adults with ASD (Tam et al. 2017). In a study of 8-month- Subjects and Stimuli
olds at low and high likelihood of developing autism (based
on older sibling diagnosis) the high-likelihood infants later Twenty-two neurotypical subjects (9 females) ranging from
diagnosed with autism (n = 14) showed reduced repetition 7.1 to 12.8 years and 13 ASD (2 females) ranging from 7.4
suppression when compared to the high-likelihood infants to 12.8 years participated in the experiments. All partici-
with typical development (n = 44) (Kolesnik et al. 2019). pants were administered the Weschler Intelligence Scales for
A longitudinal social intervention study showed greater Children—Fifth Edition (WISC-V) and scored 85 or greater
reduction in habituation times in responses to faces and in the nonverbal index. The average nonverbal index score
a more typical P400 response to faces and objects in the for the ASD group was 104.42 (with a range of 90–126)
high-likelihood group of infants whose parents accessed and compared with the NT average of 116.90 (with a range of
delivered the promoting-first-relationships (PFR) parent-led 100–135). All ASD participants met criteria for autism or
intervention than the high-likelihood infants whose parents autism spectrum disorder based on administration of the
did not receive training in the intervention, suggesting that Autism Diagnostic Observation Schedule-2 (ADOS-2),
the intervention affected neural responses related to habitua- Module 3.
tion (Jones et al. 2017). A recent study of children and young Two 6-min passive observation experiments (one audi-
adults showed that those on the autism spectrum displayed tory and one visual) were conducted with concurrent EEG
atypical responses on multiple auditory paradigms, includ- recordings for all participants. In each experiment, par-
ing less reduction in P3a response to novel sounds over time ticipants were presented with 300 trials of either auditory

13
Journal of Autism and Developmental Disorders

or visual repetitive stimuli. Each auditory stimulus in the atypical sensory behavior relative to peers. The CBCL, a
sequence was a beep of 250 Hz frequency, 116 ms dura- component of the Achenbach System of Empirically Based
tion, and 73 dB volume, presented repeatedly with an inter- Assessment (ASEBA), is a caregiver-report questionnaire
stimulus interval of 1 s. The visual stimulus was a radial that measures behavioral and emotional problems in chil-
checkerboard on a gray background with 116 ms duration dren and adolescents ages 6–18 years. The CBCL provides
and 1 s inter-stimulus interval. Order of presentation for the two composite scaled scores, The Total Syndrome and Total
auditory and visual experiments was counterbalanced across Competence Scores, which are norm-referenced and derived
participants. from subtests in the Syndrome Scales and the Competence
This study was approved by Massachusetts Institute of Scales. The Syndrome Scale composite score is comprised
Technology (MIT) Committee on Use of Humans as Experi- of 8 sub scales: anxious/depressed, withdrawn/depressed,
mental Subjects (COUHES) in accordance with all institu- somatic complaints, social problems, thought problems,
tional review board requirements. The experimental protocol attention problems, rule-breaking behavior, and aggressive
was explained in detail prior to the start of the study, chil- behavior, as well as six DSM-oriented scales consistent with
dren and their legal guardians were notified that participa- DSM-5 categories. Higher scores on the Total Syndrome
tion was voluntary, written consent was obtained from the scaled score indicates greater degree of these features. The
legal guardians, and assent from each participant. Total Competence scaled score, on the other hand, meas-
ures activities, social relations, and school performance.
Behavioral Measures Higher scores on the Total Competence scaled score indi-
cates greater frequency of engaging in these categories of
We administered the WISC-V (Wechsler 2014) and the activities. The SCQ is a caregiver-report questionnaire that
ADOS-2 (Lord et al. 2012) as well as several caregiver- measures communication skills and social functioning in
report questionnaires, including the Short Sensory Profile children ages 4 and older, with a mental age of 2 years or
Second Edition (SSP-2) (Dunn 2014), the Child Behavior greater. In the SCQ-Lifetime, the parent is asked to report
Checklist (CBCL) Total Syndrome and Total Competence whether the behaviors were ever present during the child’s
Scores (Achenbach and Rescorla 2004), the Social Com- life. In the SCQ-Current, the parent is asked to report
munication Questionnaire (SCQ) (Rutter et al. 2003), Cur- whether the behaviors were present at any point in the past
rent and Lifetime. All assessments and questionnaires were 3 months. Higher scores indicate more ASD symptomology.
administered for all participants, except the ADOS-2, which We used the above measures from the SCQ, SSP and CBCL
was only administered to those children in the study with a to examine associations between habituation measures and
previous autism diagnosis, in order to determine eligibility autistic symptomatology, sensory profiles and, in aggregate,
for inclusion in the study. The scores for these standardized the syndromic traits linked to autism, respectively.
measurements are calculated according the scoring manual
for each questionnaire. We collected data for all study par- EEG Recordings and Analysis
ticipants for the CBCL, SSP-2, and SCQ. Please see Table 1
for group averages for each measure. We recorded 32 channel electroencephalography (EEG) sig-
The SSP-2 is a caregiver report questionnaire that meas- nals with the BioSemi ActiveTwoTM electrode system, at
ures children’s sensory processing patterns in everyday a 512 Hz sampling rate. The electrodes were mounted in an
activities. Higher scores indicate increased frequency of elastic cap using a subset of the International 10/20 System

