6  Tutorial: Reproducible-fMRI on ds000102 (Flanker task)

A worked end-to-end walkthrough using the Reproducible-fMRI template on the OpenNeuro ds000102 dataset (Andy’s Brain Book canonical example). Each numbered chapter starts where the previous one ended, mirroring Andy Jahn’s pedagogical structure but using our make flow throughout.

6.1 Why ds000102?

  • Public: no IRB hassle.
  • Real: Flanker task fMRI, 26 subjects, suitable for both task GLM and group analysis.
  • Small: ~2 GB; downloads in ~30 minutes.
  • Cross-tutorial reference: Andy’s Brain Book uses it; if you’ve already done his fMRIPrep tutorial #1-#3, you have the reports already.

6.2 Chapters

  1. 01_setup_and_download.md — Clone the template, configure for local execution, download ds000102 from OpenNeuro. (5 min)
  2. 02_run_fmriprep.md — make preprocess for one subject, read the HTML report. (1-2 hours per subject)
  3. 03_qc_and_inclusion.md — Use make qc-rater to rate subjects → qc_decisions.tsv → cohort-level inclusion mask. (10-30 minutes per cohort)
  4. 04_resting_state_derivatives.md — make alff / make reho / make seed-fc on the preproc BOLD; verify outputs. (5-10 minutes)
  5. 05_task_glm.md — Author a .smdl.json, run make glm, render results into a subject report. (10-20 minutes)
  6. 06_group_analysis.md — Apply inclusion mask, run group-level FitLins/nilearn, render group report. (30-60 minutes)

Status (April 2026): Chapter 1 + 4 written; 2/3/5/6 are pending. Use the in-source examples in examples/lc-study/ (a complete end-to-end demo using the LC-pitch synthetic data) until full Brain Book parity lands.

6.3 How to run

Each chapter is a Markdown file with copy-paste-able shell blocks. For an automatable version, use the examples/lc-study/run_lc_demo.sh script as a model — produces a complete cohort QC dashboard from synthetic 5-subject BIDS in ~2 seconds.

6.4 Cross-walk to Andy’s Brain Book

Brain Book chapter Our equivalent
Tutorial #1: Download Chapter 1 above
Tutorial #2: Run fMRIPrep Chapter 2 above (make preprocess)
Tutorial #3: HTML report Chapter 3 above (make qc-rater)
Tutorial #4: Extra preproc Built into our defaults (smoothing, scaling)
Tutorial #5: 1st-level GLM Chapter 5 above (make glm RUNNER=nilearn)
Tutorial #6: Group analysis Chapter 6 above

We don’t aim to replace Brain Book — we scaffold what Brain Book teaches. Read his chapters for the conceptual narrative; run our make commands for the executable infrastructure.