Emotion Recognition Through Nonverbal Channels In Children with Autism Spectrum Disorder

Background: Sensitivity to nonverbal cues is crucial for successful social and communicative functioning. It is estimated that more than 60% of a message is expressed nonverbally (Philipott, 1983), suggesting listeners must rely on nonverbal cues, including facial expression, body posture and gesture, and vocal prosody to understand an intended message (Lieberman & Rosenthal, 2001). Although there is substantial evidence that individuals with ASD have atypical face processing abilities, we know comparatively little about the processing of other nonverbal cues in ASD.   Thus, the understanding of emotions through these less studied nonverbal cues warrants further investigation. 

Objectives: The current study had two objectives: 1) to assess group differences in accuracy of detecting emotional states through the nonverbal channels of gestures and vocal prosody, and 2) to evaluate the relationship between nonverbal emotion perception accuracy and cognitive ability.

Methods: Participants included 15 children with high-functioning ASD and 15 typically developing controls, matched on chronological age (M=10 years, 0 months; SD=1 year, 2 months) and IQ (M= 107, SD= 18), as measured by the WASI (Wechsler, 1999). Participants completed two emotion recognition tasks that required them to make a forced judgment about the emotional state (happy, sad, angry, or afraid) depicted in each stimulus. Task 1 used affective prosody using the Child Paralanguage subtest of the DANVA-2 (Nowicki & Duke, 1994). Stimuli were 24 audio clips of an actor saying the sentence: “I’m going out of the room now, but I’ll be back later,” with varying intonation to depict one of the four emotions. Task 2 used affective gestures using stimuli from Pohlig (2007). Stimuli were 24, 3-second silent video clips of an actor depicting one of the four emotions.  Facial features were concealed so gesture and body posture were the only information provided. 

Results: Group differences were examined utilizing a 2 (diagnosis; ASD or TD) x 2 (channel; gesture or prosody) analysis of variance (ANOVA). An overall effect of diagnosis was obtained, F(1,28)= 9.08, p = .005.  Follow up analyses indicated that children with ASD were impaired on both gesture, t(28)= 2.01, p=.05, and prosody, t(28)= 2.08, p<.05.  Regression analyses indicated that nonverbal ability (matrix reasoning) was related to understanding of gesture (r= .86, p= .001) and prosody (r= .43, p= .11) for children with typical development.  However, nonverbal reasoning was not related to understanding of gesture (r= -.39, p= .15) or prosody (r= -.06, p= .83) for children with ASD.

Conclusions: This study provides evidence that emotion understanding difficulties in ASD extend beyond understanding of faces to include understanding of gestures and vocal prosody.  Additionally, the results of this study suggest that children with ASD may not recruit nonverbal reasoning skills to help them recognize gestures and vocal prosody. These results suggest the importance of teaching children with ASD to understand both gestures and prosody when targeting emotion recognition during social skills and language therapies.