brainprep 🧠

Preprocessing for pediatric (1–7 yo) MRI brain data

This preprocessing tool follows the BIDS standard and writes all outputs directly into your dataset’s derivatives/brainprep/ folder. It works with any BIDS-compliant dataset that provides an code/qc/raw/exclude.yaml file. The workflow runs in three steps and supports both cross-sectional and longitudinal datasets.

Pipeline (per T1w structural volumes):

SynthStrip → N4 → affine registration to template → SynthSeg → WhiteStripe intensity normalization.

Full usage + details (exclude rules, layouts, age filters, outputs) live in the documentation.

---

Installation (conda)

conda env create -f environment.yml
conda activate brainprep

Quick check:

python brainprep.py --help

---

Workflow

1) Create the input list (create_input_txt.py)

Build a .txt list of absolute paths to the images you want to preprocess.

python create_input_txt.py /path/to/bids_root -l long --modality T1w -o to_preprocess.txt

---

2) Run preprocessing (brainprep.py)

Run the pipeline and write outputs into derivatives/brainprep/.

python brainprep.py \
  --inputs to_preprocess.txt \
  --template /path/to/template.nii.gz \
  --bids-root /path/to/bids_root \
  --dataset mydataset_name

---

3) Build the training CSV (create_dataset_csv.py)

Aggregate one or more datasets into a single CSV (paths + demographics + folds).

python create_dataset_csv.py \
  --bids-roots /path/to/hc-calgary-preschool \
  --layouts long \
  --input-lists preprocess_hc-calgary-preschool.txt \
  --age-units y \
  --dest-path-for-images /scratch/$USER/training_inputs \
  --out-csv dataset.csv

---

Project structure

brainprep/
├── brainprep.py
├── create_input_txt.py
├── create_dataset_csv.py
├── environment.yml

Documentation

Author

Andjela Dimitrijevic

How to cite

Publication coming soon...