Patterns of Face Gaze Among Infants at Risk for ASD

Background: The core deficits of Autism Spectrum Disorder (ASD) can reliably be observed and classified by two years (Lord et al., 2006; Chawarska et al., 2009). Recent research has aimed to understand whether atypical behaviors that characterize ASD in toddlerhood (e.g., limited eye contact, decreased interest in faces) emerge in the first year of life. Behavioral observations of high-risk (HR) siblings later diagnosed with ASD suggest that face gaze may be preserved early in infancy, at 6 months of age, but declines significantly by 12 months (as reviewed in Jones et al., 2014). However, as previous studies have either exclusively evaluated mother-infant interactions or have examined face gaze during a general evaluation (e.g., Mullen), this is the first study to evaluate gaze patterns in HR-infants using a standardized social interaction procedure between an examiner and infant.

Objectives: To investigate patterns of face gaze directed at an experimenter using a novel interactive procedure at 6, 9, and 12 months among HR siblings who later develop ASD (HR-ASD), HR siblings who do not receive an ASD diagnosis (HR-nASD), and typically developing low-risk children (LR-TD).

Methods: Participants included 184 infants (HR-ASD, n=21; HR-nASD, n=104; LR-TD, n=59) who completed Social Orienting Probes at 6, 9, and 12 months and were evaluated for ASD at 24 and 36 months. Trained examiners administered a series of 1-minute probes designed to elicit attention through several means: a) motherese, b) song, and c) peek-a-boo. Videotaped sessions were coded offline by blinded coders for the duration of time, in seconds, that the infant looked at the examiner’s face. Face gaze was calculated as the proportion of time spent looking at the examiner’s face during each probe (%Face). Generalized linear mixed model analyses were used to examine the fixed effects of diagnostic group, probe, infant age, and their interactions on %Face, followed by planned contrasts using the sequential Bonferroni correction.

Results: Analyses of %Face revealed a main effect of diagnosis (F(2, 1395)=4.4, p<.05; HR-ASD<HR-nASD=LR-TD), a main effect of probe (F(2, 1395)=245.3, p<.01; peek-a-boo>song>motherese), no main effect of infant age (p=.58), a significant group x age interaction (F(4, 1395)=2.7, p<.05), and no three-way interaction (p=.11). Planned comparisons indicated that groups differed significantly only at 6 months across all probes, such that HR-ASD<HR-nASD=LR-TD (p=.001; see Figure 1).

Conclusions: Results of these analyses diverge from previous literature by revealing differences in infants’ face gaze at 6 months, but not at 9 or 12, between HR-infants who are later diagnosed with ASD, unaffected HR-siblings, and LR-TD children during a standardized behavioral paradigm. The finding that ASD infants spend less time orienting to the face of an interactive partner contributes to evidence from neuroimaging (Wolff et. al., 2012), neurobehavioral (Chawarska et al., 2013; Shic et al., 2014), and neurophysiological (Elsabbagh et al., 2012) data that subtle abnormalities related to later emerging ASD can be detected as early as 6 months. This suggests a potential sensitive period early in infancy for the development of social cognition that may be highly influenced by social orienting and attention.