Average Look Duration and Its Correlation with Neurophysiological Activity in Young Children with and without Autism Spectrum Disorder

Background:

Autism spectrum disorder (ASD) is characterized by early attentional differences that emerge during the infant/toddler period. Development of novel tools that measure attentional behaviors may lead to earlier identification of children at risk for ASD. In this work we introduce a new measure related to social attention, the average look duration (ALD) to a stimulus, a measure of attention behavior while viewing dynamic stimuli. We also introduce the Relative ALD (RALD),a normalized measure indicating preference when comparing stimuli responses, e.g., to compare ALD in social vs nonsocial stimuli.

Objectives:

Our objectives are to evaluate whether preschool-age children with ASD versus typical development (TD) differ in terms of their ALD and RALD while watching social and nonsocial video stimuli; and to examine associations between these measures and signal features of simultaneously recorded electroencephalographic (EEG) activity.

Methods:

Two groups of 31 age-matched ASD and TD children, 28-81 months old, participated. Participants were presented 3 types of stimuli (social– nursery rhymes; nonsocial - toys; neutral - bubbles) lasting 1 minute each and repeated twice, with simultaneous video recording of children faces and EEG. Periods of visual attention to the stimulus were marked by raters on recorded videos. ALD(Stimulus) was computed as total time spent looking/attending at the stimulus divided by total number of looking periods. RALD(Stimulus1,Stimulus2) was computed as (ALD(Stimulus1) – ALD(Stimulus2))/(ALD(Stimulus1) + ALD(Stimulus2)). First, ANCOVA models were utilized to compare groups on the attention variables. Second, EEG data were pre-processed, resulting in 40 seconds of artifact-free data for each stimuli type. Average relative power of the EEG signal (RP), log-ratio of RP (LRP) in four frequency bands, and log-ratio of theta-beta ratio (LR_TBR) between social and toys stimuli, retrieved from frontal, central, and posterior scalp regions, were computed. Associations between RP/LRP/LR_TBR and ALD(social)/RALD(social,toys) were tested using linear models, FDR-corrected at 0.05 level. Covariates are Full Scale IQ, age, and sex.

Results:

While the ALD pattern while viewing toys and bubbles was similar in both groups, ASD children demonstrated shorter average look duration when viewing the social stimulus (Fig. 1). Strong effects of group (p<0.001), stimulus type (p<0.001), and group by stimulus interaction (p<0.05) were observed for ALD. RALD (social,toys) and RALD (social,bubbles) were significantly lower in the ASD group (p < 0.002 and p < 0.04 , respectively), while RALD (toys,bubbles) was not. Simultaneous increase of EEG theta RP and decrease in beta RP during social stimulus was positively associated with increase in RALD (social, toys) in the TD group, while no significant associations were found for the ASD group. Log-ratio of posterior TBR was positively associated with RALD (social,toys) in the TD and negatively in the ASD group (Fig.2).

Conclusions:

Average look duration was significantly lower in the ASD group across all conditions. Also, ASD was associated with relatively shorter look duration while viewing social as compared to nonsocial stimuli. Relative look duration to social vs. nonsocial stimuli was differentially associated with specific patterns of relative theta and beta EEG power for ASD and TD children.