Parameter tuning for the efficient eikonal-based computation of reactivation patterns in cardiac tissue

  • chair:Computational Cardiac Modeling
  • type:Bachelor or Master thesis
  • tutor:

    M. Sc. Stephanie Appel
    M. Sc. Cristian Barrios Espinosa

  • Download as PDF

     Computer models have become a powerful tool for understanding cardiac arrhythmias. Reaction-Diffusion (RD) models can reproduce complex mechanisms leading to reentry and atrial fibrillation (AF) and are, therefore, commonly used. However, their high computational cost make them unsuitable for numerous clinical applications. Another approach for computing propagating wavefronts are eikonal models. By combining cellular models with the eikonal equation, electrophysiology during a healthy heartbeat can be computed up to three orders of magnitudes faster compared to RD models, mainly due to less strict mesh resolution requirements. Based on this approach, a novel method was proposed: The Diffusion-Reaction-Eikonal-Alternant-Model (DREAM). The model enables the time-efficient simulation of reactivation patterns and, therefore, facilities the study of cardiac diseases such as AF. Its development and evaluation is subject of ongoing research. So, this thesis aims to optimize the parameter settings of the DREAM to fully utilize the model’s advantages and work towards potential clinical applications.

    Snapshot of the transmembrane voltage of the RD model and the DREAM model, when inducing reentry.

    Student Project
    This project focuses on improving the computational efficiency of the DREAM. The model’s hyperparamers will be characterized in relation to computational accuracy and efficiency. It will be especially important to identify the influence of dynamic changes in time, as well as spatial heterogeneity of electrophysiological characteristics related to AF, such as the action potential duration (APD) and the conduction velocity (CV). The aim is to then find ideal parameter settings for different simulation scenarios, balancing the computational accuracy and efficiency of the DREAM. Potentially this also involves an adaptation of the current code to take into account adaptive parameter changes.

    • C++ or Python programming skill are beneficial.
    • The scope of the work can be adapted based on the type of thesis.
    The weighting of individual elements can be tailored to your expertise and goals.