question about Spatial normalization with smriprep

[alihassoun1]

hello experts,

If I run smriprep with a docker for a subject as without a special option then the HTML report of this brain does not say anything about spatial normalization.

This is also not mentioned under Methods in the HTML report.

only when i run smriprep with docker again with the option --output-spaces MNI152NLin2009cAsym the HTML report of the brain shows that a spatial normalization has been performed and a picture of the spatial normalization is shown

But there are 12 nifti files when I run smriprep with --output-spaces MNI152NLin2009cAsym. that was not my goal to have so much Nifti data .

Is there really no spatial normalization when I run smriprep without the option --output-spaces MNI152NLin2009cAsym ?

[oesteban]

Is there really no spatial normalization when I run smriprep without the option --output-spaces MNI152NLin2009cAsym ?

Correct. For the sMRIPrep BIDS-App the default is not running any spatial normalization unless they are explicitly requested with --output-spaces. fMRIPrep overrides this default behavior and dMRIPrep does not.

[oesteban]

But there are 12 nifti files when I run smriprep with --output-spaces MNI152NLin2009cAsym. that was not my goal to have so much Nifti data

Can you provide a list of these files?

[alihassoun1]

yes ok thank you very much. Then my question is over. I can still send you my output but I think the output is all right because I worked with the option --output-spaces.

My output:

dataset_description.json
logs
sub-OAS30204.html
sub-OAS30204
├── figures
│   ├── sub-OAS30204_desc-about_T1w.html
│   ├── sub-OAS30204_desc-summary_T1w.html
│   ├── sub-OAS30204_ses-d0020_desc-conform_T1w.html
│   ├── sub-OAS30204_ses-d0020_dseg.svg
│   └── sub-OAS30204_ses-d0020_space-MNI152NLin2009cAsym_T1w.svg
└── ses-d0020
    └── anat
        ├── sub-OAS30204_ses-d0020_desc-brain_mask.json
        ├── sub-OAS30204_ses-d0020_desc-brain_mask.nii.gz
        ├── sub-OAS30204_ses-d0020_desc-preproc_T1w.json
        ├── sub-OAS30204_ses-d0020_desc-preproc_T1w.nii.gz
        ├── sub-OAS30204_ses-d0020_dseg.nii.gz
        ├── sub-OAS30204_ses-d0020_from-MNI152NLin2009cAsym_to-T1w_mode-image_xfm.h5
        ├── sub-OAS30204_ses-d0020_from-T1w_to-MNI152NLin2009cAsym_mode-image_xfm.h5
        ├── sub-OAS30204_ses-d0020_label-CSF_probseg.nii.gz
        ├── sub-OAS30204_ses-d0020_label-GM_probseg.nii.gz
        ├── sub-OAS30204_ses-d0020_label-WM_probseg.nii.gz
        ├── sub-OAS30204_ses-d0020_space-MNI152NLin2009cAsym_desc-brain_mask.json
        ├── sub-OAS30204_ses-d0020_space-MNI152NLin2009cAsym_desc-brain_mask.nii.gz
        ├── sub-OAS30204_ses-d0020_space-MNI152NLin2009cAsym_desc-preproc_T1w.json
        ├── sub-OAS30204_ses-d0020_space-MNI152NLin2009cAsym_desc-preproc_T1w.nii.gz
        ├── sub-OAS30204_ses-d0020_space-MNI152NLin2009cAsym_dseg.nii.gz
        ├── sub-OAS30204_ses-d0020_space-MNI152NLin2009cAsym_label-CSF_probseg.nii.gz
        ├── sub-OAS30204_ses-d0020_space-MNI152NLin2009cAsym_label-GM_probseg.nii.gz
        └── sub-OAS30204_ses-d0020_space-MNI152NLin2009cAsym_label-WM_probseg.nii.gz

[alihassoun1]

hello @oesteban,

there is no option that outputs only the normalized nifiti files ? Because actually i only need the normalized nifti files.

[oesteban]

Not at the moment. I think you can always set the output folder to be in /tmp or some other fast, transient storage and then have a final hook that copies only the files you really want to the final location.

However, the only upshot to this is saving storage. Any other aspect of the solution would count as a downside.

Instead, I'd mirror all fMRIPrep outputs to some cold, cheap storage location and cleanup the files you don't need on your workstation.

I don't think there are plans to add granular control over the outputs (or more granular it is now), and the argument to consider it must be really strong.

[alihassoun1]

ok then I know about it, thank you 👍