Non-phase-locked arrhythmogenic signal detection in high resolution ECG by phase-ordering in the wavelet domain

The aim of this study is to detect non phase-locked arrhythmogenic signals (NoPLAS) in high resolution ECG. The analysis is performed on a beat to beat basis by a novel pseudo-continuous wavelet transform technique. In order to assess the presence of NoPLAS, we propose the phase-ordering factor, which gives a representation of the phase distribution in a given time-frequency region. A mathematical model of small variable signal (NoPLAS) in a large repetitive waveform (QRS complex) has been developed and computer simulation tests have been conducted in order to assess the effectiveness of the phase-ordering factor for NoPLAS identification. The technique has been tested also on real signals recorded from normal subjects and post myocardial infarction patients prone to ventricular tachycardia (VT). The NoPLAS has been detected in specific frequency bands (60-150 Hz) in a significantly higher number of VT patients (80%) as compared with normals (15%).