Pay Attention during the Important Part: Adults with ASD Increase Their Gaze to Faces When Watching Richer Social Scenes

Background:  Reduced attention to social information is a defining characteristic of autism spectrum disorder (ASD; (APA, 2013)), and can be directly measured using eye tracking technology (Klin et al., 2002). Despite overall reductions in attention to social information, however, some older individuals with ASD appear to utilize compensatory mechanisms to “hack the social code”, particularly when cognitive reasoning skills are intact (Frith, 2004). In this study, we test whether adults with ASD show evidence of contextually modulated attention to social information (faces) using an established paradigm that distinguishes children with ASD from typically developing counterparts (Chevallier et al., 2015). We hypothesized that adults with ASD would look relatively more at people’s faces in the context of nonverbal communication (two children playing together) than in the absence of communication (two children playing in parallel). 

Objectives:  Assess contextual modulation of visual attention to social stimuli in adults with ASD.

Methods: Twenty-eight adult participants with ASD (4 female; see Table) watched a ~6-minute video comprised of 22 sequential 15.5-second scenes of children playing together (Joint condition) or separately (Parallel condition). A Tobii X120 infrared eye tracker measured participants’ gaze direction and duration. Areas of interest (AOI) were drawn around faces (social stimuli) and background objects (nonsocial stimuli). To control for individual differences in overall attention, we calculated the proportion of looking time toward each AOI relative to total looking time at the full screen. Preliminary analyses revealed no significant correlations between gaze variables and age or IQ, so we did not include these as covariates. 

Results:  A 2x2 repeated measures ANOVA revealed a significant interaction between condition (Joint, Parallel) and gaze duration to each stimulus type (Face, Background Objects), F(1,27)=53.59, p<.001, η_p²=.67 (see Figure). Planned paired t-tests showed that participants looked more to faces in the Joint condition than the Parallel condition. To assess the nature of this effect, we calculated a Social Prioritization metric for each condition. Social Prioritization was defined as the total duration of looking at faces minus the total duration of looking at background objects. A paired samples t-test comparing average Social Prioritization in the Joint vs. Parallel condition showed that participants looked twice as long at faces relative to objects in the context of joint play between actors (mean=.08) than in the context of parallel play (mean=.04). 

Conclusions:  Contextual modulation of visual attention to social stimuli is an important skill that may be relatively preserved in adults with ASD. Despite clinically significant social impairments, participants looked more at children’s faces in the context of interactive play than parallel play. This suggests that typical patterns of attention distribution may exist in older individuals with ASD, with the ability to tune into social context by modulating their gaze to capture communicative information from the face when it is present. Planned future analyses will explore whether this effect is also found in typically developing adults (data collection underway), will compare the magnitude of the modulation by diagnosis, and will assess whether contextual modulation can serve as a metric of treatment response.