Abstracts

Rationale: Hemodynamic response (HR) to IEDs has been traditionally and worldwide studied by BOLD responses in EEG-fMRI. However, since fMRI scanning imposes strict immobility inside the scanner, EEG-fMRI only allows short lasting investigations, which constrains the amount of IEDs that can be recorded. Moreover, EEG-fMRI is only sensitive to fluctuation in deoxyhemoglobin. In contrast, EEG-Near Infra-Red Spectroscopy (NIRS) is a non-invasive tool to study the neurovascular coupling (NVC) to IEDs, with better temporal resolution (10 Hz). EEG-NIRS monitors fluctuations of deoxy-hemoglobin (HbR) and oxy-hemoglobin (HbO) signals during prolonged recordings at the bedside. Our goal was to assess the integrity of the NVC at the time of IEDs using overnight EEG-NIRS recordings, as well as the interaction between NVC and sleep stages.  Methods: Three drug-resistant epilepsy patients had overnight EEG-NIRS recordings, guided by prior source localization results from simultaneous EEG-MEG. As per personalized NIRS optimal montage protocol (Pellegrino et al., 2016), NIRS sensors positions were individualized in order to maximize NIRS sensitivity to the underlying cortical region (Machado et al., 2018). For each patient, the NIRS montage covered the sources unveiled by EEG and MEG, their contralateral homologous region as well as a control region. For each patient, all IEDs were visually marked ensuring labeling was done according to the same morphology and distribution as IEDs analyzed for EEG-MEG and their corresponding HR were analyzed.   Results: For all three patients, we recorded in total 32 hours of EEG-NIRS data, including 12 during wakefulness and 20 during sleep. Preliminary results for the first patient are presented. 25 irregular bilateral polyspike-wave bursts, started on the left side, were visually marked and analyzed (figure). The average time courses of the NIRS response associated to these IEDs were characterized by an onset response occurring 5 secs after the onset of the average bursts, peaking at 12 secs and lasting 30 secs. The HR consisted in a strong negative response characterized by a large  [HbO] decrease and [HbR] increase covering bilateral frontal regions, whereas no modification from the baseline was found in the control region (left parietal).Moreover, fluctuations of NIRS signals according to wakefulness and different sleep stages (light sleep, N3, REM) were found in the power spectral density similar to Näsi T et al., (2011), particularly a decrease of very low frequency oscillations during N3 sleep. Conclusions: This preliminary result from simultaneous EEG-NIRS recordings showing a negative hemodynamic response associated to interictal polyspike-wave bursts suggests a locally altered NVC, which might be related to three main mechanisms: a) an insufficient local oxygen perfusion leading to a hypoxia, b) an increased local inhibition concomitant with the presence of slow-waves in the EEG as suggested when a negative BOLD response was found in Pittau et al., (2013), c) an abnormal control on local vasodilatation. A better understanding of these underlying mechanisms and their interaction with sleep might have clinical relevance and guide surgical targets.   Funding: No funding