Abstracts

Rationale: One third of patients with focal epilepsy are pharmaco-resistant and surgery is an option to control seizures. In one third of surgery candidates, brain areas responsible for seizure generation can be defined exclusively with invasive recordings. Among them, Stereo-EEG intracerebral exploration, where ictal and interictal discharges are recorded from the epileptogenic and the surrounding brain areas and can be analyzed to study epileptogenic network. With help of computer-assisted system in parallel with clinical evaluation, we aimed to identify and classify different ictal patterns and markers of the Epileptogenic Zone (EZ) based on Stereo-EEG analysis. Methods: To identify different ictal patterns computer-driven Stereo-EEG analysis based on time, frequency and space domains was retrospectively applied in parallel with visual inspection.Human EEG ictal patterns and electrographic biomarkers were correlated with clinical profile, neuropathology, location of the EZ and post-surgical outcome. Results: Based on the ictal transients (pre-ictal events, low-voltage fast activity, slow-wave component, asynchronous activity, bursting activity, seizure duration and post ictal depression) all ictal events from 138 patients were classified in 3 main patterns. Pattern A: long seizures (50-120s) with low-voltage fast activity onset (~120Hz) and bursting phase with following post-ictal depression at the end. Pattern B: shorter seizures (10-30s) with characteristic on-set/off-set high amplitude transients and “higher” low-voltage fast activity (140-200Hz) at the onset. And pattern C: mixed pattern B+A. The resting minority group (< 10%) was short ictal events (less than 5s) associated with high-amplitude voltage oscillations and 30-60Hz frequency increase frequently associated with spasms Pattern A was prevalently recorded from the temporal lobe while pattern B more often correlated with extra temporal brain areas. Pattern B may precede pattern A. Conclusions: The quantified analysis showed an excellent overlap between contacts within the EZ identified by expert clinicians. Proposed stereo-EEG computer-assisted analysis can reveal information hidden from visual inspection and help in interpretation of the possible epileptogenic mechanisms. Our patterns classification combined with EZ biomarkers lead to the better understanding of the epileptogenic network organization and improve EZ detection. Funding: Italian Health Ministry (RC2016)