giga-connectome

This is a BIDS-App to extract signal from a parcellation with nilearn, typically useful in a context of resting-state data processing.

You can read our JOSS paper for the background of the project and the details of implementations.

Description

Functional connectivity is a common approach in analysing resting state fMRI data. The Python tool Nilearn provides utilities to extract and denoise time-series on a parcellation. Nilearn also has methods to compute functional connectivity. While Nilearn provides useful methods to generate connectomes, there is no standalone one stop solution to generate connectomes from fMRIPrep outputs. giga-connectome (a BIDS-app!) combines Nilearn and TemplateFlow to denoise the data, generate timeseries, and most critically giga-connectome generates functional connectomes directly from fMRIPrep outputs. The workflow comes with several built-in denoising strategies and there are several choices of atlases (MIST, Schaefer 7 networks, DiFuMo, Harvard-Oxford). Users can customise their own strategies and atlases using the configuration json files.

Supported fMRIPrep versions

giga-connectome fully supports outputs of fMRIPrep LTS (long-term support) 20.2.x.

For fMRIPrep 23.1.0 and later, giga-connectome does not support ICA-AROMA denoising, as the strategy is removed from the fMRIPrep workflow.

Quick start

Pull from Dockerhub (Recommended)

docker pull bids/giga_connectome:latest
docker run -ti --rm bids/giga_connectome --help

If you want to get the bleeding-edge version of the app, pull the unstable version.

docker pull bids/giga_connectome:unstable

How to report errors

Please use the GitHub issue to report errors. Check out the open issues first to see if we're already working on it. If not, open up a new issue!

How to contribute

You can review open issues that we are looking for help with. If you submit a new pull request please be as detailed as possible in your comments. If you have any question related how to create a pull request, you can check our documentation for contributors.

Contributors

Hao-Ting Wang 🤔 🔬 💻 ⚠️	Quentin Dessain 📓 📦	Natasha Clarke 📓 💡 🐛	Remi Gau 🚇 🚧	Lune Bellec 🤔 💵	Jon Cluce 🐛	Emeline Mullier 🐛
James Kent 🐛 📖	Marcel Stimberg 📓 📖 🐛

Acknowledgements

Please cite the following paper if you are using giga-connectome in your work:

@article{Wang2025,
    doi = {10.21105/joss.07061},
    url = {https://doi.org/10.21105/joss.07061},
    year = {2025}, publisher = {The Open Journal},
    volume = {10},
    number = {110},
    pages = {7061},
    author = {Hao-Ting Wang and Rémi Gau and Natasha Clarke and Quentin Dessain and Lune Bellec},
    title = {Giga Connectome: a BIDS-app for time series and functional connectome extraction},
    journal = {Journal of Open Source Software}
}

giga-connectome uses nilearn under the hood, hence please consider cite nilearn using the Zenodo DOI:

@software{Nilearn,
    author = {Nilearn contributors},
    license = {BSD-4-Clause},
    title = {{nilearn}},
    url = {https://github.com/nilearn/nilearn},
    doi = {https://doi.org/10.5281/zenodo.8397156}
}

Nilearn’s Research Resource Identifier (RRID) is: RRID:SCR_001362

We acknowledge all the nilearn developers as well as the BIDS-Apps team

This is a Python project packaged according to Contemporary Python Packaging - 2023.