Validation of Eye-Tracking Measures of Social Disability As a Treatment Endpoint in School-Age Children with ASD

Background: One critical goal of biomedical research in autism is to develop effective treatments that improve the quality of life for individuals on the autism spectrum.  Two major obstacles currently hinder this goal, particularly for the development of treatments that target core social deficits of the condition: the heterogeneity of the autism spectrum and the social nature of the disability.  The heterogeneity means that different individuals will present with varying levels of affectedness, requiring different treatment approaches; and the social nature of the condition means that the disability itself is inherently challenging to quantify.  Given these obstacles, two important questions present themselves: How do we best quantify the degree of social disability at the core of autism?  And how do we best measure whether a treatment for social disability has been effective?  The goal of the current research project is to test the extent to which eye-tracking measures of social disability can serve as successful endpoints or "outcome measures" for the assessment of new treatments.

Objectives: This study aims to assess the psychometric validity of eye-tracking measures to be used as treatment endpoints in school-age children with ASD.

Methods: We assessed the use of eye-tracking measures, which were collected during free viewing of naturalistic videos of social interaction in terms of their general appropriateness to the condition; their content, construct, and convergent validity; their reliability, precision, and internal consistency; as well as their interpretability and patient acceptability.  To assess the range and distribution of scores found in the general population, we collected normative eye-tracking data from typically developing children (TD, mean age=9.61 years, N=42) and then also collected comparison values in a large and heterogeneous sample of children with ASD (mean age=10.55 years, N=128).

Results: Consistent with previous research, children with ASD spent significantly less time looking at the eyes than did TD peers (M(SD) = 32.3(17.8)% for ASD vs. 39.9(18.2)% for TD, F = 51.8, p<0.05).  Eye-looking reliability measures indicate good test-retest correlation within each group: r = 0.619 for TD and r =0.624 for ASD.  The measures were collected with a greater than 88% success rate and with measurement accuracy of less than 3 degrees of visual angle.

Conclusions: Eye-tracking measures of visual scanning, while watching naturalistic scenes of social interaction, can serve as effective and valid quantifiers of social disability.  In the current analyses, we focused on attention to the eyes of others; however, we expect that a more important assay of social disability will be measuring not only overall looking time to others’ eyes, but more temporally sensitive measures of looking to the right eyes at the right moments in time. Our rigorous assessment of promising eye-tracking measures will provide us with multiple dimensions with which to quantify social deficits in a naturalistic way and assess the effectiveness of treatments. By doing so, we hope to alter the course of ASD so as to afford individuals on the spectrum an improved quality of life.