Affective prosody production in autistic and typically developed adult males

Abstract

This study examined differences in production of affective prosody in adult males with autism spectrum disorder (ASD) and typically developed (TD) controls. Previous studies of children with ASD have reported increased variability in fundamental frequency (f0) in spontaneous and semi-spontaneous speech compared to TD children. A controlled set of expressive speech recordings was collected from 30 talkers (15 ASD) to measure differences between the two groups using the same lexical content. Isolated vowels, vowel-consonant-vowel (VCV) syllables, words and short phrases were elicited in five emotion contexts: angry, happy, interested, sad, and neutral. The recordings were obtained using evoked and portrayed elicitation techniques: talkers were asked to recall past emotional episodes (evoked) and role-play scripted scenarios (portrayed) specific to each emotion context. Consistent with previous work and extending the findings to adults producing the same lexical content, talkers with ASD showed increased f0 variability in each emotion context except for neutral. In addition, systematic group differences were found in acoustic properties other than f0 used to convey affective prosody—including harmonics-to-noise ratio and intensity—which were higher in each emotion context for talkers with ASD compared to TD talkers.