AI-backed sqEEG tracks epilepsy outside hospitals
An artificial intelligence-driven implant placed beneath the skin is offering new possibilities for epilepsy monitoring, replacing unreliable patient diaries with continuous, objective data.
The subcutaneous electroencephalography (sqEEG) device was developed by King’s College London’s Institute of Psychiatry, Psychology & Neuroscience (IoPPN) in collaboration with the Mayo Clinic and UNEEG medical. Published in Epilepsia and supported by the Epilepsy Foundation of America, the study tested the implant in real-world conditions over more than a year.
About the size of a pound coin, the implant sits behind the ear with a thin wire positioned near seizure sites. Using AI, it analyses brain activity and wirelessly sends the information to an external recorder. Ten adults with treatment-resistant epilepsy trialled the device for up to 15 months, logging nearly 72,000 hours of brainwave data and 754 seizures. Participants also kept seizure diaries and used fitness trackers.
The results highlighted the limitations of self-reporting: patients correctly recorded only 48% of seizures, while 27% of diary entries did not match seizure activity. By contrast, sqEEG captured both the timing and type of seizures more accurately and was reported as unobtrusive by most participants.
Professor Mark Richardson, Paul Getty III Professor of Epilepsy at King’s IoPPN, said, “It is vital that people with treatment resistant epilepsy are able to access the best possible care. This is made significantly more challenging by the fact that clinicians must rely on patient reporting to establish when episodes have taken place. Our study has been able to provide a vital and viable alternative to relying on self-reported episodes. A small tracker placed under the skin was able to detect seizures far more accurately than the participants themselves.”
Dr. Pedro Viana, lead author and Senior Clinical Research Fellow at King’s IoPPN, added, “The ability to monitor seizures in the real world, accurately collecting data on the type and timing that they occur will be an invaluable tool for clinicians moving forwards, and should hopefully have a big impact on how we approach the treatment of this life-threatening condition. While this is an important step forwards, it’s now vital that we conduct larger trials to further validate this technology, with a view to hopefully making this available to everyone in need.”
Dr. Benjamin Brinkmann, Professor of Neurology at Mayo Clinic, said, “Overall, these results demonstrate that Subscalp EEG devices are able to provide objective information about seizures and brain activity over long timeframes. This technology could provide physicians and patients with new insights about long-term patterns and help optimize their epilepsy treatment.”
Jonas Duun-Henriksen, Senior Director AI & Science at UNEEG medical, called the findings “a significant milestone,” adding, “While we’ve long heard from both people living with epilepsy and their treating physicians that they are very satisfied with our system and helped improve outcome, this is the first time a study based on more than a full year of continuous EEG recordings confirms those experiences. We’ve gained valuable quantitative insights throughout the study, and we’re especially grateful for the valuable feedback from the clinicians at King’s College London – which we’ve already started incorporating into our solution to further enhance its clinical value.
