Array Processing in Atrial Fibrillation

Abstract

Atrial fibrillation (AF) is the most commonly occurring arrhythmia in clinical practice and can have a significant impact on the current and future wellbeing of the patient. By placing an unipolar sensor array directly on the epicardium during an open-heart surgery to measure the electrical activity of the atrium, more insight about this disease can be obtained. One method to obtain these insights is by applying common signal processing models to the measured electrogram (EGM). This thesis further investigates this application and argues why the most common array processing signal models are fundamentally incompatible with this application and proposes two different signal models to rectify the identified discrepancies. The proposed signal models are subsequently analyzed and the conclusion is drawn that both novel signal models better fit the EGM signals, but one signal model in particular shows promising results. This potential is exemplified by using this signal model to formulate a novel LAT estimation technique that can compete with state of the art LAT estimation methods in terms of estimation accuracy and execution time. This result shows the potential of the proposed signal model and opens the door to explore more applications in the future.