Predicting psychosis before it occurs | The University of Tokyo

Machine-learning tool classifies MRI Brain n scans into thousands are healthy and those at risk of psychosis onset.

In training, the tool was 85% accurate at classifying the re Research hile in the final test using new data it was 73% accurate.

The tool could be helpful in future clinical settings, as while most people who experience psychosis make a full recovery. Koike the University of Tokyo ghtText__NxlGi">Brain over 2,000 _highLightTex 21 _NxlGi">JapanAbout half lass="summaryFeed_highLightText__NxlGi">73% 85% 85% ghtText__NxlGi">two four 1,02973% fourth three 1,165 three Koike Koike 15 21 first clinicians first ="summaryFeed_highLightText__NxlGi">the University of Tokyo the Graduate School of Arts and Sciences most only 30% ed_highLightText__NxlGi">Shinsuke Koike 70%