DAM: Clinical Mental Health Labels and Model Scores for 35,000 Users

Offering clinical mental health labels and audio-based model scores for 35,000 individuals, totaling 863 hours of speech data. Created by KintsugiHealth in 2026, it includes demographic metadata for validation and test sets used in model development.

Use Cases

• Validating mental health screening algorithms using clinical labels
• Auditing model fairness across demographic metadata categories
• Correlating audio-based model scores with clinical mental health outcomes

Strengths

• Includes clinical mental health labels for validation and test sets
• Derived from a large-scale collection of 863 hours of speech
• Contains demographic metadata for 35,000 individuals
• Licensed under Apache 2.0 for open use

Limitations

• Excludes raw audio files, providing only model scores and metadata
• Geographically restricted to the US region
• Does not include the full training set records

Provenance

Source

KintsugiHealth

Collection Method

Collected via phone, tablet, or web app

Freshness

Last updated February 2026.

Geography

United States