Magnetic Resonance Spectroscopy is a non-invasive tool dedicated to in vivo interrogation of brain metabolite concentrations, applicable in both healthy and diseased humans and animals. In this thesis three main problems are investigated: (a) the effects of chemical shift displacement error and corrective solutions when using the PRESS product-sequence available on General Electric (GE) 3 T MR scanners; (b) the increase in accessible spectral content at analysis of GABA and GSH signals edited using HERMES and HERCULES at 3 T; and (c) assessment for the need of spectral editing on preclinical scanners (Bruker 9.4 T) via comparison of PRESS, MEGA-PRESS, and HERMES. A general description of the chemical shift displacement was provided along with results for phantom and human investigations on corrections of metabolite concentrations. PRESS 3 T GE data was retrospectively and prospectively corrected for chemical shift displacement error using our proposed solutions. The aspect of scarce spectral content in HERMES/HERCULES edited spectra was resolved using a new LCModel-based analysis method incorporating both the Hadamard encoding and reconstruction methodologies and the spectral concatenation formalism. Multiplexed fitting and simultaneous quantification of GABA and GSH edited spectra enabled access to more robust concentration estimates. Preclinical spectral editing methods were also compared with conventional spectroscopy, while investigating GABA detection performance of PRESS, MEGA-PRESS, and HERMES. Phantom and animal studies suggested the need for spectral editing even at high field strength when studying low-concentration brain chemicals. All methods presented in this thesis are expected to further facilitate the translation and application of Magnetic Resonance Spectroscopy to a wide range of studies across human and animal cohorts. All our work was aimed at improving the accessibility and scope of Magnetic Resonance Spectroscopy research.
|Date of Award
|1 Mar 2023
|David Lythgoe (Supervisor), Diana Cash (Supervisor) & Richard Edden (Supervisor)