Development and clinical application of personalized IoT system to control the risk of mental and physical disorders of workers

2022 Fiscal Year Final Research Report

• Project/Area Number: 20H00569
• Research Category: Grant-in-Aid for Scientific Research (A)
• Allocation Type: Single-year Grants
• Section: 一般
• Review Section: Medium-sized Section 59:Sports sciences, physical education, health sciences, and related fields
• Research Institution: The University of Tokyo
• Principal Investigator: Yamamoto Yoshiharu 東京大学, 大学院教育学研究科(教育学部), 教授 (60251427)
• Co-Investigator(Kenkyū-buntansha): 北島 剛司 藤田医科大学, 医学部, 教授 (40360234) ; 吉内 一浩 東京大学, 医学部附属病院, 准教授 (70313153) ; 中村 亨 大阪大学, 大学院基礎工学研究科, 特任教授(常勤) (80419473)
• Project Period (FY): 2020-04-01 – 2023-03-31
• Keywords: Micro-randomized trial / Affective computing / 睡眠安定性 / AIoT

Outline of Final Research Achievements

The aim of this research project was to identify the vulnerable state of mental health among Japanese employees and to modify their health-related behaviors in real-world settings. By using digital devices, including a smartphone application and a wearable activity monitor, we measured employees' moods and bio-signals (i.e., physical activity, heart rate, and voice) in real-world settings. Then, we developed an AI model to predict the self-reported moods from the emotionally neutral voice data. As a result, the model could predict employees' moods reasonably. In addition, we conducted micro-randomized trial targeting instability in habitual sleep behaviors, which is known as potential factor contributing to mood disturbances. It was found that habitual sleep duration was significantly stabilized during the trial especially among the employees with unstable habitual sleep behaviors.

Free Research Field

教育生理学、生体信号処理、健康情報学

Academic Significance and Societal Importance of the Research Achievements

本研究では(1)感情中立的発話データに基づく気分推定技術の開発、および (2)睡眠習慣の安定化を目的としたMicro-randomized trialの実証実験を行った。これらの研究開発は、国際的にも実証例の少ない、あるいは、前人未到の試みであり、不調の検知・制御に関する先駆的な取り組みであったと言える。また、上述の研究成果は全て日常生活下で得られたデータを基盤として得られたものであることから、real-worldでの健康リスクの予兆検知・一次予防の実現可能性を担保するものである。