Using the CONN Toolbox on the SCC
CONN is a MATLAB- and SPM-based toolbox for computing, displaying and analyzing functional connectivity in fMRI. It covers preprocessing, denoising, quality control, first-level connectivity analyses (seed-based, ROI-to-ROI, ICA, graph-theoretic, gPPI and others) and second-level group statistics, and it writes a methods description of whatever it did.
This guide covers running CONN on Boston University's Shared Computing Cluster (SCC). It documents CONN 25b.
How this guide is organized
CONN is unlike the other tools documented here: it is driven primarily by a graphical interface, not a command line. That shapes how you run it on a cluster, so this guide covers two routes and you can use either.
| Route | You work in | Good for |
|---|---|---|
| The GUI | An interactive desktop session on the SCC | Exploring a dataset, setting up a project, reading results |
| MATLAB code | A conn_batch script, submitted as a job |
Reproducibility, large datasets, unattended runs |
They are not exclusive. A common pattern is to set a project up in the GUI, then script the parts you will repeat.
Before You Start
CONN requires MATLAB and SPM, and on the SCC those must be loaded as separate modules in a specific order. See Getting Started.
Contents
- Getting Started — loading the CONN, MATLAB and SPM modules.
- Launching CONN — starting an interactive session on the SCC.
- Submitting Jobs from the GUI — CONN's cluster settings.
- Submitting Jobs from MATLAB —
conn_batchand batch scripts. - Checking Your Output — quality control and reading results.
- Appendix — resource recommendations and version support.
Additional Resources
- CONN toolbox website
- CONN HPC configuration
- CONN support forum on NITRC
- RCS CONN software page
- RCS modules page
Questions? Contact us at help@scc.bu.edu.