Diminished Attention to Faces In 6-Month Old Infants Later Diagnosed with ASD

Background: Though several studies have shown that toddlers and young children with ASD scan and recognize faces atypically (e.g. see Chawarska & Shic, 2009), the phenotypic precursors to these atypical behaviors in infancy have not been well mapped. 

Objectives: To examine the relationship between developmental changes in face scanning and outcome in a prospective longitudinal cohort of high-risk infant siblings of children with ASD.  

Methods: Infants were presented at 6, 9, and 12 months with stimuli of adult female faces in three conditions: (1) Static: an image of a face; (2) Affect: a video of an actress smiling; (3) Speech: a video of an actress reciting a nursery rhyme.  Stimuli were preceded by a central fixation before being presented for 20s; gaze patterns were recorded with an eye-tracker.  At 24 months, high-risk siblings (N=62) were classified by a team of expert clinicians based on Mullen, ADOS, Vineland, and CSBS results into the categories: ASD (N=14), Broader Autism Phenotype (BAP; N=17), other specific Developmental Delays (DD; N=13), and No Concerns / Not Affected (NA; N=18).  A longitudinal comparison group of typically developing (TD; N=46) infants was also recruited. 

Gaze patterns were analyzed in a top down approach.  Dependent measures were: (1) ValidTime% - proportion of time during the trial the eye-tracker recorded the eye; (2) InnerTime% - proportion of time spent examining the “inner face” (i.e. excluding hair, neck, and body) relative to looking at the stimulus; (3) EyeTime% and (4) MouthTime% - proportion of time spent examining the eyes and mouth, relative to looking at the inner face; and (5) EyeMouthRatio% - Eye/(Eye+Mouth).  Statistical analyses were conducted using a top-down linear mixed models refinement approach (Diggle, 2002) with quadratic fixed and random age effects.

Results: At 6 months, infants later classified as ASD showed lower ValidTime% across conditions than BAP (p<.01), DD (p=.084), NA (p<.05), and TD (p=.05) groups.  However, by 12 months the ASD group showed higher ValidTime% than both BAP (p<.05) and TD (p<.05) groups.  No significant differences between ASD and other groups were found at 6, 9, or 12 months for InnerTime%, EyeTime%, MouthTime%, or EyeMouthRatio%.  All groups exhibited sensitivity to experimental conditions with ValidTime%: Static<(Affect=Speech); InnerTime%: Static<Affect<Speech; EyeTime%: Static>Affect>Speech; MouthTime%: Static<Affect<Speech; and EyeMouthRatio%: Static>Affect>Speech.  Profiles of non-ASD groups varied, but NA and TD groups exhibited marked similarities.

Conclusions:   Results suggest that at 6 months infants who will later express marked ASD symptomology exhibit an overall pattern of decreased attention towards face stimuli shown in multiple contexts.  This deficit lessens by 9 months.  Internal patterns of face scanning are remarkably similar to other high-risk and low-risk groups, consistent with prior studies (Young, et al., 2009).  Nevertheless, decreased attention towards faces at 6 months could suggest vulnerabilities in early social processing and result in substantial developmental consequences.  As our study is ongoing, future reports will address issues of experimental power, specificity of attentional deficits, profiles of performance in non-ASD groups, and include other methodological developments.