MR:Processing:Preprocessing

Once files have been transferred from the scanner to Hoffman and converted to .nii, they are run through standard preprocessing steps. All resultant files are stored in:
/u/home/FMRI/sbook/data/<project_id>/subjects/<subject_id>/<scan_date>/analysis/preprocessing

All files

• converted to Radiological orientation if necessary (fslorient -swaporient). Note that this changes the orientation of the original file, as all files should be stored in Radiological orientation before subsequent analysis.
• reoriented to RL PA IS orientation (fslreorient2std)

output: <filename>_rpi.nii.gz

Initial structurals (eg axial coplanar, matched bandwidth)

• skull stripped with f=0.3 (bet)

output: <init_struct>_brain.nii.gz

Main structurals (eg MP-RAGE)

• skull stripped with f=0.3 and bias field & neck cleanup (bet)

output: <main_struct>_brain.nii.gz

• Note: Sometimes, BET will skull strip incorrectly. Modifying the -f and -g may help (see: BET FAQ)
  • When BET fails to skull strip correctly, manual skull stripping may be necessary (see: MR:Processing:Preprocessing:ManualSkullStripping)

Functionals

• example functional obtained

output: <functional>_example.nii.gz

• motion corrected (mcflirt)

outputs: <functional>_mcf.nii.gz, <functional>_mcf_abs.rms, <functional>_mcf_rel.rms, <functional>_mcf.par

• mean 3d image calculated

output: <functional>_mean.nii.gz

• skull stripped with f=0.3 (bet2)

output: <functional>_mask.nii.gz

• mask motion corrected data with brain mask

output: <functional>_mcf_brain.nii.gz

DTI

• file converted from DICOM to .nii with dcm2nii

outputs: <dti>_raw.nii.gz, bvecs, bvals

Manual preprocessing

The program that performs preprocessing on Hoffman is hoffman_preproc.py. It should be in your path. Type 'hoffman_preproc.py --h' to see the usage.

Usage: hoffman_preproc.py -p <project_name> -s <subject_id> -d <date> <options>
Run basic preprocessing on raw .nii files
Example: hoffman_preproc.py -p AD -s AD_10374 -d 20100909 -m mp_rage_3 -f axial_bold_ur1_5
Assumed directory structure: /u/home/FMRI/sbook/data/<project_name>/subjects/<subject_id>/<date>
Options:
 -h, --help            show this help message and exit
 -p PROJ, --project=PROJ
                       project name
 -s SUBJID, --subjid=SUBJID
                       directory name for the subject
 -d DATE, --date=DATE  date directory name for the subject
 -m MAINSTRUCT, --mainstruct=MAINSTRUCT
                       main structural image file prefix, eg mp_rage_3
 -i INITSTRUCT, --initstruct=INITSTRUCT
                       initial structural image file prefix, eg axial_t2_4
 -f FUNC, --func=FUNC  functional image file prefix, eg axial_bold_ur1_5
 -t DTI, --dti=DTI     diffusion image file prefix, eg dti_30_dir_9
 -q                    use qsub