993-44 The Ventricular Response in Atrial Fibrillation is Not Chaotic

Abstract

In nonlinear dynamics, chaos refers to a system that is aperiodic but deterministic (nonrandom). Although the ventricular response during atrial fibrillation (AF) is commonly described as chaotic, it has yet to be demonstrated that this represents chaos in the mathematical sense. A defining characteristic of chaotic systems is sensitive dependence on initial conditions; i.e. similar sequences evolve similarly in the near future, but then diverge exponentially. We developed a nonlinear predictive forecasting algorithm to search for evidence of short-term predictability and sensitive dependence on initial conditions in recordings of 2000 ectopy-and artifact-free RR intervals obtained during routine activity in 5 pts with AF. The algorithm: 1) uses the technique of lags to reconstruct phase spaces with embedding dimensions from 3 to 10,2) the 3 nearest neighbors of a given trajectory are used to predict the evolution of the trajectory from 1 to 10 intervals into the future and 3) the correlation coefficient between predicted and actual evolution is computed for all RR intervals in the time series. The results were compared to test sequences from linear oscillators with high and low signal-noise ratios and a system with chaotic dynamics (the logistic map). Distinct from each of the test sets, the ventricular response in AF was only weakly predictable at all time scales, and did not exhibit sensitive dependence on initial conditions. The irregular ventricular response during AF is not governed by a low-dimensional chaotic attractor.