Grant: Project Grant

Alzheimer's disease (AD) is characterized by heterogeneous patient subgroups with unique progression rates and patterns, complicating diagnosis and treatment. In AD, a protein called 'tau' abnormally spread throughout the brain along brain networks, contributing to loss of brain cells and connections, and ultimately cognitive decline. As such, understanding how tau builds up and spreads is key to optimize therapy for AD. We can study network connectivity and protein accumulation using advanced brain imaging and statistical methods. In this work, we will use a novel data-driven method to study smooth spatial changes in connectivity and determine their link to pathology and cognition. Then, we will use artificial intelligence (AI) models to segregate patients into unique subgroups and predict subject-specific trajectories. Our first aim is to study how brain connectivity affects the spread of tau, in 715 participants from two large longitudinal studies. We will test if patterns of connectivity changes and protein spreading are related to patients' symptoms over time. Next, we will define unique patient subgroups using an AI subgrouping and disease staging model that will rely on multimodal imaging data. Finally, we will develop an AI model to predict patient-specific tau spread and memory decline. This study will provide an unprecedented and more complete picture of the effects of brain connectivity on tau spread. Our AI models will improve patient selection for trials, lead to more specific diagnostic criteria, and help identify high-risk individuals. Our study presents a unique and more detailed model of AD progression.