PRODUCTION OF EMOTIONAL PROSODY AND FACIAL EXPRESSIONS IN ADOLESCENTS WITH AUTISM

Background:   Studies of affective prosody production in individuals with autism spectrum disorder (ASD) have described subjective and objective differences in pitch and volume, indicating both more monotonous productions and greater expressive range in non-monotonous utterances (Shriberg et al. 2001, Hubbard & Trauner, 2007).  Macdonald et al. (1989) found similar differences in productions of affect across voice and face.    

Objectives:   To capture and analyze productions of affective vocal prosody and facial expressions among adolescents with ASD.

Methods:   Adolescents with ASD and typically developing (TD) controls (age 10-19) were asked to retell four brief, highly animated first-person stories containing happy, angry, surprised and fearful emotions presented on video tape by an adolescent actor.  We digitally recorded prosody of 17 participants with ASD and 13 TD controls and facial expressions from 15 ASD and 14 TD adolescents.  Voice information was analyzed using PRAAT software.  Three independent coders blind to diagnosis and target emotion coded facial expressions for emotion, expressiveness (flat, mild, moderate, extreme), naturalness (natural, slightly awkward, very awkward, unnatural) and eye gaze (on, away) on a second-by-second basis. 
Results:   We found slower, higher, softer voices and more monotonous productions in the ASD group.  In utterances showing noticeable emotion, vocal pitch and intensity ranges were significantly larger in the ASD group.  The ASD participants’ facial expressions were judged as expressive as their TD peers', but significantly more awkward, and with eye gaze more focussed on the camera.  Both groups displayed significantly more affect in their voices than on their faces. 

Conclusions:   Adolescents with ASD display qualitative and quantitative differences from their typical peers in vocal expressive range and naturalness of face during affective narrative productions.   Both groups have significantly higher percentages of animated vocal performances than animated facial expressions, indicating a higher threshold for affective modulation of face than voice.