TIME- AND FREQUENCY-BASED INDEPENDENT EVALUATION OF QRST CANCELLATION TECHNIQUES FOR SINGLE-LEAD ELECTROCARDIOGRAMS...

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Atrial fibrillation (AF) prevalence is increasing, necessitating improved non-invasive diagnostic methods.
• Accurate analysis of atrial activity from electrocardiograms (ECGs) requires effective removal of ventricular QRST complexes.
• Current QRST cancellation methods lack objective, physiologically-based evaluation frameworks.

Purpose of the Study:

• To develop an objective, physiologically-based evaluation approach for QRST cancellation methods.
• To compare the performance of three common QRST cancellation techniques: Average Beat Subtraction (ABS), Singular Value Cancellation (SVC), and Principal Component Analysis (PCA).
• To assess these methods using synthesized ECGs that preserve atrial-specific characteristics.

Main Methods:

• Development of a novel evaluation framework comparing cancellation methods against AF power spectra.
• Evaluation of ABS, SVC, and PCA in both time and frequency domains.
• Utilized synthesized ECG datasets with controlled QRST morphological variability.

Main Results:

• The Average Beat Subtraction (ABS) method demonstrated superior performance in scenarios with low QRST morphological variability.
• Principal Component Analysis (PCA) provided the most accurate estimation when significant QRST morphological variability was present.
• The study established a basis for objective comparison of QRST cancellation techniques.

Conclusions:

• The choice of QRST cancellation method should be tailored to the specific characteristics of the ECG signal, particularly QRST complex variability.
• PCA offers a robust solution for complex ECG signals with dynamic QRST morphology.
• The developed evaluation approach facilitates objective assessment and selection of optimal QRST cancellation strategies for AF analysis.