Subtraction artifacts and frequency (mis-)alignment in J-difference GABA editing.

Area of Science:

• Neuroimaging
• Magnetic Resonance Spectroscopy

Background:

• γ-aminobutyric acid (GABA) is a key inhibitory neurotransmitter.
• Accurate in vivo measurement of GABA is crucial for understanding neurological function and disease.
• J-difference editing is a common technique for GABA quantification, but its spectral quality can be affected by magnetic field inhomogeneities.

Purpose of the Study:

• To compare the repeatability of GABA measurements using J-difference editing before and after spectral realignment.
• To evaluate the impact of different spectral alignment approaches (independent vs. pairwise) on GABA measurement repeatability.

Main Methods:

• In vivo GABA measurements were performed in the occipital cortex, sensorimotor cortex, and dorsolateral prefrontal cortex (DLPFC).
• Two alignment strategies were compared: independent alignment of each subspectrum and pairwise alignment of on/off subspectra.
• Simulations were used to assess GABA quantification error under varying B0 drift conditions.

Main Results:

• Pairwise alignment improved group mean coefficient of variation (CV) across all regions (occipital: 0.4%, sensorimotor: 1.1%, DLPFC: 1.1%).
• Independent alignment led to subtraction artifacts and increased CV in the DLPFC by 9.4%.
• Simulations showed GABA quantification error decreased from 4.5% without alignment to <1% with optimal (pairwise) alignment.

Conclusions:

• Pairwise spectral alignment is a superior method for improving the repeatability of in vivo GABA spectroscopy.
• Independent alignment of all subspectra can introduce artifacts and result in poorer repeatability compared to non-aligned data.