Nonlinear Forecasting of RR-Intervals of Human Electrocardiograms

Abstract

The heart is a typical nonlinear oscillator. In mammals the heartbeat is controlled by the heart’s own pacemaker, called the sinoatrial node (SAN), which emits electrical pulses, that trigger the contraction of the heart muscle. The isolated SAN is a periodic oscillator¹, but in the intact heart the node is controlled by the autonomic nervous system. Para-sympathetic nerves reduce the rate and sympathetic nerves increase the rate at which the SAN’ fires’ its electrical pulses. The two types of nerves are therefore in a constant’ fight’ to control the electrical activity of the SAN resulting in substantial fluctuations in the time intervals (RR-intervals) between successive heartbeats in healthy subjects. Recent studies of these fluctuations in RR-intervals as well as of whole electrocardiograms using nonlinear methods have indicated that the dynamics of the heart is chaotic^2-4. On the other hand it has been known for more than a decade that certain heart dysfunctions and severe diabetes can be associated with a loss in the variability of the RR-intervals^{5, 6}, suggesting that during these illnesses the heartbeat becomes more regular and hence’ less’ chaotic. In this study we employ nonlinear forecasting to analyze the RR-intervals of electro-cardiograms from 120 patients who have experienced their first myocardial infarction, and 35 normal subjects. The purpose of this study was to evaluate if nonlinear forecasting could be used to characterize the dynamics of the heart and in particular if this method would reveal any differences between the normal subjects and the patients.