M.Sc. Stephanie Appel

B.Sc. Carl Enslin

Motivation

Understanding electrical excitation patterns in the heart is crucial for diagnosing and treating rhythm diseases such as atrial fibrillation (AF). Computer models that simulate cardiac electrophysiology are often used to investigate these patterns. These simulations of cardiac electrophysiology are powerful tools, helping to explore how variations in electrical activity can lead to these arrhythmias. However, they can also be extremely time-consuming due to their high computational demands.
The DREAM (Diffusion-Reaction-Eikonal-Alternant-Model) provides a faster alternative to other models without compromising the essential details required for studying heart diseases. This makes DREAM a promising tool for both research and potential clinical applications. To unlock its full potential, however, the model requires further fine-tuning.

Project Description
This project focuses on improving the computational efficiency of the DREAM by tuning its hyperparameters. The aim is to optimize simulation speed while maintaining model accuracy. Therefore, it is necessary to analyze how dynamic changes over time and variations in heart tissue commonly observed during AF - particularly in action potential duration (APD) and the conduction velocity (CV) - affect the model’s computational performance. The goal is to identify optimal configurations for different simulation scenarios of arrhythmic behavior, balancing accuracy and efficiency. This may also involve modifying the model to account for adaptive parameter changes over time.