Table 1  Group mean and, in Measure NT mean (SD) ASD mean (SD)
parentheses, standard deviations
for each of the characterization WISC-V nonverbal index 116.90 (9.73) 104.42 (12.59)
measures outlined in the
SSP-2 seeking/seeker quadrant (raw score/35) 9.00 (3.22) 18.25 (6.84)
correlational analyses
SSP-2 avoiding/avoider quadrant (raw score/45) 12.10 (4.04) 25.42 (7.95)
SSP-2 sensitivity/sensor quadrant (raw score/50) 13.19 (3.63) 33.08 (10.38)
SSP-2 registration/bystander quadrant (raw score/40) 8.24 (2.21) 24.25 (9.18)
SSP-2 sensory section raw (raw score/70) 18.00 (5.63) 39.83 (13.64)
SSP-2 behavior section (raw score/100) 24.29 (7.33) 59.25 (18.31)
CBCL total competence t-score 55.95 (5.07) 39.50 (9.99)
CBCL total syndrome t-score 41.57 (8.29) 61.42 (11.80)
SCQ Current 3.05 (2.01) 14.00 (8.34)
SCQ lifetime 2.76 (3.00) 23.00 (10.03)

13
 Journal of Autism and Developmental Disorders

sites as depicted in Fig. 1. We checked that the BioSemi The temporal ERP changes across standard repetition of
electrode offsets were < 20 mV to ensure quality of the elec- stimuli were used to operationalize habituation. The EEG
trode contact, as recommended by BioSemi’s user manual. recording was epoched between 100 ms pre-stimulus and
To detect blinks and lateral eye movements for later correc- 900 ms post stimulus and baseline normalized. We calculated
tion, electro-oculogram (EOG) electrodes were placed at the ERPs over all 32 channels for each subject in 50 trial segments
infraorbital ridge of the left eye and the lateral canthus of using a sliding window over all trials. This method provides
the right eye. The data were visually inspected prior to and a measure of ERP over trials as the experiment progresses
throughout the recording to ensure quality of the EEG signal. which can be later modelled to characterize habituation. For
Out of the 22 NT participants we collected data from, one EEG signals x(t, k) for time t, trial k where N = total number
child did not complete the auditory experiment so the data of trials and wl = window length 50, we calculated the ERP.
was excluded from the analysis. We excluded 1 ASD and N
1 NT participant from visual experiment analysis, as they x(t, w) =
1 ∑
x(t, k).I(k, w)
were unable to complete the full experiment. Participants z(w) k=1
who were unable to complete the full length of the stimulus
protocol were excluded from the analysis since their EEG N
data did not have adequate trials to complete the analysis and
∑
z(w) = I(k, w)
thus compare habituation profiles with other participants. k=1
All offline EEG analyses and preprocessing were done using
the Signal Processing Toolbox of Matlab. The left and right I(k, w) = 1, if w + wl > k ≥ w ∀(w, k)
mastoids were used for offline re-referencing of the EEG 0, otherwise
signals to the average of the mastoid signals. The EEG data
1≤k≤N
were low- and high-pass filtered to (0.5–45 Hz). Independent
component analysis was used to identify and remove ocular 1 ≤ w ≤ N − wl
artifacts. Trials with muscle artifacts were also rejected prior
We detected the amplitude of the most prominent peak
to averaging. Overall, 23.9% (mean 71.7, standard deviation
p(w), of the ERP x(t, w) over successive w and approximated
11.3) and 25.3% (mean 75.9, standard deviation 9.8) trials
the line of best fit on p(w) using the least square method.
were rejected from the NT and ASD group respectively due
Slopes (m) of the best fit line for each subject were used to
to artifacts. There were no significant differences (p < 0.05)
determine the overall change in stimulus-induced activation.
in the number of trials rejected between the ASD and NT
This approach quantifies the degree of habituation independent
group.
of the initial reaction differences across subjects.
The mean of the standardized auditory and visual habitu-
ation measures (slope, m) was used to generate a composite
sensory response score (CSR) for the participants. The inde-
pendent samples t-test was used to compare group differences
in the slopes (m) which serve as measures of auditory, visual
and CSR habituation.

Brain‑Behavior Correlations

Pearson’s r correlations were calculated to test for associa-

tions between a composite sensory response profile (derived
from the average of the ERP slopes for the auditory and visual
stimuli) and the behavioral scores on the CBCL competence
and syndromic normalized score, the SSP-2 subscores and the
SCQ. There were a total of 10 correlations.

Fig. 1  Electrode recording montage

13
Journal of Autism and Developmental Disorders

Results event-related component elicited by the impulses in the

initial section of both the auditory and visual stimulus
Auditory and Visual Event Related Potentials sequence were stronger than those evoked by the later
ones. By contrast, the amplitude of evoked responses in
Evoked responses for the auditory and visual experiments ASD participants did not show reduction between the first
were computed separately by averaging the electroenceph- and the last ERP. In fact, in some cases, as shown in Fig. 2,
alography (EEG) signals across 50 trial segments using a ASD participants exhibited an increased positive compo-
sliding window approach described in the methods sec- nent towards the end of the experiments.
tion. The gradual changes in the amplitudes of the positive The individual subject slopes m, of the best fit line of
component of ERPs in response to repetitive tones/visual the ERP amplitude were used to examine group differ-
stimuli during passive listening/watching were used to test ences in habituation. Figure 3a (left panel) shows sample
for differences in rates of habituation. Figure 2 shows the ERP peaks and the best-fit line from one NT and one ASD
first and the last auditory and visual ERPs averaged across participant in response to auditory tones at channel Fp1.
a window of 50 trials for sample NT and ASD subjects. As The NT participant shows gradual reduction in ERP ampli-
expected, the NT participants exhibited marked habitua- tude over time (negative slope), while the ASD participant
tion over the course of the experiment. The positive (P1) shows an increase (positive slope). In order to determine the
most pertinent effect of auditory habituation, we conducted

Fig. 2  First and last ERP for the Auditory and Visual experiment for show ERP computed for the first 50 repetitions, while the green
sample NT and ASD participants. a ERPs for auditory stimuli at Fp1. curves correspond to the last 50 repetitions in a 300 event sequence
b ERPs for visual stimuli at C3. In each condition, the black curves

13
 Journal of Autism and Developmental Disorders

Fig. 3  Left panels: ERP plots for the Auditory a response at Fp1 and exhibiting pronounced habituation in both auditory (Fp1) and visual
Visual b response at C3 for one NT and one ASD participant. Right (C3) experiments, in contrast to those with ASD
panels: Habituation slopes of all ASD and NT participants, the latter

significance tests on the individual subject slopes across all For the visual domain, we employed the direct analogue
electrodes. The bar plot on the right shows the habituation of the auditory analysis described above. We observed that
slopes of all ASD (n = 13) and NT (n = 21) participants at habituation in the visual ERP signals emerged similarly to
Fp1, with the latter exhibiting pronounced habituation, in the results from our auditory paradigm. Figure 3b shows the
contrast to those with ASD (t(32) = − 2.67, p = 0.012, 95% evoked potential response and slope analysis for the visual
CI [− 0.040, − 0.005]). When compared across all electrode experiment at channel C3 (NT, n = 21; ASD, n = 12). We
sites, this phenomenon has the greatest difference at channel find significant differences (t(31) = − 3.41, p = 0.002, 95%
Fp1 and in the frontal and fronto-central electrodes for audi- CI [− 0.13, − 0.033]) in visual habituation profiles between
tory stimuli. This is consistent with previous research, which our ASD and NT groups for the visual experiment. ASDs
investigated the topography of ERP changes with standard show reduced habituation as indicated by positive or less
repetition where ERP related repetition effects were only negative slopes compared to NTs. The greatest difference
observed over the fronto-central scalp (Haenschel et al. between habituation in the two groups occur in the central
2005). and centro-parietal electrodes in keeping with previous work

13
Journal of Autism and Developmental Disorders

in the domain of Brain Computer Interface (BCI) showing 0.05. Further analysis of correlation between the CSR scores
robust visual responses over the central region of the scalp and the WISC nonverbal IQ show the results are not signifi-
(Bleichner et al. 2015). Research in visual habituation in cant. (p = 0.30, r (30) = − 0.19) hence there is no relationship
infants indicates information about the stimulus acquired between habituation and IQ. Table 1 shows the means and
for habituation is processed centrally (Kavšek and Born- standard deviations of each measure for each group.
stein 2010). Furthermore, early work showed decreased con-
tingent negative variation (CNV) in these regions over the Habituation Rate Versus Autism Symptoms
time course of the experiment, due to the effects of habitu-
ation (McAdam 1966). Correlations between CSR scores and the SCQ current and
To validate our results further, we ran the significance lifetime scores obtained at the time of the experiment were
tests for both auditory and visual habituation using the calculated for the participants. We found that, across all sub-
bootstrapping method with matched participant numbers in jects, higher SCQ scores were associated with higher CSR
each group. Over 5 bootstrap iterations, all cases returned scores (i.e. lower rates of habituation) indicated by p < 0.001
significant results (p < 0.05) where the average p = 0.020 (Fig. 5).
(s.d 0.0066) for auditory habitation at Fp1and the average
p = 0.012 (s.d 0.0010) at C3 for visual habituation. Habituation Versus Behavioral Measures
As described in the methods, we computed the indi-
vidual “composite sensory response” (CSR) score of each In order to examine the relationships between altered ERP-
participant by averaging their auditory response slope at derived habituation scores and measures of psychological
electrode Fp1 and visual response slope at C3. Figure 4 well-being, we calculated correlations between the CSR
shows the significant group differences in the CSR scores scores and the CBCL, SSP-2, and SCQ Current and Life-
with t(30) = − 4.10, p < 0.001 95% CI [− 0.080, − 0.027]. time scores across all participants. The SSP-2 is used to
Since CSR is a slope-derived measure, high CSR scores cor- characterize the severity of sensory processing differences
respond to lower habituation rates. in autism or with other developmental conditions as well
as to characterize sensory preferences in children who are
Relationships Between CSR Scores and Clinical/ typically developing (Tomchek and Dunn 2007).
Behavioral Measures
Child Behavior Checklist (CBCL) CBCL syndromic score rep-
The individual CSR scores were used to determine the cor- resents parent ratings of the child’s behavior linked to diag-
relations between the habituation profiles and each of mul- nostic categories from the DSM such as anxiety, depression,
tiple behavioral measures. The WISC nonverbal IQ group and attention problems, wherein a higher score represents
difference between ASD and NT had a p value of less than a higher presence of these symptoms. There were signifi-

Fig. 4  Composite sensory

response scores from NT
(n = 20) and ASD (n = 12).
These scores merge slope
results from auditory and visual
habituation studies to derive a
composite measure of habitua-
tion response. ASD participants
generally have higher scores
relative to NT participants,
indicating reduced habituation
in the former group

13
 Journal of Autism and Developmental Disorders

Fig. 5  Relationship between Composite Sensory Response (CSR) r = 0.60, p < 0.001) and lifetime (Pearson’s r = 0.58, p < 0.001) Social
scores and ASD symptoms. There is a significant correlation between and Communication Questionnaire (SCQ) scores in the ASD group
the extent of reduction in habituation and the current (Pearson’s

Fig. 6  Relationship between CSR scores and behavioral characteristics. Sensory response is positively correlated with CBCL Syndrome score
across all subjects. The CBCL Competence score has a negative relationship with participants’ CSR scores

cant correlations between CSR scores and CBCL syndrome CSR scores derived from composite EEG habituation slope
scores derived from parent questionnaires at the time of the and sensory profiles as measured by the SSP-2 sub-scores.
experiment (Pearson’s r (30) = 0.53; p = 0.002). In contrast Figure 7 shows strong positive relationships between CSR
there was a negative correlation between CSR scores in all scores and all SSP-2 sub-scores with p < 0.05 in Pearson’s
children and the CBCL total competence score as shown in r correlation test. The sensitivity sensor quadrant, regis-
Fig. 6 (Pearson’s r (30) = − 0.46; p = 0.009). Total compe- tration bystander quadrant, sensory section raw score and
tence score represents parent ratings of the child’s perfor- behavior section raw score correlations with CSR had p
mance in academic activities, social, and school behavior, values < 0.001 and r values of 0.64, 0.60, 0.58 and 0.61
wherein a higher score represents a higher degree of compe- respectively. The seeking quadrant (r (30) = 0.45, p = 0.010)
tence in these domains. and avoider quadrant (r (30) = 0.55, p = 0.0012) also showed
significant correlation.
Short Sensory Profile (SSP‑2) Correlations were also used
to determine whether there were relationships between

13
Journal of Autism and Developmental Disorders

Fig. 7  Relationship between CSR scores and sensory profile scores. The rate of habituation is strongly correlated with the sensory scores of the
SSP across all subjects

Discussion the P1 positive polarity wave in their ERPs, supporting the

inference that these individuals have reduced habituation to
Data from our experiments have allowed us to address the repeated stimuli.
three goals we stated at the outset: (1) To examine whether The finding of reduced habituation in both the auditory
with non-social, repeating stimuli there are differences in and visual sensory modalities suggests that the phenom-
the neural habituation profiles of ASD and NT children, enon may be domain general in ASD. This has important
(2) To determine whether differences in habituation apply implications for the nature of causal mechanisms underlying
across sensory modalities, and (3) To determine whether the observed differences in habituation. It seems plausible,
neural habituation profiles correlate with clinical measures though not certain [for a discussion of sensory subtypes, see
used to characterize phenotypic features of autism. On all of work by Lane et al. (2014)], that reductions in habituation
these fronts, our empirical data indicate affirmative answers. may arise from central mechanisms that affect sensory pro-
To briefly recapitulate, in neurotypical participants, EEG cessing across modalities, rather than from modality specific
responses to a metronomic sequence of tones and repeated peripheral factors.
visual stimuli showed expected patterns of habituation, or Finally, our results show significant correlations between
attenuated neural response over time. By contrast, partici- the degree of habituation and parent reports of participants’
pants with ASD exhibited significantly reduced habituation. social communication difficulties, as measured by the
Unlike NTs, the ASD group did not show a decrement in SCQ, a tool used to screen for ASD. The extent of neural

13
 Journal of Autism and Developmental Disorders

habituation also correlated with parents’ impressions of the of language, social interactions and motor control. Given its
severity of sensory differences, as measured by the SSP-2. broad relevance, it is important to test the theory’s predic-
The CBCL competence (negatively) and syndrome scales tions empirically in well-defined domains, as is the case in
(positively) correlate with the composite sensory response the present study. However, it needs to be noted that the data
(CSR) scores, indicating that psychological symptomatol- we have found so far, while consistent with the theory, need
ogy and habituation appear to be related as well. Recently, to be strengthened further to more rigorously test the theory.
it has been hypothesized that differences in predictive abili- For instance, it will be important to examine how exoge-
ties in autism are related to the compendium of features of nous changes in stimulus predictability (say, by making the
autism (Sinha et al. 2014). Habituation to a repetitive, invari- sequence more stochastic rather than purely deterministic)
ant stimulus represents a most basic form of prediction. If, affect habituation rates in neurotypical participants. Further-
as we hypothesize, predictive abilities are dampened for more, we have not considered potential contributions from
autistic individuals, the ability to ignore irrelevant sensory attentional or memory consolidation processes that have
stimuli would lead to an overwhelming sensory experience. previously been reported to be able to modulate habituation
The findings here suggest that reduced habituation, a form (Lagemann et al. 2010; Sörqvist et al. 2012), although with
of prediction, is associated with autistic symptomology, as more complex stimuli and over longer time-scales than those
measured by the SCQ, SSP-2 and general psychological used in the present study. These factors are especially impor-
well-being, as measured by the CBCL. tant for further exploration given the known atypicalities in
These relationships between neural habituation and memory and attention in autism (Ames and Fletcher-Watson
social communication, sensory difficulties and psychologi- 2010; Williams et al. 2006).
cal symptomatology imply that that the degree of habitu- While acknowledging the aforementioned limitations of
ation impairment may be useful as one of several factors the present study, some arising from constraints on numbers
informing neuropsychological interventions that may benefit of subjects and circumscribed stimulus conditions, we are
the individual. In neuropsychology, much reliance is now encouraged by the systematicity of the data observed. The
necessarily placed on parental reports to identify and diag- basic finding of a reduction in habituation accompanying
nose, which are often problematic with regards to reliabil- autism is potentially an important explanatory factor for
ity (Mazefsky et al. 2011). The relative ease of collecting understanding the commonly observed sensory abnormali-
sensory response data using a protocol of the kind we have ties in the condition. While on the one hand, reduction in
described here makes this assay a potential candidate for habituation can lead to a sense of immersion in unrelentingly
inclusion in the early detection battery for ASD. Of course, salient stimuli and hence, an aversive response (a ‘hyper-
doing so pre-supposes that the habituation differences we sensitivity’), on the other hand, it might result in abnormally
have observed with our cohort of participants will be evi- sustained deployment of attention to specific stimuli. For
dent even much earlier in the developmental timeline, prior neurotypical individuals, stimuli that initially have positive
to the age of formal diagnosis. Despite a number of previ- valence (a pleasing jingle, an interesting video clip) eventu-
ous research studies investigating habituation in autism, the ally ‘lose their charm’ as a person habituates to them. How-
field currently largely lacks empirical data regarding how ever, a reduction in habituation would be expected to lead
habituation profiles change with age and whether children to a stimulus maintaining its appeal for extended periods of
at elevated likelihood of ASD are systematically different in time, and the individual would be seen as having an obses-
their profiles relative to those at typical likelihood. This is sive interest in the item. Additionally, the focus of attention
one of the avenues of research we are pursuing by recording on one stimulus might render it hard to disengage and attend
sensory responses to repeating stimuli from infants at low to another, leading to seeming ‘hypo-sensitivity’ to the latter.
and elevated likelihood of ASD. In this way, atypicalities in habituation can be important for
It is worth pointing out that the reductions in habitu- understanding multiple, sometimes contradictory, aspects
ation in ASD we have reported here are consistent with, of an individual’s response to his/her sensory environment.
and may be partially accounted for, by a theory for autism In summary, with the broad objective of examining sen-
we have recently proposed (Sinha et al. 2014). According sory processing in ASD and NT participants, our study
to the theory, autism is accompanied by an impairment in specifically highlights differences in habituation across the
temporal prediction. Given that stimulus predictability is a two groups, and the possible role of reduced habituation to
key modulating factor for the extent to which the stimu- simple sensory stimuli as a potential underlying mechanism
lus elicits habituation (Rankin et al. 2009), an endogenous for understanding sensory sensitivities in autism. Our data
reduction in predictive abilities would be expected to lead also demonstrate relationships between neural habituation
to a corresponding reduction in habituation. The theory has and sensory, cognitive and behavioral traits. We have shown
relevance beyond understanding sensory responses; it may that neural habituation is domain-general, at least across two
also account for autism symptoms manifested in the domains sensory modalities, and is behaviorally-relevant in autism.

13
Journal of Autism and Developmental Disorders

Acknowledgments We are grateful to the families who participated for brain-computer interfaces. An EEG you do not see? Physi-
in our research for their time and dedication to the pursuit of better ological Reports. https​://doi.org/10.14814​/phy2.12362​
understanding autism spectrum disorder. We also wish to express our Bölte, S., Schlitt, S., Gapp, V., Hainz, D., Schirman, S., Poustka, F.,
gratitude to the McGovern Institute for Brain Research Martinos Imag- et al. (2012). A close eye on the eagle-eyed visual acuity hypoth-
ing Center facility and staff, and the contributions of Lukas Vogel- esis of autism. Journal of Autism and Developmental Disorders,
sang, Rowan Cheung, Leila Denna, Emily Losordo, Alison Franco, 42, 726–733.
James Borders, Molly Jabeck, Xue Bao, Alesa Commedore, Jennifer Bonnel, A. C., Mottron, L., Peretz, I., Trudel, M., Gallun, E., & Bon-
McCleary, Thuy Duong Vuong, Anita Liu, Lily Zhang, Arushi Agar- nel, A. M. (2003). Enhanced pitch sensitivity in individuals with
wal, and Nishad Sharker. autism: A signal detection analysis. Journal of Cognitive Neuro-
science, 15, 1–10.
Author Contributions WJ designed and performed the experiments, Cascio, C., McGlone, F., Folger, S., Tannan, V., Baranek, G., & Pel-
collected and analyzed the data, and wrote the manuscript. AC con- phrey, K. A. (2008). Tactile perception in adults with autism: A
ducted outreach, screening, enrollment of participants; contributed multidimensional psychophysical study. Journal of Autism and
to the design of experiments, collected data, performed scoring and Developmental Disorders, 38(1), 127–137.
administration of behavioral measures, and reviewed the manuscript. Christensen, D. L., Bilder, D. A., Zahorodny, W., Pettygrove, S., Dur-
AJH contributed to participant enrollment and phenotyping. MK kin, M. S., Fitzgerald, R. T., et al. (2016). Prevalence and charac-
and PS contributed to study design, data analysis, and writing the teristics of autism spectrum disorder among 4-year-old children
manuscript. in the autism and developmental disabilities monitoring network.
Journal of Developmental & Behavioral Pediatrics, 37(1), 80–82.
Dabbous, A. O. (2012). Characteristics of auditory brainstem response
Funding This research was supported by the Simons Foundation
latencies in children with autism spectrum disorders. Audiological
Autism Research Initiative and a pilot grant from the MIT Simons
Medicine, 10(3), 122–131.
Center for the Social Brain.
DePape, A.-M.R., Hall, G. B. C., Tillmann, B., & Trainor, L. J. (2012).
Auditory processing in high-functioning adolescents with autism
Compliance with Ethical Standards spectrum disorder. PLoS ONE, 7, e44084.
Dunn, W. (2014). Short sensory profile manual (2nd ed.). San Antonio,
Conflict of interest The authors declare that they have no conflict of TX: Pearson Psychological Corporation.
interest. Ferri, R., Agarwal, E. M., Lanuzza, B., Musumeci, S. A., & Pennisi,
G. (2003). The mismatch negativity and the P3a components of
Consent to Participate Written consent was obtained from the legal the auditory event-related potentials in autistic low functioning
guardians, and assent from each participant. subjects. Clinical Neuropsychology, 114, 1671–1680.
Gomot, M., Giard, M. H., Adrien, J. L., Barthélémy, C., & Bruneau,
Ethical Approval This study was approved by Massachusetts Institute N. (2002). Hypersensitivity to acoustic change in children with
of Technology (MIT) Committee on Use of Humans as Experimental autism: Electrophysiological evidence of left frontal cortex dys-
Subjects (COUHES) in accordance with all institutional review board functioning. Psychophysiology, 39, 577–584.
requirements. Guiraud, J. A., Kushnerenko, E., Tomalski, P., Davies, K., Ribeiro,
H., & Johnson, M. H. (2011). Differential habituation to repeated
sounds in infants at high risk for autism. NeuroReport, 22(16),
845–849.
Haenschel, C., Vernon, D. J., Dwivedi, P., Gruzelier, J. H., &
References Baldeweg, T. (2005). Event-related brain potential correlates of
human auditory sensory memory-trace formation. The Journal of
Achenbach, T. M., & Rescorla, L. A. (2004). The Achenbach System Neuroscience: The Official Journal of the Society for Neurosci-
of Empirically Based Assessment (ASEBA) for ages 1.5 to 18 ence, 25(45), 10494–10501.
years. In The use of psychological testing for treatment planning Heaton, P., Hermelin, B., & Pring, L. (1998). Autism and pitch process-
and outcomes assessment: Volume 2: Instruments for children and ing: A precursor for savant musical ability? Music Perception,
adolescents (3rd ed.). London: Routledge. 15, 291–305.
American Psychiatric Association. (2013). Diagnostic and statistical Hudac, C., DesChamps, T., Arnett, A., Cairney, B., Ma, R., Webb, S.,
manual of mental disorders-5 American Journal of Psychiatry et al. (2018). Early enhanced processing and delayed habituation
(5th ed.). Washington D.C.: American Psychiatric Association. to deviance sounds in autism spectrum disorder. Brain and Cogni-
Ames, C., & Fletcher-Watson, S. (2010). A review of methods in the tion, 123(2018), 110–119.
study of attention in autism. Developmental Review, 30(1), 52–73. Jones, E., Dawson, G., Kelly, J., Estes, A., & Webb, S. (2017). Parent-
Ashwin, E., Ashwin, C., Rhydderch, D., Howells, J., & Baron-Cohen, delivered early intervention in infants at risk for ASD: Effects
S. (2009). Eagle-eyed visual acuity: An experimental investiga- on electrophysiological and habituaiton measures of social atten-
tion of enhanced perception in autism. Biological Psychiatry, 65, tion. Autism Research, 10(5), 961–972. https​://doi.org/10.1002/
17–21. aur.1754
Barry, R. J., & James, A. L. (1988). Coding of stimulus parameters in Kanner, L. (1943). Autistic disturbances of affective contact. Nervous
autistic, retarded, and normal children: Evidence for a two-factor Child., 2(3), 217–250.
theory of autism. International Journal of Psychophysiology: Offi- Kavšek, M., & Bornstein, M. H. (2010). Visual habituation and disha-
cial Journal of the International Organization of Psychophysiol- bituation in preterm infants: A review and meta-analysis. Research
ogy, 6(2), 139–149. in developmental disabilities, 31(5), 951–975.
Bernal, M. E., & Miller, W. H. (1970). Electrodermal and cardiac Khalfa, S., Bruneau, N., Rogé, B., Georgieff, N., Veuillet, E., Adrien,
responses of schizophrenic children to sensory stimuli. Psycho- J. L., et al. (2004). Increased perception of loudness in autism.
physiology, 7(2), 155–168. Hearing Research, 198, 87–92.
Bleichner, M. G., Lundbeck, M., Selisky, M., Minow, F., Jäger, M., Kleinhans, N. M., Johnson, L. C., Richards, T., Mahurin, R., Greenson,
Emkes, R., et al. (2015). Exploring miniaturized EEG electrodes J., Dawson, G., et al. (2009). Reduced neural habituation in the

