Electroencephalogram synchronization measure as a predictive biomarker of Vagus nerve stimulation response in refractory epilepsy: A retrospective study.
Venethia DanthineLise CottinAlexandre BergerEnrique Ignacio Germany MorrisonGiulia LiberatiSusana Ferrao SantosJean DelbekeAntoine NonclercqRiëm El TahryPublished in: PloS one (2024)
There are currently no established biomarkers for predicting the therapeutic effectiveness of Vagus Nerve Stimulation (VNS). Given that neural desynchronization is a pivotal mechanism underlying VNS action, EEG synchronization measures could potentially serve as predictive biomarkers of VNS response. Notably, an increased brain synchronization in delta band has been observed during sleep-potentially due to an activation of thalamocortical circuitry, and interictal epileptiform discharges are more frequently observed during sleep. Therefore, investigation of EEG synchronization metrics during sleep could provide a valuable insight into the excitatory-inhibitory balance in a pro-epileptogenic state, that could be pathological in patients exhibiting a poor response to VNS. A 19-channel-standard EEG system was used to collect data from 38 individuals with Drug-Resistant Epilepsy (DRE) who were candidates for VNS implantation. An EEG synchronization metric-the Weighted Phase Lag Index (wPLI)-was extracted before VNS implantation and compared between sleep and wakefulness, and between responders (R) and non-responders (NR). In the delta band, a higher wPLI was found during wakefulness compared to sleep in NR only. However, in this band, no synchronization difference in any state was found between R and NR. During sleep and within the alpha band, a negative correlation was found between wPLI and the percentage of seizure reduction after VNS implantation. Overall, our results suggest that patients exhibiting a poor VNS efficacy may present a more pathological thalamocortical circuitry before VNS implantation. EEG synchronization measures could provide interesting insights into the prerequisites for responding to VNS, in order to avoid unnecessary implantations in patients showing a poor therapeutic efficacy.
Keyphrases
- end stage renal disease
- drug resistant
- resting state
- ejection fraction
- functional connectivity
- newly diagnosed
- physical activity
- peritoneal dialysis
- working memory
- sleep quality
- prognostic factors
- randomized controlled trial
- multidrug resistant
- magnetic resonance
- magnetic resonance imaging
- multiple sclerosis
- machine learning
- computed tomography
- cystic fibrosis
- artificial intelligence
- patient reported
- big data
- anti inflammatory
- data analysis