Researchers at the University of Pennsylvania have developed an AI platform that allows patients to communicate with their medical devices, a development that could fundamentally alter how individuals manage their health.
The system, tested in a proof-of-concept study, enables a two-way exchange of information between patients and their monitoring devices, facilitating continuous learning and adaptation of disease management. This approach shifts medical devices from passive data collectors to active collaborators in patient care.
“The idea is that instead of waiting for a specialist to analyze the data collected by your device, the device could detect changes and alert you immediately, like your own personal Jiminy Cricket on your shoulder,” said senior author Brian Litt, MD, a professor of Neurology, Neurosurgery, and Bioengineering at the University of Pennsylvania.
The platform was evaluated with 13 adults undergoing continuous electroencephalogram (EEG) monitoring in Penn’s Epilepsy Monitoring Unit. Participants used a smartphone application, dubbed BrainChat, which featured a conversational AI agent. Through the app, patients could inquire about their brain activity, request data visualizations, report symptoms, and receive personalized alerts.
The system captured EEG data, analyzed it in a secure cloud environment, and presented information to patients via text, charts, and other visual formats. Simultaneously, the device could alert patients to significant changes and learn from their responses in real time. Researchers analyzed 1,307 messages exchanged during the study, with most conversations revolving around seizure events, sleep, mood, and energy levels.
Participants reported that the platform helped clarify patterns they had noticed but struggled to understand. One individual noted the system “solidify[ed] what I felt to be true between some of my episodes and emotions.” Others appreciated access to sleep information and the potential for pre-seizure warnings.
Patients could also train the system by identifying correct and false seizure detections, significantly improving its accuracy. Clinicians could also interact with the system to review assessments and provide annotations, further refining its performance through an “expert in the loop” algorithm.
Litt stated, “Patients often understand aspects of their disease that sensors cannot measure directly. We envision medical devices becoming partners in care that learn alongside patients and help them make better decisions every day.”
While the study focused on epilepsy, researchers believe the platform has broad applicability for conditions managed by cardiac devices, insulin pumps, and treatments for neuropsychiatric disorders such as Parkinson’s disease, depression, and cognitive disorders.
Future applications may involve systems that automatically adjust treatment, communicating directly with patients about symptoms and goals. The team is preparing studies to apply the platform to automatically tune deep brain stimulation devices for Parkinson’s disease and is incorporating wearables for insomnia and vital sign monitoring.
This AI-driven communication aims to transform medical devices into partners in care, continuously integrating physiological signals, patient feedback, and AI analysis for more personalized and responsive healthcare.
The research is supported by a $5.6 million NIH Pioneer Award (DP1 NS122038-01). Ongoing research will assess the platform’s impact on seizure frequency and severity, quality of life, and healthcare utilization. Plans include expanding testing to outpatient settings with advanced EEG and wearable technologies.
The platform represents a fundamental shift from current medical devices, which primarily collect data for periodic clinical review, to systems that actively collaborate with patients to manage their health.
The Penn-developed program allows patients and devices to exchange information, learn together, and adapt disease management on the fly. Participants in a proof-of-concept study undergoing monitoring for seizures could ask questions about their brain activity, receive personalized alerts, report symptoms and behaviors, and help train the system to better recognize seizure-related patterns.
