Enhancing Biophysical Models of Stroke with Artificial Intelligence

This project will combine techniques from biophysics and machine learning (AI) to enhance the predictive capabilities of a model of embolic stroke. Strokes occur when the blood supply to the brain is interrupted. This interruption often occurs when emboli (e.g. blood clots) move through and then block the vasculature. The

final resting position of the embolus is related to the damage caused by the stroke. A Monte Carlo simulation approach developed at the Open University can be used to determine the pattern of embolus resting places [1]. There are several open research directions relating to the physics and physiology of the model, including the role of patient specific vasculatures in

predicting stroke outcomes. For example, only 50% of the population have all the possible connecting blood vessels, so identifying variants and implementing them in models is important. These lead to a range of modelling and analysis options for the project. We anticipate a large modelling component to this project, and you will work on

enhancements to a recently developed stroke model, by improving both the underlying physics of embolus motion and the structure of the major vasculature to represent a wider range of the population. We anticipate that you will use and develop AI methods to automatically identify stroke lesions and the major cerebral vasculature from medical

images. You will create and manage a large data set to support and collaborate with other related projects.