An Eye-Tracking Study of Visual Attention to Human and Nonhuman Animals,Landscapes and Abstract Patterns in Young Children with Autism Spectrum Disorder

Background: Studies of visual attention deployment using eye-tracking methods have suggested various atypicalities in the gaze behavior of individuals with autism spectrum disorder (ASD), in particular in response to social stimuli, but empirical evidence that visual preferences reflect the social profile characteristic of the disorder remains controversial (Klin et al., 2002; van der Geest et al., 2002; Boraston & Blakemore, 2007)

. Research on the gaze behavior of human infants has demonstrated a preference for attractive human faces even in newborns, indicative of a possibly innate biasing mechanism toward social engagement, although recent findings have suggested that this preference extends beyond conspecifics in 3- 4 month olds (Quinn, Kelly, Lee, Pascalis, & Slater, 2008). Currently little is known about the early development of visual preferences in children with ASD, who show deficits in face processing and in social reciprocity that may be linked to abnormalities in social attention (e.g. limited attentional bias for faces and avoidance of eye-contact).  

Objectives:  The goal of this study was to investigate whether and how children with ASD differ from typically-developing (TD) children in their visual preferences for several categories of visual stimuli, besides human faces and objects, which have been the focus of prior research. Specifically, we examined whether children with ASD show content-related visual attention deployment differences compared to TD children of the same mental age, and whether attentional biases or visual preferences reflect the early social phenotypes of the ASD group.

Methods:  Twenty-seven children with ASD (ages 3-6 years) and 35 TD children in the same age range passively viewed a slide-show of 50 image-pairs presented on a TOBII 1750 eye-tracker. Images of people (adults, children, babies), paired with images of nonhuman animals, flowers, landscapes, or abstract designs were displayed side-by-side for 6 seconds, separated by a 1-second central fixation stimulus. The image pairs were matched for complexity, mean luminance and balanced (left-right) for content categories.

Results: Eye-tracking data provided evidence of group differences on several measures of content-related eye-movements. An ANOVA comparing groups on total fixation times to the 6 content categories (people, animals, landscapes, flowers, butterflies and abstract images) revealed a significant category effect  (F(5,58) = 32.3, p < .001) and a significant group by category interaction (F(5,68) = 6.42, p = .011). TD children showed a preference for images of people over all other image-categories, whereas children with ASD showed no difference in total fixation durations between nonhuman animals, humans and abstract patterns. 

Conclusions:

Differences between the TD and ASD children in patterns of spontaneous viewing behavior and visual preferences found in this study suggest that preferential attention deployment may account for aspects of the ASD social phenotype. Findings will be discussed in terms of the significance of investigating gaze behavior as an index of social attention and preferences in children with ASD.