Novel multiscale E-metric cross-sample entropy-based cardiac arrhythmia detection and its performance investigation in reference to multiscale cross-sample entropy-based analysis

Cardiac arrhythmia is a common difficulty of human cardiovascular system and can be evaluated using cardiac rate variability. Multiscale Cross Sample Entropy (MCSEn) is used as a reference to quantify cardiac arrhythmia on the basis of complexity for double-interval series at multiple scales. This measure is failed to provide complexity with reduced scale factors for large data lengths. To hypothesize this measure for two series cardiac data by using coarse-grained process, Multiscale E-metric Cross Sample Entropy (MECSEn) has been proposed and is used to measure complexity between arrhythmia subjects, named atrial fibrillation (AF) and congestive heart failure (CHF) and healthy subjects at multiple scales. Besides short series data and undefined value, MECSEn has come up with a very new concept of banishing the use of a large number of scale factors for evaluating the complexity between two different interval series across multiple scales. It makes the proposed algorithm less time consumer. Both measures have found subjects derived from AF behave as white noise and subjects derived from CHF behave as pink noise. The t test validates MCSEn and the proposed algorithm, MECSEn by providing p < 0.00001. Moreover, MCSEn and MECSEn algorithms are compared with multiscale sample entropy algorithm (MSEn) which uses single cardiac series to evaluate complexity of healthy and arrhythmia subjects.