Implementation of the Real-time Algorithms Based on the Symbolic Dynamics of a Coarse-grained Heart Rate Variability in Ubiquitous Health Care Systems

In order to perform a real-time analysis of a physiological signal, such as heart rate variability (HRV), obtained from a ubiquitous health care system, we propose some effective measures of complexity, which are basing on the symbolic dynamics. In symbolic dynamics, HRV extracted from an electrocardiogram is transformed into a symbolic sequence in the coarse-graining process where the difference of periods between consecutive heartbeats plays a threshold for determining the type of symbols in the symbolic return map. Quantification of a symbolic sequence was performed by using the generalized symbolic measures; pNNx, pWx, CNx, and CSEx in which x denotes a threshold of coarse-graining symbolization. By varying the threshold x, we comprehensively examined the symbolic dynamic parameters for normal controls with normal sinus rhythms and the presumed patients with atrial fibrillation, or congestive heart failure, and identified the optimal threshold distinguishing the normal group from the abnormal group. The range of optimal thresholds was found at 10 ∼ 25 ms, which is far below from the conventional standard of 50 ms. We demonstrate that pNNx, pWx, CNx, and CSEx are useful algorithms for the real-time analysis of HRV acquired from the ubiquitous health care system as well as for the assessment of the emergent alterations in a heart dynamics.