Research
Neurological disorders are now the leading cause of illness and disability worldwide, affecting more than 3.4 billion people (Source: WHO report). Despite the large volume of data collected from individual patients — imaging, neurophysiology, and electronic health records — treatment decisions still rely heavily on visual inspection and clinical intuition. Care can vary dramatically across centers, and many patients do not receive optimal treatment.
Our research develops rigorous, quantitative methods to guide interventions and personalize therapy, validates them at scale across institutions, and deploys them as tools for multicenter clinical trials. We integrate state-of-the-art engineering, informatics, neurology, radiology, and neurosurgery to develop practical tools that improve and standardize patient care.
Our innovations cut across traditional diagnostic boundaries. We develop rigorous methodological foundations in multimodal data integration, normative modeling to characterize heterogeneity across patient populations, human-in-the-loop AI, digital twins for personalized therapy, and scalable infrastructure to support discovery in epilepsy, movement disorders, brain injury, neurodegeneration, and beyond. Our approach applies wherever clinicians must synthesize imaging, neurophysiology, and clinical evidence to make individualized treatment decisions.
Our research is guided by a broader vision to create a learning health system in which clinical care and biomedical research continuously inform one another. We develop the infrastructure, methods, and governance needed to support this cycle. We build cloud platforms and federated learning methods that bring together multimodal clinical data across institutions, harmonize these data at scale, and return validated tools to clinical sites. Our goal is to build a connected health system where data, expertise, and experience from one center can be readily shared to inform and improve care for patients at any other center.
Focus Area
Epilepsy
Epilepsy is our primary focus area. Our research spans the full epilepsy care pathway. We aim to personalize and improve epilepsy therapies. We develop quantitative methods to diagnose epilepsy in patients with first-time seizures or diagnostic uncertainty. We also predict treatment response, including identifying patients who may become drug-resistant and could benefit from earlier surgical evaluation.
For the one-third of the 70 million people with epilepsy who are medication-resistant, our research focuses on guiding surgical planning. In these patients, our work centers on three questions: Which therapy should we pursue? Where should we intervene? How should we integrate multimodal evidence? We develop mechanistic network models, normative atlases, and digital twins. These tools can help move epilepsy care beyond visual inspection and clinical intuition. Our goal is to support surgical planning, electrode targeting, and outcome prediction with rigorous quantitative methods.
Brain Injury
Traumatic brain injury is highly heterogeneous. Injury location, severity, recovery trajectory, and long-term outcomes can vary widely across patients. Our work develops quantitative tools to track disease progression and recovery in individual patients. We integrate longitudinal multimodal data to stratify patients by injury severity, progression, and outcome. We also identify biomarkers that predict recovery trajectory, comorbidity risk, and cognitive decline. These comorbidities include post-traumatic epilepsy and dementia. We design these tools to support clinical trials of disease-modifying therapies through precise patient stratification and outcome prediction.
Neurodegeneration
Our work in neurodegenerative disorders includes Alzheimer’s disease and primary progressive aphasia. We focus on two complementary areas. First, we characterize disease progression. We use spatial normative modeling to create patient-specific readouts of neurodegeneration. These models help us study how language and other cognitive networks reorganize as disease progresses. This approach moves beyond group-average diagnostics toward individualized disease trajectories across the lifespan. Second, we study therapeutic interventions. We investigate non-invasive brain stimulation, including transcranial magnetic stimulation and transcranial direct current stimulation. We use these approaches as candidate treatments for cognitive and language symptoms in neurodegenerative disease. These techniques can modulate cortical activity in targeted networks.
Movement Disorders
Our work in Parkinson’s disease and essential tremor centers on adaptive deep brain stimulation. We integrate preoperative imaging with intracranial neurophysiology recorded during DBS surgery. We use these data to build normative models of structure–function coupling. These models guide individualized biomarker selection and move us toward closed-loop neuromodulation tailored to each patient’s neuroanatomy.
Brain Tumors
Neuro-oncology research is an emerging direction in our lab. We apply network neuroscience methods to study how tumor location and infiltration disrupt functional brain networks. We use these models to inform individualized surgical planning.