Dwi_fod_msmt_csd pipeline¶

Estimate DWI fiber orientation distribution (FOD)¶

The aim of this pipeline is to estimate the orientation of the fibers in each voxel using constrained spherical deconvolution (CSD). To perform CSD, it is necessary to first estimate a “response function” (RF) which will be used as a kernel for deconvolution. In this pipeline, we estimate a RF for each tissue type (using msmt_csd algorithm), so the input diffusion should contain at least three unique b-values to estimate three tissue compartments.

The following steps are done:
- Estimate different response functions for the three different tissue types: white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF)
- In every voxel, estimate the orientation of all fibers crossing that voxel.
- Correct the FOD for global intensity differences. This step helps to make the FODs comparable between several subjects (group study preparation)

Pipeline insight

Dwi_fod_msmt_csd pipeline combines the following bricks:

- Response function estimation
- Estimation of Fiber Orientation Distributions (FOD)
(default values: algorithm = “msmt_csd”, get_predicted_signal = True)
- Intensity Normalization

Mandatory inputs parameters

in_dwi (a string representing an existing file)
Preprocessed diffusion image (valid extensions: [.mif]).
ex. '/home/username/data/raw_data/DWI_denoised_unringed_dwifslpreproc_unbias.mif'

brain_mask (a string representing an existing file)

DWI brain mask image.

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

Outputs parameters:

voxel_final selection (a pathlike object or string representing a file)
Image showing the final voxel selection for the estimate response functions
ex. '/home/username/data/derived_data/DWI_denoised_unringed_dwifslpreproc_unbias_response_voxels.mif'
predicted_signal_file (a pathlike object or string representing a file, optional)
Predicted signal following the estimation of the FOD.
ex. '/home/username/data/derived_data/DWI_denoised_unringed_dwifslpreproc_unbias_odf_predicted_signal.mif'
wm_odf_norm (a pathlike object or string representing a file)
Normalized WM fibre orientation distributions
ex. '/home/username/data/derived_data/DWI_denoised_unringed_dwifslpreproc_unbias_wm_odf_norm.mif'
gm_odf_norm (a pathlike object or string representing a file, optional)
Normalized GM fibre orientation distributions
ex. '/home/username/data/derived_data/DWI_denoised_unringed_dwifslpreproc_unbias_gm_odf_norm.mif'
csf_odf_norm (a pathlike object or string representing a file, optional)
Normalized CSF fibre orientation distributions
ex. '/home/username/data/derived_data/DWI_denoised_unringed_dwifslpreproc_unbias_csf_odf_norm.mif'