̽����ѡ

Disentangling the heterogeneity in Alzheimer’s disease (AD) will lead to a better understanding of the underlying pathophysiological mechanisms of disease, enabling precision medicine approaches. Therefore, the overall aim of this project is to unravel the complexity within AD and understand the overlap with other disorders. This will pave the way for early diagnosis and successful intervention/clinical trials. This is an innovative multimodal neuroimaging project taking advantage of machine learning to study large cohorts. As an example, we have developed the first Bayesian hierarchical clustering method, modeling longitudinal information using a clear timescale (example: time from clinical diagnosis, not just baseline scan). This enables us to separate disease staging from distinct subtypes using different imaging modalities for the first time. This will be crucial for the characterization of disease pathways and subtypes and their underling ethiology. Previous studies have  focused on defining and describing subtypes. We will also attempt to translate findings and confirm that a deeper phenotyping is crucial for both diagnosis and management of patients in the clinic as well as for obtaining an optimal response to treatment/intervention. If the project is successful, the alignment of disease modeling and the development of accurate tools for prediction will enable precision medicine approaches to facilitate the personalization of treatment.