Dementia is a subject that is well worth doing research about. Worldwide, at least 50 million people have some form of dementia and in the time it takes to read this sentence out loud, another person somewhere in the world has been diagnosed with it. Worldwide dementia care is estimated to cost upwards of US$1 trillion. 70% of dementia is caused by Alzheimer's disease (AD). Better diagnostics in AD will improve the lives of the patients and their families. It is exactly what this project aims to achieve.
The brain is a network. Unraveling the structure of this network is vital for AD diagnostics. In the first part of the project we will use Bayesian methods to discover this structure. In the second part of the project we will apply Graph Neural Networks (GNN) on this inferred network. GNNs are a class of deep learning methods that have been adapted to leverage the structure and properties of graphs and have gained incredible success. Combining Bayesian methods (subproject 1) with GNNs (subproject 2) will enable us to capture model uncertainty which is crucial for the clinical assessment of patients. Although our main focus is AD diagnostics, our research can be applied in any field that deals with a high number of variables, such as economics and genetics.