Relationship of Atypical Prosodic Features to Acoustic Startle Measures in Children with Autism Spectrum Disorders and Those with Typical Development

Background:   People with autism spectrum disorders (ASD) has long been known to have atypical prosodic features in their speech. Yet their prosodic difference from typically developing (TD) people as well as its physiological background, has not been well-understood. Recently, using an intonational phonological framework, we found that ASD participants can produce lexically and syntactically intact prosody, on the other hand, ASD children produced many more various atypical prosodic and non-prosodic features in spontaneous speech than TD children, which appeared to be associated with the core symptoms of ASD in social, communicative interaction difficulty.  As the ability to produce prosody is acquired during infancy, investigation of the physiological aspect of prosodic atypicality in ASD might increase understanding of the biological mechanisms underlying the development of ASD characterisitics.

Objectives:  The objectives of this study is to investigate the relationship between intonational phonological framework of prosody and physiological indexes of acoustic startle response (ASR), which is a promising endophenotype of developmental and psychiatric disorders.

Methods:  Participants were 11 children diagnosed with high functioning ASD, and 12 TD children. Spontaneous speech of participants was recorded while each child talked with an experimenter about common topics for 5-10 minutes. Three trained phoneticians listened to the speech, and marked the sections that were judged “atypical.” The atypical sections were then annotated using the X-JToBI scheme, and, were classified into subtypes of prosody [utterance too long, too much emphasis (stress, pitch), sudden speech rate changes, pitch range too narrow, inappropriate boundary pitch movement (pitch contour, type), and, others]. As for acoustic startle measurements, mean ASR magnitudes to acoustic stimuli presented at 65 dB to 105 dB in increments of 10 dB were analyzed. Average peak startle latency (PSL) and startle modulations (habituation and prepulse inhibitiion) was also examined.

Results:  Compared to TD control, children with ASD had more atypical prosodic section for prosody subtype of “utterance too long” and “too much emphasis”. And, the total number of atypical prosodic sections was also significantly larger in ASD compared to TD. As for ASR measures, ASD children had greater ASR magnitude to small stimuli intensity of 75 and 85 dB, and, more prolonged PSL compared to TD children.

The total number of atypical prosodic sections and the number of atypical prosodic section for prosody subtype of “too much emphasis” and “others” significantly correlated to startle measures of PSL as well as ASR magnitude to large stimuli intensity of 95 and 105 dB.

Significant correlations to PSL were also found in the number of atypical prosodic section for prosody subtype of “utterance too long” and “sudden speech rate changes.”

Conclusions:  Our results suggest that various atypical prosody subtype of ASD in spontaneous speech is associated in part with the mechanism of ASR latency delay. Considering the fact that PSL in TD is usually 60-80 msec, atypical neural mechanism to acoustic stimuli around this latency window might contribute to the development of prosody production as well as the core symptoms of ASD in social, communicative interaction difficulty.