GABA-edited MEGA-PRESS at 3 T: Does a measured macromolecule background improve linear combination modeling?

Area of Science:

• Neuroimaging
• Magnetic Resonance Spectroscopy (MRS)

Background:

• J-difference edited gamma-aminobutyric acid (GABA) signals in MRS are often contaminated by co-edited macromolecule (MM) signals, reported as "GABA+."
• MM signals are broader and less characterized than metabolites, typically approximated by Gaussian models.
• Experimentally measured MM signals are recommended but under-studied in edited MRS.

Purpose of the Study:

• To develop and evaluate improved methods for modeling macromolecule (MM) signals in GABA-edited magnetic resonance spectroscopy (MRS).
• To compare a new multi-Gaussian MM parameterization and experimental MM basis functions against the conventional single-Gaussian approach.

Main Methods:

• Acquired GABA-edited MEGA-PRESS data with pre-inversion in 13 healthy controls.
• Derived an experimental MM basis function and a new multi-Gaussian parameterization from acquired data.
• Compared single-Gaussian, multi-Gaussian, and experimental MM models using a public MEGA-PRESS dataset of 61 participants.

Main Results:

• Both experimental MM and multi-Gaussian models showed significantly reduced fit residuals compared to the single-Gaussian model (p=0.034, p=0.031).
• Experimentally derived models estimated a larger MM fractional contribution to the GABA+ signal (58%) compared to the single-Gaussian approach (50%).

Conclusions:

• Single-Gaussian parameterization is insufficient for modeling edited MM signals in GABA-edited MRS.
• Experimentally derived GABA+ spectra and their parameterized replicas enhance the accuracy of GABA+ spectral modeling.