Skip to content Skip to navigation

Multi-modal deep learning for automated seizure localization

A major challenge in the practice of neurology is the functional localization of epileptic seizure.  Our team is leveraging exciting recent advancements in artificial intelligence (AI) technologies to develop an automated seizure detection and localization system based on deep neural networks, electroencephalography (EEG) data, and real-time video with the goal to dramatically increase neurologist diagnostic capabilities while improving quality of care. Current software for automated seizure detection is slow, inaccurate, and rarely precise enough for clinicians to rely upon, and there are no approaches for automated functional localization of seizure, which is a key component in differential diagnosis and treatment plan formation. We are creating a uniquely rich resource to accelerate our developments that includes large datasets describing the seizure type, location, length, and severity of seizure on the appropriate sensing modalities (EEG and video). We will leverage our encouraging preliminary work in automated seizure detection from EEG and our team’s experience developing deep learning applications based on large, curated medical datasets to develop an automated system for seizure detection and functional localization that could fill a market need by ensuring high-quality automated monitoring while providing additional clinical insight to neurologists. Our project will address major unmet clinical needs in seizure disorders and substantially improve quality of care by enabling more accurate automated seizure detection, enhancing the diagnostic performance highly trained neurologists, and providing neurological services in currently underserved markets. Our work will also establish a unique interdisciplinary team of Stanford that includes neurologists, data scientists, and computer scientists who will develop and translate new AI methods to advance care in seizure disorders.


Team Members: 

Jared Dunnmon (Postdoctoral Research Fellow) 
Khaled Saab (PhD student, Electrical Engineering)

Funding Type: 
Award Year: