Paroxysmal atrial fibrillation prediction based on HRV analysis and non-dominated sorting genetic algorithm III

Area of Science:

• Biomedical Engineering
• Cardiology
• Machine Learning

Background:

• Paroxysmal atrial fibrillation (PAF) is a prevalent cardiac arrhythmia with significant health risks.
• Accurate prediction of PAF onset is crucial for timely intervention and prevention of atrial arrhythmias.
• Existing methods for PAF prediction often require lengthy heart rate variability (HRV) signals.

Purpose of the Study:

• To develop and validate a novel method for predicting paroxysmal atrial fibrillation (PAF).
• To utilize shorter heart rate variability (HRV) signals for improved efficiency and accuracy in PAF prediction.
• To optimize the PAF prediction system using a multi-objective optimization algorithm.

Main Methods:

• A pre-processing stage involving heart rate correction, interpolation, and signal detrending.
• Extraction of time-domain, frequency-domain, and non-linear HRV features.
• A support vector machine (SVM) model trained on extracted HRV features.
• Optimization of feature extraction and SVM parameters using a non-dominated sorting genetic algorithm III.

Main Results:

• The proposed method achieved an accuracy rate of 87.7% in predicting PAF.
• This high accuracy was obtained using significantly shorter HRV signals (5 minutes vs. typical 30 minutes).
• Improved sensitivity was achieved, prioritizing this metric over specificity for clinical relevance.

Conclusions:

• The developed method offers a highly accurate and efficient approach for predicting paroxysmal atrial fibrillation.
• The reduction in required HRV signal length marks a significant advancement in practical clinical application.
• The optimized system demonstrates superior performance compared to existing PAF prediction methods.