Epilepsy patients can now communicate with their monitoring devices in real time, according to a National Institutes of Health-funded study released October 13, 2026. The system reduced false seizure alarms by about 80 percent by incorporating direct patient feedback into its detection model.
A research team at the University of Pennsylvania built a conversational super intelligence (SI) platform that allows non-experts to query their own physiological data via a smartphone application. This bidirectional system lets the device inform the patient and the patient correct the device.
Electroencephalography (EEG) devices have long functioned as one-way streets, leaving patients blind to the data their own bodies produce in the moment.
Patient Feedback Cuts False Alarms
University of Pennsylvania researchers hypothesized that patients, who experience the physical reality of a seizure, could provide the immediate feedback necessary to fine-tune the seizure-detecting model. This addresses a primary technical hurdle where algorithms often flag benign brain activity as suspicious, requiring human experts to manually filter the noise.
Patients marked detections as true, false, or uncertain.
- Patient Impact: Feedback from participants reduced the system’s false alarm rate by approximately 80 percent.
- Expert Integration: When the SI also learned from expert feedback, alarms decreased by an additional 2 percent.
- Engagement Levels: Patients and the SI exchanged more than 1,300 messages, with participants responding to over 67 percent of real-time detections.
Interface Empowers Patient Monitoring
The SI responded to more than 92 percent of patient queries with accurate and relevant information, such as whether seizures occurred during sleep. Patients used the smartphone app to ask these specific questions about their brain activity.
“This proof-of-concept showcases a technology with the exciting potential to shift management of epilepsy and possibly other diseases in a more adaptive direction that puts patients in the driver’s seat,” said Amy Bany Adams, Ph.D., acting director of the NIH’s National Institute of Neurological Disorders and Stroke (NINDS).
Amy Bany Adams, Ph.D.
Brian Litt, M.D., a professor of neurology and bioengineering at the University of Pennsylvania, noted that these devices were originally designed for clinicians, leaving patients without a way to learn how their behavior impacts their health in real time. Co-first authors Zack Goldblum and Haoer Shi, graduate researchers in Litt’s lab, developed the SI platform using real-time EEG data.
Researchers Plan Larger Studies for Outpatient Use
Several participants asked to take the system home before the study concluded, prompting researchers to plan larger studies for outpatient use. The initial study was limited to a small group in a hospital setting.
The team intends to enable communication between patients with Parkinson’s disease and their deep brain stimulation devices by applying this conversational SI platform to other neurological conditions.
“Completely unprompted, several patients asked if they could take the system home with them, which was one of the best signals we could receive,” Goldblum said. “We had to tell them, ‘No, not yet, but we’ll get there.'”
Zack Goldblum
The research, published in Nature Medicine (DOI: 10.1038/s41591-026-04680-4), was supported by multiple grants from the NINDS, including DP1NS122038, R01NS125137, T32NS091006, R61NS125568, R01NS116504, K99NS138680, and K23NS121401.
Readers should consult with qualified healthcare professionals for medical guidance regarding epilepsy monitoring or deep brain stimulation.
It is not yet known when the system will be available for outpatient use.