13
 Journal of Autism and Developmental Disorders

amygdala and social impairments in autism spectrum disorders. Perry, W., Minassian, A., Lopez, B., Maron, L., & Lincoln, A. (2007).
The American Journal of Psychiatry, 166, 467–475. Sensorimotor gating deficits in adults with autism. Biological Psy-
Kolesnik, A., Ali, J., Gliga, T., Guiraud, J., Charman, T., Johnson, M., chiatry, 61(4), 482–486.
et al. (2019). Increased cortical reactivity to repeated tones at 8 Plichta, M. M., Grimm, O., Morgen, K., Mier, D., Sauer, C., Haddad,
months in infants with later ASD. Translational Psychiatry, 9, 46. L., et al. (2014). Amygdala habituation: A reliable fMRI pheno-
https​://doi.org/10.1038/s4139​8-019-0393-x type. NeuroImage, 103, 383–390.
Lagemann, L., Okamoto, H., Teismann, H., et al. (2010). Bottom-up Rankin, C. H., et al. (2009). Habituation revisited: An updated and
driven involuntary attention modulates auditory signal in noise revised description of the behavioral characteristics of habituation.
processing. BMC Neuroscience, 11, 156. Neurobiology of Learning and Memory, 92(2), 135–138.
Lane, A. E., Molloy, C. A., & Bishop, S. L. (2014). Classification Rutter, M., Bailey, A., & Lord, C. (2003). Social Communication Ques-
of children with autism spectrum disorder by sensory subtype: tionnaire. Los Angeles, CA: Western Psychological Services.
A case for sensory-based phenotypes. Autism Research, 7(3), Sinha, P., Kjelgaard, M. M., Gandhi, T. K., Tsourides, K., Cardinaux,
322–333. A. L., Pantazis, D., et al. (2014). Autism as a disorder of predic-
Lawson, R. P., Aylward, J., White, S., & Rees, G. (2015). A strik- tion. Proceedings of the National Academy of Sciences, 111(42),
ing reduction of simple loudness adaptation in autism. Scientific 15220–15225.
Reports, 5(1), 16157. Sörqvist, P., Nöstl, A., & Halin, N. (2012). Working memory capacity
Leekam, S. R., Nieto, C., Libby, S. J., Wing, L., & Gould, J. (2007). modulates habituation rate: Evidence from a cross-modal audi-
Describing the sensory abnormalities of children and adults with tory distraction paradigm. Psychonomic Bulletin & Review, 19(2),
autism. Journal of Autism and Developmental Disorders, 37(5), 245–250.
894–910. Tam, F. I., King, J. A., Geisler, D., Korb, F. M., Sareng, J., Ritschel,
Lord, C., Rutter, M., DiLarore, P., Risi, S., Gotham, K., & Bishop, S. F., et al. (2017). Altered behavioral and amygdala habituation in
(2012). Autism diagnostic observation schedule—second edition high-functioning adults with autism spectrum disorder: An fMRI
(ADOS-2). Los Angeles, CA: Western Psychological Services. study. Scientific Reports, 7(1), 13611.
Marco, E. J., Hinkley, L. B. N., Hill, S. S., & Nagarajan, S. S. (2011). Thompson, R. F., & Spencer, W. A. (1966). Habituation: A model
Sensory processing in autism: A review of neurophysiologic find- phenomenon for the study of neuronal substrates of behavior. Psy-
ings. Pediatric Research, 69(8), 48–54. chological Review, 73(1), 16–43.
Martineau, J., Roux, S., Garreau, B., Adrien, J. L., & Lelord, G. (1992). Tomchek, S. D., & Dunn, W. (2007). Sensory processing in children
Unimodal and crossmodal reactivity in autism: Presence of audi- with and without autism: A comparative study using the short
tory evoked responses and effect of the repetition of auditory sensory profile. The American Journal of Occupational Therapy,
stimuli. Biological Psychiatry, 31(12), 1190–1203. 61(2), 190–200.
Mazefsky, C. A., Kao, J., & Oswald, D. P. (2011). Preliminary evidence Webb, S. J., Jones, E. J. H., Merkle, K., Murias, M., Greenson, J., Rich-
suggesting caution in the use of psychiatric self-report measures ards, T., et al. (2010). Response to familiar faces, newly familiar
with adolescents with high-functioning autism spectrum disor- faces, and novel faces as assessed by ERPs is intact in adults with
ders. Research in Autism Spectrum Disorders, 5(1), 164–174. autism spectrum disorders. International Journal of Psychophysi-
McAdam, D. W. (1966). Slow potential changes recorded from human ology, 77(2), 106–117.
brain during learning of a temporal interval. Psychonomic Sci- Wechsler, D. (2014). WISC-V administration and scoring manual. San
ence, 6, 435–436. Antonio, TX: Psychological Corporation.
Miller, L. J., Reisman, J. E., McIntosh, D. N., & Simon, J. (2001). An Williams, D. L., Goldstein, G., & Minshew, N. J. (2006). The profile
ecological model of sensory modulation: Performance of chil- of memory function in children with autism. Neuropsychology,
dren with fragile X syndrome, autistic disorder, attention-deficit/ 20(1), 21–29.
hyperactivity disorder, and sensory modulation dysfunction. In:
Understanding the nature of sensory integration with diverse Publisher’s Note Springer Nature remains neutral with regard to
populations (pp. 57–88). jurisdictional claims in published maps and institutional affiliations.
Montagu, J. D., & Coles, E. M. (1966). Mechanism and measurement
of the galvanic skin response. Psychological Bulletin, 65(5), 261–
279. https​://doi.org/10.1037/h0023​204
O’Riordan, M. A., & Passetti, F. (2006). Discimination in autism
within different sensory modalities. Journal of Autism and Devel-
opmental Disorders, 36, 665–675.

13