The fluctuation of RR intervals (RRI) on an electrocardiogram (ECG) is called heart rate variability (HRV). HRV reflects the autonomic nerve activity; thus, HRV analysis has been widely used for health monitoring. However, RRI and HRV features are easily affected by artifacts, which deteriorates the health monitoring performance. Hence, a methodology for dealing with RRI fluctuation disturbed by artifacts needs to be developed to realize precise health monitoring. This work proposed a method for RRI modification from a predictive perspective by treating the RRI data affected by artifacts as missing values and using their preceding data to predict the disturbed RRI data. The proposed method utilized the N-BEATS model, a cut-edged model in the field of time series forecasting, to perform RRI modification. The results show that the N-BEATS model successfully predicts the contaminated RRI data, and the root mean square error (RMSE) of the N-BEATS modified RRI is much smaller than other methods and has a good generalization to perform well on different test individuals. The proposed N-BEATS-based method will help to achieve accurate health monitoring based on heart rate variability in the future.
Heart rate variability can be continuously monitored by a wearable device that analyzes abnormalities and sends notifications so that the patient and those around them can be informed in advance of a seizure and safety measures can be taken. This study can reduce the effect of artifacts on HRV and improve the performance of HRV analysis.
氏名 | コース | 研究室 | 役職/学年 |
---|---|---|---|
CHEN Hongtao | システム科学コース | Human Systems Laboratory | 博士1回生 |
加納 学 | システム科学コース | Human Systems Laboratory | 教授 |
加藤 祥太 | システム科学コース | Human Systems Laboratory | 助教 |