Zijian Dong

97% of 184 high-impact biomedical AI studies compare cross-validated models with tests assuming fold independence; in 69% of re-analyzable studies a significant comparison does not survive correction.

This study examines how voluntary and enforced exercise affect brain structure, function, and behavior in Alzheimer's disease models using neuroimaging techniques.

WMH-DualTasker is a weakly-supervised deep learning model that segments white matter hyperintensities and predicts visual ratings, outperforming supervised methods on benchmark datasets.