Early Vocal Development in Infants At Risk of Autism: Prosody and Social Interaction

Background: Prosodic deficits in autism, including monotonic or exaggerated patterns of intonation, have been recognized since the earliest descriptions of the syndrome, but the onset, development, and origin of atypical intonation patterns have never been properly determined. In early infancy, the fundamental frequency contour of the infant's voice, although initially reflexive, is known to be shaped by prenatal experience of the language environment, loses those properties during the first year of life as volitional control is gained, then converges gradually on native language patterns again in response to engagement with the social world. Conversely, mothers shape their intonation patterns to attract infant attention, specifically by exaggerating changes in fundamental frequency, and infants are sensitive to such changes in intonation. Thus, the development of intonational exchanges throughout childhood may provide a useful index of social attunement between infant and caregiver and its potential derailment in autism.

Objectives: The goal of this study is to track the development of intonation in the first two years of life in infants at risk of autism, concurrently with the development of vocal interaction between infant and caregiver, in order to detect the earliest manifestation of any disruption of prosody, and test the hypothesis that the disruption of expressive prosody in infants with ASD is associated with an earlier derailment of mechanisms of social engagement.

Methods: As part of an ongoing study, we recruited 4 low-risk infants, with no family history of autism, and 4 high-risk infants, with older siblings already diagnosed with autism. Using a miniature digital audio recording device (LENA Foundation) sent out to families in the mail and worn by each child all day, we made day-long audio recordings of each child's language environment at monthly intervals from 2 months onwards. Using automatic speech recognition technology, we extracted and labeled sequences of utterances containing interactions between infant and caregiver, and calculated the fundamental frequency contour, utterance duration, and relative timing between utterances for all of our labeled segments. Using Functional Data Analysis to time-align all fundamental frequency contour shapes for infant and caregiver at each monthly time point, we were able to quantify developmental changes in intonation and also the timing statistics describing vocal interactions.

Results: Consistent with previous studies, all of our infants showed evidence of a development in intonation and vocal turn-taking over the first 12 months, with higher mean F0 and greater F0 variability in our high-risk sample, as well as a reduction in F0 modulation relative to controls. In at least one of our high-risk infants, this was preceded by reduced vocal interaction.

Conclusions: Preliminary results suggest that prosodic deficits in ASD may begin within the first year of life, preceded by atypical patterns of social vocal engagement.