Skip to main content

Running CIVET on ChRIS at FNNDSC

CIVET is a T1 MRI processing pipeline (similar to FreeSurfer) developed at the Montréal Neurological Institute (MNI). This tutorial is a beginner-level walk-through of how to run CIVET on ChRIS at the FNNDSC.

1. Log in to ChRIS​

Go to http://chris-next.tch.harvard.edu:2222/login and log in using your FNNDSC account username and password.

2. Create a feed​

Data are imported into ChRIS in feeds.

  1. Click "Create Feed" in the top-right corner.
  2. Choose the "Generate Data" option.
  3. Select the pl-neurofiles-pull plugin, which pulls data from /neuro/labs/grantlab/research into ChRIS.
  4. Specify which path to pull as the --path parameter.
  5. Keep clicking "Next" until the feed is created.

Screenshot of creating a feed in ChRIS

The path you choose must be a directory containing exactly one NIFTI file called *_T1w.nii. Any unrelated files (such as *.json, *._T2w.nii) will also be uploaded to ChRIS but ignored by the pipeline.

3. Run the pipeline​

Click into the feed you just created, and right-click the only circle in the top-left quadrant. Select "Add a pipeline."

Screenshot of feed view

Search for the "CIVET" pipeline and select it.

danger

Carefully review the pipeline's options before you run it. Make sure that each step of the pipeline is configured to run on "galena," our private compute cluster. You may also choose "NERC," a public cloud, as long as your data does not contain PHI.

Screenshot of pipeline configuration

Once the pipeline is running, you can see each of its steps and their statuses. You can see that the pipeline selects the *_T1w.nii file, renames it to *_t1.nii, then converts it to MINC.

Screenshot of the pipeline as its running

4. Wait a couple hours​

CIVET takes 4-8 hours to run to completion.

Next Steps​

You now know how to run CIVET using the ChRIS user interface. Next steps might include visualization and downstream analyses of CIVET outputs, or scaling up to run this pipeline on multiple subjects.