Local synchronization indices for rotors detection in atrial fibrillation: A simulation study

Abstract

Stable rotors have been proposed as mechanisms that maintain atrial fibrillation which is the most common arrhythmia worldwide. The information of intracardiac electrograms (EGMs), recorded through multielectrode arrays, is used to characterize the electrical conduction dynamics and thus to identify rotors or reentrant propagating waves. Most of the methods of EGMs processing are based on the assessment of the individual properties of each EGM signal. Additionally, synchronization indices have been proposed to evaluate the properties of the conduction patterns by means of multivariate analysis. However, the problem of rotor detection through EGMs remains open. We evaluate the behavior of four local synchronization indices using computational simulations of different conduction patterns and the corresponding EGMs in 2D models. The results show that phase synchronization exhibits better performance than correlation, coherence, and mutual information for detecting rotors under different fibrillatory patterns. We also show that this approach outperforms a previously reported technique based on entropy analysis of individual EGM. Synchronization maps using phase-locking values calculated from adjacent EGM highlight the vicinity of the core of stable rotors, even in the presence of multiple wavefronts and wave breaks. Therefore, phase-locking maps can be a useful tool for characterizing rotors during atrial fibrillation episodes.