Parsing Heterogeneity In Autism Spectrum Disorder Using Blink Inhibition as a Measure of Social Engagement

Background: Heterogeneity in autism spectrum disorder (ASD) is an obstacle to advancements in identifying and treating causes of the disorder.  Measures capturing the core underlying features of ASD, such as reduced engagement with socially adaptive stimuli, may provide a means for parsing phenotypic heterogeneity in ASD.  Our laboratory developed a novel approach for quantifying viewers’ moment-by-moment engagement with dynamic stimuli by measuring patterns of blink inhibition during a natural viewing task.  This method capitalizes on the fact that people spontaneously inhibit eye-blinks when processing salient visual stimuli to minimize the loss of visual information that occurs when blinking.  Using this method, we demonstrated that children’s visual engagement with dynamic stimuli led to predictable patterns of blink inhibition: typically-developing (TD) toddlers inhibited eye-blinks when viewing highly affective content and exhibited increased eye-blinking during portions of low affective content.  In contrast, the blink rate of toddlers with ASD did not vary with respect to social content. While these findings suggest reduced engagement with socially relevant stimuli in ASD, the clinical utility of these results for parsing heterogeneity in ASD and identifying homogeneous subgroups remains unknown.

Objectives: (1) To assess whether reduced social engagement, as measured by blink inhibition, correlates with social disability in ASD, and (2) to assess the utility of this measure for identifying homogeneous subgroups in ASD.

Methods: Eye-tracking data were collected from school-age children with ASD (n = 49) and matched typically-developing peers (n = 26) while watching movies of social interaction.  Visual engagement in individuals with ASD was indexed by patterns of blink inhibition relative to salient movie events, defined as movie frames during which the TD group inhibited their blinks.  Our dependent measures were (1) an individual’s decrease in blink rate relative to salient events, and (2) the latency at which an individual decreased their blink rate following salient events.

Results: A subgroup of children with ASD (n = 15) exhibited relatively normative patterns of blink inhibition, as indexed by a significant decrease in blink rate following salient events.  Within this subgroup, both percentage decrease in blink rate and latency of blink inhibition correlated strongly with social ability, measured by the ADOS (r’s =.5-.8, p’s < .05).  Greater decrease in blink rate and shorter latency of blink inhibition were associated with greater social ability.   Preliminary analyses identified a second subgroup of children (n=7) with disrupted patterns of visual engagement, as indexed by a significant increase in blink rate following salient events.  These children had significantly lower IQ and Vineland scores compared with the first subgroup. Ongoing analyses are aimed at further identifying and understanding altered patterns of visual engagement in subgroups of ASD.

Conclusions: Visual engagement with dynamic social stimuli, as indexed by patterns of blink inhibition, provides a promising measure for parsing heterogeneity in ASD and improving quantified phenotypic characterization of children.  Future analyses, aimed at investigating the type of stimuli that engage subgroups of children with ASD will be an important step towards identifying alternate learning strategies and ways of experiencing the world in ASD.