Physiologically motivated detection of Atrial Fibrillation

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Atrial Fibrillation (AF) is the most prevalent cardiac arrhythmia, affecting millions globally.
• AF significantly elevates risks of mortality, stroke, and heart failure, particularly in older individuals and those with comorbidities.
• Developing advanced, cost-effective algorithms for AF detection is crucial for personalized health systems.

Purpose of the Study:

• To introduce a novel algorithm for Atrial Fibrillation detection.
• To leverage three key physiological characteristics of AF: heart rate irregularity, absence of P-waves, and presence of fibrillatory waves.
• To enhance AF detection accuracy through multi-lead ECG analysis.

Main Methods:

• Extraction of discriminative features from 12-lead electrocardiograms (ECG) based on AF characteristics.
• Utilizing a Support Vector Machine (SVM) classification model for distinguishing AF from non-AF episodes.
• Comparison of the proposed 12-lead ECG algorithm against a previous single-channel approach.

Main Results:

• The algorithm achieved a sensitivity of 88.5% and a specificity of 92.9% in AF detection.
• Identification of fibrillatory patterns from 12-lead ECG significantly improved algorithm performance.
• The 12-lead approach demonstrated superior performance compared to the single-channel method.

Conclusions:

• The proposed 12-lead ECG-based algorithm effectively identifies Atrial Fibrillation.
• Leveraging specific AF physiological features enhances diagnostic accuracy.
• This algorithm shows promise for integration into low-cost, personalized cardiovascular health monitoring systems.