Detection of Subtypes of Prodromal Alzheimer's Disease through Deep Learning-Based Vascular Signatures

Project Summary:

Wisconsin faces a critical health challenge, with 10.9% of seniors affected by Alzheimer's disease (AD) and 22% experiencing mild cognitive impairment, often a precursor to AD. With the aging population expected to grow, the resulting healthcare and economic strain will increase statewide. Early detection and accurate classification during the prodromal stage of AD offers the best chance for effective intervention, yet current diagnostics have limitations. Traditional methods, focusing on amyloid, tau, and neurodegeneration biomarkers, can overlook up to 30% of patients with significant vascular contributions to cognitive decline, leading to delayed treatment and poorer outcomes. This project addresses this gap by harnessing advanced technology to detect neurovascular dysfunction, which often emerges before traditional AD biomarkers. By integrating multi-modal neuroimaging, computational modeling of blood flow, and deep learning, the project team aims to automatically classify prodromal AD through neurovascular signature analysis. This approach will help identify high-risk patients earlier and recognize unique vascular subtypes that could respond to tailored treatments.

Leveraging the team's expertise in AI-driven neuroimaging and computational brain modeling, they aim to transform the diagnosis of prodromal AD in Wisconsin, enabling earlier and more accurate detection. The framework lays the groundwork for personalized interventions, promising better outcomes for thousands of families and easing the state’s healthcare burden through timely, targeted treatment.