An anatomy-guided whole-heart framework for computational modelling of the complete human cardiac conduction system

Introduction Computational models of cardiac electrophysiology are increasingly used to investigate arrhythmia mechanisms and evaluate therapeutic interventions, yet most existing frameworks omit or oversimplify the cardiac conduction system, limiting their physiological fidelity and clinical applicability. Methods We present a subject-specific, anatomy-guided computational model of the complete human cardiac conduction system (CCS) integrated within a four-chamber heart and torso geometry derived from magnetic resonance imaging. The model comprises a three-dimensional sinoatrial node with discrete exit pathways, interatrial conduction pathways including Bachmann’s bundle and posterosuperior connections, an atrioventricular node with explicit fast and slow pathway anatomy, His bundle, left bundle branch with trifascicular configuration, right bundle branch, and fascicle-anchored Purkinje networks. The unknown ramification topology of the Purkinje network was approximated by a partitioned subendocardial surrogate with region-specific isotropic conduction velocities. Whole-heart electrophysiology was simulated using a reaction-eikonal formulation with region-specific ionic models, enabling near real-time computation of cardiac activation and the associated 12-lead electrocardiogram. Results Simulated activation sequences reproduced physiologically consistent atrial and ventricular depolarisation patterns, including right atrial breakthrough along the crista terminalis, Bachmann’s bundle-mediated interatrial conduction, dual-pathway atrioventricular nodal activation, and near-synchronous ventricular depolarisation initiated at the fascicle-myocardial junctions. The simulated 12-lead electrocardiogram (ECG) demonstrated close qualitative agreement with the measured recording, with particularly strong agreement in QRS morphology and overall correspondence in P-wave morphology. Discussion This framework provides an anatomy-guided and computationally tractable platform for investigating conduction system physiology, bundle branch block phenotypes, and conduction system pacing strategies.