12  References

Standards, tools, and key papers that inform this template’s design.

12.1 Standards

12.1.1 Brain Imaging Data Structure (BIDS)

BIDS normalizes folder structure, filenames, and metadata so labs can exchange MRI datasets without ad-hoc README files. This template uses BIDS for raw data layout, derivatives naming, and statistical model definitions (.smdl.json).

12.1.2 BIDS Apps

  • Paper: Gorgolewski et al. (2017). “BIDS Apps: Improving ease of use, accessibility, and reproducibility of neuroimaging data analysis methods.” PLoS Computational Biology, 13(3):e1005209.
  • Registry: bids-apps.neuroimaging.io

Container-based analysis tools (fMRIPrep, MRIQC, XCP-D, FitLins) that read BIDS input and write BIDS derivatives. The template wraps these via run_*_hpc.sh launchers with automatic container resolution.

12.2 Tools

Tool Version Purpose Reference
fMRIPrep 25.x Preprocessing Esteban et al. (2019) Nature Methods
MRIQC 24.x Quality control Esteban et al. (2017) PLoS ONE
XCP-D 0.10.x Denoising (rest/FC) Mehta et al. (2024) Imaging Neuroscience
FitLins 0.11.x BIDS Stats Models execution fitlins.readthedocs.io
GLMsingle 1.x Single-trial beta estimation Prince et al. (2022) eLife
nilearn 0.11.x Python neuroimaging analysis Abraham et al. (2014) Frontiers in Neuroinformatics
DataLad 1.x Version-controlled data management Halchenko et al. (2021) JOSS

12.3 Methodology

12.3.1 Better Code, Better Science

Playbook for applying software engineering habits (version control, testing, packaging, documentation) to neuroscience analysis. This template follows BCBS principles: code/data separation, pinned environments, automated testing, documentation as a feature.

12.3.2 Reproducible Neuroimaging Frameworks

Framework Focus Reference
Neuroscout Automated fMRI analysis platform de la Vega et al. (2022) NeuroImage
Neurodesk Portable analysis environment Renton et al. (2024) Nature Methods
HALFpipe Interactive reproducible analysis Waller et al. (2022) Human Brain Mapping

12.4 Citation

If you use this template, cite it via the CITATION.cff at the repository root:

Olsson, E. (2025-2026). Reproducible-fMRI: An open-source template framework
for reproducible multimodal neuroimaging analysis. Cognitive & Neural Computation
Lab, UC Irvine. https://github.com/CNClaboratory/Reproducible-fMRI

12.5 Demo Datasets for Tutorials

Two OpenNeuro datasets are recommended for tutorials, smoke tests, and demos. They are public, small enough to download fast, and used by other community tutorials so users can compare outputs.

Dataset Size Subjects Tasks Use case
ds000102 ~2 GB 26 Flanker (cognitive control) Andy’s Brain Book canonical example. Use when teaching task GLM end-to-end.
ds000114 ~3 GB 10 Multiple (motor, finger-foot-lips, line-bisection, covert-verb) Smaller subject count; good for fast smoke tests on multiple tasks.

Download with aws s3 sync --no-sign-request:

# Flanker (Brain Book canonical)
aws s3 sync --no-sign-request s3://openneuro.org/ds000102 ~/data/ds000102

# Multi-task (BIDS-Examples-style)
aws s3 sync --no-sign-request s3://openneuro.org/ds000114 ~/data/ds000114

Cross-walk: when you have one of these datasets, every step in docs/GETTING_STARTED.md works as documented. Andy’s Brain Book Tutorial #2 walks through fMRIPrep on ds000102 and is a natural complement to our make preprocess flow.

12.6 Tutorials and Pedagogy

These external resources teach concepts our template assumes you already know. Recommended reading paths for new users:

Resource When to read
Andy’s Brain Book — fMRIPrep Tutorial First time running fMRIPrep. Six numbered chapters: Download → Run → Read HTML report → Extra preproc → 1st-level GLM → Group analysis. Each chapter has a YouTube video.
Andy’s Brain Book — fMRIPrep Tutorial #3 (HTML report) Read this before approving QC for any subject. Six-screenshot walkthrough of the fMRIPrep visual report.
Andy’s Brain Book — 1st-Level Analysis Before authoring your first BIDS Stats Model.
Brainhack School — fMRIPrep 101 Alternative to Brain Book; more concise. Good for one-day onboarding sprints.
Princeton Brainhack Handbook — fMRIPrep HPC-flavoured walkthrough; uses Singularity + SLURM, similar to our HPC stack.
Russell Poldrack — Better Code Better Science Why we structure code, tests, and docs the way we do. The template’s docs consolidation pattern came from here.