Collaborative Research: FDT-BioTech: Advancing Alzheimer's Disease Understanding and Treatments through Digital Twin Modeling

About This Grant

This project develops a new framework for digital twin modeling of Alzheimer’s disease (AD), combining clinical data, biomedical research, and advanced computational methods to support personalized medicine. A digital twin is a computational replica of an individual’s health state, enabling the prediction of disease progression and the evaluation of treatment options tailored to the patient. The project contributes to national efforts in healthcare innovation by addressing the urgent need for a better understanding, prediction, and treatment of Alzheimer’s disease, which affects millions of Americans. This work also advances the broader field of personalized medicine by demonstrating how digital twin tools, powered by large language models, machine learning, and causal inference, can accelerate discovery and improve health outcomes. In addition, the project supports interdisciplinary collaboration across artificial intelligence, mathematics, and medicine, while offering new training opportunities for students in data science, modeling, and biomedical research. This project builds a unified modeling framework for population-based and personalized digital twins of AD. The approach uses large language models (LLMs) to extract causal networks of AD biomarkers from scientific literature and combines this with clinical data to generate personalized predictions. Conformal prediction techniques are applied to quantify uncertainty in model outputs, and optimization under limited data is achieved by integrating gradient-based learning with LLM-guided parameter search. The digital twin models simulate disease trajectories and support digital clinical trials. Treatment planning is formulated as a Markov Decision Process and solved using deep reinforcement learning to identify optimal, individualized therapeutic strategies. The framework integrates causal modeling, machine learning, generative AI, and decision theory, advancing both the science of Alzheimer’s disease and the computational tools for biomedical digital twin development. While centered on AD, the methods are generalizable and contribute broadly to AI-enabled modeling under data constraints in biomedical research. This award by the Division of Mathematical Sciences in the Mathematical and Physical Sciences Directorate is jointly supported by the Office of Advanced Cyberinfrastructure in the Computer and Information Science and Engineering Directorate. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.