Early Social Cognitive Development and Behavioural Signs of Autism in Very Preterm Infants

Background:  Children born preterm or at low birth weight are at an elevated risk for autism spectrum disorder (ASD)(Johnson et al., 2010). Toddler parent report measures (e.g. M-CHAT) have revealed very high positive screening rates amongst very preterm (VP) infants (approximately 25%), and direct measures of early symptomology (e.g. Autism Observation Scale for Infants or AOSI) have indicated scores in a similar range to infant siblings later diagnosed with autism. However, given the frequency of cognitive, motor, language and sensory impairment in VP populations, it is difficult to infer whether these elevated scores truly reflect an early autism phenotype. This study seeks to investigate whether early behavioural signs of autism in VP infants (using the AOSI) are associated with other social cognitive features of the broader autism phenotype (e.g. differential neural responses to direct gaze) and/or developmental delay.

Objectives:  To evaluate group differences between very preterm and full term infants in early symptomology related to ASD (AOSI) at 6 and 12 months gestationally corrected age (GCA) and infants' neural responses (event-related potentials [ERP]) to direct versus averted gaze at 6m. Also, to examine associations between AOSI score, gaze ERP and infants’ general cognitive, motor and language development (Bayley III) at 12m.

Methods:  Early behavioural signs of ASD were measured in VP infants (25-31 weeks GCA) and full term controls (37-42 weeks) with no family history of ASD at both 6 and 12 months using the AOSI assessment. Total scores and total ‘marker’ counts (i.e., items scored non-zero) were computed for each child. At 6m, infants’ ERPs were recorded in response to viewing faces with eye gaze directed toward versus away from the infant. Analyses included various relevant components in the ERP (P1, N290 and P400). Developmental functioning (cognitive, motor and expressive/receptive language) was measured at 12 months for both preterm and full-term infants using the Bayley III assessment.

Results:  Preliminary results from 15 full term and 9 VP infants indicate that relative to full term controls, VP infants show more early behavioural signs (higher Total Scores and more Markers) of ASD on the AOSI. Higher AOSI scores amongst VP infants are associated with lower developmental functioning scores on the Bayley III, but not with abnormal evoked responses to dynamic gaze shifts. 

Conclusions:  Within the first year, preterm infants show more behavioural signs associated with ASD on the AOSI, demonstrating scores in a similar range to high-risk siblings later diagnosed with ASD. However, higher AOSI scores in VP infants may reflect more non-specific behaviours associated with developmental delay, rather than a true early autism phenotype. Further follow-up will be important to distinguish whether behavioural abnormalities predict later ASD symptoms/diagnoses and/or non-ASD related impairments in this VP cohort.