Detection of Complex Fractionated Atrial Electrograms Using Recurrence Quantification Analysis

Abstract

Atrial fibrillation (AF) is the most common cardiac arrhythmia but its proarrhythmic substrate remains unclear. Reentrant electrical activity in the atria may be responsible for AF maintenance. Over the last decade, different catheter ablation strategies targeting the electrical substrate of the left atrium have been developed in order to treat AF. Complex fractionated atrial electrograms (CFAEs) recorded in the atria may represent not only reentry mechanisms, but also a large variety of bystander electrical wave fronts. In order to identify CFAE involved in AF maintenance as a potential target for AF ablation, we have developed an algorithm based on nonlinear data analysis using recurrence quantification analysis (RQA). RQA features make it possible to quantify hidden structures in a signal and offer clear representations of different CFAE types. Five RQA features were used to qualify CFAE areas previously tagged by a trained electrophysiologist. Data from these analyzes were used by two classifiers to detect CFAE periods in a signal. While a single feature is not sufficient to properly detect CFAE periods, the set of five RQA features combined with a classifier were highly reliable for CFAE detection.