Edited in vivo NMR spectroscopy, coupled with multiparametric MRI to locate the epileptogenic zone in an animal model of focal epilepsy

For the 30% of patients with drug-resistant epilepsy, mainly with focal epilepsy, surgical resection of the epileptogenic zone (EZ) can be proposed with an average efficiency of 50-80%. Precise identification of the EZ is essential for effective resection.

Despite a wide panel of non-invasive preoperative explorations, intracranial EEG (iEEG) is often necessary. However, poor spatial sampling of individual electrodes and/or poorly guided implantation can result in the resection of an insufficient brain area. Therefore, the development of non-invasive methods to better identify the EZ and guide electrode implantation remains a challenge.

Metabolomics based on nuclear magnetic resonance spectroscopy (MRS) could address this problem in mesiotemporal lobe epilepsy (MTLE), the prototype of surgical epilepsy. In this study, the KA-MTLE mouse model, highly predictive of human MTLE, with a typical EEG profile and hippocampal sclerosis visible on the anatomical MRI, was used. By ex vivo MRS, the metabolic profile of the EZ in this model was determined, and GABA emerged as the most discriminating metabolite with an increase of around 100%. Thus, GABA appears to be a specific biomarker of the ZE in KA-MTLE mice.

In this thesis, a NMR spectroscopic imaging method was developed to map GABA. The results obtained were coupled with multiparametric MRI data to assess the relevance of combining these two methods for the spatial localisation of the EZ. Finally, a preliminary biological study was conducted to try and establish a link between epilepsy and the measured increase in GABA.