Looking and Language in Autism Spectrum Disorder: A Family Study

Background: Converging evidence suggests that individuals with autism spectrum disorder (ASD) display atypical visual attention patterns that predict greater social impairment (Rice et al., 2012; Sasson et al., 2008; Klin et al., 2012). Unaffected relatives also demonstrate subtle differences in visual processing (Adolphs et al., 2008, Dalton et al., 2007). This study examines whether visual attention patterns relate to social language differences that characterize ASD, and which have been observed to express among unaffected relatives as well, though in milder form (e.g., Losh et al., 2008). Specifically, we examined visual attention during narration in ASD and first-degree relatives to investigate whether differences in perceptual strategies might relate to social communicative differences in ASD and subtler differences in social communication style in parents (i.e, the broad autism phenotype or BAP). Examining links between looking and speaking may inform how visual attention shapes the content and quality of social language.  

Objectives: To examine relationships between visual attention and social language for individuals with ASD and their parents during two narrative tasks.

Methods: Thirty-four high functioning individuals with ASD (IQ>80) and 24 age-matched controls, as well as 147 parents of individuals with ASD and 61 control parents completed two narrative tasks presented on an eye-tracker: 1) a 24-page wordless picture book (PB), and 2) a series of six static images of emotionally ambiguous scenes drawn from the Thematic Apperception Test (TAT). During the PB task, participants narrated each page as it was presented on the eye-tracker. For the TAT task, individuals viewed six images from the TAT for eight seconds each. After each image was removed, participants were asked to tell a story about the image. Visual attention for narrative tasks was quantified as the proportion of fixations to different aspects of the images (e.g., social, background elements, white-space); Latent Semantic Analysis (LSA), a computational linguistic tool, measured narrative quality. Additionally, parents completed BAP assessments.

Results: Individuals with ASD demonstrated decreased attention to background elements during their narration of both tasks (p < .05). Whereas the ASD parent group did not differ from controls on the PB task, during the TAT BAP(+) parents looked significantly more at faces and significantly less at background elements of the scenes (ps < .05). Greater attention to non-social stimuli was positively correlated with semantic quality across narrative tasks in the ASD group and controls (rs  > .4), as well as in parents (rs > .3).

Conclusions: This study is the first to assess the relationship between visual attention and narrative production in individuals with ASD and their parents. Results suggest that both individuals with ASD and their parents demonstrate different patterns of visual attention that may underlie differences in complex language use. In particular, reduced attention to the background of the scenes may result in narratives less rich in semantic content. Overall, this study provides further evidence of atypical visual processing as an index to genetic liability in ASD that may help explain the roots of differences in social language in ASD and the BAP.