Multifractal analysis of human synchronous 12-lead ECG signals using multiple scale factors

Abstract

Life is one of the most complex nonlinear systems and heart is the core of this lifecycle system. Electrocardiogram (ECG) signals taken from healthy adult subjects have been found to characterize multifractality. In this paper, multiscale analysis method was introduced to find the most effective parameters for expressing the complex dynamic characteristics during heart electrical activities. We then investigated the multifractal singularity spectrum area of synchronous 12-lead ECG signals from healthy human subjects and those with different clinical syndromes. The spectrum areas have spatial distribution along with scale factors, which is higher in the middle and lower on both sides and is not related to data length. The statistical results suggest the arithmetical mean value of the area of the 12 leads ECG signals is obviously small for myocardial infarction (MI) sufferer and large for healthy young, while the dispersing degree of the area of the 12 leads ECG signals is apparently large for MI sufferer and small for healthy young. As for the other crowds (e.g., the ischemia sufferer), these two values are almost of middle magnitude. Through those individual discrepancies, we can find some effective approaches to distinguish among healthy persons and the heart diseased.