Adults with and without Autism Differ In Their Emotional Responses to Non-Social Images Related to Circumscribed Interests

Background: Both preschool and school-aged children with autism spectrum disorders (ASDs) exhibit restrictive and perseverative patterns of visual attention to objects related to circumscribed interests (CI; e.g., trains; Sasson et al, 2008; Sasson et al 2010). Similarly, recent fMRI findings demonstrate that adults with ASD, but not matched controls, are characterized by hypoactivation in neural reward circuits in response to monetary incentives but hyperactivation to images related to CI (Dichter et al, in press), a finding that reveals a potential neurobiological mechanism of restricted interests in ASD. While taken together these studies suggest that objects related to CI are disproportionately rewarding for ASD children and adults, a need remains for better understanding the subjective experience underlying heightened salience and reward value of CI in ASD.

Objectives:  To determine whether adults with ASD differ from Typically Developing (TD) adults in their emotional ratings of CI, non-CI, and social images.

Methods: A series of 40 “High Autism Interest” (HAI), 34 “Low Autism Interest” (LAI) and 40 social images were developed based upon previous studies of CI (South et al, 2005, Sasson et al 2008). 214 TD Adults and 51 adults with ASD provided arousal and valence ratings (Lang, 1995) for each image using 9-point Likert scales.

Results:  A mixed model ANOVA with stimulus type (HAI vs. LAI vs. social) and emotion rating (arousal vs valence) as the within group factors and group (TD vs ASD) as the between group factor produced a significant three way interaction between stimulus, rating and group (F (2, 262) = 14.59, p < .01, h_p²=.10). Follow-up tests revealed that this interaction was driven by similar arousal ratings by both groups across the three stimulus types, but group differences in valence ratings: ASD adults provided higher valence ratings for HAI objects than TD adults, while the reverse pattern was found for social images and LAI objects. Next, correspondence analysis was used to examine the relationship between ratings levels (e.g., middle, high, low) by each group on all stimuli. For the ASD group, “very positive” ratings were assigned almost exclusively to specific subcategories of HAI objects (e.g., video game and electronic equipment), while the TD group more evenly assigned high valence ratings across the three stimulus categories.

Conclusions: The present study reports differences in subjective emotional ratings of CI, non-CI and social images in ASD. While the groups did not differ in how emotionally arousing they found the three image categories, the ASD group rated CI images as more pleasant and other object types and social images as less pleasant than TD individuals. These findings provide additional evidence that certain categories of non-social objects are preferred in ASD and provide support for the development and dissemination of a standardized set of non-social images related to CI in autism that are suitable for studies of incentive motivation salience in ASD.