Abstracts

To clarify the specific role of the insular cortex and to study ictal electroclinical patterns within temporolimbic networks during seizures involving the temporal lobe, using intracranial recordings. We analysed stereoelectroencephalographic (SEEG) recordings of the insular cortex and of temporo-perisylvian regions in 19 patients suffering from drug-resistant partial epilepsies involving at least the temporal lobe. All SEEG investigations were individually designed according to previously acquired electroclinical and MRI data. The insular cortex was recorded because of evidence of an early perisylvian involvement or rapid extratemporal diffusion. The insular electrode was inserted with the help of a computer-driven robot using an oblique trajectory avoiding the sylvian fissure, in order to reduce the risks of bleeding and to try to improve the spatial sampling of the insular cortex along antero-posterior and dorso-ventral axis. In a first and most important group of 10 patients, we have highlighted an anteromesio-polo-insular pattern, corresponding to the spread of mesiotemporal discharges towards the insula, via the temporal pole and the amygdala. These patients were successfully treated by temporal resections sparing the insula, wether or not they had insular interictal discharges. In a second group (n=4), the insula belonged to the seizures onset zone (n=2) or was affected by prominent interictal discharges asynchronous from temporal interictal discharges. Among these 4 patients, 2 were operated-on with a poor post-operative outcome resulting from a temporal lobe surgery. In the 5 remaining patients, the seizures affected mainly mesial, lateral and posterior temporal regions, whereas the insular cortex was belatedly and/or slightly involved. Our findings suggest that the insular cortex seems to constitute a major relay of mesiotemporal lobe discharges, presumably via the spreading through the amygdala and the temporal pole. In order to avoid failure of the surgery, intracranial recordings remain sometimes necessary to exclude insular onset during temporolimbic seizures including an early perisylvian involvement or rapid extratemporal spreading. Depth recordings of the insular cortex through an oblique trajectory seems to be safe and useful in term of spatial sampling.