Vocal output characteristics of speech have previously been identified as possible cues to the assessment of suicide risk, and there is evidence that certain vocal parameters may be used as suicidal discriminators. The acoustic properties of male speech samples comprised of individuals carrying diagnoses of depression, suicide risk, and remission were analyzed and statistically compared.
The male sample contained 10 high-risk suicidal patients, 13 depressed patients, and 9 remitted patients. Acoustic analyses of voiced power distribution were performed on speech samples extracted from audio recordings collected from the patients during clinical interviews. Features derived from the power spectral densities were found to be powerful discriminators of class membership in both studies of interview and reading passage sessions. The results support theories that identify psychomotor disturbances as central elements in depression and suicidality.
Cite as: Yingthawornsuk, T., Keskinpala, H.K., France, D., Wilkes, D.M., Shiavi, R.G., Salomon, R.M. (2006) Objective estimation of suicidal risk using vocal output characteristics. Proc. Interspeech 2006, paper 1321-Mon3CaP.4, doi: 10.21437/Interspeech.2006-231
@inproceedings{yingthawornsuk06_interspeech,
author={T. Yingthawornsuk and H. Kaymaz Keskinpala and D. France and D. M. Wilkes and R. G. Shiavi and R. M. Salomon},
title={{Objective estimation of suicidal risk using vocal output characteristics}},
year=2006,
booktitle={Proc. Interspeech 2006},
pages={paper 1321-Mon3CaP.4},
doi={10.21437/Interspeech.2006-231}
}