Characterization of the Electrophysiological Characteristics of Chronic Atrial Fibrillation for Efficient Simulations.

Atrial biophysical simulations have the potential to enhance outcomes by enabling the simulation of pharmacological and ablative strategies. However, the high computational times associated with such simulations render them unsuitable for diagnostic purposes. To address this challenge, discrete models such as cellular automata (CA) have been developed, which consider a finite number of states, thus significantly reducing computational times. Yet, there is a pressing need to determine whether CA can replicate pathological simulations with accuracy. The analysis of simulations under different degrees of electrical remodeling shows an expected increase of Action Potential Duration (APD) with the previous Diastolic Interval (DI) interval, indicating short-term memory of atrial cardiomyocytes: shorter APD0 provoked shorter APD+1, and previous DI has a similar effect on APD+1. Independent prediction using both APD0 and DI was found to provide a far better estimation of APD+1 values, compared to relying on DI alone (p<<0.01). Finally, the CA models were able to replicate reentrant patterns and cycle lengths of different states of atrial remodeling with a high degree of accuracy when compared to biophysical simulations. Overall, the use of atrial CA with short-term memory allows accurate reproduction of arrhythmic behavior in pathological tissue within a clinically relevant timeframe.Clinical Relevance- Discrete electrophysiological models simulate pathological self-sustained arrhythmias in diagnostic times.