Missing RRI interpolation for HRV analysis using locally-weighted partial least squares regression.

Area of Science:

• Biomedical Engineering
• Physiological Signal Processing
• Machine Learning in Healthcare

Background:

• Heart Rate Variability (HRV) analysis, derived from R-R interval (RRI) fluctuations in electrocardiograms (ECGs), reflects autonomic nervous system function.
• HRV is crucial for health monitoring services like stress estimation and seizure prediction, but ECG artifacts can cause missing RRI data.
• Accurate interpolation of missing RRI data is essential for reliable HRV analysis.

Purpose of the Study:

• To propose and evaluate a novel missing RRI interpolation method using just-in-time (JIT) modeling.
• To enhance the accuracy of HRV analysis by addressing data gaps caused by ECG artifacts.
• To compare the performance of the proposed JIT-based method against traditional static interpolation techniques.

Main Methods:

• Development of a missing RRI interpolation method based on just-in-time (JIT) modeling.
• Application of locally weighted partial least squares (LW-PLS), a JIT modeling technique, for RRI interpolation.
• Validation using real-world RRI data collected from healthy individuals.

Main Results:

• The proposed JIT-based method, utilizing LW-PLS, demonstrated improved interpolation accuracy for missing RRIs.
• The JIT approach effectively handled data gaps that would challenge static interpolation methods.
• Case studies with real RRI data confirmed the enhanced performance of the proposed method.

Conclusions:

• Just-in-time (JIT) modeling, specifically LW-PLS, offers a superior approach for interpolating missing RRI data compared to static methods.
• This improved interpolation accuracy can lead to more reliable HRV-based health monitoring.
• The proposed method holds potential for advancing the precision of physiological signal analysis.