Parts | Days | Selection | Search | Updates | Downloads | Help
DY: Fachverband Dynamik und Statistische Physik
DY 45: Poster Session: Nonlinear Dynamics, Pattern Formation, Data Analytics and Machine Learning
DY 45.12: Poster
Thursday, September 8, 2022, 15:00–18:00, P2
Reconstructing spatiotemporal chaos in three-dimensional excitable media based on surface data — •Inga Kottlarz1,2, Sebastian Herzog1,3, Roland Stenger1,2, Baltasar Rüchardt1,4, Stefan Luther1,4,5, and Ulrich Parlitz1,2,4 — 1Max Planck Institute for Dynamics and Self-Organization, Göttingen, Germany — 2Institute for Dynamics of Complex Systems, Georg-August-Universität Göttingen, Göttingen, Germany — 3Department for Computational Neuroscience, Third Institute of Physics - Biophysics, University of Göttingen, Göttingen, Germany — 4German Center for Cardiovascular Research (DZHK), partner site Göttingen, Robert-Koch-Str. 42a, 37075 Göttingen, Germany — 5Institute of Pharmacology and Toxicology, University Medical Center Göttingen, Robert-Koch-Str. 40, 37075, Göttingen, Germany
The cardiac muscle is an excitable medium that can exhibit complex dynamics, including spatiotemporal chaos associated with (fatal) cardiac arrhythmias. While mechanical motion within the myocardium can be observed with ultrasound, there are no noninvasive techniques (to date) to measure the electrical state within the tissue. To overcome this limitation of observable quantities, we address the task of predicting the electrical state inside the heart from surface data using data-driven reconstruction by means of artificial neural networks. We study the feasibility of this approach in a homogenous and isotropic excitable medium with spatiotemporal dynamics in three spatial dimensions, applying and comparing different deep learning methods (i.e. LSTM, Convolutional Autoencoder, ...).