Societal expectations for social-emotional behavior differ across sexes; however, diagnostic definitions of autism do not account for this when delineating "typical" versus "atypical." This study examines sex differences in autism in one behavior associated with strong gender biases: smiling. Computer vision was used to quantify smiling in 60 autistic (20 female) and 67 neurotypical (25 female) youth during conversations. Effects of sex and diagnosis were examined on degree of smiling, smile prototypicality, changes in smiling, and impact of smiling on interaction quality. Sex differences in smiling persisted across diagnosis groups: females smiled more than males, and their smiles were more prototypical. Autistic youth smiled less, and less prototypically, than neurotypical youth, with no sex by diagnosis interactions. In autism, the association between smile activity and interaction quality approached statistical significance, seemingly driven by autistic males but not females. Findings are consistent with population trends for females to smile more during social exchanges and "display rules" requiring more positive expressivity from females. Autism has historically been defined based on differences between autistic and neurotypical males. Failure to acknowledge sex-based differences in social-emotional behavior may leave some females appearing to have fewer autistic traits, increasing their risk of being under-identified and misunderstood.
Keywords: autism; computer vision; facial expression; sex differences; smiling.
This study uses automated computerized methods to measure facial expression, namely smiling, in autistic and neurotypical males and females as they converse naturally with an unfamiliar social partner. Results show that typical population-level sex differences in smiling also exist in autistic youth; females smile more and more prototypically than males. Failure to acknowledge these sex-based differences in social-emotional norms and behavior may leave autistic females at increased risk for being misdiagnosed and misunderstood.