Developmental Trajectory of Selective Attention to a Talking Face in Infants at-Risk for ASD

Background: Infants’ gaze behavior during face-to-face speech communication changes through the first year of life. Specifically, infants’ attention to a talking face shifts from the eyes to the mouth between 4 and 8 months of age. Mouth-gazing behavior after 8 months old is reported to predict later language development, possibly because this mouth gazing may reflect the ability of canonical babbling, which is crucial for speech production. However, such gaze behavior toward a talking face has not been thoroughly studied for infants at-risk for autism spectrum disorders (ASD). Their atypical language development may be reflected by atypical gaze patterns, which may be an early marker of ASD.

Objectives: The present study longitudinally examined infants’ attention to a talking face among 6- to 18-month-olds to reveal differences in attention patterns between typically developing infants (TD) and infants at-risk for ASD (RA). Furthermore, because selective attention to a talking face may differ depending on the language, the present study also aims to reveal its development in the case of a Japanese-speaking environment.

Methods: Participants (25 RA and 20 TD) were followed up to record their eye gaze at 6, 9, 12, and 18 months of age. However, all the infants were not able to complete or participate in all the sessions number of participants differ depending on age. Each infant’s eye movement while viewing a video of a woman who was speaking to the infant (20 seconds x 2), was recorded using the Tobii X120 Eye Tracker system. We created three areas of interest (AOI) on the female face–around the eyes, mouth, and entire face–and calculated the proportion of looking time (total looking time at the eyes or mouth / total looking time at the face) using an arcsine transformation. We also performed developmental testing and administered a language questionnaire at each recording point.

Results: The RA showed a different developmental pattern of eye gaze behavior compared to the TD. Firstly, the Japanese TD showed a similar developmental shift for mouth watching as in American infants; namely, their attention shifted to the mouth from the eyes at 9 months old. However, among the RA, such a developmental shift did not occur until 18 months old. There was a considerable difference at this age between the TD and RA. While the RA shifted their attention to the mouth from the eyes, the selective attention of the TD was directed toward the eyes at 18 months old. These results were statistically confirmed by a linear mixed-effects model with group, age, and the interaction of these two factors as effects and participants as random effects.

Conclusions: Although the TD and RA showed a similar eye gaze pattern at 6 months old with selective attention to the eyes, their developmental trajectories were different. The RA’s delayed attentional shift (at about 9 months old) to the mouth from the eyes while watching a talking face may be chiefly related to their slower language acquisition, particularly delayed speech production as assumed from their language profiles.