Engineering Clinical-Grade Biomedical Data Systems
Wednesday, February 18, 2026
2:30 PM-4:00 PM
BIOMED Seminar
Title:
Engineering Clinical-Grade Biomedical Data Systems: Multimodal Integration, Informatics Infrastructure, and Deployable AI
Speaker:
Dominique Duncan, PhD
Associate Professor of Neurology, Bioengineering, Informatics, and Neurosurgery
Perelman School of Medicine
School of Engineering and Applied Science
University of Pennsylvania
Details:
Biomedical research now routinely generates large, heterogeneous datasets spanning neuroimaging, electrophysiology, biosignals, wearables, and electronic health records. Despite advances in machine learning, most analytic approaches remain dataset-specific and fail to translate across institutions or into clinical workflows. The core barrier is not model performance, but the absence of scalable biomedical data infrastructure capable of supporting reproducible computation and deployable decision-support.
This talk presents an engineering and informatics framework for building clinical-grade biomedical data systems. I will describe methods for multimodal data harmonization, feature representation, and cross-site integration, including automated pipelines for acquisition standardization, quality control, and provenance tracking. Algorithmic approaches such as signal processing, geometry-based modeling, and interpretable machine learning applied to neuroimaging and physiological data will be discussed.
A major focus will be the design of research data archives that support collaborative analysis across institutions. I will present architectures for storing and querying multimodal datasets, linking imaging and biosignal data to clinical records, and enabling reproducible analyses through shared computational environments.
Biosketch:
Dominique Duncan, PhD, is an Associate Professor of Neurology, Bioengineering, Informatics, and Neurosurgery at the University of Pennsylvania. Her research focuses on biomedical data infrastructure and multimodal data integration for neurological diseases. She develops scalable informatics platforms that standardize, harmonize, and analyze large heterogeneous datasets, including neuroimaging, electrophysiology, and clinical records.
Dr. Duncan's laboratory designs algorithms and computational tools using machine learning, signal processing, and geometry-based modeling, embedded within interactive analytic environments that enable reproducible and collaborative research. Her work emphasizes translation: transforming large-scale biomedical data into deployable clinical decision-support systems and operational research platforms.
Contact Information
Carolyn Riley
cr63@drexel.edu