Generalized hurst exponents as a tool to estimate Atrial Fibrillation organization from the surface ECG

Atrial Fibrillation (AF) is the most common supraventricular arrhythmia found in clinical practice. The aim of the present work has been the application of the generalized Hurst Exponent of order 2, H(2), to study AF organization from the surface ECG. H(2) relates to the behavior of the autocorrelation function of a time series, thus measuring long-term statistical dependencies that could be used to estimate AF organization. Since the spontaneous termination of paroxysmal AF is related to the arrhythmia organization, the AF Termination Database available at Physionet has been analyzed to test the method's ability in the discrimination of organization-related events. The performance of H(2) was compared with two noninvasive established AF organization metrics, the dominant atrial frequency (DAF) and Sample Entropy (SampEn). H(2) yielded better classification accuracy results (95.0%) than DAF and SampEn (both 88.3%). Moreover, statistically significant differences were found between non-terminating and terminating groups and also between soon terminating and terminating groups. As conclusion, H(2) can be considered as a promising tool for the non-invasive study of AF organization.