Abstracts

Rationale: Ictal central apnea (ICA) is a seizure semiological sign seen in forty percent of patients with epilepsy, only in focal epilepsy and more frequently in temporal lobe epilepsy. We first described ICA as a first clinical sign in mesial temporal lobe seizures using visual analysis of multimodal stereotactic electroencephalography (SEEG) and respiratory monitoring. Stimulation studies have found that mainly amygdala but also limbic/paralimbic temporal regions elicit central apnea and have been suggested as symptomatogenic zones of ICA. In this study, we wished to characterize ICA in all types of focal epilepsies and identity which regions are involved in ICA generation by using a quantitative approach. ICA recognition may help anatomo-electro-clinical localization of clinical seizure onset to known symptomatogenic areas.

Methods: We studied video of SEEG recorded seizures in patients undergoing additional respiratory monitoring (thoraco-abdominal belts, pulse oximeter) in the epilepsy monitoring unit. SEEG signal, bipolar montaged, was obtained from 40 seizures in 20 patients. Twelve cortical regions present in all our subjects were studied. We classified ICA as early-onset ICA (within six seconds of the seizure onset) and late-onset ICA (after six seconds of seizure onset). Using the Morse wavelet transform, we obtained time-frequency maps centered on two gamma frequency bands: 30 to 60 Hz and 60 to 150 Hz. Baseline normalization, via z-score, was applied to highlight the brain regions that have a significantly different activity with respect to baseline at ICA onset. Finally, for the group analysis we averaged across time-frequency maps of each seizure in our two groups.

Results: Out of all regions, amygdala, hippocampus, fusiform, lingual and supramarginal gyri, insula and opercular cortex displayed a significant difference (z-score, p< 0.01) in gamma activity at ICA onset with respect to baseline (Figure 1 and 2). In the group of early-onset ICA (Figure 1A and 2A), amygdala showed the earliest activation followed by hippocampus, lingual, supramarginal and posterior fusiform gyri. On the other hand, in the group of late-onset ICA, amygdala was not different at the time of ICA onset (Figure 2A and 2B). However, late-onset ICA was associated with increased gamma activity in hippocampus, opercular cortex, lingual and the anterior fusiform gyri (Figure 2A and 2B).