Abstracts

Rationale: VNS is a well-known add on treatment for patients with refractory epilepsy. It has been shown, that triggering the stimulation at seizure onset increases the response rate from 35% to 59% compared to continuous stimulation (1). Since the timing of the stimulation is crucial, a reliable method to detect seizures is needed. We examined the spiking behavior in the vagus nerve electroneurogram (VENG) in anesthetized rats during acutely induced pentylenetetrazol (PTZ) seizures. Methods: Ten male Wistar rats 370g +-42g were used. Anesthesia was induced using 60mg/kg Ketamine and 7mg/kg Xylazine i.p. and maintained by 60mg/kg Ketamine i.p. Three epidural stainless-steel electrodes were implanted in the skull ([-]:AP:-6mm, ML:-1.5mm, [+]:AP:+2mm,-2mm, [GND]:AP: +10mm, ML: +0mm ) to record EEG. The vagus nerve was isolated, desheated and either placed on a 3-contact hook electrode made of 200µm Ag wire (n=8) or placed in a cuff electrode (x-wide contact micro-cuff, Microprobes, Gaithersburg, USA) (n=2). 10mg/ml PTZ solution was infused in the tail vein at 0.5ml/min/kg. All signals have been recorded using self-made amplifiers and MATLAB 2018a software (MathWorks, Natick, USA). Spike sorting was performed using a Matlab application adapted from the Osort spike detector (2). Results: Seizures were recorded in 10 animals (mean PTZ dosage: 56.2mg/kg +-19.2mg/kg, n=8). Four animals were excluded due to a suboptimal recording because of an insufficiently hydrated vagus nerve (1/4) or a too noisy VENG recording (3/4). Six animals showed respiratory related burst activity in the VENG recording, similar to previously reported VENG data (3).Tonic-clonic seizures were observed within 6.5-30.2 minutes after the start of the PTZ-infusion. New spike cluster, different from baseline activity, could be detected 3.8-15.4 minutes before the tonic-clonic stage in 5/6 animals (Fig. 1). Two experiments were performed to prove the genuine character of the VENG recorded spike clusters: 1) In one animal (1/6), the vagus nerve was silenced by the application of 0.05ml 2% lidocaine solution to the desheated nerve. As a result, the respiratory-related burst activity vanished, and no new spike cluster could be identified during the seizure evolution. 2) In 2 animals, a second hook electrode was placed on the sternohyoid muscle as a control for EMG artifacts. Spike trains extracted from the EMG control and the VENG recording were poorly correlated (Pearson's r: 0.033 +- 0.011, p <<0.001). Moreover, in the EMG control, the signal amplitude was decreased by factor 10. This indicates that the recorded VENG spikes did not reflect any EMG artifacts.  Conclusions: Altogether, our results show the occurrence of a specific VENG spike activity early on during induced acute seizures in anesthetized rats. Recording of VENG may be useful to detect seizure onset and provide a basis for future closed-loop vagus nerve stimulation (VNS) systems.  Funding: The research was supported by a grant from the Walloon Region-Pôle de Compétitivité Mecatech445 (Projet NEUROPV).