Abstract
Introduction: Atrial fibrillation (AF) is the most frequently occurring atrial arrhythmia. Fibrillation intervals are used to characterize AF and have a potential to determine the strategy for AF therapy regardless of its etiology. However, measuring local fibrillation intervals requires an extensive period of time when using visual methods, and such methods are also subject to data bias. Recently several methods for measuring the atrial fibrillation cycle length (AFCL) using computers have been developed. Atrial activation during AF was considered to be random or chaotic. From recent studies, however, it was found that AF has various degrees of organization, and therefore activation during AF can be analyzed by spectral analysis. Several kinds of spectral analysis have become available for characterization of AF, and fast Fourier transform analysis is most frequently used. Autocorrelation function also can be used for analysis of AF intervals and could give us further information, such as minimum AFCL. Further, the frequency analysis of AF may be useful to predict the effect of antiarrhythmic drugs, cardioversion and catheter ablation for AF. This review was designed to investigate the analysis of AF intervals and to assess methods, interpretation and clinical implications.
