Highlights from EEC 2026
The European epilepsy community gathered in Athens for one of the most important scientific events of the year. From September 5-9, the 16th European Epilepsy Congress brought together clinicians, researchers, and industry leaders to discuss the latest advances in epilepsy diagnosis, treatment, and research.
In case you missed it, below is a selection of a few hot topics covered during this 2026 edition.
Artificial intelligence use in clinics : how far are we?
In an inspiring talk entitled“From research promise to clinical implementation”, Prof. Sandor Beniczky (Copenhagen, Denmark) highlighted a persistent challenge:“The gap between scientific innovation and patient benefit remains substantial.”
Automated methods can already detect and characterize increasingly sophisticated neurophysiological biomarkers, offering new opportunities to support the localization of the epileptogenic zone and clinical decision-making.
Recent research illustrates how rapidly these technologies are progressing. For example, Sun and colleagues developed MORGOTH, an EEG foundation model designed to perform a broad range of EEG interpretation tasks rather than detect a single predefined feature. Developed using EEG recordings from multiple institutions and diverse patient populations, MORGOTH achieved expert-level performance across 17 EEG findings, with relatively consistent performance in external validation. (Sun et al., 2026). The study points towards a new generation of AI-based EEG tools: rather than focusing on a single task, they could support clinicians across multiple aspects of EEG interpretation.
Yet strong research performance does not automatically translate into clinical impact, and no FDA-cleared software is currently validated for spike detection in SEEG, nor indicated for seizure or HFO detection, and human supervision remains indispensable for safe clinical use.
This gap between technological performance and real-world adoption is precisely the challenge discussed by Teo, Beniczky and colleagues in a recent Nature Reviews Neurology Perspective. They argue that neurology has reached an inflection point: demonstrating that an AI algorithm works is no longer enough. The challenge is to translate promising algorithms into safe, sustainable clinical benefit at scale (Teo et al., 2026).
This requires moving beyond algorithm development towards rigorous validation, regulatory approval and real-world clinical implementation. The authors notably emphasize the importance of prospective validation and of evaluating technologies across multiple sites and countries.
For intracranial EEG in particular, the challenge is therefore no longer only to develop better algorithms, but to demonstrate that they can be robustly validated, safely implemented and ultimately deliver meaningful clinical benefit to patients.
References
Epilepsy surgery : patient selection is key to determine the tool
Among the many interesting sessions on surgery planning and invasive methods, an inspiring talk by Pr Christian Dorfer (MUV, Vienna, Austria) reviewed surgical techniques (resection/disconnection, LITT, RTCF, neuromodulation), and explained that resective epilepsy surgery should not be abandoned to the preference of newest devices. Instead, patient selection should determine the tool, and not local preferences or ideology. The future of invasive intervention is hybrid : open surgery shall be more selective and better planned. Earlier referral and faster evaluation are key.
Importantly, surgical evaluation must not remain exceptional, and the patient benefit must not be mistaken with cure. It is critical to choose the least burdensome treatment with the best realistic chance of seizure freedom. If seizure freedom is not reachable, then patient safety, development and quality of life should prevail, after which comes burden and severity reduction as a third target.
The Human Intracerebral EEG Platform
Pr Philippe Ryvlin (CHUV, Lausanne, Switzerland) shed light on the Human Intracerebral EEG Platform (HIP), a secure, cloud-based trusted research environment (TRE) developed within the Human Brain Project, and now integrated into the EBRAINS ecosystem. Designed to support the sharing, curation, and analysis of intracerebral EEG data across institutions, HIP brings together datasets, analytical tools, and collaborative workspaces within a common framework. It operates under the principes of the Five Safe framework : safe people, safe projects, safe settings, safe data, safe outputs.
By enabling harmonized and reproducible multicenter research, the platform helps accelerate discoveries in epilepsy and neuroscience while preserving data security and governance.
There are currently iEEG from over 1500 patients, with the plan to reach 5000. Such number would account for the large heterogeneity in epileptic lesions and networks, SEEG scheme and resection locations.
Two on-going research projects rely on HIP datasets : NeuroFAME and NEWRON.
Stay tuned for more updates on HIP in future editions of our newsletter !
