American scientists have revealed plans to create a mobile application that will detect if a person is suffering from depression based on changes in their voice tone. According to researchers at the University of Maryland, speech patterns change when a person becomes depressed. Individuals with depression struggle to think quickly, and their speaking rate slows down with longer pauses compared to those who are not depressed.
This voice detection app, which utilizes deep learning—a type of machine learning based on artificial neural networks—could help identify such traits, which are often subtle. Mental health professionals may recommend the app to their patients, who would send video and audio updates about their mood from home, which the app will then evaluate. This would help both patients and those around them stay informed about changes that may threaten their mental health and lives.
The project is led by Carol Espey Wilson, a professor of electrical engineering and computer science at the University of Maryland. During the 180th meeting of the Acoustical Society of America, scheduled for June 8-10, Wilson will discuss how a person's mental health condition is reflected in their speech gesture patterns. The unnamed app, still in early planning stages, is expected to be suggested by therapists to help monitor patients between therapy sessions.
Professor Espey Wilson stated, "Ideally, therapists will give the app to patients who are suffering from major depressive disorder when they are stable or only experiencing mild depression, allowing them to potentially use the app regularly to track their mental health status. This will alert the appropriate people if the app detects a significant increase in depression severity."
Users will log into the app on their smartphones, which will then ask some basic questions about how they felt physically and emotionally in the past week. Users will answer verbally, and their speech will be recorded. The smartphone can also capture video of the user while they talk, if they choose to turn on the camera.
If concerning changes are detected, the relevant parties, such as the therapist, the patient, and the healthcare provider, will be notified, according to the Daily Mail.
