Abstracts

Rationale: Ictal hypoxemia has been reported in several studies and may contribute to sudden unexpected death in epilepsy (SUDEP). Ictal hypoxemia occurs often in patients with localization-related epilepsy and may be prolonged. Scalp EEG studies have shown than seizures originating from the temporal lobe are more prone to ictal desaturation than extratemporal seizures. However, the neural bases of ictal desaturation in temporal lobe seizures remain elusive. We aim to describe the brain networks underlying ictal desaturations using intracranial EEG.  Methods: Thirty-seven patients with medically intractable seizures who underwent Stereotactic EEG (SEEG) in our tertiary center were included in the study. 18 patients presented ictal desaturation with a  saturation below 90 % during the course of the seizures while 19 patients had no desaturation. Depth electrodes targeting temporal structures (amygdala, anterior hippocampus, temporal pole, temporal neocortex), anterior and posterior insula, frontal cortex (orbito-frontal cortex, anterior cingulate, frontal operculum) were implanted in all patients. We analyzed SEEG signals with indices quantifying the seizure-onset zone (Epileptogenicity Index EI), ictal propagation (broadband signal energy), pairwise functional connectivity (imaginery part of the Coherency for broadband signals) during different parts of the seizures lasting 10 seconds each (baseline, seizure onset, middle of seizure, end of the seizure, post seizure). We evaluated whether seizures leading to desaturation differed from seizures without desaturation using linear mixed effect models taking into account the heterogeneity of the data.  Results: Ictal desaturation developed mostly at the end of seizures : mean latency 59.3 s after seizure onset. Seizure duration did not differ between seizures with and without ictal desaturation. The EI was higher in the anterior hippocampus, amygdala and temporal pole than other structures but did not differ between seizures with and without desaturation. Broadband signal energy was higher in all structures at the middle and end of seizures with ictal desaturation. Functional connectivity was also higher in all structures of interest at the middle and end of seizures with ictal desaturation.   Conclusions: Ictal desaturation in temporal lobe epilepsy is related to a late ictal involvement of a widespread network involving medial temporal structures, neocortical temporal cortex, frontal cortex and insula. The intensity of propagation and connectivity within this network is thus the determinant factor of ictal desaturation. No specific brain structure is specifically related to ictal desaturation.  Funding: No funding