LACK OF EVIDENCE FOR LOW-DIMENSIONAL CHAOS IN HEART-RATE-VARIABILITY

Introduction: The term chaos is used to describe erratic or apparently random time-dependent behavior in deterministic systems. It has been suggested that the variability observed in the normal heart rate may be due to chaos, but this question has not been settled. Methods and Results: Heart rate variability was assessed by recordings of consecutive RR intervals in ten healthy subjects using ambulatory ECG. All recordings were performed with the subjects at rest in the supine position. To test for the presence of nonlinearities and/or chaotic dynamics, ten surrogate time series were constructed from each experimental dataset. The surrogate data were tailored to have the same linear dynamics and the same amplitude distribution as the original data. Experimental and surrogate data were then compared using various nonlinear measures. Power spectral analysis of the RR intervals showed a 1/f pattern. The correlation dimension differed only slightly between the experimental and the surrogate data, indicating that linear correlations, and not a ''strange'' attractor, are the major determinants of the calculated correlation dimension. A test for nonlinear predictability showed coherent nonlinear dynamic structure in the experimental data, but the prediction error as a function of the prediction length increased at a slower rate than characteristic of a low-dimensional chaotic system. Conclusion: There is no evidence for low-dimensional chaos in the time series of RR intervals from healthy human subjects. However, nonlinear determinism is present in the data, and various mechanisms that could generate such determinism are discussed.