Examining visual attention in autism spectrum disorder: An exploratory analysis of instructional stimuli on young children

Autism spectrum disorder (ASD) impacts developmental trajectories of approximately 1 in 59 children in the United States. Intervention greatly improves developmental outcomes for young children with ASD, particularly evidence-based interventions directly targeting impairments in communication and reciprocal social interaction. Robot-assisted interventions are one treatment approach with an increasing research foundation. The increase in robots as a technological component of interventions for children with ASD points to a need for more comparative research to define how children with and without ASD respond to robots in an instructional capacity. Thus, the present study sought to evaluate visual attention responses to robot- (RDI) and human-delivered instructions (HDI) in groups of young children with and without ASD. RDI and HDI stimuli encompassed motor imitation and intraverbal language, two instructional targets often used in early intervention programming for young children with ASD. Ten children with ASD (Mean age = 29 months, 8 male) and eight children without ASD (Mean age = 35 months, 4 male) participated in assessment and experimental sessions for the study. Participants completed behavioral and diagnostic assessments (i.e., Mullen Scales of Early Learning (MSEL); Verbal Behavior Milestones Assessment and Placement Program (VB-MAPP); Autism Diagnostic Observation Schedule, Second Edition (ADOS-2)) examining pre-existing problem-solving, fine motor development, language, and motor imitation. Assessments also established diagnostic categorization of ASD or non-ASD using standardized measures. Multiple aspects of visual attention were evaluated using a Tobii TX300 eye-tracker, including: 1) Time to first fixation; 2) Total fixation duration; 3) Total fixation count; 4) Total visit duration; and 5) Percentage fixated. Multilevel regression analyses were completed to compare responses across group (i.e., ASD v. Non-ASD), instructional delivery (i.e., RDI v. HDI), and baseline performance on the VB-MAPP. Results revealed significant differences across all visual attention metrics, indicating different patterns of viewing instructional stimuli across group, delivery, and pre-existing skills. In general, children with ASD showed slower times to first fixation and lower levels of total fixation duration for HDI stimuli. Children across both groups showed higher (i.e., longer) fixation duration and fixation counts for RDI stimuli. More limited baseline behavioral repertoires were a significant covariate for children with ASD, particularly for total visit duration. Percentage fixated values were also significantly different across groups of children, with children with ASD showing lower fixation percentages -especially for HDI stimuli. Results confirmed different patterns of visual attention when comparing children with and without ASD viewing robot and human stimuli. Detailed consideration of these results directs future research into the implementation of robots as components of early intervention. Matching types of instructional or intervention targets to delivery method and baseline behavioral performance will promote more effective robot-assisted intervention and inclusive (i.e., children with and without ASD) educational settings.