Post-Processing of R Wave Detection to Enhance Heart Rate Variability Analysis.

Area of Science:

• Biomedical Engineering
• Physiological Monitoring
• Wearable Technology

Background:

• Heart rate variability (HRV) is a crucial biomarker for autonomic nervous system function.
• Accurate R-R interval (RRI) calculation from ECG is essential for HRV analysis.
• R wave misdetections in ECG pose a significant challenge to reliable HRV monitoring.

Purpose of the Study:

• To develop a method for identifying misdetected R waves in ECG signals.
• To improve the accuracy of RRI editing for enhanced HRV analysis.
• To support high-precision HRV-based health monitoring using wearable devices.

Main Methods:

• Utilized ECG amplitude and detected R wave position for misdetection identification.
• Compared three machine learning algorithms and three preprocessing techniques.
• Evaluated classification performance using Matthew's correlation coefficient.

Main Results:

• Non-linear machine learning algorithms combined with min-max or Z-score normalization demonstrated superior performance.
• Achieved a Matthew's correlation coefficient of 0.86 in identifying misdetected R waves.
• The proposed method effectively utilizes post-detection information for RRI editing.

Conclusions:

• This study presents a reliable method for detecting misdetected R waves, crucial for accurate HRV analysis.
• The findings enhance the feasibility of using wearable devices for daily life health monitoring.
• Improved RRI editing contributes to more trustworthy HRV-based health assessments.