Longitudinal Development of Social Visual Engagement in Infants Later Diagnosed with ASD

Background: From birth, typically-developing infants preferentially attend to the social signals, such as eye gaze, of their caregivers (Haith et al., 1977). Recent findings from our laboratory revealed that infants later diagnosed with autism spectrum disorder (ASD) exhibit decline in eye fixation from 2 until 24 months of age and that the decline is already underway within the first 6 months after birth. In contrast, typically-developing infants show an increase in eye fixation from 2 to 6 months (Jones & Klin, 2013). These findings represent the earliest known indicators of social disability in infancy. If confirmed in larger samples, these findings have the potential to inform efforts at early identification and intervention prior to the emergence of overt symptoms of social disability in ASD. 

Objectives: Measure growth charts of social visual engagement from 2-24 months in TD infants and infants later diagnosed with ASD.  

Methods: 106 infants were enrolled as risk-based cohorts: N = 40 at low risk (22 males) and N = 66 at high risk for ASD (41 males). Risk status was based on having either a full biological sibling with ASD (high) or on not having ASD among 1^st, 2^nd, or 3^rd degree relatives (low). Diagnostic status was ascertained at 36 months. Of the HR sample, at outcome 13 received a diagnosis of ASD (9 male); 13 showed symptoms of the Broader Autism Phenotype (BAP, 11 male), and 40 (20 male) were confirmed non-ASD. Infants were shown scenes of naturalistic caregiver interaction as in Jones & Klin (2013). Eye-tracking data were collected at 10 time points (months 2, 3, 4, 5, 6, 9, 12, 15, 18, and 24). Longitudinal looking profiles for this cohort were compared with those of the earlier sample (2013) for eyes, mouth, body, and object regions within and between diagnostic group categories. Cohorts 1 and 2 were also combined for larger sample analyses. 

Results: Analyses of the ASD replication cohort by functional ANOVA are consistent with earlier results: longitudinal looking profiles for eyes, mouth, body, and object were not significantly different between cohorts 1 and 2 (all F< 2.4, P > 0.14) (Fig.1). Analyses of the TD replication cohort showed no significant differences from cohort 1 in eye or object fixation (all P > 0.5), but did differ from cohort 1 in mouth (F = 4.95 P = 0.032) and body (F = 5.55, P = 0.024) fixation. Across outcome groups, typically-developing children show increasing eye fixation; non-ASD siblings show increasing eye fixation; siblings with subthreshold (BAP) symptoms show neither increasing nor decreasing eye fixation, and infants later diagnosed with ASD show declining eye fixation. Similar spectrum effects are observed for body fixation (Fig. 2). 

Conclusions: Our results replicate earlier findings showing that infants later diagnosed with ASD show decline in eye fixation from 2-to-24 months. Spectrum effects emerging in the first 6 months give early indication of social disability in infants later diagnosed ASD. Future analyses will include longitudinal looking profiles for female infants with TD (35) and ASD (7) outcomes.