This tutorial introduces machine learning applied to functional MRI data, with hands-on exercises using nilearn. It covers functional connectivity, predictive modelling, and interpretation of ML results in a neuroimaging context.
Contents¶
Functional connectivity with nilearn — computing and visualizing connectomes from fMRI data
Machine learning with nilearn — building and evaluating predictive models
Exercises: ADHD classification — applying the methods to a real research question
Dataset¶
The tutorial uses the NeuroDev dataset — a developmental functional MRI dataset preprocessed and packaged by Elizabeth Dupre specifically for this tutorial. This dataset has since been integrated as one of the main tutorial datasets in nilearn itself.
History and acknowledgments¶
This tutorial has grown through several iterations and owes its existence to a wide community of contributors.
| Year | Event | Instructors |
|---|---|---|
| 2020 | Brainhack School | Jacob Vogel — original lecture developed for Jean-Baptiste Poline’s course QLSC 612 at McGill University |
| 2022 | MAIN Educational Workshop | Hao-Ting Wang and Yasmin Mzayek — Montreal Artificial Intelligence and Neuroscience (MAIN) conference |
| 2024 | MAIN Educational Workshop | Hao-Ting Wang and Himanshu Aggarwal — MAIN conference |
| 2026 | PSY3019 — Université de Montréal | Material adapted for an undergraduate psychology course by Lune Bellec |
We are grateful to all instructors, teaching assistants, and students who contributed feedback and improvements across these iterations.
License¶
This material is shared under the terms of the LICENSE file in this repository.