Artificial intelligence applications in intracranial electroencephalography for epilepsy management: Insights from AI in Epilepsy 2025.

Intracranial electroencephalography (iEEG) provides direct recordings of cortical activity and remains central to the presurgical evaluation of drug-resistant focal epilepsy. However, the high dimensionality and complexity of iEEG data increasingly challenge conventional visual interpretation. Artificial intelligence (AI) methods have accelerated quantitative analyses in electrophysiology, supporting localization, outcome prediction, and therapeutic decision-making in epilepsy. This article reviews recent advances in the application of AI to iEEG in epilepsy and in particular stereo-EEG, although the concepts discussed are broadly relevant across iEEG modalities. Although many of these approaches have shown encouraging results, the current evidence remains largely preliminary, with most studies being retrospective, single-center investigations that lack independent external validation. AI-based integration of iEEG with multimodal magnetic resonance imaging shows promise for improving surgical planning in drug-resistant focal epilepsy. Concurrently, early studies suggest that AI-driven iEEG analysis may enhance epileptogenic zone identification by detecting subtle electrophysiological signatures, network interactions, and temporal dynamics. These multimodal frameworks emphasize distributed epileptic networks rather than isolated focal abnormalities. AI is also reshaping how iEEG-guided stimulation is designed and interpreted. Data-driven approaches have the potential to optimize stimulation parameters, predict network responses, and clarify mechanisms underlying diagnostic and therapeutic stimulation, laying the groundwork for adaptive, patient-specific neuromodulation. Beyond short-term epilepsy monitoring, AI applications increasingly incorporate continuous sensing, longitudinal network modeling, and integration with implantable neurotechnology platforms. Although these developments require prospective multicenter studies with external validation before widespread clinical implementation, advances in multimodal data integration, interpretable modeling, and continuous sensing position AI-enhanced iEEG as an emerging platform for precision brain network medicine for epilepsy.