Sook‐Lei Liew

This study compares data-driven and theory-based biomarkers for predicting chronic motor outcomes after stroke, finding data-driven models perform better, especially when using structural disconnection measures.

This study finds that corpus callosum microstructure predicts bimanual motor performance in chronic stroke survivors, with associations beyond primary sensorimotor regions.

This study finds that smaller ipsilesional hippocampal volume is associated with worse upper limb sensorimotor impairment in chronic stroke patients, with a stronger effect in women.

This study shows that global brain health, measured by brain age, influences post-stroke sensorimotor outcomes and mediates the effect of lesion damage.

This study finds that smaller spared subcortical nuclei are linked to worse sensorimotor outcomes after stroke, varying by time since stroke and type of impairment.

This study uses machine learning to analyze 608 participants and identifies two factors improving neurofeedback performance: pre-training no-feedback runs and training patients over healthy participants.

This study compares a new convolutional neural network-based hippocampal segmentation method (Hippodeep) with existing methods on stroke MRIs, finding Hippodeep more accurate in quality control and spatial accuracy.

This study investigates whether pre-training brain activity can predict success in real-time fMRI neurofeedback learning across 24 studies and 401 participants, finding a weak correlation but no consistent predictors.

This study compares the effects of high-definition tDCS over the motor hotspot and premotor cortex on motor network excitability, finding greater responsiveness and fMRI complexity changes with premotor stimulation.

The paper compares automated lesion segmentation methods for chronic stroke T1-weighted MRI data, finding LINDA performs best overall while Gaussian Bayes has highest recall.

The paper identifies distributed brain structural patterns that predict motor cortical excitability and inhibition in healthy volunteers.

The paper introduces ATLAS, a dataset of 304 T1-weighted MRIs with manual lesion segmentations to improve lesion segmentation algorithm accuracy.