Project Grant 2533996

This $224,994 Project Grant award from the National Science Foundation's (NSF) Computer and Information Science and Engineering (CISE) program supports the development of a new framework for digital twin modeling of Alzheimer's disease (AD). The project aims to combine clinical data, biomedical research, and advanced computational methods to create personalized digital twins capable of predicting disease progression and evaluating treatment options for individual patients.

The key products or services to be delivered under this award include:

• Building a unified modeling framework that uses large language models to extract causal networks of AD biomarkers from scientific literature and integrate this with clinical data to generate personalized disease trajectory predictions.
• Applying conformal prediction techniques to quantify uncertainty in the model outputs and using gradient-based learning with LLM-guided parameter search to optimize under limited data.
• Simulating disease trajectories through the digital twin models and formulating treatment planning as a Markov decision process to be solved using deep reinforcement learning to identify optimal, individualized therapeutic strategies.

This work is expected to advance the field of personalized medicine by demonstrating how digital twin tools powered by AI can accelerate discovery and improve health outcomes for Alzheimer's disease, which affects millions of Americans.