Xia Wu

TetraFuse improves medical image classification by integrating four domains to balance accuracy and efficiency, achieving high performance with reduced computational demand.

This paper introduces a neural framework to separate subject-specific features from transient brain states in EEG signals, showing improved accuracy in biometric recognition.

This paper shows that small effect sizes in resting-state fMRI studies lead to reproducibility issues, even after multiple comparison corrections, and recommends meta-analysis of un-thresholded data for better results.