Running the sample script (process_data.sh) using sct_run_batch#

In your working directory, you should have the batch processing script, a config file, and a data directory containing three subjects in BIDS dataset structure. You can double-check this using the ls command:

ls
# Output should be: data/  config.yml  process_data.sh  README.txt
ls data
# Output should be: sub-01  sub-03  sub-05

In order to apply the process_data.sh script to the multi-subject dataset, we can use an SCT tool called sct_run_batch. This command loops across each subject directory and launches the script individually for each subject.

Running this command will take 10-30m depending on the capabilities of your processor.

sct_run_batch -script process_data.sh -config config.yml
Input arguments:

  • -script : The path to the script you wish to run. This script should be written to accept a single argument (a path to the subject folder). That way, sct_run_batch can launch the script once for each subject, passing along each subject folder as an argument to the script.

  • -config : A config file containing input arguments for sct_run_batch. Inside this file, you will find:

    • path-data : Path to the folder containing the BIDS dataset. sct_run_batch will look for subject folders inside, so that it can loop across subjects.

    • subject-prefix : The prefix that each subject folder has.

    • path-output : Path to save the output to. This path is what determines the output directories for results, QC reports, and logs.

    • path-segmanual : Path to the folder containing manually-corrected segmentations.

    • jobs : Number of jobs for parallel processing.

    • itk-threads : Number of jobs for ANTs routines.

Output files/folders:

  • ./output/results/ : A folder containing the processed data, as well as CSV files containing computed metrics (CSA, and MTR in WM). This is equivalent to the derivatives/ folder of BIDS-compliant datasets.

  • ./output/qc/ : A folder containing the Quality Control (QC) reports for the commands in the pipeline.

  • ./output/log/ : A folder containing the log files for each subject. This is where you would go to troubleshoot any errors or issues that occur during the processing.

Note

We strongly recommend using a config file as shown here (rather than passing arguments directly to sct_run_batch) because it helps with reproducibility.

To learn about the full list of options that are available to tweak in your config file, run sct_run_batch -h.