qsiprep.workflows.anatomical.volume module

Anatomical reference preprocessing workflows.

Anatomical reference preprocessing workflows

qsiprep.workflows.anatomical.volume.init_anat_preproc_wf(num_anat_images, num_additional_t2ws, has_rois, anatomical_template, do_biascorr=True, t2w_do_biascorr=True, name='anat_preproc_wf')[source]

Build a workflow that runs the anatomical preprocessing stages of qsiprep.

This includes:

  • Creation of a structural template (AC-PC aligned)

  • Skull-stripping and bias correction

  • Tissue segmentation

  • Normalization

(Source code)

Parameters:

  • num_anat_images (int) – Number of anatomical images available in the chosen modality

  • num_additional_t2ws (int) – If anat modality is T1w and there are available T2ws that can be used by DRBUDDI, how many are there?

  • has_rois (bool) – Are there lesion ROI files?

  • anatomical_template (str) – Template specification of the form <template>[+<cohort>].

  • do_biascorr (bool, optional) – Whether to apply N4 bias correction to the T1w(?) or not. Default is True.

  • t2w_do_biascorr (bool, optional) – Whether to apply N4 bias correction to the T2w or not. Default is True.

  • name (str, optional) – Name of workflow (default: anat_preproc_wf)

Inputs:

  • t1w – List of T1-weighted structural images

  • t2w – List of T2-weighted structural images

  • roi – A mask to exclude regions during standardization (as list)

  • subjects_dir – FreeSurfer SUBJECTS_DIR

Outputs:

  • t1_preproc – Bias-corrected structural template, defining T1w space

  • t1_brain – Skull-stripped t1_preproc

  • t1_mask – Mask of the skull-stripped template image

  • t1_seg – Segmentation of preprocessed structural image, including gray-matter (GM), white-matter (WM) and cerebrospinal fluid (CSF)

  • t1_tpms – List of tissue probability maps in T1w space

  • t2_preproc – List of preprocessed t2w files

  • t1_2_mni – T1w template, normalized to MNI space

  • t1_2_mni_forward_transform – ANTs-compatible affine-and-warp transform file

  • t1_2_mni_reverse_transform – ANTs-compatible affine-and-warp transform file (inverse)

  • to_template_affine_transform – The full affine to the anatomical template that acpc_transform (its rigid component) was extracted from. Maps the pre-AC-PC anatomical reference; compose with acpc_inv_transform to map AC-PC space linearly into the template space (used by SyNb0).

  • t1_resampling_grid – Image of the preprocessed t1 to be used as the reference output for dwis

qsiprep.workflows.anatomical.volume.anat_biascorrect_enabled(image_files=None)[source]

Decide whether N4 bias correction should run on these anatomical images.

--anat-biascorrect governs anatomicals only; --dwi-biascorrect governs the DWIs and never reaches this path.

auto inspects the BIDS ImageType metadata for NORM, which is how Siemens (among others) flags that intensity normalization was already applied on the console. Console normalization does not necessarily remove the need for N4, which is why n4 stays the default; auto and none are for deliberately skipping it.

N4 is skipped only when EVERY input image is marked normalized. A mixed set still needs correction to be merged sensibly, and an image whose metadata is missing is treated as un-normalized – the conservative direction, since running N4 unnecessarily is milder than skipping it when it was needed.

qsiprep.workflows.anatomical.volume.init_anat_derivatives_wf(anatomical_template, has_t2w=False) → LiterateWorkflow[source]

Set up a battery of datasinks to store derivatives in the right location.

qsiprep.workflows.anatomical.volume.init_anat_normalization_wf(anatomical_template, has_rois=False) → LiterateWorkflow[source]

Build a workflow that registers the anatomical reference to the template.

This workflow performs registration from the original anatomical reference to the template anatomical reference.

(Source code)

Parameters:

  • anatomical_template (str) – Name of the template the anatomical reference is registered to.

  • has_rois (bool, optional) – Whether Registration should account for regions to exclude

Inputs:

  • anatomical_reference – Bias-corrected anatomical reference image (head)

  • brain_mask – Brain mask restricting the affine registration that the AC-PC transform is extracted from

  • nonlinear_brain_mask – Brain mask restricting the nonlinear registration to the template. The same as brain_mask unless --force no-csf-synthstrip is set.

  • roi – A mask to exclude regions during standardization (as list)

Outputs:

  • to_template_nonlinear_transform – Composite transform (SyN warp and affine) to the template

  • to_template_rigid_transform – Rigid (AC-PC) component of the full affine to the template

  • to_template_affine_transform – The full affine to the template that the AC-PC transform was extracted from (maps the pre-AC-PC anatomical reference; compose with from_template_rigid_transform to map AC-PC space linearly into the template space)

  • from_template_nonlinear_transform – Inverse of to_template_nonlinear_transform

  • from_template_rigid_transform – Inverse of to_template_rigid_transform

  • out_report – Reportlet visualizing the spatial normalization

qsiprep.workflows.anatomical.volume.init_anat_preproc_wf(num_anat_images, num_additional_t2ws, has_rois, anatomical_template, do_biascorr=True, t2w_do_biascorr=True, name='anat_preproc_wf')[source]

Build a workflow that runs the anatomical preprocessing stages of qsiprep.

