Nonparametric, nonlinear modeling of physiological systems

The issue of mathematical modeling of nonlinear physiological systems is addressed here. Two different nonparametric methods of identifying nonlinear dynamic physiological systems with arbitrary nonlinearities are presented and their physiological implications as well as their virtues and deficiencies are discussed. The first approach is based on Wiener´s theory of nonlinear system identification by use of a Guassian white noise (GWN) test input. The second approach is based on nonparametric differential mapping that uses information about the state variables and leads to phase space representation. The latter method is shown to be more efficient than Wiener´s approach in identifying nonlinear dynamic behaviors of physiological systems