METHODS TO QUANTITATE ATRIAL ELECTRICAL ACTIVITY USING ELECTROGRAMS ACQUIRED DURING ATRIAL FIBRILLATION

Abstract

Herein, commonly used quantitative bioengineering methods that have been developed to analyze fractionated electrograms recorded from the surface of the atria during atrial fibrillation (AF) are described. Techniques were categorized as time-domain and frequency-domain methods. The main time-domain method is peak counting. Its variations based on preprocessing and thresholding are discussed. The main frequency-domain method is spectral analysis. Two spectral estimators, the discrete Fourier transform (DFT) and the new spectral estimator (NSE) are described. The ability of each estimator to detect the main periodic component of fractionated atrial electrograms is compared. Several spectral parameters that are used for analysis of atrial electrograms including the dominant frequency (DF), dominant amplitude (DA) and mean spectral profile (MP) are defined. Mean values of these parameters are compared in paroxysmal versus persistent AF fractionated electrograms based upon the results of several studies. Time-domain methods are shown to work best for analysis with deterministic, not fractionated atrial electrograms. For fractionated atrial electrograms, frequency-domain methods are often used. The DF, DA and MP spectral parameters are significantly different in paroxysmal versus longstanding persistent AF recordings. The DF and the DA are significantly higher, and the MP is significantly lower, in persistent AF electrogram recordings. The higher DF and DA parameter values reflect substrate remodeling in persistent AF, which increases the stability of the electrical activation pattern. The lower MP value in persistent AF reflects the lower spectral noise floor, indicative of a less complex and more periodic pattern of electrical activity.