Virtual bipolar and laplacian electrodes for activation map construction in ECGi

Activation map computation represents an important post-processing step in ECGi.---We sought to determine if inverse reconstructed potential gradient in the direction normal to the epicardial surface (virtual bipolar electrode) and the laplacian of the potential along this surface (virtual laplacian electrode) would enhance non-invasive activation maps compared to the standard approach of using reconstructed unipolar electrograms (virtual unipolar electrode). After analytical derivation, virtual unipolar, bipolar and laplacian electrograms (EGMs) were computed for all epicardial nodes. Local activation time (LAT) at each node was determined using the maximal downstroke (maximal negative ds/dt) for the 3 EGMs. The techniques were evaluated in-silico using simulated data of 7 different pacing sequences and on clinical data from 12 different activation patterns. The use of virtual bipolar EGMs for activation map construction resulted in a significant relative reduction of error of 8% compared to unipolar EGMs on clinical data. A similar albeit non-significant trend existed in simulated data. Virtual Laplacian electrode use resulted in a nonsignificant enhancement of activation maps for both datasets.