High-dimensional time irreversibility analysis of human interbeat intervals

Abstract

Time irreversibility is a subject of increasing interest in biomedicine. However, measurements of this property have been limited to a two-dimensional state space and few effective methods for high-dimensional time irreversibility testing have been reported. Using a strategy based on multiple bidimensional tests, we propose an extension of the traditional low-dimensional method for the estimation of time irreversibility in a high-dimensional state space. The method is applied to both physiological and synthetic interbeat (RR) intervals and the results indicate that time irreversibility should be an intrinsic characteristic of human RR intervals and is complementary to the conventional properties that can be measured in the time and frequency domain. Furthermore, a loss of time irreversibility is detected when the method is applied to patients with congestive heart failure (CHF). Compared with the traditional time irreversibility test, the proposed method is not only capable of quantifying time irreversibility in a more reliable manner, but is also effective for analysis of short records of only a few minutes' length, which would be more useful clinically.