Studying Social Attention in Autism Spectrum Disorders: Stimulus Type Matters

Background: Autism spectrum disorders (ASD) are characterized by social and communication deficits, which we have hypothesized to be caused by atypicalities in the areas of social attention and responsiveness to social reward.  Prior studies show that individuals with ASD attend to faces less than typically developing controls (TDCs), a pattern that is often coupled with an unusually high interest in certain non-social objects (e.g., trains).  Infrared eye tracking technology has proven to be an effective method for examining social perception and social preferences in ASD, but different paradigms are more or less successful in revealing social attention deficits.  We hypothesize that the nature of social stimuli shown in an eye tracking study has an impact on participants’ gaze patterns, and that highly ecological, dynamic stimuli are optimal to uncover meaningful group differences in social attention.

Objectives: Compare the effectiveness of 3 different eye tracking tasks in distinguishing ASD and TDC participants on the basis of visual attention to face and object stimuli. The three tasks differed primarily on the ecological validity of the social stimuli: Static vs. Dynamic vs. Interacting faces. This variation in ecological relevance enabled us to investigate patterns of attention modulation across social and nonsocial stimuli within and between TDC and ASD groups.

Methods: 66 children with an ASD (mean age = 11.78 years) and 22 TD children (mean age = 14.9 years) participated in three passive viewing tasks while eye gaze data was collected using a Tobii X120 system.  The first task, a “Static Visual Exploration” task, showed still images of objects and people; the second, a “Dynamic Visual Exploration” task, simultaneously showed four dynamic video clips of individual faces and objects; the third, an “Interactive Visual Exploration” task, showed video clips of children playing.

Results: A repeated measures ANOVA was constructed to determine whether patterns of eye gaze to social vs. non-social stimuli of interest vary by diagnostic group and by task. A three-way interaction was found between Task (Static, Dynamic, Interactive), Stimulus-Category (social, non-social), and Diagnostic Group (ASD, TDC). Post hoc two-way ANOVAs within each task revealed that the Group by Stimulus-Category interaction was not significant for the Static task, reached trend levels for the Dynamic task, and was highly significant in the Interactive task.

Conclusions: Eye tracking is an increasingly popular method for evaluating visual behavior across multiple fields of study, including research in autism spectrum disorders.  It is important to note, however, that not all eye tracking paradigms provide equally good metrics of social attention in ASD.  Dynamic stimuli appear to be better than static images for measuring social response, and highly ecological paradigms (depicting actual interactive scenes) appear to be optimal.