A hands-on, reproducible guide to analysing functional MRI — from a raw scanner export all the way to a thresholded group activation map and task-based connectivity. The material grew out of the MSense Lab seminar series and is updated as tools and best practices evolve.

Who it is for. Master’s and early-PhD researchers who have fMRI data (or soon will) and want to understand not just which button to press but why each step matters. We assume some Python and basic statistics; everything MRI-specific is built up from the ground.

What it covers. The tutorial walks through the whole task-fMRI pipeline, one step at a time, with runnable code (FSL, Nilearn, Nipype) and worked examples:

How to use it. Read top to bottom for a full course, or open a single chapter as a standalone reference — each is self-contained. The companion seminar slides (Session 1 Slides, Session 2 Slides) give the same material in a lecture format, and the Server and Software section covers the computing setup.

A note on philosophy. We favour reproducible, well-documented choices over black-box defaults. Wherever a decision changes the result — a confound set, a smoothing kernel, a correction method — we try to lay out the trade-off so you can justify it in your Methods section. Same data plus the same documented choices should give the same map.