Fixation Durations During Static Scene Viewing in 6-8-Month-Old High-Risk Infants Relates to ADOS Scores At 36 Months

Background: Research into the early development of attention in Autism Spectrum Disorders (ASD) has mainly used experimental paradigms such as the gap-overlap task. However, it how remains poorly understood how differences in reaction time latencies on these tasks relate to altered behaviour in more naturalistic contexts, where previous research has suggested that contrasts between ASD and typical development may be most acute (e.g. Speer et al., 2007).

Here, we report a series of analyses that looked at spontaneous fixational eye movement behaviours during the unconstrained viewing of a static scene. These analyses offer a method for bridging the gap between experimental and more naturalistic assessments of attention. 

Objectives: To investigate spontaneous fixational eye movement behaviours in the early development of ASD.

Methods: We measured spontaneous fixational eye movement patterns during viewing of a static scene consisting of a mixture of faces and non-social objects in 6-month-old infants, 45 of whom were at high familial risk of developing ASD (HR) and 47 at low familial risk (LR) (cohort 1). Viewing data was recorded at 50/60Hz using a Tobii 1750/T120 eyetracker. We administered an identical test to an independent cohort of 40 HR and 43 LR infants aged 8 months (cohort 2). Cohort 1 was followed up for outcome characterisation at 36 months, using the Autism Diagnostic Observation Schedule (ADOS).

Analysis of fixational eye movement behaviours was performed using previously published Matlab scripts (Wass et al., 2012). Fixations were identified as periods in which the eye remains static (defined as a period of >100ms with no change in reported position of gaze >35 degrees per second during that period). 

Results: For cohort 1, we found that HR infants showed significantly shorter fixation durations than LR infants, even when possible confounds such as number of usable fixations and eyetracker data quality were accounted for. An identical analysis conducted on data from cohort 2 found the same group difference. Between-group differences in fixational eye movements were found to be consistent across fixations targeted at social and at non-social objects. 

We found that shorter fixation durations at 6 months related to higher Autism Diagnostic Observation Scores scores at 36 months, after confounding variables were controlled for. 

Comparison of results from the gap-overlap task showed a significant correlation between shorter fixation duration and faster saccadic orienting times in the baseline (non-competition) condition. 

Conclusions: These results are consistent with the results from Talk 2, in which faster reaction times under non-competition conditions were identified in HR vs LR infants. The findings suggest that abnormalities can be identified in spontaneous attentional orienting behaviours in ASD during the first year of life. 

One possible explanation for these findings is increased vigilance, modulated by short-term changes in arousal (cf e.g. Aston-Jones et al., 1981; de Barbaro et al., 2012).