研究課題/領域番号 |
16H07469
|
研究機関 | 東京大学 |
研究代表者 |
梁 滋路 東京大学, 大学院工学系研究科(工学部), 特任助教 (10782807)
|
研究期間 (年度) |
2016-08-26 – 2018-03-31
|
キーワード | sleep / personal informatics / wearable fitness tracker / EEG |
研究実績の概要 |
After a thorough literature review on wearable home sleep tracking technologies, a small-scale validation study was conducted to understand the functionality and usability of current consumer sleep tracking technologies. In order to understand the accuracy of these devices, we compared typical consumer sleep tracking devices, such as Fitbit and wearable EEG eye mask Neuroon, against clinical device SLEEP SCOPE and subjective sleep experience. The results show that the accuracy of wearable EEG devices is low due to the immaturity of this technology. In comparison, the accuracy of wearable wristbands is promising, though still not up to clinical standard. Also, our participants generally found it easy to wear a Fitbit during sleep, but they reported difficulty in falling asleep wearing a EEG device. Our validation results suggest that wearable wristbands may be a good alternative to wearable EEG for home sleep monitoring. The outcomes of the study was published in several domestic and international conferences.
|
現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
At the strategy level, we found that wearable EEG devices may not be a good candidate for home sleep tracking, and our validation results suggest that Fitbit may be a good alternative. Therefore, the main focus of this year is to develop new algorithms for improving the accuracy of Fitbit.
At the tactic level, we have conducted a pilot study to collect sleep data from a small cohort. Ethics permission was obtained for running large-scale experiment for data collection. Collaborations with Dr.Bernd Ploderer of the Queensland University of Technology was established, and we have designed our user study plan together. All devices that are needed for the experiment have been purchased. We plan to start our experiment soon to collect rich qualitative and quantitative data.
|
今後の研究の推進方策 |
(1) Run a large-scale experiment to collect sleep data using consumer sleep trackers and clinical sleep monitor. These data will be used to develop new algorithms for improving the accuracy of Fitbit. Thirty health adults will be recruited for the experiment. Each participant will be asked to wear a Fitbit, a Neuroon, and a SLEEP SCOPE for 3 nights. After the experiment they will be invited for an 1-hour interview. (2) Develop new algorithms for accurate sleep tracking in home setting using multi-modal approach. The basic idea is to integrate physiological signals from multiple sources, including body movements, heart rate, brain wave. Machine learning techniques will be applied to establish mathematical models for mapping these physiological signals to sleep quality.
|