Discrete wavelet analysis of the signal-averaged high-resolution electrocardiogram

The purpose of this study was the evaluation of the discrete wavelet transformation (DWT) applied to signal-averaged high-resolution ECG signals with the aim of separating healthy volunteers and patients with coronary artery disease (CAD) and inducible ventricular tachycardia. Especially, the usefulness of the DWT to the determination of the QRS-duration (QRSD), which is the main parameter of the time-domain. Late potential analysis, was under investigation. Using DWT processing, the classification performance (leaving-one-out simulation) based on QRS duration increased by approximately 10% in accuracy (in each detail signal) as compared to the standard VLP analysis (51 healthy volunteers vs. 44 patients with coronary artery disease). This was due to an increase of sensitivity. The authors conclude that the type of bandpass filtering realized by DWT improves the extraction of pathological intra-QRS micropotentials in patients with malignant arrhythmias.