Bispectral analysis of intracardiac electrograms

Abstract

Previous work has shown that the average magnitude-squared coherence (MSC) differentiates ventricular tachyarrythmias and discriminates between fibrillatory and nonfibrillatory rhythms. The MSC quantifies the extent to which linear dynamics explain time series covariability. The magnitude-squared bicoherence (MSBC) does the same for bilinear dynamics. Unlike the MSC, which must be computed from two time series, the MSBC can be calculated from a single time series. The average magnitude-squared autobicoherence is computed for a collection of intracardiac electrograms. Results indicate that it has capabilities similar to average MSC in differentiating ventricular tachyarrythmias and discriminating between fibriallatory and nonfibrillatory rhythms. The MSBC might be used in implantable devices for cardiac monitoring and intervention to provide cross-validitory and localized information on cardiac dynamics. Moreover, its quantification of nonlinear dynamics may improve understanding of the intricate electrical organization of the heart.<>