Music pseudo-bispectrum detects ECG ischaemia

Abstract

Up to 30% of patients with suspected or known coronary artery disease are unable to perform an adequate exercise stress test due to poor physical condition. It is beneficial to be able to detect ischaemic heart diseases when these do not manifest themselves as ST depression or elevation. In this paper, a subspace-based MUSIC algorithm is used to examine normal and abnormal episodes from the same patient. The analysis reveals abnormal peaks in both of these episodes as opposed to the frequency analysis of normal episodes taken from normal records. Results presented include 46 records from the MIT-BIH databases. High resolution is obtained using the MUSIC algorithm compared to the maximum entropy method (MEM). The accuracy, sensitivity and specificity of the proposed algorithm are 82.8%, 87% and 90% respectively. This leads to the possibility of the detection of ischaemia without the need for an exercise test.