Paper
Novel Sleep Stage Estimation Method for Children Using Body Movement
- Authors:
- Ikuko Mohri; Yuko Ohno; Sachiko Shimizu; Shima Okada; Masako Taniike; Masaaki Makikawa; Masaaki Makikawa
- Abstract
- Sleep dysfunction among children is the subject of increasing research attention. However, current methods for assessing daily childhood sleep are problematic, and there are no techniques for estimating sleep stage from body movement alone. Here we propose a new method for estimating sleep stages in children using video analysis of body movement. We examined eight healthy children using polysomnography (PSG) and video analysis, comparing the results of the PSG examination with the findings of our new estimation method. The linear discriminant function method was used to predict the sleep stage from body movements. The continuous time of non-movement and the moving average amount of body movement were included as independent variables. The three sleep stages, Awake, Light & (Rapid Eye Movement) REM, and Slow Wave Sleep (SWS), were included as response variables. The results confirmed that the transition to a different sleep stage was accompanied by changes in body movement. Overall, the estimates generated by our method showed strong agreement with PSG results. The average agreement rate for sleep stage was approximately 67% for Wake, 73% for Light & REM, and 66% for SWS. The results revealed that the transition between sleep stages can be successfully estimated by body movement during sleep, providing a promising approach for developing a simple sleep disorder screening system that can be used at home.
- Keywords
- Body Movement; Linear Discriminant Analysis; Video Image Processing; Gross Movements
- StartPage
- 61
- EndPage
- 68
- Doi
- 10.5963/PHF0202002