Hongjian He

The paper tracks five-year longitudinal data of children learning abacus-based mental calculation to show how sustained practice reconfigures whole-brain states and enables complex skill acquisition.

The paper develops an individualized TMS targeting approach for ASD based on functional connectivity to the posterior cingulate cortex, showing symptom reduction in a case series.

This study evaluates the reliability and validity of multi-band multi-echo fMRI for mapping individual differences in spontaneous brain activity, showing improved reliability and detectable differences in resting states.

AmygdalaGo-BOLT improves amygdala segmentation accuracy and efficiency in pediatric MRI scans across multiple centers and ages.

SACNet corrects susceptibility artifacts in dMRI scans using a registration network with neuroanatomical constraints, showing improved performance across diverse datasets.

This paper introduces a deep learning framework to harmonize multisite neuroimaging data by removing site effects while preserving biological features, showing improved performance over traditional methods.