Biostatistics Seminar: Learning Retinotopic Maps: Diffusion Models, Variational Uncertainty, and Bayesian Inference on Cortical Surfaces

An introduction to a statistical framework for improving retinotopic mapping using variational uncertainty estimation, hierarchical Bayesian models, and optimal-transport geometry. These methods enhance boundary delineation, surface alignment, and cross-subject inference from clinical-grade fMRI. The work illustrates how modern statistical tools can advance the precision and reliability of human visual-cortex mapping.

Presenter: 
Yalin Wang, School of Computing and Augmented Intelligence