An AI initiative is standardising epilepsy surgery to assist the 50 million individuals affected by epilepsy.
The dataset standardises clinical metadata and establishes benchmark tasks that link AI outputs to postoperative seizure outcomes.
Researcher Dr Yipeng Zhang, from the Department of Electrical and Computer Engineering at the University of California, Los Angeles, said: ‘If we want AI to aid in surgical decisions, we need frameworks that enable results to be compared across hospitals.’
Known as Omni-iEEG, the framework combines pre-surgical brain recordings from eight epilepsy centres, involving 302 patients and 178 hours of data, to support those affected.
About one-third of people with epilepsy worldwide experience seizures that cannot be controlled by medication.
For many of these patients, surgical removal of seizure-causing brain tissue offers the best chance of long-term relief.
Identifying that tissue heavily relies on intracranial electroencephalography (iEEG), which captures high-resolution recordings of brain activity from implanted electrodes.
Over the past decade, researchers have developed AI systems to assist in analysing iEEG recordings, especially in detecting high-frequency oscillations (HFOs) and signal patterns linked to seizure-generating regions.
Many studies show promising results.
However, most AI systems in this field are trained on data from a single hospital or research centre.
Differences in recording protocols, labelling conventions, and clinical definitions make it challenging to compare results across institutions or to establish whether findings are applicable elsewhere.
Zhang said: ‘The field has focused heavily on improving AI accuracy. Without shared evaluation standards, it’s hard to know whether systems will perform reliably outside the original study setting.’
His earlier work focused on refining pathological HFO detection and contributed to the development of PyHFO, a research tool used by independent groups studying seizure-related brain activity.
He pointed out that improving individual systems is only part of the challenge.
Instead of just assessing whether an algorithm can detect abnormal signals, the framework evaluates whether the brain regions identified by AI are associated with better surgical outcomes.
Regulatory agencies have increasingly emphasised reproducibility and cross-site validation in medical AI.
Experts argue that multi-centre benchmarks might become essential before such systems can be routinely integrated into surgical planning.
As AI continues to progress in clinical research, some suggest that future developments may rely less on novel algorithms and more on shared standards that enable reliable validation.
For epilepsy surgery, where decisions are irreversible, and precision is measured in millimetres, this shift could have significant implications.
Zhang et al concluded: ‘We anticipate that our effort will not only support more reproducible benchmarking but also catalyse new methods, biomarkers, and clinical insights that advance both machine learning and epilepsy treatment.’
logy is regarded as a crucial step towards further integrating artificial intelligence into operating theatres. Additional validation and certification should facilitate broader clinical adoption.
