Facial Emotion Recognition and Gender Categorization Abilities As Predictors of Social Functioning in Adults with High-Functioning Autism

Background:  Individuals with autism display deficits in various face processing skills, including facial emotion recognition (Rump et al., 2009) and facial gender categorization (Newell et al., 2010).  Face processing abilities have important implications for the development of fundamental social interaction skills (Tonks et al., 2007).  However, no known studies to date have tested whether these face processing abilities relate to social behavior among adults with high-functioning autism (HFA).  

Objectives:  The present study examined the extent to which facial emotion recognition and gender categorization abilities relate to social behavior in adults with HFA.  It was expected that better performance in facial emotion recognition and gender categorization tasks would be associated with lower social impairment as reflected in scores on the Social Responsiveness Scale (SRS). 

Methods:  Adults with HFA (n=26; M_age=23.8Y) completed a facial emotion recognition task that consisted of ambiguous morphed faces along neutral to target emotion continua.  The target emotions were anger, fear, sadness and disgust.  In addition, participants (n=28; M_age=23.6Y) completed a facial gender categorization task of faces that ranged in their typicality of gender (e.g., more or less masculine looking men).  All participants were administered the Autism Diagnostic Observation Scale (ADOS), the Wechsler Abbreviated Scale of Intelligence (WASI; Wechsler, 1999) and the Social Responsiveness Scale (SRS). 

Results:  Results indicate that emotion recognition accuracy for ambiguous faces was negatively associated with the overall SRS score (r=-.49, p=.01).  Interestingly, the SRS subscale of social motivation had the highest correlation with facial emotion recognition performance (r=-.54, p<.01), indicating that as social motivation deficit scores increased, facial emotion recognition accuracy decreased.  With regard to gender categorization, as adults with HFA’s accuracy scores for more typical gendered faces increased, overall SRS scores decreased (r=-.46, p=.01).  Additionally, participants who had higher autism mannerisms SRS subscale scores tended to be less accurate at gender categorization for more typical faces (r=-.46, p=.02) but not for emotion recognition (r=-.31, p>.05).  

Conclusions:  Both facial emotion recognition ability and facial gender categorization ability were associated with social behavior among adults with HFA.  Specifically, higher accuracy at these face processing tasks was related to lower social impairment scores.  This supports the applicability of face processing skills to actual social functioning and potentially the development of fundamental social interaction skills.  Surprisingly, higher social motivation was related to better facial emotion recognition performance.  Although the causal direction of this relationship cannot be ascertained from the present study, the role of social motivation on socially salient cognitive skills was an interesting finding that requires further investigation.  In addition, the association between autism mannerisms and poorer facial gender categorization ability suggests that prototypical face processing deficits are indicative of autism symptomatology.