Process Data¶
This guide walks through a complete DCEPrep run from start to finish using Docker.
Prerequisites¶
- Docker installed and running (see Installation)
- MATLAB and FreeSurfer license files available
- BIDS-organized data (see BIDS Data Setup)
Step 1 — Point to your data¶
Edit run_docker.sh line 24 to point to your BIDS rawdata directory:
Step 2 — Launch the container¶
The container starts with DCEPrep's working directory as the current directory and your data directory mounted inside.
Step 3 — Run preprocessing¶
| Flag | What it does |
|---|---|
-d /data/rawdata | Path to your BIDS rawdata folder (required) |
-b | Bias field correction |
-Z | Z-axis slice normalization |
-m | Head motion correction |
For a first run on new data, add -c to start with a clean derivatives folder:
Step 4 — Run analysis¶
| Flag | What it does |
|---|---|
-d /data/rawdata | Path to your BIDS rawdata folder (required) |
-s | Skip cases whose Ktrans maps already exist |
Add -f to enable FreeSurfer white matter parcellation for subregion analysis:
Step 5 — Review outputs¶
After the analysis completes, find your reports in each subject/session's derivatives folder:
derivatives/
└── sub-01/
└── ses-01/
├── case_report.html # Per-case QC report
├── dce/
│ ├── sub-01_ses-01_Ktrans.nii
│ └── sub-01_ses-01_vp.nii
└── anat/
└── sub-01_ses-01_space-DCEref_T1map.nii
A population_report.html is generated at the top of the derivatives tree once all cases finish.
Common first-run example¶
This is a typical first run with all major corrections enabled and comparison mode to keep outputs labeled:
# Preprocess with bias correction, z-normalization, and auto AIF
./preprocess_all.sh -d /data/rawdata -b -c -Z -A -C fullPrep
# Analyze with FreeSurfer parcellation, skipping already-processed cases
./DCE_all.sh -d /data/rawdata -s -f -C fullPrep