Mutual Information Function in Respirocardial Coordinations of Healthy Human Neonates in Quiet and Active Sleep

The complexity of heart rate fluctuations (HRF) is based on several interacting physiological mechanisms operating on different time scales. No prominent time scale for HRF complexity analysis is given a priori. The aim of this work is to discriminate the active and quiet sleep of healthy full-term neonates by quantitatively assessing respirocardial coordination dynamics using the recently introduced complexity parameter mutual information function (MIF). Representing the different time scales of information flow in autonomic nervous system, MIF carries information on a wider scope of complex interdependencies than complexity estimators previously known. Our hypothesis was therefore that MIF discriminates sleep states by comprehensively characterizing complex coordinations of HRF and respiratory movements (RM). RM and ECG-derived HRF of 6 healthy full-term neonates (4±1 days of life) were studied. As standard measures characterizing sleep states linear parameters were calculated (power spectra, coherence, auto- and cross-correlation functions). As non-linear parameters of HRF and RM, auto- and cross-MIF were analyzed. All results were statistically tested for their discriminatory power and non-linearity. Confirming our hypothesis we were able to discriminate active and quiet sleep states in all individual cases using one single global time scale parameter of HRF total auto-MIF. We assume that the vagal influence in healthy human neonates mediates mostly complex (linear and non-linear) HRF properties, whereas the sympathetic effect is mainly responsible for linear HRF properties. With the character of the MIF parameters deployed in mind, this finding would explain our success in discriminating the sleep states. Remarkably, HRF complexity was larger in quiet than in active sleep. Complex respirocardial interdependencies cannot be identified completely by the local time scale MIF parameters alone. New information is gained when total MIF values are also considered. This result confirms the relevance of global measures of information flow for a comprehensive discrimination of complex systems. Sleep state-related changes of MIF parameters extend the possibilities of interpreting the underlying physiological processes of complex respirocardial coordination dynamics.