This includes:

  • Creation of a structural template (AC-PC aligned)

  • Skull-stripping and bias correction

  • Tissue segmentation

  • Normalization

(Source code)

Parameters:

  • num_anat_images (int) – Number of anatomical images available in the chosen modality

  • num_additional_t2ws (int) – If anat modality is T1w and there are available T2ws that can be used by DRBUDDI, how many are there?

  • has_rois (bool) – Are there lesion ROI files?

  • anatomical_template (str) – Template specification of the form <template>[+<cohort>].

  • do_biascorr (bool, optional) – Whether to apply N4 bias correction to the T1w(?) or not. Default is True.

  • t2w_do_biascorr (bool, optional) – Whether to apply N4 bias correction to the T2w or not. Default is True.

  • name (str, optional) – Name of workflow (default: anat_preproc_wf)

Inputs:

  • t1w – List of T1-weighted structural images

  • t2w – List of T2-weighted structural images

  • roi – A mask to exclude regions during standardization (as list)

  • subjects_dir – FreeSurfer SUBJECTS_DIR

Outputs:

  • t1_preproc – Bias-corrected structural template, defining T1w space

  • t1_brain – Skull-stripped t1_preproc

  • t1_mask – Mask of the skull-stripped template image

  • t1_seg – Segmentation of preprocessed structural image, including gray-matter (GM), white-matter (WM) and cerebrospinal fluid (CSF)

  • t1_tpms – List of tissue probability maps in T1w space

  • t2_preproc – List of preprocessed t2w files

  • t1_2_mni – T1w template, normalized to MNI space

  • t1_2_mni_forward_transform – ANTs-compatible affine-and-warp transform file

  • t1_2_mni_reverse_transform – ANTs-compatible affine-and-warp transform file (inverse)

  • to_template_affine_transform – The full affine to the anatomical template that acpc_transform (its rigid component) was extracted from. Maps the pre-AC-PC anatomical reference; compose with acpc_inv_transform to map AC-PC space linearly into the template space (used by SyNb0).

  • t1_resampling_grid – Image of the preprocessed t1 to be used as the reference output for dwis

qsiprep.workflows.anatomical.volume.init_anat_reports_wf(anatomical_template) → LiterateWorkflow[source]

Set up a battery of datasinks to store reports in the right location.

qsiprep.workflows.anatomical.volume.init_anat_template_wf(num_images, do_biascorr=True) → LiterateWorkflow[source]

Build a workflow that generates a canonically oriented structural template.

This workflow generates a canonically oriented structural template from input anatomical images.

(Source code)

Parameters:

  • num_images (int) – Number of anatomical images

  • do_biascorr (bool, optional) – Whether or not to apply N4 bias correction. Default is True.

Inputs:

anatomical_images – List of structural images

Outputs:

  • template – Structural template, defining T1w space

  • template_transforms – List of affine transforms from template to original images

  • out_report – Conformation report

qsiprep.workflows.anatomical.volume.init_dl_prep_wf(name='dl_prep_wf') → LiterateWorkflow[source]

Prepare images for use in the FreeSurfer deep learning functions.

qsiprep.workflows.anatomical.volume.init_output_grid_wf() → LiterateWorkflow[source]

Generate a non-oblique, uniform voxel-size grid around a brain.

qsiprep.workflows.anatomical.volume.init_synthseg_wf() → LiterateWorkflow[source]
qsiprep.workflows.anatomical.volume.init_synthstrip_wf(do_padding=False, unfatsat=False, no_csf=False, name='synthstrip_wf') → LiterateWorkflow[source]

Skull strip an image with SynthStrip, or with niimath’s surface stripper under –sloppy.

no_csf selects SynthStrip’s --no-csf model, which trims CSF and dura from the brain border. Ventricular CSF stays inside the mask either way.

Under --sloppy the node is niimath -skullstrip (AFNI’s surface method, 1.5 s, 90 MB, Dice 0.96 against SynthStrip on a 1 mm T2w) instead of the torch runtime; no_csf has no counterpart there and is ignored.

qsiprep.workflows.anatomical.volume.init_t2w_preproc_wf(num_t2ws, do_biascorr=True, name='t2w_preproc_wf')[source]

Build a workflow that preprocesses additional T2w images.

If T1w is the anatomical contrast, you may also want to process the T2ws for worlflows that can use them (ie DRBUDDI).