Absence of Lateralization of Visual Scanning to Varying Facial Affect in Toddlers with Autism

Background: The ability to derive socially relevant information from faces is fundamental to interpersonal communication and reciprocal social interaction.  Studies in typically-developing individuals have demonstrated that emotional facial affect is expressed on the face in a laterally asymmetric manner: using chimeric faces, in which an image of the face is bisected vertically so that its left or right half is mirrored about the midline to create a new image, left side composites are rated as more emotionally intense than right side composites.  These findings, however, have not been extended to the case of dynamic facial stimuli, nor have they been extended to test the effect of this lateralization on the way that viewers look at faces.  We hypothesized that visual scanning of faces exhibiting dynamic and varying facial affect would reflect these lateral asymmetries in expression, with the central tendency of fixation locations shifting toward the side of the face showing greatest emotional expression.

Objectives: This study examines lateralization of visual scanning in typically developing toddlers and toddlers with autism during viewing of dynamic faces displaying a range of naturally-occurring intensity and valence of affect.

Methods: Two-year-olds with autism spectrum disorders (age = 2.28 (0.58) years; gender = 11M, 4F; n = 15) and control children matched on age-, verbal-, and nonverbal function (age = 2.03 (0.68) years; gender = 24M, 12F; n = 36) watched video scenes of female actors playing the role of caregiver, while eye-tracking data were collected.  Dynamic facial expressions in the caregiver videos were quantitatively ranked by adult observers, naïve to the aims of the study, for intensity and valence of affect using a modified analytic hierarchy process of paired comparisons.  Then, using the ratings of affect as a regressor, we examined lateralization of visual fixation data in relation to varying degree of naturally-occurring facial affect.

Results: Preliminary results suggest that both emotional intensity and valence of dynamic faces alter visual fixation patterns and lateralization of gaze in two-year-old, typically-developing children. In examples of positive and high intensity social affect, typically-developing two-year-olds show increased visual scanning of the female actors’ left hemifaces. In two-year-olds with autism, however, facial expressions have little impact on visual scanning, and no lateralization trends are apparent.  Across varying social affect, toddlers with autism looked less at the eyes of others, while looking more at others’ mouths.

Conclusions: Differential attention to faces, particularly in conditions of changing affect, is critical for extracting information about the intentionality of others.  Failure to do so suggests an altered path for learning about the surrounding world, with potentially profound impact on subsequent social development.  Over the course of development, failing to reallocate visual resources in a manner that is contingent with changing facial affect is likely to exacerbate increasingly atypical neural specialization, altering the formation of the social mind and brain.