Audiovisual Synchrony Predicts Level of Visual Fixation On Mouth and Eyes in 2-Year-Olds with Autism

Background:  Previous studies found that 2-year-olds with autism spectrum disorders (ASD) looked less at the eyes and more at the mouths of approaching adults, as compared with both typically-developing (TD) and with non-autistic, developmentally-delayed (DD) controls. A parallel study showed that toddlers with ASD failed to give preferential attention to point-light displays of human biological motion, again in contrast with TD and DD controls.  Instead, viewing by the ASD group was predicted by level of audiovisual synchrony (AVS): increased fixation was given to stimuli in which change in movement was accompanied by synchronous change in sound. In control children, viewing was unrelated to AVS. These results raised the hypothesis that, in toddlers with ASD, increased fixation on the mouth and decreased fixation on eyes may be due to increased attention to the synchrony of lip movements and speech sounds.

Objectives: The primary goal of this study is to determine whether the reduced visual fixation on eyes and increased fixation on mouths in toddlers with ASD is due to an imbalance in preferential attention to social versus physical contingencies.  Secondarily, if this relationship exists, this study aims to elucidate its developmental progression.

Methods: Physical contingencies were identified by quantifying AVS in naturalistic, child-directed caregiver videos.  AVS was defined as simultaneous change in motion and change in sound amplitude, measured within localized regions of interest (ROIs) comprising eyes, mouth, body, and object areas.  Our hypothesis was that for toddlers with ASD, but not for typically-developing children, visual fixation to each ROI would be positively correlated with level of AVS.

Results: We found that, in toddlers with ASD, visual fixation to faces is strongly correlated with level of AVS.  In the ASD group, level of fixation on the mouth across different movie clips was positively predicted by level of AVS (r² = 0.40, p < 0.01).  Surprisingly, we found that fixation on the eye region in children with ASD was also highly correlated with AVS (r² = 0.47, p < 0.01).  Visual fixation in typically-developing children was not significantly correlated with AVS in either facial region (p > 0.05).  In response to our secondary aim, preliminary results suggest that sensitivity to AVS also varies developmentally, with a peak in sensitivity at 24 months.

Conclusions: These results suggest that toddlers with ASD look at different locations of the face as a function of the physical contingencies embedded therein: they vary their attention to eyes and mouth in direct relation to synchronous change in motion and sound.  In contrast, typically-developing toddlers do not appear to vary their fixation patterns on the basis of these physical contingencies; instead, other studies (Edwards et al, IMFAR 2009) suggest that TD toddlers vary their fixation patterns in relation to social cues, such as the level of intensity of facial affect.  This contrast implies that children with autism take a markedly different avenue for learning about the actions of others than their typically-developing peers, and may be a target for new diagnostic and therapeutic approaches in the future.