Self-organizing network simulation of cardiac electrical dynamics.

Network provides a low-dimensional representation of the heart through a sparse adjacency matrix, which ushers in a new opportunity to conduct cardiac simulation. We discovered that a self-organizing network encodes and resembles complex heart geometry. This, in turn, helps characterize the structure-function relationship of the heart through network theory. However, very little has been done to investigate the simulation of electrical activity on a self-organizing network. Thus, this paper presents a new self-organizing network approach for simulating cardiac electrical dynamics. We formulate and solve reaction-diffusion equations on the self-organizing network to simulate the propagation and turbulent behavior of electrical waves. The proposed methodology is evaluated and validated on both 2D cardiac tissues, consisting of healthy and infarcted cells, and the whole heart. Experimental results show that the proposed approach not only yields a compact network representation that resembles the heart geometry but also provides an effective simulation of spatiotemporal dynamics when benchmarking with traditional finite element method simulations.