Cognitive Control of Visual Attention in ASD in Response to Social and Non-Social Stimuli

Background:  Narrowed interests, perseverative patterns of attention and reduced visual exploration have been linked to restricted, repetitive behaviors (RRBs) in Autism Spectrum Disorder (ASD). Circumscribed interests (CIs) in ASD, are unusually intense interests in a narrow range of subject areas, and these activities (i.e., collecting, manipulating, reading, conversing about their interest) often lead to functional impairments. Eye-tracking research investigating responses to categories of images reflecting CI that capture attention during passive viewing tasks has found that children and adults with ASD display an attentional bias towards certain categories of nonsocial images (e.g. train, automobiles, electronic devices, computers). This bias has been conceptualized to reflect increased visual salience of nonsocial images relative to social images (e.g. faces) and other, commonplace, nonsocial information (e.g. furniture, clothing, dishes).

Objectives:  This eyetracking study aimed to extend findings regarding atypical patterns of attention in children with ASD by investigating the cognitive control over visual attention in response to social and non-social images.

Methods:  

40 individuals ranging in age from 9-18 years old participated. Of those, 19 individuals had a diagnosis of ASD (mean age(stdev)= 13.92(±3.05), 16 male) and 21 were typically developing controls (mean age (stdev)= 14.06(±2.79), 19 male).  Participants completed a visual saccade paradigm where, on each trial, first a central crosshair was presented followed by a social or nonsocial image to the left or right side of center.  The task included both a prosaccade condition (“look towards a peripheral target when it appears”) and an antisaccade condition (“look to the opposite side of the screen when a peripheral target appears”). Stimuli varied between social images (smiling faces) and non-social images previously shown to be related to CI’s in ASD (“high interest” images) and commonplace, everyday objects (“low interest” images).

Results:  The primary dependent measure was the difference between antisaccade and prosaccade errors, reflecting the influence of cognitive control over visual attention.  Children and adolescents with ASD demonstrated an increased overall rate of error relative to typically developing controls, t(38)=2.21, p<.03.  A repeated measures analysis of variance did not reveal stimulus or a group x stimulus interaction was revealed (p’s>0.20). Between groups t-tests were carried out to investigate group differences in saccade errors in response to each of the 3 stimulus categories. Relative to typically developing controls, the ASD group made an increased number of errors in response to both non-social stimulus categories- “high interest” (t(38)=2.23, p<.03) and “low interest” images (t(38)=2.64, p<.01). Groups did not differ in saccade errors in response to faces (t(38)=1.09, p<.28). Within the ASD group, error rate was significantly related to social-communication impairments and restricted and repetitive behaviors calculated from the ADOS-2. Exploratory comparisons across children with ASD and a small cohort with another development disorder characterized by repetitive behaviors (pediatric OCD; n=9, mean age(stdev)=15.43(±1.01)), will also be presented.

Conclusions:  

Impairments in cognitive control of visual information appear to be domain general and a powerful predictor of ASD symptom severity.  This study provides support for the use of visual attention and oculomotor behavior to quantify impairments in ASD.