Authors :
Presenting Author: Elena Acedo Reina, Msc – Université catholique de Louvain
Enrique Germany Morrison, PhD – Post Doctoral Fellow, Université catholique de Louvain; Romain Raffoul, M.Eng – PhD Student, Université Libre de Bruxelles; Antoine Nonclercq, PhD – Principal investigator, Université Libre de Bruxelles; Riëm El Tahry, MD, PhD – Neurologist, Principal Investigator, Clinique universitaire Saint-Luc
Rationale: Seizures produce autonomic symptoms, mainly sympathetic but also parasympathetic in origin. Within this context, the vagus nerve (VN) is a key player as it carries information from the different organs to the brain and vice versa. For this reason, exploiting the vagal neural traffic with respect to seizures might offer a novel way to detect seizures and develop a closed-loop Vagus Nerve Stimulation (VNS).
This project aims to develop a VENG-based detection algorithm to detect seizures in an acute kainic acid rat (KA) model. Methods: Six male Wistar rats (296,8
±50,2g) were used. Anaesthesia was induced using 100mg/kg Ketamine and 7mg/kg Xylazine i.p. and maintained by half concentration of the initial mixture.
Three epidural stainless-steel electrodes were implanted on the frontal cortex ([+]: AP:+2mm, ML:±3mm, [-]: AP:+6mm) to record EEG. Three ECG Lab-made Tungsten electrodes were implanted according to Eindhoven’s lead II. Finally, a tripolar Micro-Cuff electrode was implanted around the left cervical portion of the VN. Vagus nerve electroneurogram (VENG) was recorded for 20 minutes during baseline, ictal and postictal periods. Seizures were induced by an injection of KA (0.4µg/0.2µl saline; 0.1µl/min) in the right hippocampus ([RH]: AP:-5.6mm, ML/DV: 4.5mm). The VENG seizure algorithm detects acute changes in the respiratory component of the signal, using a ratio of two sliding windows (Foreground (Fg) 5 sec/Background (Bg) 90 sec) of the mean power of the dominant respiratory component of the VENG. Whenever a change in the dominant respiratory frequency component occurs, the Fg/Bg ratio will decrease. The seizure onset detection was set at a Fg/Bg threshold of 0.5. A logistic regression was performed to distinguish VENG detected and non-detected seizures with heart rate changes and individual EEG band spectrum evolution as predictive variables. Results: Twenty-three seizures were recorded with a mean duration of 51,21 ± 19 seconds. Thirteen seizures were correctly detected by the VENG-based detection algorithm (Fig.1). The detected seizures showed a different ictal EEG pattern compared to non-detected seizures, with a trend in higher beta power (p-value = 0,058), while non-detected seizures showed a significant increase in slower theta band (p-value= 0,0173) and a similar trend in delta band (p-value= 0,0517). In contrast, ictal heart rate changes were not predictive for correct VENG seizure detection (p-value= 0,747).
Conclusions: Kainic, acid induced temporal seizures can be detected through VENG recording, which may offer new opportunities for development of closed loop stimulation. Whenever VENG did not detect the seizures, ictal EEG showed less rapid activity, potentially indicating a less severe seizure and different seizure spread to autonomic central centers. Further work will be performed to reproduce these results in awake epileptic animals.
Funding: Study funded by the Queen Elisabeth Medical Foundation (F.M.R.E.) and Walloon Excellence in Life Sciences and Biotechnology (WELBIO)