Adini Y, Bonneh YS. Levanon Y, Dean-Pardo O, Lossos L.
Background: Children with autism spectrum disorder who can speak often show abnormal voice quality and speech prosody. The underlying abnormal mechanisms are currently unknown and it is yet unclear if they stem from a high-level deficit in communication or alternatively related to basic speech mechanisms such as involved in controlling its spectral content.
Objectives: compare and analyze the speech spectra of young autistic children and normal controls and investigate their differences.
Methods: we recorded 82 children (41 autistic, 41 controls) ages 4 to 6 years (mean 5) while naming a sequence of daily life pictures pointed by the experimenter for 60 sec in a quite room in their preschools. We computed the power spectra of the speech recordings, averaged across time and normalized for each child in the range of 40 to 2000 Hz.
Results: The group averages of the spectra of autistic children and controls differed significantly in three spectral regions, around 300 Hz (N>A), 400 Hz (A>N) and 1100 Hz (A>N; p<0.006 in all cases). In addition, the autistic spectra were more uniform and less fluctuating, which we quantified by spectral analysis of the speech spectra curves for each child. We found that high frequency component of this analysis was higher in the controls and could predict if the child is autistic by ~85% correct using a simple threshold. No difference was found between boys and girls but the high functioning autistics differed from the other autistics.
Conclusion: The speech spectrum of young autistic children differs from that of normal controls. The more uniform spectrum we observed is in agreement with the often observed monotonic or machine-like speech in autism. Since the development of non-uniform speech spectrum is likely to involve auditory feedback, this could imply abnormal interaction between speech reception and production in autism. Further work is needed to develop the speech spectrum as a tool for early diagnosis.