Quantifying Convergent Validity between Parent-Reported Developmental Concerns and Eye-Tracking-Based Measures of Social Visual Engagement

Poster Presentation
Friday, May 3, 2019: 5:30 PM-7:00 PM
Room: 710 (Palais des congres de Montreal)
A. I. Mendez, T. A. Ponzo, A. Klin, C. Klaiman, S. Shultz and W. Jones, Marcus Autism Center, Children's Healthcare of Atlanta and Emory University School of Medicine, Atlanta, GA
Background: Parent-report plays a crucial role in the diagnostic landscape for autism spectrum disorder (ASD). Caregivers have the most in-depth experience of their children’s development — offering information that might not be easily obtained within shorter-duration clinician-based assessments. Here, we investigate the association between performance-based, objective measures of a child’s behavior, collected via eye-tracking, and parent-reported concerns on one of the most widely-used developmental screeners, the Ages & Stages Questionnaire - 3rd Edition (ASQ). We investigate whether parent’s concerns are associated with a child’s social visual engagement when viewing scenes of peer social interaction. Social visual engagement is a central means through which children define the course of their own developmental trajectories (Constantino et al., 2017): by engaging with certain aspects of their social environment, children gather unique information that alters their learning. In ASD, variation in social visual engagement is strongly associated with levels of clinician-measured social-communicative competence.

Objectives: To investigate the relationship between parent-reported developmental concerns and performance-based, objective measures of children’s social visual engagement via eye-tracking in a large, heterogeneous, clinically-referred sample.

Methods: Parent-reported developmental concerns and eye-tracking data were collected from N=146 consecutive referrals (Mage=28.23 months, SD=6.89) to a community-based clinic. Children were referred because of general or ASD-specific developmental concerns. All children received a gold-standard diagnostic evaluation, including the ADOS-2, developmental history, and developmental functioning measures (ASD n=115, non-ASD developmental delays n=14, no clinically-verified delays n=17). Parent-reported developmental concerns were collected through the ASQ. Eye-tracking data were collected while toddlers watched video scenes of children in daycare settings and were quantified by measuring percentage of visual fixation to faces and objects. Spearman correlation was used to test for associations between parent-reported concern and visual fixation to regions-of-interest.

Results: The data show significant associations between measures of social visual engagement and parents’ concerns about their children’s social-emotional and communication development, with increased attention to faces predicting fewer concerns (r=0.451, p<0.001 and r=0.330, p<0.001, respectively; see Figure 1) and increased attention to objects predicting greater concerns (r=-0.452, p<0.001 and r=-0.266, p=0.003, respectively; see Figure 1). Specifically, fixation on faces was significantly associated with parents’ concerns about whether their child points, follows directions, plays pretend, plays with dolls, and talks like other children (all p<0.05). These associations appear specific to social-emotional and communication concerns, rather than to general developmental concerns, as fixation on faces was unrelated to parental concern about their child’s ability to walk, jump, use stairs, turn doorknobs, or copy a line (all p>0.05).

Conclusions: Results indicate significant moderate convergent validity between eye-tracking-based measures of social visual engagement and parent-reported developmental concerns — consistent with prior research measuring the extent to which measures of social visual engagement are associated with clinician-administered assessments (Jones & Klin, 2013; Jones, Carr, Klin, 2008). Follow-up work will replicate the current results in an independent sample (N=105) and will test the extent to which the relationships between measures can be used to help clarify a challenge in the use of parent-report measures: biases that can manifest as inconsistently reported concerns.