Nonsocial Attentional Bias in Adolescents with ASD Is Not Influenced By the Inversion Effect

Background: Circumscribed interests are a common expression of repetitive behaviors in Autism Spectrum Disorders (ASD); these behaviors tend to be nonsocial in nature and often reflect a disproportionate amount of attention and engagement, compared to social interactions. Our lab previously developed a task to quantify aspects of circumscribed attention for nonsocial information in visual arrays containing both social and nonsocial images (Visual Exploration Task, VET). Using this task, we have previously reported attentional patterns in ASD that are biased toward nonsocial information (e.g., increased number and duration of fixation) and away from social information. However, the mechanism underlying this bias remains unclear.

Objectives: Inverting visual information can lead to attention changes and recognition impairments, most commonly for holistically processed images, such as faces, but extend in some individuals to domains of individual expertise (e.g., cars, birds). The purpose of this task was to assess the influence of inversion on social and nonsocial information in adolescents with ASD, compared to typically developing (TYP) peers.

Methods: Attentional parameters were measured for samples of adolescents with ASD (N = 15; mean age = 13.9 years) and TYP (N = 17; mean age = 13.8 years) for both an upright and an inverted version of the VET eye-tracking task. In this passive viewing task, participants viewed upright and inverted visual arrays, containing both social and nonsocial images. Visual attention was assessed between array types for a) perseveration: average length of time on an item; b) exploration: proportion of items explored; c) detail orientation: average number of discrete fixations; and d) duration: average length of an individual fixation.

Results: For upright arrays, participants with ASD displayed increased nonsocial detail orientation (F = 7.4, p = .008), a trend for increased nonsocial perseveration (F = 3.02, p = .078), and decreased social perseveration (F = 3.7, p = .058), exploration (F = 17.1, p < .0001), and duration (F = 10.4, p = .002) compared to TYP participants. For inverted arrays, ASD maintained increased nonsocial perseveration (F = 4.6, p = .042) and detail orientation (F = 7.0, p = .038), relative to TYP; however, the groups showed no differences for social variables (all ps > .05).

Conclusions: As expected, we found that inversion of images altered attentional patterns to social images in the TYP group but not the ASD group. Further, the nonsocial attentional bias pattern seen in participants with ASD was not impacted by inversion of the images. Overall this pattern of results suggests a robust nonsocial bias in ASD that may reflect enhanced motivation to view nonsocial images, even in a context where doing so is more difficult.