Your medical device may soon talk back

Most medical devices are silent witnesses. They record, store and wait for a specialist to read the data weeks later. A University of Pennsylvania team now wants devices that answer questions the moment a patient asks them.

The proof-of-concept study, published in Nature Medicine, enrolled 13 adults undergoing continuous EEG monitoring in Penn's Epilepsy Monitoring Unit. Together they exchanged 1,307 messages with the system, mostly about seizures, sleep, mood and energy.

How does it work?

The platform has three parts:

• Software that captures EEG data from electrodes in the brain
• A secure, HIPAA-compliant cloud environment that analyzes the signals in near real time
• A smartphone app with an AI conversational agent, similar to the technology behind ChatGPT

Patients ask about their brain activity, request charts, report symptoms and answer surveys three times a day. The device can also message them first when it spots a change. Patients flag which detected seizures were real and which were false alarms, and the system improves. Clinicians can review its assessments too, through an "expert in the loop" algorithm that limits their input to the most valuable annotations.

One participant said the system helped "solidify what I felt to be true between some of my episodes and emotions."

Why does it matter?

Senior author Brian Litt describes the aim as a device that works like a personal Jiminy Cricket, and one that learns what matters to each patient. That is a real shift from periodic clinician review to continuous collaboration.

The team sees uses well beyond epilepsy: cardiac devices, insulin pumps, and treatments for Parkinson's disease, depression, OCD and addiction. Studies are already planned to tune deep brain stimulation for Parkinson's automatically. And if personalized systems need only rare clinical supervision, long-term outpatient use becomes practical.

Still, this is a small inpatient study. Future trials will test whether the approach reduces seizure frequency, improves quality of life and cuts emergency visits.

The context

For the GCC, the timing is notable. Saudi Arabia's Vision 2030 and the UAE's digital health strategies both push toward preventive, data-driven, patient-centered care, and the region has a growing burden of chronic conditions such as diabetes. Devices that interpret data and talk to patients could ease pressure on specialist capacity and support remote care. But regulators and hospitals would need clear rules on data residency, privacy and clinical accountability before adopting AI that adjusts treatment on its own.

The work is backed by a $5.6 million NIH Pioneer Award and builds on Litt's wider effort to develop brain-computer interfaces that sense, communicate, learn and adapt.