Apoorva S. Kelkar

This paper develops a machine learning model to predict post-stroke aphasia speech performance using multimodal data, achieving high accuracy with explainable features.

The paper uses support vector regression to predict individual cognitive control behaviors from functional connectivity between brain networks during tasks and rest.

This study uses simulated attacks on brain networks to assess how stroke lesions affect network properties and aphasia severity, finding that global network measures explain some variance in aphasia symptoms beyond lesion volume.

The paper uses GluCEST MRI to detect glutamate concentration changes in the motor cortex after cTBS, finding decreased ipsilateral signal but no consistent contralateral or sham effects.

The paper investigates how structural disconnections in the posterior medial frontal cortex affect speech error monitoring in post-stroke aphasia patients.

This study examines how post-stroke bypasses in brain networks predict variance in language functions, finding that bypass length and overlap predict lexical production and auditory comprehension.