Nonlinear methods in heart rate variability

In recent years, analysis of heart rate variability (HRV) has become a common tool for prediction of cardiac mortality. The universal rule is that a reduced variability is a signature for enhanced risk. However, the clinical value of the usual methods (the calculation of averages, standard deviations and variances of RR interval subsets, analysis of the power spectra in frequency domain) is very limited: false positive results cut the positive predictive value down to about 30%. We applied a net information flow method of determination of complexity in deterministic chaos (associated with nonlinear dynamics) to 24 hour recordings of beat-to-beat heart-rate. It is likely that analysis of HRV by the means of nonlinear dynamics may open a way for a new classification of HRV.