Physics-informed machine learning for functional MRI modeling

Abstract:

Extracting functional tissue properties from magnetic resonance imaging data is a complex endeavor. Yet, it offers unique insights into physiological and pathological processes non-invasively. The measured MRI signal is a function of a multitude of effects affecting the local magnetic field, and current models have only limited capacity to consider interactions. This talk will focus on the development of novel deep learning-based frameworks which encode physical laws and architectural tissue properties to predict the MRI signal dynamics.

About Lukas:

Lukas is a postdoctoral researcher and radiology resident at the German Cancer Research Center Heidelberg. His research focuses on modeling and development of functional MRI techniques as well as statistical methods for forecasting immunotherapy response of solid tumors.