Enhanced multi-source data analysis for personalized sleep-wake pattern recognition and sleep parameter extraction
- PDF / 2,899,468 Bytes
- 21 Pages / 595.224 x 790.955 pts Page_size
- 70 Downloads / 159 Views
ORIGINAL ARTICLE
Enhanced multi-source data analysis for personalized sleep-wake pattern recognition and sleep parameter extraction Sarah Fallmann1
· Liming Chen1 · Feng Chen2
Received: 30 October 2019 / Accepted: 19 August 2020 © Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract Sleep behavior is traditionally monitored with polysomnography, and sleep stage patterns are a key marker for sleep quality used to detect anomalies and diagnose diseases. With the growing demand for personalized healthcare and the prevalence of the Internet of Things, there is a trend to use everyday technologies for sleep behavior analysis at home, having the potential to eliminate expensive in-hospital monitoring. In this paper, we conceived a multi-source data mining approach to personalized sleep-wake pattern recognition which uses physiological data and personal information to facilitate fine-grained detection. Physiological data includes actigraphy and heart rate variability and personal data makes use of gender, health status, and race information which are known influence factors. Moreover, we developed a personalized sleep parameter extraction technique fused with the sleep-wake approach, achieving personalized instead of static thresholds for decisionmaking. Results show that the proposed approach improves the accuracy of sleep and wake stage recognition, therefore offering a new solution for personalized sleep-based health monitoring.
1 Introduction Sleep stages are traditionally detected based on electroencephalogram (EEG), electrooculogram (EOG), and electromyogram (EMG) data. The process requires trained technicians to visually inspect data and score usually 30s intervals based on guidelines, mainly the Rechtschaffen and Kales (R&K) [30] method or the American Academy of Sleep Medicine (AASM) [1]. The specific sleep stages can be divided into S1, S2, S3, and S4, Rapid eye movement (REM), and wake [30] or N1, N2, and N3, REM, and wake [1], where N3 is integrating S3 and S4 stages. Recently, data analysis approaches to automatically detect sleep stage are introduced to decrease the trained scoring technicians’ workload; this has been proven helpful and is Sarah Fallmann
[email protected] Liming Chen [email protected] Feng Chen [email protected] 1
School of Computer Science and Informatics, De Montfort University, Leicester, UK
2
School of Computing, Ulster University, Belfast, UK
already adopted in certain settings. The approach to test against one technician is limited as technicians’ scoring is subjective, i.e., having an individual component in the scoring process; therefore, the agreement is not always present [19]. This means that machine learning approaches help improve automation, but have also potentially learned one rater’s style. To collect sleep behavior data at home, and provide an unbiased data source for doctors, sensor technology is applied for detecting sleep stages [6, 9, 13]. Sleep parameters such as sleep onset latency (SOL) help in interpreting the overall sleep qualit
Data Loading...