A research partnership between Mentis Care and Emory University is exploring whether transformer-based AI models can interpret EEG data across both hospital and wearable systems.
The Push for Neurological Monitoring Beyond Hospitals
Across the UK health and social care system, there is growing interest in technologies that allow long-term conditions to be monitored outside hospitals. Remote monitoring is already used for heart failure, diabetes and respiratory disease, but neurological conditions have proved far harder to track in everyday environments.
Epilepsy illustrates the challenge. The condition affects around 50 million people worldwide and many patients still struggle to achieve stable seizure control. Monitoring typically takes place in specialist clinical settings using electroencephalography (EEG), meaning treatment decisions are often based on limited data captured during short hospital visits.
A new research collaboration between Mentis Care and Emory University aims to address that gap by developing artificial intelligence capable of interpreting EEG signals across both clinical systems and simplified wearable devices.
The two-year programme will focus on creating a transformer-based “foundation model” designed to detect and potentially predict seizures using a wide range of EEG configurations.
Difficulty Moving EEG to Everyday Care
EEG measures the electrical activity of the brain through electrodes placed on the scalp. In clinical environments this often involves more than 20 electrodes arranged according to a standard configuration.
Such systems provide highly detailed information but are impractical for long-term use outside hospital settings. Wearable EEG devices typically use far fewer sensors, which can reduce signal quality and make automated detection more difficult.
Another challenge is signal noise. Everyday activities such as touching the scalp, moving, or adjusting electrodes can generate electrical patterns that resemble seizure activity.
These factors have historically limited attempts to create reliable home-based seizure monitoring tools.
The research programme led by Dr Samaneh Nasiri at Emory’s School of Medicine aims to address this by developing what researchers describe as a “channel-agnostic” AI model. Instead of relying on a fixed number of electrodes, the model is designed to adapt to different EEG configurations while maintaining performance.
Dr Nasiri said advances in AI architecture are opening new possibilities for analysing brain signals.
“Foundation models are redefining what’s possible in healthcare, and nowhere is that more apparent than in seizure detection. By adapting across channel configurations, they enable high-performance detection even in reduced-channel wearable formats.”
Relevance for Digital Health and Community Care
While the project is being conducted in the United States, the underlying technology reflects wider global trends in digital health.
Health systems, including the NHS, are increasingly exploring remote monitoring tools to support people with long-term conditions in their homes and communities. The NHS Long Term Plan highlights digital monitoring as a way to reduce hospital admissions and enable more proactive care.
Neurological conditions have been slower to benefit from these approaches because of the complexity of brain data. Reliable wearable EEG analysis could therefore open new possibilities for community-based neurological care.
Continuous seizure monitoring could provide clinicians with richer datasets about how patients respond to medication or lifestyle factors. Over time this may support more personalised treatment decisions.
It could also support people who wish to live independently by providing alerts to carers or healthcare professionals when seizures occur.
Broader Implications for Brain Health Research
Researchers involved in the project say advances in large-scale EEG analysis could also benefit the wider study of neurological conditions.
Dr Gari Clifford, chair of biomedical informatics at Emory University and professor of biomedical engineering at Emory and Georgia Tech, said EEG remains one of the most valuable tools for understanding brain activity.
“The EEG is the brain’s stethoscope that listens to its electrical conversations — not just for epilepsy, but for sleep, memory, movement disorders, and many other neurological conditions.”
Large AI models trained on extensive datasets may help researchers identify patterns in brain activity that are difficult to detect using traditional analytical methods.
For the UK care sector, the long-term significance lies in whether complex conditions such as epilepsy can eventually be monitored reliably outside hospital environments.
If that becomes possible, digital health tools could support earlier interventions, improve data available to clinicians and help people manage neurological conditions while living at home.
Innovative Care NEWS 
Leave a comment