Jingwen Cheng

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.

The paper evaluates how different censoring strategies affect cortical parcellations and TMS targets in rs-fMRI, finding that lenient censoring yields better results than strict censoring.

This paper develops a tree-based algorithm to identify personalized TMS targets for depression and anxiety by optimizing functional connectivity and scalp proximity.