Modeling the human heart conduction network in 3D using DTI Images

Accurate modeling of the human heart’s conduction system is essential for simulating electrophysiological behavior and solving the Forward Problem in cardiac studies. Traditional methods rely on manual identification of activation points or trial-and-error construction of Purkinje networks, often lacking anatomical completeness and precision. This study introduces a novel approach using Diffusion Tensor Imaging (DTI) to extract and model the three-dimensional ventricular conduction system. Three models are developed: two manual models based on early activation zones and trabecular muscle structures, and a third model derived directly from DTI data using the Diffusion Volume (DV) quantity. The DTI-based model enables automatic identification and visualization of the Purkinje network, distinguishing it from Myocardium and coronary arteries. The resulting activation sequence closely matches known anatomical and electrophysiological data. The proposed method offers a data-driven, anatomically consistent solution to model the heart’s conduction system and improve simulation accuracy.