Dissociating Content-Influenced Changes From Maturational Changes in Oculomotor Function in Infants with Autism Spectrum Disorders

Background: Research regarding vision in autism spectrum disorders (ASD) has produced two separate bodies of literature related to oculomotor function—one using traditional saccade-eliciting paradigms (e.g. gap/overlap task), another using natural viewing paradigms. Together, they suggest that basic mechanisms of oculomotor function appear to be largely intact in autism, but that content- and context-dependent attentional biases offer evidence of distinct differences between individuals with ASD and their typically-developing (TD) peers. Previous research has provided converging evidence that individuals with ASD attend preferentially to non-social aspects of the environment. However, it remains unclear when, developmentally, this bias emerges, and whether it is due to motivational factors that guide attention or to developmental changes in response to physical factors that capture attention. The present study will chart basic oculomotor response while viewing naturalistic as well as abstract stimuli during the first two years of life, in order to shed light on how both exogenous and endogenous attentional systems affect visual behaviors in infants who develop ASD.

Objectives: The current study will chart the longitudinal development of visual fixation responses and saccadic eye movements during natural viewing of social scenes and during prosaccades to peripheral targets, and will compare these changes in infants who develop ASD and their TD peers.  

Methods: Fixation and saccades were identified from data in a longitudinal study using eye-tracking equipment to examine the viewing patterns of naturalistic scenes in infants at high and low risk for developing ASD. Eye-tracking data were collected at 2, 3, 4, 5, 6, 9, 12, 15, 18, and 24 months while infants viewed videos of actresses engaging in child-directed caregiving behaviors, toddlers interacting with each other in playground settings, and geometric animations. Diagnoses were given at 36 months, assigning infants into ASD (n=13) and TD groups (n=51). The following properties and content of eye movements were analyzed cross-sectionally and longitudinally, and then compared between groups: fixation duration; frequencies of saccade and fixations; saccade latency and accuracy; relationship between saccade amplitude and duration; and relationship between saccade velocity and amplitude.  

Results: Preliminary analyses suggest that while basic properties of saccades and fixations undergo developmental change, they do not differ between infants with ASD and their TD peers. Moreover, there are clear indicators that saccadic properties for both groups are influenced by content, demonstrating the emergence of endogenous control of saccades in early infancy. However, these groups differed in when and where they looked. Infants with ASD did not show difficulty disengaging but were more likely to saccade between non-social aspects of the scene.  

Conclusions: Basic oculomotor circuitry appears to develop normally in individuals with ASD. Properties of eye movements reflected task-specific differences. This suggests that discrepancies in viewing patterns between toddlers with ASD and their TD peers are not the result of oculomotor impairments, but rather reflect differences in what aspects of a social scene are most salient to them. Our data provide converging evidence pointing to top-down rather than low-level visual factors influencing dynamic visual engagement between these two groups.