Inducibility Tests of Atrial Fibrillation by Automata-based Efficient Simulations

This study extends an existing cellular automaton (CA) model to simulate atrial electrophysiology across diverse stages of Atrial Fibrillation (AF) by integrating tissuespecific properties. The CA model, initially developed by the authors, undergoes meticulous training using biophysical simulations across varying degrees of electrical remodeling within a limited atrial tissue domain, accounting for tissue differentiation and anisotropic propagation through pacing simulations.Validation involved comprehensive comparisons with realistic 3D atrial model to assess pro-arrhythmic behaviors.Results demonstrate remarkable congruence between the extended CA model and the biophysical solver, with CA exhibiting an 80% accuracy rate in predicting AF inducibility and a notable decrease in computing time compared to the conventional biophysical solver: 1.96 sec vs 126.75 sec for a 284.578 nodes geometry.In conclusion, the extended CA emerges as an efficient and valid alternative for simulating atrial electrophysiology across various stages of AF, providing a valuable tool for personalized therapy planning through digital twin simulations and facilitating more targeted and effective treatment strategies.