Cardiac arrhythmia spectral analysis of electrogram signals using Fourier Organization Analysis

Implantable Cardioverter Defibrillator (ICD) are devices which give relevant information about tachyarrhythmias in patients. The classical spectral approaches to assess cardiac electrograms (EGM) from ICD, Dominant Frequency Analysis (DFA) and Organization Analysis (OA) have often discarded relevant information in the spectrum, such as the harmonic structure. We propose a description, called Fourier Organization Analysis (FOA), for characterizing the spectral and organization features of EGM. FOA includes two stages, a first step involving fundamental frequency estimation, and a second organization analysis step, using Least Squares projection onto a signal space of sinusoids possibly containing fluctuations. The algorithm was first tested on synthetic EGM from a simple model simulation. Then, a data base was analyzed which included 14 episodes, namely, 5 Sinus Rhythm, 8 Supraventricular Tachycardia, 8 Ventricular Tachycardia, and 7 Ventricular Fibrillation records. Each episode had two EGM recordings, monopolar and bipolar. FOA showed high accuracy when estimating fundamental frequency for all rhythms, and also allowed to establish a coherent comparison between them. We conclude that FOA yields a more compact and adequate framework for analyzing ICD stored EGM.