Wavelet Analysis of Cardiac Electrical Activity Signals

Abstract

To solve important problems of cardiovascular monitoring, effective algorithms for computer processing of electrocardiogram signals (ECS) should be developed on the basis of nonlinear dynamic analysis. ECS can be represented as electric excitation of the conducting nerve network of the heart (CNNH) in the form of solitons of different sizes, taking into account their polarization along the main CNNH branches. Detailed information on the electrical activity in all parts of the four-chamber heart is contained in the self-similar fractal scale-invariant CNNH structure. With the help of wavelet transform, it is possible to represent the structure of the process of excitation of CNNH segments as a system of local extrema of the wavelet diagram of ECS. The wavelet spectrum of ECS has a fractal structure in the form of self-similar waves with scaling 1/f. Each of these waves reflects the excitation of the corresponding CNNH segment. Wavelet representation of the ECS can be used as a tool for detecting various cardiovascular diseases by visualizing skeleton functions of the ECS wavelet transform.