Neural and perceptual processing of emotional speech prosody in typically developed children and in children with autism spectrum disorder

Human voice conveys information about a speaker’s emotions and intention via speech prosody; changes in the speaker’s intonation, stress, rhythm, and tone of voice. The ability to recognize and interpret speech prosody is essential for successful social communication and has been associated with social competence throughout childhood. Autism spectrum disorders (ASD) are characterized by impaired social communication and social interaction skills as well as repetitive patterns of behavior. Semantic-pragmatic language deficits, including deficient speech prosody production and comprehension, are often found to be impaired in individuals with ASD. The present thesis investigated processing of words and speech prosody in school-aged typically developed children and two groups of children with ASD: those with accompanying language impairments (children with ASD [LI]) and those with no accompanying language impairments (children with ASD [no LI]). To this end, auditory event-related potentials (ERPs) were recorded for a Finnish word uttered with different emotional connotations (neutral, scornful, sad, and commanding). Two of the thesis studies included a behavioral prosody discrimination task, and in one study, facial electromyographic (facial EMG) reactions were recorded for the above-mentioned speech stimuli. In typically developed children, changes in speech prosody elicited mismatch negativity (MMN) / late discriminative negativity (LDN) responses, demonstrating that the auditory system of school-aged children automatically detects prosodic changes in speech. Further, these prosodic changes in speech activated an involuntary attention mechanism in typically developed children, as reflected by a P3a component. However, no reliable facial EMG reactions were found for these non-attended prosodic changes in speech. Both groups of children with ASD had diminished ERP responses to words, suggesting that initial stages of sound encoding were deficient in children with ASD, but these processes were more impaired in children with ASD (LI) than in children with ASD (no LI). In both groups of children with ASD, MMN/LDN responses to the scornful stimulus were diminished, suggesting abnormal auditory discrimination mechanisms in children with ASD. In addition, P3a responses were diminished and atypically distributed in children with ASD, suggesting that these children have difficulties in orienting to speech sound changes. Finally, children with ASD (no LI) were slower in behaviorally discriminating prosodic changes in speech compared to the typically developed children. Taken together, these and the ERP results show that processing of natural speech prosody is impaired in children with ASD at various information processing levels, including aberrant discrimination of and orienting to, as well as sluggish responding to prosodic changes. These speech processing deficits might contribute to the observed difficulties in comprehending another person’s emotional state based on his/her tone of voice in individuals with ASD.