Sleep stage estimation by specifying latent structure of heartbeat data

2016 Fiscal Year Final Research Report

• Project/Area Number: 15K12105
• Research Category: Grant-in-Aid for Challenging Exploratory Research
• Allocation Type: Multi-year Fund
• Research Field: Soft computing
• Research Institution: The University of Electro-Communications
• Principal Investigator: TAKADAMA Keiki 電気通信大学, 大学院情報理工学研究科, 教授 (20345367)
• Project Period (FY): 2015-04-01 – 2017-03-31
• Keywords: ソフトコンピューティング / 機械学習 / 睡眠段階推定 / 人工知能

Outline of Final Research Achievements

This research focuses on the sleep stage estimation method without connecting any devices to human body, and improves its estimation capability even in bad health condition, aging process, and sleep disorder. Since a sleep rhythm or deepness is affected by health condition, aging, and/or sleep disorder, this research proposed the approach of mining the latent structure of heartbeat data (called the sleep property) to specify the common sleep property as a combination of the frequency components of the ultradian rhythm (i.e., a cycle of non-REM and REM sleep) derived from the heartbeat data. To investigate the effectiveness of the proposed approach, we conducted the human subject experiments which revealed that the improved sleep stage estimation method based on the proposed approach is robust to bad health condition, adapts to aging, and copes with sleep disorder.

Free Research Field

マルチエージェント，進化計算，機械学習