Computes the lesion volumetry and location of a subject.

For a better understanding of the pipeline as well as the files it uses/produces, please report to the pipeline schema les_voloc_pipeline.png

• First argument must be the subject id.
• recon-all.sh must already have been launched.
• Files must be organized in BIDS format.
• If no option is specified, the '-all' option is assumed.
• Run "source source setup_env.sh" before executing this script.

     les_voloc $SUBJECT_ID [-all|-p|-s|-a|-h]

• -all   Performs all the pipeline.
• -p     Steps 1-2  : Processes data (MPRAGE and FLAIR) to prepare for samseg lesion localisation.
• -s     Step 3     : Runs samseg.
• -les   Step 4     : Makes the binarized lesion mask.
• -a     Steps 5-10 : Runs analyses and makes output db files.
• -h     Prints this help.

• anat/sub-${SUBJECT_ID}_FLAIR.nii.gz
• anat/sub-${SUBJECT_ID}_MPRAGE.nii.gz

• sub-${SUBJECT_ID}_lesions.csv
• sub-${SUBJECT_ID}.csv

1. Normalize FLAIR (<1min)
2. Register MPRAGE to normalized FLAIR (<5min)
3. Run samseg segmentation. Produces lesion probability mask. (<1h)
4. Binarize the lesion probability mask up to a certain threshold (<1min)
5. Transform the Freesurfer segmentation (product of recon-all) into the normalized space (Samseg space) (<5min)
6. Transform the binarized lesion mask (product of samseg) into Freesurfer space (<3min)
7. Merge the brain segmentation files with the lesion masks in both samseg and Freesurfer spaces (<2min)
8. Lesion labelling, volumetry and location (<15min)
9. Recompute Freesurfer volumetry based on the new segmentation file (<30min)
10. Make subject-xxx.csv (<1min)

Maxence Wynen - UCLouvain maxencewynen@gmail.com 2020