When Are Adolescents with Autism Spectrum Disorder Better At Emotion Recognition Than Their Peers?

Background:  There is mixed evidence as to whether there is a deficit in recognising emotional expressions in Autism Spectrum Disorder (ASD). Variation occurs not only in the findings from different research groups but across different types of experimental stimuli. Emotion recognition in ASD has been investigated across different media - still photographs, moving video footage and voice recordings. There is also evidence to suggest that when there is a deficit in recognising emotion in human faces by people with ASD, this deficit does not extend to recognising emotion in cartoon representations of emotion. In this study we compare emotion recognition in ASD in all three types of emotion expression media (photo, video, voice) across both human stimuli (such as a photo of a human face) and animated stimuli (such as a cartoon face).

Objectives:  To identify emotion recognition performance in ASD upon photo, video, and voice stimuli across both human and animated representations.

Methods:  Participants were 37 adolescents (age 11-14) with a diagnosis of ASD (33 male, 4 female). 42 males and 39 females served as typically developing, age-matched controls. The first factor, ‘Representation’, had two levels: human and animated. The second factor was Media: photograph, video and voice. There were therefore six categories of stimuli (2 Representations x 3 Media), each of which contained six stimuli, totalling 36 stimuli. All participants were exposed via a computer display to all stimuli (in a randomised order) and had to identify the emotion being expressed in the stimuli. First we examined in Representation impacted upon emotion recognition, then the effect of each Media in isolation.

Results:  For all analyses there were no significant differences between male and female control groups. For the Representation factor, there were no significant differences between ASD and control groups for the animated stimuli. For the human stimuli, however, the control groups, but not the ASD group, demonstrated significantly enhanced emotion recognition when compared to the animated stimuli. For the photo media in isolation, there was a significant interaction: Controls significantly outperformed the ASD group on human faces and the ASD group significantly outperformed controls on the animated faces. For the video media in isolation, controls were significantly more accurate than the ASD group for both types of representation. For the voice media in isolation, there were no group differences in recognising animated sounds, but controls had significantly enhanced recognition of emotion in human voice, whereas the ASD group did not.

Conclusions:  For photographic and voice media, the emotion recognition deficit in ASD was only evidenced for human, not animated, stimuli. The ASD group did not demonstrate an advantage for recognising emotions within human faces and voices, whereas the control groups did. The ASD demonstrated an advantage over controls for recognising emotion in animated (cartoon) still images. This advantage did not extend to video media and may reflect different strategies being employed to recognise human and animated representations of emotion.