P wave detector with PP rhythm tracking: evaluation in different arrhythmia contexts

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Accurate P wave detection in electrocardiograms (ECGs) is vital for diagnosing cardiac arrhythmias.
• Existing methods often struggle with complex arrhythmias beyond normal sinus rhythm, atrial flutter, and fibrillation.
• P wave detection is challenging due to signal variability and interference during various arrhythmias.

Purpose of the Study:

• To present a novel knowledge-based algorithm for automatic P wave detection in ECGs.
• To develop a self-adaptive algorithm capable of handling diverse arrhythmias by tracking PP rhythm.
• To evaluate the algorithm's performance on challenging arrhythmia datasets.

Main Methods:

• Development of a knowledge-based algorithm incorporating domain expertise for P wave identification.
• Implementation of a self-adaptive mechanism to personalize detection to individual patients.
• Testing the algorithm on the MIT-BIH arrhythmia database, including records with ventricular and supra-ventricular arrhythmias.

Main Results:

• The algorithm achieved an overall sensitivity (Se) of 96.60% and precision (Pr) of 95.46% across all records.
• For normal sinus rhythm, performance reached Se = 97.76% and Pr = 96.80%.
• Demonstrated robust performance in Mobitz type II (Se = 72.79%, Pr = 99.51%), trigeminy, and bigeminy, outperforming some advanced techniques.

Conclusions:

• The knowledge-based P wave detector demonstrates significant potential for improving arrhythmia diagnosis.
• The algorithm's self-adaptive nature and PP rhythm tracking enhance its robustness in complex cardiac conditions.
• Domain knowledge integration proves effective in advancing signal processing for challenging ECG analysis.