Using Speech Acoustics to Reveal Motor Disruptions in Psychosis

Project: Research project

Project Details


The goal of this project is to investigate the feasibility of using speech acoustics as a clinical biomarker in individuals at clinical high risk (CHR) for developing psychosis. There is evidence that disruptions to cortico-cerebellar circuits in individuals experiencing attenuated psychosis symptoms impact motor control of the face and limbs. This proposal would be the first study to examine whether these motor disruptions in high-risk populations also affect the complex motor control required for speech. In Aim 1 an instrumental approach will be used to investigate the acoustic correlates of psychosis risk. Specifically, speech data will be collected to investigate fine-grained acoustic properties of vowels and consonants in simple repetition tasks as well as during more naturalistic conversational speech. The speech of CHR young adults will be compared to age-matched healthy controls to discover if there are group differences in the speech acoustics that allow us to classify speech samples into healthy and clinical groups. To enable fast, reliable analysis, machine learning-based algorithms will be used to measure the acoustics speech properties of interest. In Aim 2, the speech properties measured in Aim 1 will be compared to other behavioral measures, in order to discover if they correlate with several measures of cerebellar dysfunction (posture control, procedural learning, and motor timing) that are known to occur in CHR individuals. These measures will provide convergent validity for these novel speech measures. Cognitive capabilities which are related and unrelated to speech and motor control will also be assessed, to provide specificity and divergent validity to these measures. In Aim 3, the links between speech features and changes in symptom severity will be assessed at two time points, connecting changes in speech motor control to longitudinal changes in the progression of the symptoms over 12 months. These investigations may reveal speech as a novel and easily-collected biomarker enabling early detection of psychosis risk.
Effective start/end date3/21/1912/31/21


  • National Institute of Mental Health (5R21MH119677-02)


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.