Synthetic Data Generation to Investigate the Dynamics of Wave-Fibrosis Interactions in Cardiac Arrhythmias

Fibrosis represents a key factor in the development and maintenance of cardiac arrhythmias. Computational models may help improving our understanding of wave-fibrosis interactions in excitable tissues, offering a controlled environment to dissect the effects of different factors and mechanisms on wave propagation dynamics. This study describes a computational framework, which combines an atrial tissue spheric model, including a fibrotic tissue generator, with a frequency-domain analysis of the synthetic signals generated by the model. The framework was applied to characterize the dynamics induced by different fibrosis patterns in an electrically-remodelled tissue, characterized by different arrhythmic mechanisms: i.e., re-entrant activity in the presence (with focal source, WFS) and in the absence (no focal source, NFS) of a localized focal source. 150 fibrotic models with different densities and dimensions of fibrotic elements were generated, atrial electrical activity was simulated in the WFS and NFS conditions, and 144 synthetic bipolar electrograms were generated for each model and condition. 85 models produced stable activity in both the WFS and NFS conditions and the corresponding electrograms were analysed in the frequency-domain to extract rate, regularity, and coupling indices. Generalized linear models were used to investigate the relationship between frequency-domain indices and fibrosis parameters. In both the WFS and NFS conditions, the rate displayed a statistically-significant decrease (p<0.0001), and the regularity and coupling a statistically-significant increase (p<0.001) with increasing fibrosis density. However, the increase in regularity and coupling was much smaller in the WFS than NFS condition, suggesting that the interaction of the focal source with the re-entrant waves augmented the complexity of propagation patterns and reduced fibrosis stabilizing effects. These results suggest the complexity of wave-fibrosis interactions and their dependence on the arrhythmic mechanisms. Our framework may represent a useful tool to understand fibrosis effects and to translate propagation dynamics into clinically-measurable markers.