Objectives: To examine dynamic visual scanning in relation to an ethological inventory of natural behaviors during video scenes of social interaction, and to then identify specific events (e.g. facial expressions, vocalizations, movements) that either elicit or fail to elicit convergent visual scanning in individuals with ASD relative to nonautistic benchmarks.
Methods: Eye-tracking data were collected from adolescents and young adults with ASD (mean age = 16.67 (3.92) years; n = 21) and TD controls matched on age and verbal function (mean age = 16.86 (4.5); n = 17) while viewing video scenes of realistic social interactions. We used kernel density estimation to quantify the level of convergence of visual scanning at each moment in time for both groups in order to obtain measures of relative salience. Ethograms were constructed for each video, for which the onset and offset of specific events were characterized on a frame-by-frame basis, and used to examine moments when significant group-differences occurred.
Results: Preliminary analyses suggest that between-group differences in visual scanning were greatest when more actors were onscreen without camera movement, during which visual scanning by individuals with ASD exhibited considerably diminished convergence on faces that were salient to TD individuals. However, when individuals with ASD demonstrated higher convergence on faces, we found that (1) the face/s occupied greater total screen area, or (2) the actor/s made higher amplitude vocalizations and/or particular body motions. Consistent with past research, the ASD group looked more at the mouth and body regions than the eyes when looking at faces.
Conclusions: During viewing of naturalistic social situations, groups of TD and ASD individuals demonstrate significantly different patterns of dynamic visual scanning. The nature of these group differences seems largely mediated by both physical and contextual factors, with individuals with ASD at their greatest disadvantage at times when more actors were present on screen and when the visual environment was more cluttered. This appears to increase the attentional demands required by viewers, both to rapidly reallocate attention to relevant events as they unfold in time and to ignore behaviorally irrelevant distractors.
See more of: Clinical Phenotype
See more of: Symptoms, Diagnosis & Phenotype