Differential Scanning of Core Facial Features In 12- and 18-Month-Old High Risk Infants

Background:  A hallmark feature of autism spectrum disorder (ASD) is impairment in social processing.  To investigate these difficulties empirically, numerous studies have examined behavioral measures of face processing in ASD individuals and found differences in scanning patterns that include decreased time scanning the core features of the face, particularly the eyes and mouth.  Recently this finding was extended to ASD children as young as 2-years-old.  Prospective work with infant siblings of children with ASD has begun testing whether differences in face processing might be seen early on in development, asking both whether this could be an early risk marker for infants who later develop ASD, and whether differences found in high-risk infants could be part of the broader autism phenotype.

Objectives:  With prior work in ASD children and adults pointing to decreased scanning of core facial features, the present study aimed to ask whether these differences in face processing will be found earlier in development and how they might change over time in a group of infants at risk for ASD.

Methods: A Tobii eye-tracker was used to monitor eye gaze while infants were presented with side-by-side images of their mother’s face and a stranger’s face displaying a neutral expression for 10s.  Twenty-eight infants were tested at both 12- and 18-months-of-age: 19 infants at high-risk for ASD (HRA; by virtue of having at least one older sibling with ASD) and 9 low-risk infants (LRC).  Eye-tracking data captured duration of looking to regions of interest, including the full images of mother and stranger and the eye and mouth regions within each image.  Analyses examined a) duration of looking to core features of the face (sum of eyes and mouth) and b) percentage of time on core features as a function of total time on the image (sum of eyes and mouth divided by total time on image).

Results: A 2 (identity; mom, stranger) x 2 (age; 12-months, 18-months) x 2 (group: HRA, LRC) mixed model ANOVA examined duration of time on core features and found a main effect of identity, F(1, 26) = 4.78, p = .04, with infants showing more time on the eyes and mouth in their mother than the stranger.  LRC also spent more time on the core features overall (M = 1950ms) as compared to HRA (M = 1497ms), but this was only a trend (p = .16).  A second analysis examined the percentage of time on core features and found a main effect of group, F(1, 26) = 5.84, p = .02:  HRA infants spent significantly less time scanning the eyes and mouth when looking at these images (M = 52%) as compared to LRC infants (M = 67%).  No other main effects or interactions were significant.  

Conclusions:  The present work finds that differences in face scanning that are found in ASD children and adults are also seen in at-risk infants during the second year of life.  Future work with HRA infants will examine individual trajectories of face processing in relation to ASD outcome.