ATLANTA ― Artificial intelligence (AI) gives specialists a powerful tool to improve outcomes in people with epilepsy, including identifying those who might be candidates for surgery or are missing test results, according to findings from a study presented at the American Epilepsy Society Annual Meeting.
“Research like this demonstrates the growing potential of responsible artificial intelligence to enhance epilepsy care,” said Howard P. Goodkin, MD, PhD, president of the American Epilepsy Society. “Instead of replacing the epilepsy specialist, AI acts as a partner, enhancing human expertise in medicine by tracking complex medical information, identifying gaps and prompting action to ensure care remains on course over time.”
AI can be particularly helpful in epilepsy care, which Is complex and often fragmented.
“Patients see multiple specialists and undergo many tests and their treatment plans evolve over years, so it’s easy for crucial details or opportunities for surgical evaluation to be missed,” said P. David Adelson, MD, lead author of the study, Steve A. Antoline Endowed Chair for Children's Neurosciences and vice chair, West Virginia University (WVU) Rockefeller Neuroscience Institute, Morgantown. “Large language AI models continuously read, organize and interpret medical records to help health teams stay a step ahead. They have the potential to transform not just epilepsy care, but any long-term condition that requires coordination across specialties.”
The researchers integrated the Autonomous Registry and Analytics (AURA) platform, which utilizes AI, into WVU Medicine’s electronic health record. The registry includes 3,348 individuals with epilepsy (1,176 children and 2,172 adults).
During a 90-day prospective evaluation, AURA assessed clinicians’ notes, imaging reports, assessments and documentation for 820 scheduled patient visits with neurologists. AURA identified 88 patients (11%) who met the
criteria for drug-resistant epilepsy but who had not been referred for surgery. American Academy of Neurology guidelines recommend that all people with drug-resistant epilepsy be evaluated for surgery, although ultimately not everyone is a candidate.
AURA also highlighted care gaps in patients’ records, including:
- Outdated or missing magnetic resonance imaging (MRI) tests (54%).
- Missing neuropsychological evaluations (91%).
- Missing electroencephalographs, or EEGs (35%).
Across the full registry, folic acid deficiencies were identified in 81% of women of childbearing age. Researchers also determined that the technology increased post-surgical outcome documentation from 1% to 70.9%. Health care teams reported more comprehensive evaluations, better collaboration in care planning and faster recognition of patients who might benefit from surgery. These improvements should lead to shorter delays, fewer missed opportunities and better outcomes for patients with drug-resistant epilepsy, Dr. Adelson said. The findings demonstrated that health systems can use this form of AI responsibly to “connect the dots” across years of documentation, something humans simply don’t have the time or bandwidth to do consistently.
Researchers evaluated several AI methods before implementation and found that only carefully designed, clinically validated models, such as AURA, met the accuracy needed for real-world use. The system was developed and tested in close collaboration with clinicians, entirely within the hospital’s secure environment, and with strict protections for patient privacy. The study shows that with the right design and oversight, AI can be both trustworthy and transformative without compromising safety or confidentiality, Dr. Adelson said.
To prevent bias and hallucinations, AURA employs a multi-agent framework — a team of AI agents working together — to respond to prompts, rather than relying on a single model. Each AI agent is trained to perform a specific task, such as identifying surgical candidates or checking for safety concerns. A separate AI judge then reviews the outputs and compares them to the source data, providing feedback until the information is accurate and fully grounded in the medical record. The teamwork between agents and the precise prompting guides enabled the level of accuracy.
“It’s like having an intelligent research assistant that never gets tired. It keeps the care progress aligned and complete, and makes specialists more efficient, while ensuring that no patient is overlooked because of information gaps or time restraints,” Dr. Adelson said. “This allows specialists to focus on interpreting results and planning treatment, helping extend expert-level care to more patients, even in hospitals with limited specialist coverage.”
*** AES 2025 news releases may contain updated data that does not match what is reported in the abstract.
For more information, visit the American Epilepsy Society online at aesnet.org. Join the AES social conversation today by following @AmEpilepsySoc on X and use the hashtag #AES2025.