Information flow enables specific identification of particular physiological controllers based on their time scales

Complex physiological systems, such as the cardiovascular system, controlled by the autonomic nervous system, and the gait and posture system, controlled by the sensomotor nervous system, incorporate several controllers acting at different time scales. We investigated the question, whether the information flow as a function of prediction time enables the identification of those contributing controllers within the entire complex system behavior.

Information flow functions assess the predictability of a time series. They were calculated based on mutual information over prediction time horizon tau. The importance of considering appropriate time scales tau was introduced by a simulation study of an excited pendulum. Based on heart rate data, complex autonomic dysfunctions were assessed by using autonomic information flow (AIF) in an experimental study of rats and in clinical studies of patients with multiple organ dysfunction syndrome (MODS) and after cardiac arrest (CA). Motor dysfunction in elderly suffering from low back pain was assessed by gait information flow (GIF) based on back movements.

In rats, AIF over one heart beat period was reduced due to anesthesia, AIF over vagal time scale was increased due to vagotomy. In both, MODS and CA patients, vagal time scale AIF was reduced, but longterm AIF was increased. In low back pain patients, short loop GIF was increased, but walking period-related GIF was reduced.

Information flow functions allow the identification of particular controllers and their interdependencies within complex physiological systems.