It was investigated how postnatal development of cardiac control can be described by linear methods and by new application of non-linear methods of time series analysis.

Sixteen healthy and term newborns were studied during their first six months of life. Mean heart rate (HR) was calulated and heart rate variability analysis - standard deviation (SD) was performed on the instantaneous heart rate (IHR) time series. The largest Lyapunov exponent (LLE) and the correlation dimension (D2) were calculated and verified by means of the surrogate-data-test.

ANOVA statistics revealed age dependency for the parameters HR, SD, LLE, and D2 for both active and quiet sleep. HR, SD, and LLE showed significantly higher values in active sleep in comparison to quiet sleep in most of the measurements. In contrast, D2 was significantly lower during active sleep for all ages.

For all ages and sleep stages positive LLE - a hall-mark of chaos - were found, indicating sensitivity to initial conditions. The positive LLE and the found non-integer D2 indicate that there is a non-linear component in the heart rate control. During the first six months, the developmental changes in both the linear and non-linear parameters were not linearly related to age. This may reflect a multitude of influences on the cardiovascular system, due to distinct structural maturation in this period of life.