Detection of nonlinearity in the healthy heart rhythm

Sequences of interbeat intervals from two groups of subjects, 24 in relaxed and 10 in sleeping condition were analyzed by the nonlinear predictor method as well as by a method, proposed recently, able to directly estimate the time series nonlinearity. The nonlinear predictability of the R-R intervals is tested by using surrogate data. The results obtained with both methods show that nearly all the sequences exhibit a statistically meaningful degree of nonlinearity. This raises the question whether such nonlinearity encodes information about the physiological condition of the examined subjects.