A03 Acoustic Analysis

High-definition mental health care using rich information embedded in sound

Whispers, sighs, sobs, screams, falsettos, speechlessness, calls, murmurs, and more…
As we have a rich variety of vocabularies, the voices we emit can also inform the small details of the movements in our hearts.
Experienced psychotherapists can freely take advantage of these voices or sounds as a valuable intervention method. They can also notice about the patients’ mental states even including the slightest fluctuation in emotion.
We believe that if we could use digital technologies to similarly take advantage of sound and apply them to various scenarios, we would be able to understand one another better and support one another better.
Masaya Ito
National Center of Neurology and Psychiatry
National Center for Cognitive Behavior Therapy and Research
Manager, Department of Research and Development
Link to members

Our Project

The AI technology behind computer performance, sensor technologies, and data communication is evolving rapidly, and processing and analysis technologies for handling big acoustic data are being developed. For instance, AI can now detect abnormal noise from factories, highways, and tunnels.

It can also process noises made by livestock to detect early signs of infectious diseases. These AI technology-based services are becoming more available day by day. Even in the human communication field, some call centers have implemented emotion recognition technologies.
Given these acoustic analysis technologies, we thought that we might be able to study sound itself as psychotherapy research. Therefore, our project uses the recordings from the clinical trials of psychotherapy, previously carried out. And we aim to use this AI technology to extract the acoustic features to identify mental states and predict psychotherapy outcomes.