PhD Student Université Catholique de Louvain Brussels, Belgium
Rationale: Acute vagus nerve stimulation (VNS) can reduce seizure duration and severity. EEG desynchronization may be an underlying mechanism explaining this effect. Animal studies proved VNS-induced acute desynchronization, and human studies investigated chronic rearrangements in EEG connectivity after VNS. However, a biomarker of the acute VNS effects on brain electrophysiology is still lacking. We hypothesized that acute VNS modulates: i) the magnitude of ongoing EEG oscillations; ii) the connectivity measures of phase synchronization, such as Coherence, Phase Locking Value (PLV) and Phase Lag Index (PLI). In addition, we studied whether these measures can differentiate responders (R) from non-responders (NR) ( > or < 50% seizure frequency reduction at last follow-up). Methods: The epilepsy surgical database of St. Luc University Hospital, Brussels, was retrospectively searched for consecutive patients who: (1) received a VNS implant between 2008 and 2019; (2), underwent a video-EEG with ≥19 electrodes + electrocardiogram (ECG) after VNS implantation; (3) had VNS active during the video-EEG. ECG derivations were screened for the presence of stimulation artifacts, to distinguish EEG epochs with VNS ON vs. VNS OFF. For each patient, ten calm waking EEG epochs and ten stage I/II sleep epochs were selected. Epochs were chosen as ten s with the lowest number of artefacts. For each epoch, Hilbert transform was performed to estimate the signal envelope in the delta, theta, alpha and beta frequency bands. Coherence, PLI and PLV were also calculated for each frequency band, according to longitudinal and transversal channel pairings. The values of single epochs were averaged within each patient, and compared between VNS ON and OFF conditions using Wilcoxon paired test. Results were analyzed either merging or separating: i) awake from sleep EEG; ii) R from NR. Non-paired Mann Whitney U test was used to compare the absolute values of Coherence, PLV and PLI between R and NR. Significance was set at p< 0.05. Results: We report preliminary results from 12 patients (8 females, 4 males; 6 R, 6 NR). VNS output current ranged between 0.75 and 2 mA, and VNS ON time duration between 14 and 30 s. With respect to the magnitude of ongoing EEG oscillations, the analysis did not show significant ON vs. OFF changes in the analyzed frequency spectrum, either in awake or sleep EEG. Among the connectivity measures, in the awake EEG, PLI during VNS ON was found to be significantly decreased, in the theta (p= 0.02) and alpha (p= 0.03) bands, as compared to VNS OFF periods. Coherence and PLV did not show significant ON vs. OFF changes. When separating R and NR in two subgroups, no significant changes were found to distinguish R from NR, for any of the investigated measures. Conclusions: Our preliminary results show that VNS induces an acute EEG desynchronization (i.e., decreased PLI) in the theta and alpha bands during wakefulness. PLI decrease in theta/alpha bands might be a potential afferent biomarker of the acute effect of VNS effects on brain function. Further validation on a larger sample is required to search for potential EEG markers that help to distinguish R from NR. Funding: Please list any funding that was received in support of this abstract.: No specific funding was received in support of this abstract.