Poster #45 - Computer Vision Based Quantification of Reduced Interpersonal Facial Affect Synchrony in Autism Spectrum Disorder

• In Event: 3-106 - Poster Session 11
  In Poster Session: PS 11 Section - Developmental Disabilities

Integrative Statement

Background: Humans nonconsciously synchronize behaviors with one another as they interact. Research suggests that this interpersonal synchrony is a socially adaptive process with developmental and evolutionary significance (Feldman, 2007; Lakin et al., 2003). One behavioral domain in which interpersonal synchrony is well-established is facial affect; seeing facial expressions causes matching responses within the observer’s own face, which may facilitate emotional processing and reciprocity (Winkielman et al., 2009). Given social-emotional reciprocity deficits in autism spectrum disorder (ASD), we examined whether children with ASD are less likely to achieve facial affect synchrony with social partners, and whether facial affect synchrony is associated with other aspects of social functioning known to be affected in ASD.

Objectives: Examine the degrees to which children with and without ASD engage in facial affect synchrony with familiar and unfamiliar partners; measure the association between facial affect synchrony and standardized measures of typical and atypical social functioning.

Methods: Participants were 20 well-characterized children with ASD (ages 9-16) and 17 age-, gender- and IQ-matched typically developing children (TDC). Children participated in cooperative conversation tasks with their mothers (familiar condition) and research assistants (unfamiliar condition). Each person’s face was recorded and analyzed frame-by-frame for the presence and intensity of positive affect (i.e., smiles) using an open-source automated facial analysis toolkit (OpenFace; Baltrusaitis et al., 2018). Synchrony was calculated across each conversation using windowed cross-correlation (Boker et al., 2002). Correlation coefficients at positive time lags represented the adult responding to the child, and coefficients at negative time lags represented the child responding to the adult. Children’s mothers also completed standardized measures of autism symptomatology, adaptive social skills, and empathy.

Results: Positive affect synchrony was analyzed with mixed design ANOVA, separately by synchrony direction. In the “adult responding to child” direction, there was a significant main effect of diagnosis, F(1,32)=5.98, p=.02, ηp2=.16; adults synchronized less to children with ASD than to TDC. Neither the within-subjects familiarity effect nor the interaction term was significant. In the “child responding to adult” direction, there was a significant diagnosis by familiarity interaction, F(1,32)=5.94, p=.02, ηp2=.16. Children with ASD showed significantly less synchrony than TDC in the unfamiliar condition, t(32)=-3.60, p=.001, d=1.23, whereas groups did not differ in the familiar condition. The degree to which children synchronized with unfamiliar partners correlated with their scores on measures of autism symptomatology, r(32)=-.63, p<.001, adaptive social skills, r(32)=.74, p<.001, and empathy, r(32)=.45, p<.01.

Conclusions: Children with ASD exhibit reduced interpersonal positive affect synchrony, particularly in response to unfamiliar partners. Thus, they appear less likely to engage in a social process that is thought to be automatic and critical from early in life. Furthermore, both familiar and unfamiliar adults demonstrate reduced synchrony in response to children with ASD, suggesting that the impact of ASD on synchrony is bidirectional. Significant associations between facial synchrony and measures of everyday social functioning implicate this as an important social process, worthy of further study in typically- and atypically-developing children. Disruptions in this process may be a mechanism underlying core social-emotional reciprocity deficits in ASD.

Authors

• Casey J. Zampella, Children's Hospital of Philadelphia
  Presenting Author

• Loisa Bennetto, University of Rochester
  Non-Presenting Author

• John D. Herrington, Children's Hospital of Philadelphia
  Non-Presenting Author