CPAC Output Quality Assessment

Software Support
1

I have just run a subject through the default CPAC pipeline and got the following two quality images.

Does this look like a segmentation issue to you all? Or is this a typical output?

sub-BNI_ses-1_desc-dsegAxial_quality
sub-BNI_ses-1_desc-dsegAxial_quality1041×700 224 KB

2

Second image (sagittal):

sub-BNI_ses-1_desc-dsegSagittal_quality
sub-BNI_ses-1_desc-dsegSagittal_quality1041×543 214 KB

3

Hi @asd_tms, Thank you for your query.
Could you please answer following questions for me so that I can help you better?

  1. What are the (list) outputs in the ‘output’ directory? Do you have segmentation tissue masks?
  2. Could you try overlaying the tissue mask over the T1w in a viewer and see if that looks okay?
  3. Could you elaborate on what the issue is or your main concern?
    Thank you,
    Biraj
4

Hi:

Here are the outputs in the output directory, from my anat folder:

sub-BNI_ses-1_desc-brain_mask.nii.gzsub-BNI_ses-1_desc-head_T1w.nii.gz
sub-BNI_ses-1_desc-preproc_T1w.nii.gz
sub-BNI_ses-1_from-MNI152NLin6ASym_to-T1w_mode-image_desc-linear_xfm.nii.gz
sub-BNI_ses-1_from-MNI152NLin6ASym_to-T1w_mode-image_desc-nonlinear_xfm.nii.gz
sub-BNI_ses-1_from-MNI152NLin6ASym_to-T1w_mode-image_xfm.nii.gz
sub-BNI_ses-1_from-MNI152NLin6Sym_to-T1w_mode-image_desc-linear_xfm.nii.gz
sub-BNI_ses-1_from-MNI152NLin6Sym_to-T1w_mode-image_desc-nonlinear_xfm.nii.gz
sub-BNI_ses-1_from-MNI152NLin6Sym_to-T1w_mode-image_xfm.nii.gz
sub-BNI_ses-1_from-T1w_to-MNI152NLin6ASym_mode-image_desc-linear_xfm.nii.gz
sub-BNI_ses-1_from-T1w_to-MNI152NLin6ASym_mode-image_desc-nonlinear_xfm.nii.gz
sub-BNI_ses-1_from-T1w_to-MNI152NLin6ASym_mode-image_xfm.nii.gz
sub-BNI_ses-1_from-T1w_to-MNI152NLin6Sym_mode-image_desc-linear_xfm.nii.gz
sub-BNI_ses-1_from-T1w_to-MNI152NLin6Sym_mode-image_desc-nonlinear_xfm.nii.gz
sub-BNI_ses-1_from-T1w_to-MNI152NLin6Sym_mode-image_xfm.nii.gz
sub-BNI_ses-1_label-CSF_desc-preproc_mask.nii.gz
sub-BNI_ses-1_label-CSF_mask.nii.gz
sub-BNI_ses-1_label-CSF_probseg.nii.gz
sub-BNI_ses-1_label-GM_desc-preproc_mask.nii.gz
sub-BNI_ses-1_label-GM_mask.nii.gz
sub-BNI_ses-1_label-GM_probseg.nii.gz
sub-BNI_ses-1_label-WM_desc-preproc_mask.nii.gz
sub-BNI_ses-1_label-WM_mask.nii.gz
sub-BNI_ses-1_label-WM_probseg.nii.gz
sub-BNI_ses-1_space-MNI152NLin6ASym_desc-brain_mask.nii.gz
sub-BNI_ses-1_space-MNI152NLin6ASym_desc-head_T1w.nii.gz
sub-BNI_ses-1_space-MNI152NLin6ASym_desc-preproc_T1w.nii.gz
sub-BNI_ses-1_space-MNI152NLin6ASym_label-CSF_mask.nii.gz
sub-BNI_ses-1_space-MNI152NLin6ASym_label-GM_mask.nii.gz
sub-BNI_ses-1_space-MNI152NLin6ASym_label-WM_mask.nii.gz

Additionally, I get several .json files and the two images I included in the original post.

  1. I think the overlay of the tissue mask looks okay
  2. My main concern is related to the skull stripping. I’m using the default method (FSL-BET) but wondering if people generally have better results with AFNI-3dskullstrip

Thanks!

5

Thank you @asd_tms for your reply and further clarification.
Skull Stripping quality varies/depends on both tool and data so, I would recommend you to try few of the skullstripping options available, on a couple of subjects first, to compare the quality.
You can also fork these methods so that you get all outputs from different methods on a single run, which is helpful for comparison.
Would you mind sharing your pipeline.yml file also?