Predictability and information-theoretic measures successfully recognized the different nature of high-order interactions in physiological data.
Predictability and information-theoretic measures can effectively distinguish between functional connections and statistical dependencies in complex physiological networks.
Physiological network systems are characterized by the presence of complex interactions between two or more subsystems, leading to the emergence of synergistic and redundant modes of interplay. As the nature of these high-order interactions (HOIs) is largely unknown, recent research has been focused on distinguishing between high-order mechanisms (HOMs) and high-order behaviors (HOBs), respectively related to the presence of functional connections or statistical dependencies within the system. This work proposes the use of predictability and information-theoretic measures of net synergy-redundancy balance as indicative of HOMs and HOBs, respectively. These measures were empirically validated on simulated settings and then tested on physiological data, i.e., interbeat interval, mean arterial pressure, arterial compliance, and respiratory time series, to shed light on the influence of cardiorespiratory dynamics on the vascular activity in the supine resting state. Our results confirm the usefulness of interaction predictability and interaction information measures in recognizing the different nature of HOIs.
Barà et al. (Mon,) conducted a other in Physiological network systems. Predictability and information-theoretic measures was evaluated on Net synergy-redundancy balance. Predictability and information-theoretic measures successfully recognized the different nature of high-order interactions in physiological data.