Disclosing Critical Voice Features for Discriminating between Depression and Insomnia-A Preliminary Study for...

Area of Science:

• Psychiatry
• Speech Science
• Biomedical Engineering

Background:

• Depression and insomnia are closely linked, sharing symptoms and causal relationships.
• Current diagnostic methods lack practical biological markers to effectively differentiate between depression and insomnia.
• Distinct treatment approaches are required, highlighting the need for accurate discrimination.

Purpose of the Study:

• To identify critical vocal features for discriminating between depression and insomnia.
• To explore the potential of voice analysis as a diagnostic tool for these conditions.
• To establish objective biomarkers for differentiating mental health states.

Main Methods:

• Recruited four patient groups: severe depression, moderate depression, insomnia, and chronic pain disorder (CPD).
• Extracted 384 voice features using the openSMILE software from recorded speech.
• Applied analysis of variance (ANOVA) to identify significant differences in voice features across patient groups.

Main Results:

• Significant relationships were found between patient status and specific voice features.
• Distinct voice feature patterns emerged for severe depression, moderate depression, insomnia, and CPD patients.
• Identified critical voice features with potential for diagnostic discrimination.

Conclusions:

• Voice analysis shows promise for developing models to discriminate between depression and insomnia.
• Further research with larger patient cohorts is needed to validate these findings.
• Future studies aim to develop quantitative methods for diagnosis based on vocal biomarkers.