Effect of different ventricular arrhythmia origin on cardiac sounds variability using m-mode signals representation

In recent years, interest has been paid to the identification of the arrhythmia anatomical origin in ventricular tachycardia from available information in the patient, including ECG morphology and electrograms stored in implantable devices. In this work, we scrutinize the differences in morphology from cardiac sounds. For this purpose, we compared the spatial and temporal similarity in cardiac sounds recorded during electrical intracardiac stimulation, in terms of the so-called M-mode representation for cardiac signals. A protocol was set in 5 patients referred to electrophysiological study. Cardiac pacing was made at 400 ms from different points in the left ventricle. During stimulation, we recorded: (1) the ECG signals from 12 leads; (2) the spatial coordinates for the pacing site, by using an intracardiac navigation system; (3) the heart sounds, by using a high-fidelity recording system. After visual-supervised correction, M-mode maps represented an aligned and clearly represented the variations due to distance and substrate in a consistent view. The same signal processing was followed for the cardiac sounds, showing that the ECG-based alignment was not sufficient, and that the sound signal represents a very much lower signal to noise ratio domain.