SCAIFIELD MRSI analysis

This tool performs the SCAIFIELD MRSI analysis

Prerequisite

• Data must already be in BIDSlike format
• Some third party software
  • LCModel
  • FSL

Usage

just call

python mrsi_analysis.py [-h] [--path PATH] [--site SITE] [--sub SUB] [--ses SES]
  --path PATH  path #BIDS folder (default: None)
  --site SITE  site (default: None)
  --sub SUB    subject name (default: None)
  --ses SES    session name (default: None)

Output

You can find the output in

path/derivatives/site/sub/ses/mrsi/lcm

path/derivatives/site/sub/ses/mrsi/maps

• lcm
  • LCModel input and output of each voxel within brain mask
• maps
  • Metabolites' maps (in .nii format)
  • FWHM/SNR maps

Working principle

1. Read in all .dcm / .IMA files
   • Sort for z-Position (code borrowed from suspect)
2. Appy Hamming filter
3. Create dummy nii file (in MRSI image space)
4. Use MPRAGE to create brain mask
   • register to MRSI image space (only quaternions)
5. Perform LCModel quantification for all voxels in brain mask
6. Read in LCModel files and create maps

TODO:

• Pseudo single-voxel