Interactions Between Gaze Following and Disengagement At 7 and 13 Months in Infants At High Risk for Autism Spectrum Disorder

Background: Emerging findings from high-risk studies suggest that infants who develop autism spectrum disorder (ASD) show early impairments in the processing of both social and non-social stimuli (Bedford et al., 2012; Zwaigenbaum et al., 2005). Although ASD is defined by social-communication impairments and restricted and repetitive behaviours, the majority of cognitive theories of ASD posit a single underlying factor, which over development has knock-on effects across domains. In order to examine the developmental interactions between measures of social and non-social attention, we looked at the relationship between gaze following and visual disengagement across the first year of life.

Objectives: In this study concurrent and longitudinal links between gaze following behaviours and disengagement were examined in low-risk controls and high-risk infants at two time points, 7 and 13 months. We aimed to test 1) whether these abilities are independent and remain separate from early on in development; or 2) whether they become separate over time.

Methods: Participants were 54 infants at high risk for ASD and 50 low-risk controls recruited through the British Autism Study of Infant Siblings (BASIS). In the gaze following task, Tobii eye-tracking was used to record gaze behaviour. Infants were shown 12 short videos of a female model turning to look at one of two objects. Infants’ proportion of congruent versus incongruent first looks and their looking time to the congruent object were calculated. In the gap-overlap task a central stimulus was presented to gain the infants’ attention followed by the onset of a peripheral stimulus either at the same time as the central one disappeared (baseline) or while the central stimulus was still on the screen (overlap). Data were video coded and disengagement was calculated as reaction time in overlap – baseline trials.

Results: An autoregressive cross-lagged model was run on the whole sample and provided a good fit to the data. At 7 months there was a significant relationship between disengagement and both measures of gaze following behaviour, first look and looking time. Faster disengagement in the gap-overlap task predicted increased looking to the congruent object, but a higher proportion of incongruent (versus congruent) first looks. No significant longitudinal relationships between the measures were found from 7 to 13 months nor was there a concurrent relationship between the measures at 13 months. The majority of children who later developed ASD at 3 years showed impairments in either gaze following or disengagement at 13 months, with only a small proportion of children having difficulties with both abilities.

Conclusions: This study finds that attentional disengagement is related to gaze following behaviours early on in development, at 7 months, with a possible speed/accuracy trade-off, in which faster disengagement gives the infant more time to look at the target object, but leading to an increase in orienting errors. By 13 months, the age at which these measures relate independently to ASD outcome, they have become de-correlated , perhaps suggesting that the neural regions subserving these behaviours follow different patterns of specialisation across development (Johnson, 2001).