TECHNICAL CONSIDERATIONS FOR 1H - MAGNETIC RESONANCE SPECTROSCOPY (1H-MRS) MEASUREMENT OF SPINAL CORD GAMMA-AMINOBUTYRIC ACID (GABA)

Abstract

Neurometabolites in the spinal cord play a significant role in biomechanics and neural plasticity. However, due to low concentrations and the limited geometry of the spinal cord, neurometabolites are not investigated with imaging modalities or in vivo. There- fore, the goal of this 1H-Magnetic Resonance Spectroscopy (1H-MRS) research was to assess a variety of techniques that aim to improve collection and quantification of in-vivo neurometabolite data in the cervical spinal cord using a Magnetic Resonance Imaging (MRI) scanner. This is important as the cervical spine can reveal information on the upper limb and its neuronal innervation nature. In this research, a phantom study was introduced as a proof of concept to ensure data acquisition based on spinal geometry can be repeatable and accurate. T-tests were performed to identify if 1H-MRS analysis was accurate and significant (P<0.05). MR spectral fit quality was assessed using two MRS fitting software packages (LCmodel and Tarquin) to see which more appropriately fit the data, based on standard and known concentrations in phantoms. T-test results found that both software packages can significantly (P<0.05) provide useful information on the spinal cord content concentrations, although software packages use di erent estimation methods. Furthermore, physiological motion such as breathing and pulsatile flow in the vessels that surround and flow through the spinal cord (i.e. CSF, blood) contribute to the highest levels of noise in the MRS signal. Therefore, a study that included the inves- tigation of main magnetic field (B0) and RF transmit field (B1+) inhomogeneities was conducted to evaluate how physiological noise (Pulsatile flow (PPG) and Respiratory flow (RESP) data) a ects data quality. Research found that there exists a relationship between PPG data and the B0 field (Correlation of 0.71) whereas RESP data and the B1+ field also yielded a relationship (Correlation of 0.62). Additional noise reduction techniques were also investigated to determine further improvement in spinal cord MRS data acquisition to best understand neurometabolite behaviour in this challenging tissue.