A new AI-based tool can detect signs of moderate to severe depression by analyzing speech patterns. This technology could help improve depression screening in primary care.
A study published in The Annals of Family Medicine tested the tool using over 14,000 voice samples from adults in the U.S. and Canada. Participants answered the question, “How was your day?” in free-form speech for at least 25 seconds.
The AI analyzed speech features like pauses, hesitations, and tone. It then compared the results to the Patient Health Questionnaire-9 (PHQ-9), a standard depression test. A PHQ-9 score of 10 or higher indicated moderate to severe depression.
The AI tool gave three possible results:
- Signs of Depression Detected
- No Signs of Depression Detected
- Further Evaluation Recommended (for uncertain cases)
The model was trained with 10,442 voice samples and tested on 4,456 samples. It correctly identified depression in 71% of cases and ruled it out in 74% of non-depressed individuals.
These findings suggest that AI could support doctors in diagnosing depression, making screening easier and more accessible.
Source: medicalxpress
