Attention to Social and Nonsocial Events In Children with Autism Spectrum Disorders: The Role of Stimulus Variability

Background:  

Autism spectrum disorders (ASDs) are characterized by impaired social-communicative functioning and decreased preferences for social events (Dawson et al., 1998, 2002). Because social events differ from nonsocial events along various dimensions, the basis for enhanced preferences for nonsocial over social events in ASD is unclear. For example, social events typically involve more stimulation, contingent feedback, and affective content, as well as greater complexity, variability, and unpredictability than nonsocial events.

Objectives:  

The present study: 1) assessed preferences for social vs. nonsocial events, while partially controlling for complexity, unpredictability, and variability (hereafter referred to as “variability”) across event type, and 2) examined whether the level of variability impacted visual preferences for children with ASDs and typical development (TD). We predicted that children with ASDs would prefer nonsocial over social events and events of lower variability than TD children. 

Methods:  

Participants included 12 children with ASDs and 12 TD children, all 1- to 4-year-old males, roughly matched on nonverbal mental age (NVMA; Mullen, 1995; ASD 3.15 years; TD: 3.44 years). Children viewed six films with synchronous soundtracks. The social events depicted a woman speaking and the nonsocial events depicted dynamic, multicolor, oscilloscopic images with music. The social and nonsocial events were designed to depict low, moderate, and high levels of variability and to be approximately matched for variability across event type. Variability levels were created by manipulating pattern length and number of repetitions of the pattern within each 30 s video. Each video was played twice for a maximum of 30-s per trial or until the child looked away for at least 2-s. The dependent variables indexing interest were looking time to each video expressed in seconds (LT), looking time to each video expressed as a proportion of the child’s looking time to all videos (PLT), and number of disengagements per minute from each video.

Results:  

A 2 x 2 x 3 repeated measures ANOVA with diagnostic group (ASD vs.TD) as the between subjects factor and event type (social vs. nonsocial) and variability level (high, moderate, low) as repeated measures factors was conducted for each measure. Results indicated a main effect of event type for each measure (ps < .05) with greater interest in the nonsocial than the social events. Moreover, a diagnostic group by event type interaction was obtained for each measure (ps < .05), indicating that the preference for nonsocial events was carried by children with ASDs. They showed significantly greater interest in nonsocial than social events whereas TD children showed no preferences for one event type over another. Surprisingly, no main effects or interactions involving variability for any of the measures emerged for either diagnostic group (ps>.10).

Conclusions:  

Regardless of the level of stimulus variability, children with ASDs preferred nonsocial to social events whereas TD children showed no preferences. These findings support previous research demonstrating social orienting deficits and enhanced preferences for nonsocial events in children with ASDs. They suggest that interest in nonsocial events may be robust despite manipulations of variability, complexity, and predictability.