Differentiation between bipolar versus unipolar depression by AI‐based multimodal behavioral analysis
Differenzierung zwischen bipolarer und unipolarer Depression durch KI-basierte multimodale Verhaltensanalyse
Psychiatry and Clinical Neurosciences · 2026-11 · 80 (11) · 925-934
DOI: 10.1111/pcn.70124 ↗Auf einen Blick
Bei Erwachsenen in Japan unterschied ein KI-Modell bipolare von unipolarer Depression mit 78 % Genauigkeit. Die AUC betrug 0,76.
Fragestellung
Kann eine KI-gestützte multimodale Verhaltensanalyse bipolare von unipolarer Depression unterscheiden?
Methode
Rekrutierung Erwachsener mit bipolarer Störung oder Major Depression in Japan; Interviewaufnahmen; Diagnosen per Structured Clinical Interview for DSM-5-Research Version; KI-Analyse von Gesicht, Blick und Stimme.
Ergebnisse
379 Personen beendeten die Interviews, 366 wurden analysiert. Diagnoseunterscheidung mit vierfacher, dreimal wiederholter Cross-Validation: Accuracy 78 %, AUC 0,76.
Schlussfolgerung der Autor:innen
Die Integration multimodaler Verhaltensindikatoren mit KI-Modellierung ermöglicht eine klinisch bedeutsame Differenzierung zwischen bipolarer und unipolarer Depression.
Kernaussagen
- KI-basierte Analyse unterschied bipolare von unipolarer Depression
- Accuracy 78 %, AUC 0,76
- Gesichts-, Blick- und Stimmmerkmale trugen zur Diagnose bei
Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.
Kurz gefasst
Fachartikel in der Zeitschrift Psychiatry and Clinical Neurosciences, online erschienen am 11.09.2026. Verfasst von Hidenori Yamasue und 13 weiteren Autor:innen. Beteiligte Einrichtung: Department of Psychiatry, Hamamatsu University School of Medicine, Hamamatsu, Japan. Hauptthemen laut MEDLINE: Bipolar Disorder, Major Depressive Disorder, Artificial Intelligence, Depressive Disorder. Die Arbeit nennt 47 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Japan Agency for Medical Research and Development.
Themen: Depression
Abstract (Originalsprache)
Background
Given their differing pathophysiology and treatment approaches, the primary aim of the current study was to develop a methodology accurately distinguishing bipolar from unipolar depression.
Methods
Adults with bipolar disorder or major depressive disorder were recruited nationwide in Japan and assessed via recorded interviews. During the interview, diagnoses were established using the Structured Clinical Interview for DSM‐5‐Research Version. Depressive and manic symptom severity was assessed using the Hamilton Depression Rating Scale (HDRS) and Young Mania Rating Scale (YMRS), respectively. Artificial intelligence (AI)‐based integrative, quantitative analyses of behavioral characteristics, including facial expressions, eye gaze, and voice prosody, were conducted on the recorded video.
Results
Between August 2022 and March 2024, 379 participants completed interviews, of whom 366 were analyzed (120 bipolar I, 93 bipolar II, 122 major depressive, and 74 persistent depressive disorder). AI‐based diagnostic differentiation between bipolar and unipolar depression using four‐fold, three‐times repeated cross‐validation achieved an overall accuracy of 78%, with an area under the curve of 0.76. Key features contributing to bipolar diagnosis included greater temporal instabilities in neutral facial expression, eye gaze, sad facial expression, and root mean square amplitude of speech. HDRS could be predicted in unipolar and bipolar depression, with influential predictors of depression including sleep inefficiency, autistic traits, variability in sad facial expression, and comorbid anxiety disorder. YMRS could be predicted with major predictors including comorbid anxiety disorder, sleep inefficiency, and variabilities in sad facial expression and eye movement.
Conclusions
Integrating multimodal behavioral indicators with AI‐based modeling enables clinically meaningful differentiation between bipolar and unipolar depression.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Psychiatry, Hamamatsu University School of Medicine, Hamamatsu, Japan.
- United Graduate School of Child Development, Osaka University, Kanazawa University, Hamamatsu University School of Medicine, Chiba University and University of Fukui, Suita, Japan.
- I'mbesideyou, Tokyo, Japan.
- Center for Clinical Research, Hamamatsu University Hospital, Hamamatsu, Japan.
- Department of Child and Adolescent Psychiatry, Hamamatsu University School of Medicine, Hamamatsu, Japan.
- Hamamatsu University School of Medicine, Hamamatsu, Japan.
ORCID-Profile: Hidenori Yamasue, Yosuke Kameno, Tomoyasu Wakuda, Tadafumi Kato
Förderung
- Japan Agency for Medical Research and Development
Zitieren
APA
Yamasue, H., Nanegaonkar, S. R., Kiyama, Y., Reddy, M. V. R., Ando, K., Kamiya, S., Inoue, J., Sakurai, A., Nakano, T., Sato, A., Kameno, Y., Wakuda, T., Odagiri, K., & Kato, T. (2026). Differentiation between bipolar versus unipolar depression by AI‐based multimodal behavioral analysis. Psychiatry and Clinical Neurosciences, 80(11), 925-934. https://doi.org/10.1111/pcn.70124
Vancouver
Yamasue H, Nanegaonkar SR, Kiyama Y, Reddy MVR, Ando K, Kamiya S, et al. Differentiation between bipolar versus unipolar depression by AI‐based multimodal behavioral analysis. Psychiatry and Clinical Neurosciences. 2026;80(11):925-934. doi:10.1111/pcn.70124
BibTeX
@article{Yamasue2026,
author = {Yamasue, Hidenori and Nanegaonkar, Sanskar Rahul and Kiyama, Yumi and Reddy, Mamindla Vishwa Raghava and Ando, Kotaro and Kamiya, Shozo and Inoue, Jun and Sakurai, Asuka and Nakano, Tahichi and Sato, Amane and Kameno, Yosuke and Wakuda, Tomoyasu and Odagiri, Keiichi and Kato, Tadafumi},
title = {Differentiation between bipolar versus unipolar depression by AI‐based multimodal behavioral analysis},
journal = {Psychiatry and Clinical Neurosciences},
year = {2026},
volume = {80},
number = {11},
pages = {925--934},
doi = {10.1111/pcn.70124}
}
EUVIMED