New bioRxiv Preprint Introduces AI Framework for Cross-Species Pupillometry
Murat Yildirim, PhD, a member of Case CCC's Cancer Imaging Program, is the senior author of NeuroPupil: A generalization-first framework for scalable and biologically informative cross-species pupillometry, a bioRxiv preprint describing NeuroPupil, a deep learning framework for high-precision pupil tracking in both mice and humans.
The study identified training and model design strategies that improved the accuracy and reliability of pupillometry across diverse datasets. NeuroPupil was tested in multiple disease settings, including glioblastoma (GBM), where it preserved biologically meaningful pupil dynamics and improved detection of disease-associated signatures.
These findings highlight the potential of AI-powered pupillometry to advance biomarker discovery and noninvasive monitoring of brain function, with possible applications in cancer research and neuro-oncology.
Additional contributors to the study include Justin Lathia, PhD, Co-leader of Case CCC's Molecular Oncology Program, the late Charis Eng, PhD, and first author Kemal Ozdemirli.