Linearity assessment of heart rate time series

Presents a new technique based on the combination of a polynomial model estimation and a statistical criterion to examine the non-linear nature of the heart rate signal. The method is applied to signals coming from control subjects and heart transplant patients under spontaneous and controlled respiration. The authors show that controlled respiration adds significant linearity to the heart rate signal compared to the spontaneous one. Furthermore, it seems to introduce more linearity to control subjects than to heart transplant patients. Finally, examining the influence of a standard resampling algorithm on the heart rate characteristics, the authors demonstrate that this kind of pre-processing clearly injects linearity into the structure of the heart rate signal.