Linear and/or nonlinear behavior of heart rate? A comparative study using Lyapunov exponents, fractal dimension, conditional entropy and spectral analysis

To investigate linear and nonlinear components of heart rate variability over the 24-hour period in healthy subjects, we analyzed RR interval series, derived from 70 recordings in free-living conditions (37 ± 4 years, 34 males) by different nonlinear approaches (largest Lyapunov exponent (LLE) computation, quantifying exponential divergence of trajectories in the phase space; corrected conditional entropy (CCE) estimating the amount of information in the signal, and pattern fractal analysis (PFD), able to quantify the complexity of the time series), and by autoregressive spectral analysis, too. The LLE, PFD, and CCE showed a strong correlation, whereas nonlinear parameters were significantly related even with total power, VLF, LF, and HF power of the spectra. Thus, common physiological phenomena seem to modulate both linear and nonlinear behavior of heart rate in healthy human beings.