Complex dynamics assessment in 24-hour heart rate variability signals in normal and pathological subjects

Abstract

Long term regulation of beat-to-beat variability involves a different kind of control. Parametric models provide quantitative indices which measure short time regulating action of the autonomic nervous system. In the long period instead, nonlinear contributions can be put into evidence by a chaotic deterministic approach. For heart rate variability (HRV) series collected in the 24 hours in 14 normal subjects and 28 subjects with cardiovascular pathologies (11 severe heart failure, 11 essential hypertensive and 6 heart transplant), we extract some parameters which are reputed to be invariant characteristic of system attractor: fractal dimension, Kolmogorov entropy and Lyapunov exponents. Geometric representations in the state space, such as delay maps and phase space plots, describe system trajectories through the singular value decomposition method. All these parameters confirm the existence of nonlinear dynamics in HRV signals and show different values for normal and pathological subjects: in particular we notice a reduction of the complexity of the discrete series when passing from normal to pathological subjects.<>