Systematic Error in the Measurement of [GABA]/[Cr] Ratio Using Methyl Resonance of Creatine

Purpose: GABA is a very important inhibitory neurotransmitter, which is present at a very low concentration in the brain (~1 mM). Special spectral editing is necessary at clinical field strengths in order to detect GABA by magnetic resonance spectroscopy. Jdifference spectroscopy is a very efficient way of spectral editing GABA signal and is widely used. It is a common practice to report [GABA]/[Cr] ratio by measuring the areas under the 3.01 ppm CH2 (C-4) GABA peak in the edited spectrum and the 3.03 ppm CH3 creatine (Cr) peak in the unedited spectrum. We show that this way of measuring [GABA]/[Cr] ratio will introduce a systematic error, and we propose the use of the 3.93 ppm CH2 Cr peak for this purpose. Methods: Both phantom and in vivo scans were performed using a 3 tesla whole body Siemens Total Imaging Matrix Trio system with a circularly polarized transmit/receive head coil. The phantom consisted of a solution of GABA and Cr, each having a 10 mM concentration. MEGA PRESS scans were performed by prescribing a 1×1×1 cm voxel in the phantom and 2×2×2 cm voxel in the sensorimotor cortex in in vivo scans. A water based interleaved navigator was implemented as well in the in vivo scans to assess subject motion, and discard motion corrupted part of data. The scan parameters were TR=2700 ms, TE=68 ms, editing pulse frequency = resonance frequency of the GABA CH2 (C-3) resonance. Macromolecule correction was done in the in vivo scan by (i) placing the off-resonance pulse symmetrically opposite of the 1.7 ppm macromolecule resonance, and (ii) running a metabolite nulling scan. The areas of the 3.01 ppm GABA peak (IGABA) in the final edited spectrum, 3.03 and 3.93 ppm Cr peaks (ICr) in the spectra prior to