Skip to content

Commit

Permalink
Create using-clairvoy
Browse files Browse the repository at this point in the history
  • Loading branch information
francopestilli authored Jul 25, 2024
1 parent 00a6f06 commit 9056605
Showing 1 changed file with 38 additions and 0 deletions.
38 changes: 38 additions & 0 deletions docs/tutorial/using-clairvoy
Original file line number Diff line number Diff line change
@@ -0,0 +1,38 @@
# Clairvoy Homepage

## Welcome to Clairvoy.

Clairvoy is an experimental App that runs on brainlife.io. The Clairvoy App utilizes a machine learning, prognostic model to predict whether someone who has experienced a concussion is at risk for slow recovery (defined as not cleared for full return to sport at greater than 28 days). The dMRI and T1 image should be collected at 24-48 hours after a diagnosed concussion. This experimental tool is entirely based on the publication of Bertó et al. 2024 (Neuroimage: Clinical, accepted for publication). Clairvoy was funded by the U.S. Department of Defense, award W81XWH-20-1-0717, PI: Nicholas L Port. The dataset analyzed by Bertó, et al. is from the NCAA/DOD CARE Consortium and was jointly funded by the NCAA and DOD, award W81XWH-14-2-0151, PI: Thomas W McAllister.

Please keep in mind that Clairvoy is an experimental app. We encourage clinicians and scientists to try Clairvoy and appreciate any and all feedback on how to refine the app. Drs. Port and Pestilli are actively seeking funding to conduct a clinical trial to validate Clairvoy, but, at the present time, Clairvoy has not been validated for clinical use. Please provide feedback directly to Dr. Port: [email protected]. Thank you for trying Clairvoy and helping us refine the application.


## Directions

The Clairvoy App lives inside the Brainlife.IO platform, an easy-to-use, web-based, open cience data analysis platform focused on neuroimaging. Free and open to anyone in the world, Brainlife has over 1,000 users across dozens of countries. Brainlife offers free access to the super-computer built into the platform. If you become a regular user, you have the option to attach your own supercomputer resources to BrainLife, and this may significantly speed up your analyses. Before using Cairvoy, you will need a free Brainlife login. We recommend reading the “getting started” help pages at this link: https://brainlife.io/docs/user/started/. There are also many helpful videos on the Brainlife youtube channel ( https://www.youtube.com/@brainlifeio/videos ).

Clairvoy uses the web-based ezBIDS tool (built into Brainlife) to import Dicom files, convert them automatically to NIfTI format, and then create a BIDS project (Brain Imaging Data Structure). You can learn more about ezBIDs at this link (https://brainlife.io/docs/using_ezBIDS/) and watch a video at this link(https://www.youtube.com/watch?v=KvhIHxzHsl4). The dMRI and T1 protocols can be found in Berto, et al. 2024 and Nencka, et al. 2018 (Brain Imaging Behav 12:1121-1140).

Step-by-step directions for Clairvoy:

A narrated video of how to use Clairvoy is available at [this link](https://youtu.be/TvVmsXITP_0).

- Step 1. Upload files into a BIDS project
Login into ezBIDS with your free Brainlife account.
The homepage will now open to the “Upload Imaging Data” on the left side tab. Drag and drop, or browse, for the files you wish to upload. Zip’ing the files may increase upload speed.
Fill in the relevant details of the Dataset Description: Dataset Name (eg. “DWI_workflow”, Authors (eg. your name). Some elements will be extracted from the DIACOM, eg. Subject and Session. If the data is not deidentified, you may wish to change the subject name to a private subject id number. In BIDS Datatype, choose IntendedFor to #1dwi/dwi. It is also best to deface at this stage, allowing for the data to be shared in the future. Choose quickshear method and run. When finished, visually confirm the brain is still in the image and select “Use Defaced.”
In Participants Info, you may want to choose the sex and age of the subject (this is not HIPAA-protected information), then select Finalize to convert to BIDS.
It is considered best practice to save the download configuration/template to your computer. Next, choose “send to Brainlife.IO and run DWI Pipeline.” Finally, choose submit and the pipeline will start.

- Step 2. Switch to Brainlife
Switch from ezBIDS to Brainlife. In Projects, choose your Dataset name (e.g. DWI_workflow).
The Pipeline tab should show the pipeline is “active” and “processing.”
Wait for completion. Time to completion depends on the resources. If you have private access to a super computer, completion may take less than an hour. If you use a public Brainlife supercomputer, the pipeline may take 24 hours to complete.

- Step 3. Run the Clairvoy App
After completion of the pipeline, choose the Archive tab for your project, then select “tractmeasures” and choose “Stage to process.” Give the process the name “Clairvoy” and choose submit.
In the Preprocess tab, choose Clairvoy and click ‘Submit App.’
Choose Clairvoy App and click submit. This should only take a minute or so to complete.
When finished, the output will say either: “This injury is at high risk of developing PPCS” in red letters, or “This injury is predicted to recover in under 28 days” in green letters.

[Here is a narrated YouTube video about how to run Clairvoy](https://youtu.be/TvVmsXITP_0)

0 comments on commit 9056605

Please sign in to comment.