Related Experiment Video
Updated: Sep 11, 2025
05:48
Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception
Published on: August 9, 2024
1.6K
A Machine Learning-Based Case-Control Study on Suicide Risk Identification: Integrating Acoustic and Linguistic
Qunxing Lin1, Jianqiang Zhang1, Weijie Wang1
1Digital Mental Health and Risk Identification and Control Lab, Department of Psychology, School of Public Health, Southern Medical University, Guangzhou, China.
Depression and Anxiety
|August 18, 2025
Summary
Speech analysis shows promise for identifying suicide risk in patients with major depressive disorder (MDD) or bipolar disorder (BD). Acoustic and linguistic features, especially from negative emotional speech, achieved 77.82% accuracy in assessing risk.
Area of Science:
- Psychiatry
- Computational Linguistics
- Speech Analysis
Background:
- Suicide is a major global health concern.
- Current suicide risk assessment relies on clinical judgment and scales, which are difficult to implement.
- Vocal and linguistic features are increasingly explored for suicide risk identification.
Purpose of the Study:
- To investigate the efficacy of speech-based methods for assessing suicide risk.
- To analyze acoustic and linguistic features in patients with major depressive disorder (MDD) or bipolar disorder (BD).
- To evaluate the impact of emotional valence and stress on speech-based risk assessment.
Main Methods:
- Two-phase study involving 90 patients with MDD or BD.
- Phase 1: Used question-answer materials with positive, neutral, and negative emotional valences.
- Phase 2: Incorporated stress factors into speech data collection.
Main Results:
- A model combining acoustic and word frequency features from negative emotional valence materials achieved 77.82% accuracy.
- Speech data collected under stress provided better insights into participants' psychological states and suicide risk.
- The study demonstrated the potential of speech analysis in suicide risk assessment.
Conclusions:
- Speech analysis, particularly using acoustic and linguistic features from emotionally negative and stressed speech, shows significant potential for suicide risk assessment.
- These findings support the development of objective, scalable tools for suicide prevention.
- Further research is needed to validate and broaden the application of these speech-based methods.
