Improving discriminality in heart rate variability analysis using simple artifact and trend removal preprocessors

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Heart Rate Variability (HRV) analysis is crucial for assessing cardiac function.
• HRV data frequently contains artifacts and trends that compromise analysis accuracy.

Purpose of the Study:

• To develop a straightforward and effective preprocessing method for cleaning HRV sequences.
• To enhance the reliability of Low Frequency (LF) and High Frequency (HF) component calculations for Autonomic Nervous System (ANS) assessment.

Main Methods:

• Applied a thresholding filter to eliminate HRV artifacts within a defined range.
• Utilized a wavelet filter to remove ultra and very low frequency components representing trends.
• Employed a K-Nearest Neighbors (KNN) classifier for congestive heart failure (CHF) detection.

Main Results:

• Preprocessed HRV data exhibited more separable features in power spectral density analysis compared to raw data.
• The KNN classifier demonstrated significant performance improvement in differentiating CHF from normal sinus rhythms (NSR) using preprocessed HRV features.
• The proposed method achieved superior CHF recognition with a simpler approach than existing literature.

Conclusions:

• The developed preprocessing technique effectively removes artifacts and trends from HRV data.
• Improved HRV analysis using this method enhances the accuracy of ANS regulation assessment.
• This simple yet effective preprocessor offers a promising solution for improved diagnostics, particularly in CHF detection.