Nonlinear interpretation of respiratory sinus arrhythmia in anesthesia.

A non-parametric method is presented for modelling nonlinear dynamic mechanisms of respiratory sinus arrhythmia (RSA) in anesthesia caused by positive pressure ventilation. RR interval sequences are shown with Tsay's linearity test to contain both short-term and long-term nonlinear components, which cannot completely be modelled with optimal linear methods. The nonlinear approach is based on Wiener's theory for broad-band random input signal. The input-output model is formed for tracheal pressure and RR interval sequence. Second-order and third-order nonlinearities in RSA fluctuation are found and demonstrated.