Improved reproducibility of γ‐aminobutyric acid measurement from short‐echo‐time proton MR spectroscopy by linewidth‐matched basis sets in LCModel

• Published in: NMR in biomedicine Vol. 37; no. 2; pp. e5056 - n/a
• Main Authors: Xiao, Ying, Lanz, Bernard, Lim, Song‐I, Tkáč, Ivan, Xin, Lijing
• Format: Journal Article
• Language: English
• Published: England Wiley Subscription Services, Inc 01.02.2024; John Wiley and Sons Inc
• Subjects: basis set; Brain; Brain - metabolism; Coefficient of variation; Cortex (motor); GABA; gamma-Aminobutyric Acid - metabolism; Homeostasis; Humans; LCModel; Magnetic fields; Magnetic resonance spectroscopy; Metabolites; Prefrontal cortex; Proton Magnetic Resonance Spectroscopy - methods; Protons; Reproducibility; Reproducibility of Results; short‐TE proton MRS; Spectra; spectral linewidth; Spectroscopy; Spectrum analysis; Thalamus; γ-Aminobutyric acid; GABA; basis set; short-TE proton MRS; LCModel; reproducibility; spectral linewidth
• ISSN: 0952-3480; 1099-1492; 1099-1492
• DOI: 10.1002/nbm.5056

γ‐Aminobutyric acid (GABA), as the primary inhibitory neurotransmitter, is extremely important for maintaining healthy brain function, and deviations from GABA homeostasis are related to various brain diseases. Short‐echo‐time (short‐TE) proton MR spectroscopy (1H‐MRS) has been employed to measure GABA concentration from various human brain regions at high magnetic fields. The aim of this study was to investigate the effect of spectral linewidth on GABA quantification and explore the application of an optimized basis‐set preparation approach using a spectral‐linewidth‐matched (LM) basis set in LCModel to improve the reproducibility of GABA quantification from short‐TE 1H‐MRS. In contrast to the fixed‐linewidth basis‐set approach, the LM basis‐set preparation approach, where all metabolite basis spectra were simulated with a linewidth 4 Hz narrower than that of water, showed a smaller standard deviation of estimated GABA concentration from synthetic spectra with varying linewidths and lineshapes. The test–retest reproducibility was assessed by the mean within‐subject coefficient of variation, which improved from 19.2% to 12.0% in the thalamus, from 27.9% to 14.9% in the motor cortex, and from 9.7% to 2.8% in the medial prefrontal cortex using LM basis sets at 7 T. We conclude that spectral linewidth has a large effect on GABA quantification from short‐TE 1H‐MRS data and that using LM basis sets in LCModel can improve the reproducibility of GABA quantification. Short‐TE proton MR spectroscopy (1H‐MRS) has been used to measure γ‐aminobutyric acid (GABA) concentration. This study focused on reducing the spectral linewidth effects on GABA quantification by using the spectral‐linewidth‐matched basis sets in LCModel, which enabled improving the reproducibility of GABA measurement.