Estimation of high-density activation maps during atrial fibrillation

The study of activation maps using multi-electrode arrays (MEA) can help to understand atrial fibrillation (AF) mechanisms. Activation mapping based on recorded unipolar electrograms (u-EGM) rely on the local activation time (LAT) detector, which has a limited robustness, accuracy, and generally requires manual post-edition. In general, LAT detection ignores spatiotemporal information about activation and conduction conveyed by the relation between signals of the MEA sensor. This work proposes an approach to construct activation maps by simultaneous analysis of u-EGMs from small clusters of MEA electrodes. The algorithm iteratively fits an activation pattern model to the acquired data. Accuracy was evaluated by comparing with audited maps created by expert electrophysiologists from a patient undergoing open-chest surgery during AF. The estimation error was -0.29 ± 6.01 ms (236 maps, 28369 LATs) with high correlation (ρ = 0.93). Therefore, activation maps can be decomposed into local activation patterns derived from fitting an activation model, resulting in smooth and comprehensive high-density activation maps.