Skip to Main Content


Delayed Cerebral Ischemia After Subarachnoid Hemorrhage

The Kim Lab studies complications after subarachnoid hemorrhage. One major complication is a type of delayed stroke called delayed cerebral ischemia, which occurs after bleeds in the brain caused by an aneurysm rupture. This complication leads to even worse disability in this population. We are interested in using brain physiology and brain imaging measures to identify the patients at highest risk for this complication. Our goal is to use these biomarkers to develop treatments that reduce the impact - and ultimately prevent - this complication from developing.

Post-Traumatic Epilepsy

Epilepsy is a complication that develops after traumatic brain injury in up to 20% of patients. Despite knowing the increased epilepsy risk that patients have after trauma, there are no treatments available to prevent this complication from developing. We aim to combine EEG and MRI to identify patients at high risk for post-traumatic epilepsy. By identifying these features, we can begin to develop and test preventative treatments targeted at this high-risk population.

Post-Ischemic Stroke Epilepsy

Similar to post-traumatic epilepsy, ischemic stroke also increases the risk of epilepsy. Most ischemic stroke patients get neuroimaging, but few get EEG studies after their stroke. By focusing on available neuroimaging characteristics that increase the risk for post-ischemic stroke epilepsy, we may be able to target our EEG studies to that population to develop a multi-modal prediction algorithm for post-ischemic stroke epilepsy prediction.

NLP-Derived Neuroimaging Abnormality Identification

Neuroimaging text reports carry rich information that describes the abnormalities observed in a brain scan. With the increase in large-scale, multi-center research, the labor required to manually review each of these scans to classify abnormalities becomes progressively untenable. Training natural language processing (NLP) algorithms to assist with this automated identification would significantly expedite and expand our ability to assess imaging outputs on a large scale.