• Citer

Recording, modeling and analysis of the vagus nerve activity to improve epilepsy therapy

par Smets, Hugo
Président du jury Gorza, Simon-Pierre
Promoteur Nonclercq, Antoine
Co-Promoteur El Tahry, Riëm
Publication Non publié, 2022-04-25

Thèse de doctorat

• Accès en ligne
• Détails
• Contenu
• Statistiques

Résumé :

Epilepsy is one of the most common neurological diseases globally, affectingaround 50 million people with a prevalence of 6.38 per 1000 people. The epilepticpopulation experience a risk of premature death three times higher than the generalpopulation. Vagus nerve stimulation (VNS) is an adjuvant treatment for refractorypatients not eligible to surgery. Automated seizure detection controlling on-demandVNS would improve the efficiency of the treatment. The seizure detection iscurrently mainly based on electroencephalogram (EEG). The recording vagalneural activity might offer an alternative convenient method for early seizuresdetection to control the stimulation of the same nerve.Before being able to propose such enhancement in the therapy, tools are mandatoryto record and analyze the vagus nerve activity. The treatment of epilepsy is the mainapplication targeted in this work. However, the tools presented here are generic andcan be used in many domains of neurology.Recording spontaneous vagus nerve activity is a challenge and requires an adaptedrecording setup. A recording setup for acute and chronic vagus nerve activitymonitoring in rat models has been developed, implemented, and validated. Theamplification system was evaluated with a test bench and with a phantommimicking the nerve impedance. It was then used in an acute recording setup andevaluated during acute experiments by recording evoked compound actionpotentials (CAPs), then spontaneous vagus nerve activity related to respiration.Finally, the recording platform was developed with specific care for the chronicenvironment, allowing the animal to move freely while maintaining good signalquality.Using the proposed recording setup, we have shown the feasibility of a long-termchronic recording of the vagus nerve activity. The impact of the light-darkvariations on the vagus nerve activity (i.e., spike frequency) is taken as anillustration. The light-dark variations derived from the vagus nerve activity werecompared with the EEG slow waves recordings, an established signal for circadianstudies. The role of the vagus nerve in the circadian rhythm was previously onlyshown by indirect measurements, while our study shows for the first time the light-dark variations of the vagus nerve spike activity by direct means. The recordinglength required to characterize an accurate measurement of the 24-h cycle has beenevaluated.To better analyze the vagus nerve activity, a 3D finite element model (FEM) hasbeen developed. The model integrates the propagation of an action potential (AP)along a nerve fiber and reproduces the recording environment including theisolating cuff and the electrodes. The contribution of an AP to the recorded signalcan be modeled as in an in vivo recording. A methodology has been developed toidentify the type of fiber based on their propagation velocity, i.e., directlyproportional to their diameter. B-fibers could be identified and distinguished fromA-fibers, and an optimized electrode configuration has been proposed dependingon the activity required to be identified. A histogram of active fibers of healthy ratsduring a recording under anesthesia has been realized.Finally, the same FEM was used to improve stimulation electrodes. Variouselectrode design parameters were varied and their impact on the associatedstimulation threshold current were derived. The stimulation current is reduced bychoosing a quasi-tripolar configuration over a bipolar configuration, increasing thecuff length and decreasing the electrode width of the stimulation electrodes. Theseresults were validated during in vivo experiments on rabbits, by comparing threeelectrode designs, which have been chosen based on the results of the simulations.Altogether, this thesis contributes to understand the vagus nerve activity, throughrecording, modeling, and analysis. In the frame of this work, experiments have beenundertaken on healthy animals, as a necessary first step to demonstrate thefeasibility of the developed methodology for further use in fundamental and appliedresearch in neurology. In particular, this thesis opens ways to understand themechanisms underlying epilepsy and improve the treatment of epilepsy. In additionto epilepsy, the growing interests for neurostimulation and peripheral nervemonitoring are among the areas of research in which this work will be beneficial.