Pathology Dynamics

Pathology dynamics is the study of how disease processes initiate, interact, and evolve over time across biological scales—from molecular signals to systems-level function and behavior. By modeling trajectories rather than static snapshots, we can forecast progression, identify windows for prevention or therapy, and personalize decisions that improve outcomes, especially in neurological disease.

Guided by this dynamic view and led by Principal Investigator Cassie Mitchell, PhD, we study how disease unfolds over time and how to intervene. Our artificial intelligence applications and computational models span biomedical systems and specialties—neurology, cardiovascular disease, cancer, infectious disease, autoimmune disease, and more. Our core focus is on neurological disease, including Alzheimer’s disease, amyotrophic lateral sclerosis, Parkinson’s disease, and spinal cord injury, where we aim to transform diagnosis, prognosis, and therapy selection.

We develop innovative computational approaches that elucidate pathology dynamics to tackle the 3 C’s: identification of disease causes, development of cures, and optimization of patient care. Our goal is to map pathology dynamics from molecules to behavior to enable earlier detection, personalized risk prediction, and data-driven clinical decisions.

We build explainable artificial intelligence, literature-informed discovery pipelines, and computational modeling to fuse complex biomedical data into actionable, predictive insights that accelerate early-stage clinical research and real-world translation. Using natural language processing (NLP), knowledge graphs, large language models, and multimodal machine learning, we mine biomedical literature and clinical text to build computable knowledge; link mechanisms to phenotypes; and integrate imaging, electrophysiology, wearables, omics, and cognitive/behavioral measures to model individual disease trajectories. Across all projects, we prioritize clinically grounded, explainable AI, FAIR/open data practices, and rigorous validation to ensure models are trustworthy and translatable.