Abstract

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, with long-term persistence (LTP-AF) posing a significant clinical challenge. The underlying mechanisms involve complex alterations in atrial electrophysiology and structure, particularly the role of fibrosis and patient-specific anatomical variations. This study aimed to develop a novel, patient-centered computational model to compare excitation conduction in normal atrial tissue against the heterogeneous substrate of LTP-AF. High-resolution, patient-specific atrial geometries were reconstructed from MRI data, incorporating realistic distributions of fibrotic tissue. The model integrated detailed cellular electrophysiology, focusing on membrane ion channel dynamics and gap junction function, to simulate electrical wave propagation. Quantitative assessment revealed a significant, localized reduction in conduction velocity and increased wavebreak incidence in the LTP-AF substrate, particularly in regions of high fibrosis density. This structural and electrical remodeling created a highly heterogeneous substrate, facilitating the formation and stabilization of multiple, rapidly-rotating re-entrant wavelets. This computational approach provides a powerful, non-invasive tool for investigating the complex interplay between cellular membrane function, structural remodeling, and macro-scale conduction abnormalities in AF, offering a platform for personalized risk stratification and optimizing ablation strategies.