Wigner distribution's cross terms characterization to detect patterns of Ventricular Late Potentials

Abstract

Thousands of deaths can be prevented detecting Ventricular Late Potentials (VLPs) in patients who have suffered infarcts heart attacks. The current method to detect VLPs based on time domain analysis presents low predictive values. VLPs are very difficult to detect because have amplitudes between 5 and 20 μV and their spectrum is difficult to distinguish from noise spectrum. The Wigner distribution (WD) is used to analyze pseudo-stationary signals, such as the VLPs. The WD generates artifacts known as cross terms (CTs), which are considered undesirable in some cases, however can be used to infer the presence of very small signal terms, in this case VLPs terms. The purpose of this research is to characterize the CTs generated by VLPs presence to improve their detection and prevent the sudden cardiac death.