Artificial intelligence (AI) is rapidly reshaping the landscape of epilepsy care and research, and this transformation calls for coordinated, interdisciplinary leadership. A Special Interest Group (SIG) on AI in Epilepsy within EpiCARE aims to provide a structured forum to align clinical needs, innovation, and methodological rigor. It aims to foster shared standards, critical evaluation of algorithms and new methodologies, and responsible translation of AI tools into routine care across diverse healthcare systems.
The SIG has three main topics of interest at the moment:
Clinical care in epilepsy uses video, EEG, and other biosignals for seizure diagnosis, classification, and follow-up in the epilepsy monitoring unit. However, these resources are limited, demand highly trained personnel, and are time-consuming. Recent advances in sensors, embedded electronics, and machine learning have enabled continuous, real‑world data acquisition from patients’ daily lives. Moreover, signal processing is very important during the evaluation of intracranial EEG data and aims to facilitate the correct identification of the epileptogenic zone. The SIG aims to promote rigorous clinical validation, meaningful performance metrics, and integration of wearables and signal-processing tools into clinical workflows, ensuring that these technologies move beyond pilot studies to become reliable tools for patient care.
Neuroimaging is essential to adequately define the etiology of many epilepsies, and is the second cornerstone of our activities. AI methods applied to MRI, EEG, PET, and multimodal imaging have demonstrated the potential to uncover subtle epileptogenic abnormalities and improve presurgical evaluation. However, variability in data acquisition and analysis remains a major challenge. The SIG works to foster multicenter collaboration, data harmonization, and reproducible AI pipelines that can support robust clinical decision‑making.
A key concern across all imaging applications is model transparency and trust. As clinicians and researchers, we recognize the limitations of black‑box algorithms in sensitive clinical contexts. The SIG actively encourages the development and evaluation of explainable and uncertainty‑aware AI approaches, ensuring that imaging‑based predictions are interpretable, auditable, and aligned with known neurobiological mechanisms.
AI in epilepsy genetics represents another rapidly expanding area of interest. Machine learning techniques are increasingly used to integrate genomic data with clinical, electrophysiological, and imaging phenotypes. The SIG supports collaborative efforts to improve variant interpretation, address rare epilepsy syndromes, and reduce bias in genetic models, paving the way toward more precise and personalized therapeutic strategies.
While these three domains currently form the SIG’s core, our scope is intentionally broad. Emerging topics such as closed‑loop neuromodulation, digital twins, new biomarkers, federated learning, and AI‑supported clinical trials are welcomed as the field evolves. The SIG is designed to adapt, acting as an incubator for innovative ideas while maintaining a strong clinical and scientific foundation. Furthermore, we aim to develop multidomain collaborations with other EpiCARE SIGs and WGs.
By joining the AI in Epilepsy SIG, you can contribute to a community committed to excellence, rigor, and impact. Our collective aim is to ensure that advances in artificial intelligence translate into measurable benefits for patients, clinicians, and healthcare systems—through collaboration, critical evaluation, and shared leadership in this transformative area.
The AI in epilepsy group focuses on three main areas of our daily work: clinical care, research and education.
If you are interested to voluntarily contribute to this SIG activities, please use the contact form to let us know.
