Home

Documentation

GitHub

Dwi_preprocessing pipeline

DWI preprocessing using a reverse-phase encoded b0

The aim of this pipeline is to preprocess a multishell diffusion data before to analysed it. In this pipeline, an image with b=0 volumes with opposite phase encoding should be provided for the purpose of EPI distortion correction.

The following steps are done:
- denoise the image and estimate the spatially varying noise map
- remove Gibb’s ringing artefacts
- motion and distortion correction (EPI-distrortion correction, B0-field inhomogeneity correction, Eddy-current and movement distortion correction )
- bias field correction

A binary mask of the brain is also estimated.

The outputs are in “.mif” format.

Pipeline insight

Dwi_preprocessing pipeline combines the following bricks:
- Conversion into MIF format (if NIfTI provided)
- b0 extraction
- b0 averaging
- b0 concatenation
- Denoising
- Unringing
- Motion distorstion correction (Topup and eddy correction) (default values: rpe_options = “pair”, eddy_options = ” –slm=linear –data_is_shelled”)
- Bias correction
- Brain mask estimation
DWI preprocessing pipeline

Mandatory inputs parameters

  • in_dwi (a string representing an existing file)

    Diffusion image to preprocess (valid extensions: [.nii, .nii.gz, ‘.mif’]). If a NIfTI is supplied, bvec and bval files will be found automatically.

    ex. '/home/username/data/raw_data/DWI.nii'
  • in_dwi_pe_dir (ap, pa, lr, rl, default value is ap, optional)
    Phase encoding direction of the in_diw image:
    - ap : Anterior to posterior
    - pa: Posterior to anterior
    - lr: Left to right
    - rl: Right toleft
    ex. ap
  • in_dwi_ro_time (a float, optional)

    Total readout time of in_dwi image (in seconds).

    ex. 2.0
  • in_b0_reverse (a string representing an existing file)

    b=0 volumes with opposing phase-direction which is to be used exclusively by topup for estimating the inhomogeneity field (valid extensions: [.nii, .nii.gz, ‘.mif’]).

    ex. '/home/username/data/raw_data/b0_reverse.nii'

Outputs parameters:

  • brain_mask

    DWI brain mask image.

    ex. '/home/username/data/derived_data/DWI_denoised_unringed_dwifslpreproc_unbias_brain_mask.mif'
  • preproc_dwi

    DWI corrected image.

    ex. '/home/username/data/derived_data/DWI_denoised_unringed_dwifslpreproc_unbias.mif'