Direct Mapping from Body Surface Potentials to Cardiac Activation Maps Using Neural Networks

Cardiac mapping is an important diagnostic tool to understand cardiac arrhythmias caused by an electrical disorder. One of its procedures is the activation mapping which consists on generating a map of the depolarization sequence. This map is constructed from the electrograms (EGMs). These latter signals are either recorded invasively through catheterization or non-invasively reconstructed using inverse methods. Our purpose is to evaluate the performance of artificial neural networks on reconstructing activation maps non-invasively directly from body surface potentials (BSPs) without providing the electrograms. We suggest a basic architecture of a neural network. The method is evaluated using simulated atrial data.