Usage

The KePrep preprocessing workflow takes as input a BIDS-appropriate dataset. Other than the BIDS dataset, the user can specify a variety of options to customize the preprocessing workflow.

Being primarily a python package, the way to use the pipeline is to populate the keprep.config module with the desired options before running the pipeline itself.

There are two main ways to customize the configuration of the pipeline:

  1. By modifying the keprep.config module directly.:

    from keprep import config
config.execution.bids_dir = '/path/to/bids/dataset' # path to the BIDS dataset
config.execution.output_dir = '/path/to/output/directory' # path to the output directory
config.execution.participant_label = ['01', '02', '03'] # list of participant labels to process

  2. By passing a dictionary of options to the from_dict method.:

    from keprep import config

inputs = {
    "bids_dir": "/path/to/bids/dataset", # path to the BIDS dataset
    "output_dir": "/path/to/output/directory", # path to the output directory
    "participant_label": ["01", "02", "03"] # list of participant labels to process
}

config.from_dict(inputs)

Configuration Options

Execution Settings

(accessible through config.execution)

  • bids_dir: Path to the BIDS dataset.

  • output_dir: Path to the output directory.

  • participant_label: List of participant identifiers to be preprocessed.

  • anat_derivatives: Path where anatomical derivatives are found.

  • bids_database_dir: Directory containing SQLite database indices for the input BIDS dataset.

  • reset_database: Reset the SQLite database (default: True).

  • fs_license_file: Path to the FreeSurfer license file.

  • fs_subjects_dir: FreeSurfer’s subjects directory.

  • log_dir: Path to the directory containing execution logs.

  • log_level: Output verbosity (default: 25).

  • work_dir: Path to the working directory where intermediate results will be available.

Nipype Settings

(accessible through config.nipype)

  • nprocs: Number of processes (compute tasks) that can be run in parallel (multiprocessing only).

  • omp_nthreads: Number of CPUs a single process can access for multithreaded execution.

  • stop_on_first_crash: Whether the workflow should stop or continue after the first error (default: True).

Workflow Settings

(accessible through config.workflow)

  • anat_only: Execute the anatomical preprocessing only (default: False).

  • dwi2t1w_dof: Degrees of freedom of the DWI-to-T1w registration steps (default: 6).

  • dwi2t1w_init: Initialization method for DWI-to-T1w coregistration (default: register).

  • cifti_output: Generate HCP Grayordinates (accepts 91k or 170k).

  • do_reconall: Whether to perform FreeSurfer’s recon-all (default: False).

  • longitudinal: Run FreeSurfer recon-all with the -logitudinal flag (default: False).

  • skull_strip_fixed_seed: Fix a seed for skull-stripping (default: False).

  • skull_strip_template: Change default brain extraction template (default: OASIS30ANTs).

  • hires: Run FreeSurfer recon-all with the -hires flag.

  • skull_strip_t1w: Skip brain extraction of the T1w image (default: force).