Combining Viewing Patterns and Socio-Emotional Insight Questions in Dynamic Social Scenes in Children with Autism Spectrum Disorder (ASD)

Background:  Given that social impairments are one of the core features in Autism Spectrum Disorders (ASD), it is not surprising that a lot of research focused on emotional processing in individuals with ASD. Despite mixed findings, most researchers using stimuli with a high ecological validity (e.g., complex emotions embedded in a social context) found evidence for problems with emotion processing in children with ASD. Since the pioneer study of Klin et al. (2002), many studies focused on emotion processing strategies using eye-movement recording. Despite the evidence for atypical scanning patterns provided by Klin et al. (2002), not all researchers found major differences in viewing style, or they found evidence for more subtle differences between both groups.

Objectives:  We wanted to provide insight into the socio-emotional processing of children with an Autism Spectrum Disorder (ASD), using facial expressions within a dynamic complex social context.  Therefore, we combined two lines of research in the emotion processing field, by examining the relationship between viewing patterns and socio-emotional insight, using complex social scenes.

Methods:  Two groups of 10-to-14-year old boys without intellectual disabilities (IQ >= 70), group-wise matched for age, TIQ, VIQ and PIQ, participated in this experiment: an ASD group of children with a diagnosis based upon a multidisciplinary assessment according to DSM-IV-TR PDD-criteria and a typically developing (TD) group, representative for the general population. Five dynamic episodes, selected from a Dutch-spoken soap series for adolescents, were shown, while eye-movements were sampled monocularly at 500 Hz using an Eyelink II system (Pupil Only Mode). At the end of each episode, a questionnaire was used to test the children’s understanding of the social-emotional events.

Results:  We did not find group differences in global scanning parameters (in none of the five episodes): there were no differences concerning fixation count, fixation duration, or saccade amplitude. In addition, four dynamic regions of interest (ROI) were created: face, body, eyes, and mouth. We could not provide evidence for differences in viewing time in the selected ROIs for one video clip analyzed so far.

Conclusions:  We compared viewing patterns of children with and without ASD when watching dynamic social scenes. No evidence for difference in global scanning parameters was found. Preliminary results showed no differences in proportion viewing time in 4 ROIs: face, body, eyes, and mouth. Future analyses could comprise more subtle scanning parameters, such as first fixations, mouths on speaking and non-speaking persons, and alternations between socially interacting individuals. Moreover, separate analyses will focus on different emotional expressions and different fragments (e.g., more subtle socio-emotional aspects on which qualitatively strange responses were made).