Paroxysmal atrial fibrillation recognition based on multi-scale Rényi entropy of ECG

Abstract

Background: Atrial fibrillation (AF) is a common type of arrhythmia disease, which has a high morbidity and can lead to some serious complications. The ability to detect and in turn prevent AF is extremely significant to the patient and clinician. Objective: Using ECG to detect AF and develop a robust and effective algorithm is the primary objective of this study. Methods: Some studies show that after AF occurs, the regulatory mechanism of vagus nerve and sympathetic nerve will change. Each R-R interval will be absolutely unequal. After studying the physiological mechanism of AF, we will calculate the Rényi entropy of the wavelet coefficients of heart rate variability (HRV) in order to measure the complexity of PAF signals, as well as extract the multi-scale features of paroxysmal atrial fibrillation (PAF). Results: The data used in this study is obtained from MIT-BIH PAF Prediction Challenge Database and the correct rate in classifying PAF patients from normal persons is 92.48%. Conclusions: The results of this experiment proved that AF could be detected by using this method and, in turn, provide opinions for clinical diagnosis.