Arrhythmia classification using two intercardiac leads

A real-time, two lead intracardiac electrogram (ICEG) classification algorithm is presented. The algorithm combines timing and morphology analysis of atrial and ventricular channels of the ICEG signal. Timing analysis is used to identify most tachycardias, but for situations where timing cannot be used to identify the tachycardia, a neural-network-based morphology classification algorithm is used. The algorithm achieved 99.6% correct classification on a database of ICEG signals containing 11290 QRS complexes from 61 patients. An implementation of the algorithm in VLSI is described