Assessment of dynamic autonomic changes with posture using instantaneous entropy measures

Dynamic analysis provides a powerful methodological framework for characterizing physiological systems. In particular, complex heartbeat dynamics related to autonomic control mechanisms are known to change at each moment in time, and complexity measures have been proven to have prognostic value in both health and disease. Nevertheless, an instantaneous measure of complexity for cardiovascular time series (or any other series of stochastic physiological “events”) is still missing. In this study we introduce a mathematical framework serving instantaneous complex estimates of heartbeat dynamics to characterize different activities, tasks, and/or pathological states. In particular we propose new definitions of inhomogeneous point-process approximate and sample entropy where the discrete events are modeled by probability density functions characterizing and predicting the time until the next event occurs as a function of past history. These definitions are built on our previous work employing Laguerre expansions of the Wiener-Volterra autoregressive terms to account for long-term memory. We demonstrate an exemplary study on heartbeat data gathered from healthy subjects undergoing postural changes such as stand-up, slow tilt, and fast tilt. Results show that instantaneous complexity is able to effectively track the complex autonomic changes as they are affected by different postural changes.