Effect of Missing Inter-Beat Interval Data on Heart Rate Variability Analysis Using Wrist-Worn Wearables

Area of Science:

• Biomedical Engineering
• Physiological Monitoring
• Wearable Technology

Background:

• Wrist-worn devices commonly offer continuous heart rate monitoring via photoplethysmography.
• Measuring continuous heart rate variability (HRV) using beat-to-beat pulse intervals remains a challenge for wearables.
• Motion artifacts significantly disrupt pulse interval measurements from photoplethysmography.

Purpose of the Study:

• To investigate the impact of missing beat-to-beat interval data on HRV analysis accuracy.
• To evaluate the robustness of different HRV parameters to data gaps caused by motion noise.
• To provide recommendations for acceptable data loss durations in wearable HRV analysis.

Main Methods:

• Simulations using electrocardiogram RR intervals from 39 subjects with randomly removed data.
• Real-world data collection of pulse intervals from 20 subjects using a wrist-worn device over 24 hours.
• Analysis of errors in time and frequency domain HRV parameters based on the duration of missing data.

Main Results:

• Mean NN interval and RMSSD demonstrated the highest robustness against missing pulse interval data.
• Errors in HRV parameters were directly proportional to the duration of missing data.
• Frequency domain parameters were often uncalculable during daily activities due to extensive data loss, except during sleep.

Conclusions:

• Mean NN and RMSSD are the most reliable HRV metrics for analysis with wearable devices experiencing motion-induced data gaps.
• The duration of missing data significantly impacts HRV analysis accuracy.
• Recommendations for maximum allowable missing data durations are crucial for reliable HRV assessment using wrist-worn devices.