Inferring Comprehension of Social Interaction from Eye Tracking in Neurodiverse Individuals in Rural Settings

Abstract

Background: Research shows deficiencies in social abilities in Autism Spectrum Disorder (ASD) are related to impaired Theory of Mind (ToM), further negatively impacted by language impairment. Both challenges are especially pronounced in rural populations. Given that ~25-35% of children with ASD are minimally verbal or nonspeaking, we employed a low-verbal ToM task, using eye tracking to measure attention, to better represent ASD population for ToM evaluation. Methods: Colle et al.’s (2007) low-verbal ToM task was adapted as 26 15-60 second videos, shown to participants while their eye gaze and non/verbal answers were recorded. Videos feature a Hider, hiding a toy Bear under one of two boxes out of view of participant, and a Communicator, pointing to the location of Bear as determined by their perspective. The pre-test and control conditions taught the participant that Communicator is reliable when they observe the placement of Bear, and established whether the participant could remember the switched locations of Bear. In the false-belief condition (FB), Hider switched the box locations while Communicator was absent. Upon returning, Communicator incorrectly indicated the location of Bear – as based on their point of view which did not match reality. To show intact ToM, participant must identify the correct location of Bear. We used Gazepoint eyetracking with analysis in RStudio. Results: We evaluated 30 child participants from rural and micropolitan Montana, ages 1.9-17 years, 11 with ASD, 7 without ASD but with language impairment or ADHD, and 12 neurotypical. 7 ASD and neurotypical adults participated too. Analysis of the correct scores on the behavioral answers (verbal or nonverbal pointing) separated by ASD vs. non-ASD diagnosis revealed a weaker comprehension of FB in the child ASD group (45%) compared to non-ASD child group (73%). A statistical model including the whole child group indicated a significant effect of nonverbal IQ and overall language skills (as measured by a parental questionnaire of CCC-2) on the knowledge of ToM. However, ToM performance was not affected by chronological age or diagnoses. Eye-tracking results for 20 participants aligned with the behavioral results in our data driven grouping: 4 adults and 13 children (across diagnoses groups) passed FB on behavioral answers, and 3 (all ASD) failed. The Pass groups showed more looks to Hider and target locations, whereas the Fail group looked more to Communicator and Hider. An additional 6 participants failed at FB, but had poor eye tracking recordings, and hence could not be analyzed. Conclusion: We found little evidence of association between the diagnosis of ASD with successful completion of the nonverbal ToM task. Eye-tracking was useful for studying attentional patterns to social interaction in pass vs. fail groups, noting the challenges with recording of good eye tracking data.