Ratings of Facial Attractiveness by High-Functioning Individuals with Autism

Background: Although differences in face perception abilities have been widely studied by autism researchers, there is limited research on whether individuals with autism perceive facial attractiveness as typically developing individuals do. White, Hill, Winston, and Frith (2008) found that adults with Asperger Syndrome were less accurate than matched control adults at judging facial attractiveness, especially if the stimuli were the same sex as the participant. Yet researchers have not considered whether difficulty judging facial attractiveness extends to individuals diagnosed with autism.                                                                                                                                                                                                             

Objectives: The aim was to determine if there are differences in the perception of facial attractiveness in high-functioning individuals with autism versus typically developing individuals.                                                                                                                                                                                                               

Methods: Participants with autism were high-functioning adults (18 – 45 years).  Participants without autism were matched to the autism group on age, gender, verbal, performance, and full scale IQ.  Sixty color face stimuli were presented individually on a screen with equal numbers of male and female faces.  Participants were asked to rate the attractiveness of each face on a 7 point scale. Face stimuli were previously rated by an independent group of raters (i.e., a class of undergraduate students). Stimuli were selected based on the class ratings so as to present an equal range of attractive, average, and unattractive faces.    

Results: Analyses revealed that mean attractiveness ratings by the autism group and control group were correlated with the class ratings of attractiveness. However, results indicated that whereas the control group was highly correlated with the class, the autism group was only moderately correlated with the class.  Additional analyses of the highest and lowest rated faces indicated that the moderate correlation between the class and the autism group was driven only by their agreement for the unattractive faces.  There was much less agreement for average and attractive faces between the autism group and the class ratings compared to the control group.                                                                                               

Conclusions: Attractiveness ratings by the control group were highly correlated with the ratings by the class reflecting agreement for perception of attractive versus average versus unattractive faces.  In contrast, attractiveness ratings by the autism group were only moderately correlated with the ratings by the class largely due to agreement for judging unattractive faces.  Unlike the control group, the autism group did not appear to distinguish attractive from average faces.  Given that perception of facial attractiveness emerges early in development as evidenced by typically developing 2-month-old infants’ preference to look at attractive versus unattractive faces (Langlois, et al. 1987), it is remarkable that adults with autism have even some difficulty judging attractiveness.  Prior research has demonstrated that perception of facial attractiveness is driven primarily by facial feature typicality (e.g., Rubenstein and Langlois, 2002).  Faces with average or typical features are considered attractive, and faces with atypical or distinctive features are considered unattractive.  The current results suggest that, perhaps, individuals with autism never abstract a representation of how faces vary with respect to typicality, which could explain why they easily judge faces as unattractive, but have more difficulty judging faces as attractive or average.