A Machine Learning Based Approach for Localization of Atrial Tachycardia Origin

Ablation procedures targeting Atrial Tachycardia (AT) can be drastically facilitated if the origin of the abnormality is located in advance using electrocardiogram (ECG) signals. The ECG recordings contain so called P waves, which represent overall summary waves generated by the atrial depolarization. Previous work has shown the possibility to predict the origin of localized AT (excluding flutters) based on P wave morphology segmented from the ECG. The present study aims to develop a machine learning algorithm that detects the likely origin of localized AT based both on P wave characteristics extracted from ECG and signals recorded in the Coronary Sinus (CS).