Automated ROI-Based Labeling for Multi-Voxel Magnetic Resonance Spectroscopy Data Using FreeSurfer

Area of Science:

• Neuroimaging
• Magnetic Resonance Spectroscopy (MRS)
• Computational Neuroscience

Background:

• Advanced analysis of multi-voxel MRS is vital for accurate neurotransmitter quantification.
• Existing region of interest (ROI)-based labeling for multi-voxel MRS data is limited.
• Neurotransmitter distribution varies across different tissue types, necessitating precise localization.

Purpose of the Study:

• To develop an automated ROI-based labeling tool for 3D multi-voxel MRS data.
• To improve the efficiency and reliability of metabolite quantification in MRS.
• To address the need for precise ROI selection in complex brain imaging analyses.

Main Methods:

• Acquisition of 3D-MRS imaging data with varying spatial resolutions and MEGA-editing.
• Extraction of brain region masks from T1-weighted structural images using FreeSurfer.
• Comparison of automated labeling with manual labeling and single voxel selection for reliability testing in subcortical regions.

Main Results:

• Automated ROI-based labeling demonstrated high consistency with manual labeling (ICC > 0.8).
• Methods relying on spatial averaging within gray matter (GM) showed less variation than uncorrected single voxel selection.
• Significant reduction in hands-on time and elimination of inter-rater bias were observed.

Conclusions:

• An automated ROI-based analysis approach for 3D multi-voxel MRS data has been successfully developed.
• This automated tool enhances efficiency and reduces subjectivity in MRS data analysis.
• The method is applicable to various types of 3D multi-voxel MRS data, facilitating accurate metabolite quantification.