Multilead estimation of T-wave alternans in the ECG using Principal Component Analysis

Several methods have been proposed to automatically detect and estimate T-wave alternans (TWA) in the electrocardiogram. All of them operate on a single-lead basis, and one major drawback is their low sensitivity to low-level alternans. Our group proposed a novel multilead TWA analysis scheme that combines Principal Component Analysis (PCA) with a technique based on the Generalized Likelihood Ratio Test (GLRT). In this work, we evaluate the estimation performance of the proposed scheme, and we compare it to a single-lead analysis scheme. Simulation results show that the proposed multilead scheme allows the estimation of TWA with a SNR 25 dB lower than a single-lead scheme for a given level of accuracy.