Preclinical ORYX-MRS: a toolbox for proton magnetic resonance spectroscopy data analysis in rat brain

Magnetic resonance spectroscopy (MRS) is a powerful method for investigating brain metabolism, but existing software tools are primarily designed for human studies and lack support for preclinical applications. This thesis introduces Preclinical ORYX-MRS, a MATLAB-based graphical toolbox designed for processing single voxel 1H-MRS data from small-animal models, specifically designed for MR Solutions (Guildford, UK) vendor formats. The toolbox integrates raw data loading, coregistration, atlas-based registration, tissue segmentation, spectral quality evaluation, and ROI-based metabolite quantification into a simplified workflow. Its performance was assessed using different datasets of rat brains. Accurate voxel localization, maintaining spatial integrity during atlas registration, automatic gray matter, white matter, and CSF segmentation, and processing of both raw .mrd files and LCModel outputs (.coord, .table) are demonstrated in the results. The toolbox's ability to provide anatomically contextualized metabolite profiles was further demonstrated using ROI analyses. Preclinical ORYX-MRS provides data processing and analysis support for preclinical MRS in contrast to current software mainly tailored for clinical MRS. However, it has some limitations, like validation on a small dataset, restriction to single-voxel MRS data, and current support for only MR Solutions data format. Multi-voxel MRS imaging (MRSI) data analysis, multi-vendor support, and longitudinal statistical analysis are potential future developments of the toolbox. In conclusion, preclinical ORYX-MRS offers a user-friendly, vendor-specific, and streamlined platform for small-animal MRS, bridging the gap between tools designed for clinical studies and the special requirements of preclinical MR research.