News Releases & Research Results Development of an advanced artificial intelligence technology to find depression using brain networks - Advance toward the clinical application of an effective brain network marker across multiple imaging sites -

News Releases & Research Results


The results of research and development conducted by Researcher Ayumu Yamashita of the Brain Information Communication Research Laboratory Group, Advanced Telecommunications Research Institute International (ATR), Hiroshima University, The University of Tokyo, Showa University, Kyoto University, Yamaguchi University, and RIKEN.

The key results of R&D are as follows:

  • An artificial intelligence technology was utilized to develop the world's first brain network marker for depression, effective across multiple imaging sites, based on functional magnetic resonance imaging (fMRI) data.
  • Specifically, the harmonization method developed last year was utilized to integrate fMRI data acquired at multiple sites as homogeneous large-scale data (from a total of 1,584 cases) with site differences removed. Then, a machine learning technique, an artificial intelligence technology, was applied to develop a brain network marker for depression, in order to identify depressed patients among healthy individuals based on their brain networks. The brain network marker allowed effective identification of depressed patients among healthy individuals at an approximately 70% accuracy even with completely independent data collected at different sites, demonstrating its effectiveness across multiple sites.
  • An application for approval was submitted for the depression brain network marker in 2021, aiming for approval by the end of 2022.

This project was conducted with the support of the Strategic Research Program for Brain Sciences by AMED.

The results were published in the American scientific journal PLOS Biology on December 8.


Yamashita A., et al. Generalizable brain network markers of major depressive disorder across multiple imaging sites PLOS Biology
DOI: 10.1371/journal.pbio.3000966


Last updated 12/08/20