default output grid: BOLD or template grid?

[snastase]

My understanding according to the documentation is that specifying, e.g., --output-spaces MNI152NLin2009cAsym without a :res will by default output the BOLD data in the native BOLD grid (synonymous to :res-native; this is what I want). However, running version 20.0.2 seems to yield preprocessed outputs in the 2 x 2 x 2 mm template grid rather than the original 3 x 3 x 4 BOLD grid. My fMRIPrep invocation is:

singularity run --cleanenv \                                                       
    --bind $bids_dir:/data \                                                       
    --bind /scratch/snastase/narratives/work \                                     
    --bind /jukebox/hasson/templateflow:/home/fmriprep/.cache/templateflow \       
    /jukebox/hasson/singularity/fmriprep/fmriprep-v20.0.2.simg \                   
    --skip_bids_validation \                                                       
    --participant-label sub-$1 \                                                   
    --nthreads 4 --omp-nthreads 4 \                                                
    --longitudinal --bold2t1w-dof 6 \                                              
    --medial-surface-nan \                                                         
    --output-spaces anat func fsaverage fsaverage6 fsaverage5 \                    
                    MNI152NLin2009cAsym MNI152NLin6Asym \                          
    --fs-license-file /data/code/fs-license.txt --notrack \                        
    --use-syn-sdc --write-graph --work-dir /scratch/snastase/narratives/work \   
    /data /data/derivatives participant

Here's a screencap of nib-ls output where you can see that the preprocessed BOLD outputs in BOLD and T1w spaces (--output-spaces anat func) are in the original 3 x 3 x 4 BOLD resolution, where as the outputs in MNI space (--output-spaces MNI152NLin2009cAsym MNI152NLin6Asym) have inherited the 2 x 2 x 2 template grid (despite me not specifying :res-2).

Maybe this is the correct behavior and I'm just confused about the documentation...? I can try re-running with explicitly specifying :res-native and see if that displays the correct behavior.

[oesteban]

Yes, your interpretation is correct. We'll look into this and come back to you ASAP.

[oesteban]

I'm not seeing that exact behavior, although I'm using the cutting edge development version:

$ fmriprep /oak/stanford/groups/russpold/data/openfmri/ds000005 derivatives/ participant --participant-label 01 -vv --nthreads 8 --omp-nthreads 8 -w work --skip-bids-validation --cifti-output --mem 32G --fs-subjects-dir /oak/stanford/groups/russpold/data/openfmri/derivatives/ds000005/freesurfer-6.0.1/ &> fmriprep2.log

$ nib-ls derivatives/fmriprep/sub-01/func/*.nii.gz
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-01_space-MNI152NLin2009cAsym_res-native_boldref.nii.gz             float32 [ 63,  74,  49]      3.12x3.12x4.00
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-01_space-MNI152NLin2009cAsym_res-native_desc-aparcaseg_dseg.nii.gz float32 [ 63,  74,  49]      3.12x3.12x4.00
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-01_space-MNI152NLin2009cAsym_res-native_desc-aseg_dseg.nii.gz      float32 [ 63,  74,  49]      3.12x3.12x4.00
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-01_space-MNI152NLin2009cAsym_res-native_desc-brain_mask.nii.gz     float32 [ 63,  74,  49]      3.12x3.12x4.00
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-01_space-MNI152NLin2009cAsym_res-native_desc-preproc_bold.nii.gz    int16  [ 63,  74,  49, 240] 3.12x3.12x4.00x2.00
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-02_space-MNI152NLin2009cAsym_res-native_boldref.nii.gz             float32 [ 63,  74,  49]      3.12x3.12x4.00
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-02_space-MNI152NLin2009cAsym_res-native_desc-aparcaseg_dseg.nii.gz float32 [ 63,  74,  49]      3.12x3.12x4.00
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-02_space-MNI152NLin2009cAsym_res-native_desc-aseg_dseg.nii.gz      float32 [ 63,  74,  49]      3.12x3.12x4.00
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-02_space-MNI152NLin2009cAsym_res-native_desc-brain_mask.nii.gz     float32 [ 63,  74,  49]      3.12x3.12x4.00
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-02_space-MNI152NLin2009cAsym_res-native_desc-preproc_bold.nii.gz    int16  [ 63,  74,  49, 240] 3.12x3.12x4.00x2.00
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-03_space-MNI152NLin2009cAsym_res-native_boldref.nii.gz             float32 [ 63,  74,  49]      3.12x3.12x4.00
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-03_space-MNI152NLin2009cAsym_res-native_desc-aparcaseg_dseg.nii.gz float32 [ 63,  74,  49]      3.12x3.12x4.00
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-03_space-MNI152NLin2009cAsym_res-native_desc-aseg_dseg.nii.gz      float32 [ 63,  74,  49]      3.12x3.12x4.00
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-03_space-MNI152NLin2009cAsym_res-native_desc-brain_mask.nii.gz     float32 [ 63,  74,  49]      3.12x3.12x4.00
derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-03_space-MNI152NLin2009cAsym_res-native_desc-preproc_bold.nii.gz    int16  [ 63,  74,  49, 240] 3.12x3.12x4.00x2.00

@mgxd, do you have any ideas of when we might have addressed this issue?

[mgxd]

interesting - i'm also running the cutting edge version (with nipreps/niworkflows#492 applied) and I see the same problem as @snastase

~/code/fmriprep-testing$ nib-ls ds005/derivatives/fmriprep/sub-01/func/*.nii.gz
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-01_boldref.nii.gz                                   float32 [ 50,  50,  34]      4.00x4.00x4.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-01_desc-brain_mask.nii.gz                            uint8  [ 50,  50,  34]      4.00x4.00x4.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-01_desc-preproc_bold.nii.gz                         float32 [ 50,  50,  34,  16] 4.00x4.00x4.00x2.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-01_space-MNI152NLin6Asym_boldref.nii.gz             float32 [ 91, 109,  91]      2.00x2.00x2.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-01_space-MNI152NLin6Asym_desc-aparcaseg_dseg.nii.gz  int16  [ 91, 109,  91]      2.00x2.00x2.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-01_space-MNI152NLin6Asym_desc-aseg_dseg.nii.gz       int16  [ 91, 109,  91]      2.00x2.00x2.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-01_space-MNI152NLin6Asym_desc-brain_mask.nii.gz      uint8  [ 91, 109,  91]      2.00x2.00x2.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-01_space-MNI152NLin6Asym_desc-preproc_bold.nii.gz   float32 [ 91, 109,  91,  16] 2.00x2.00x2.00x2.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-02_boldref.nii.gz                                   float32 [ 50,  50,  34]      4.00x4.00x4.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-02_desc-brain_mask.nii.gz                            uint8  [ 50,  50,  34]      4.00x4.00x4.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-02_desc-preproc_bold.nii.gz                         float32 [ 50,  50,  34,  16] 4.00x4.00x4.00x2.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-02_space-MNI152NLin6Asym_boldref.nii.gz             float32 [ 91, 109,  91]      2.00x2.00x2.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-02_space-MNI152NLin6Asym_desc-aparcaseg_dseg.nii.gz  int16  [ 91, 109,  91]      2.00x2.00x2.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-02_space-MNI152NLin6Asym_desc-aseg_dseg.nii.gz       int16  [ 91, 109,  91]      2.00x2.00x2.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-02_space-MNI152NLin6Asym_desc-brain_mask.nii.gz      uint8  [ 91, 109,  91]      2.00x2.00x2.00
ds005/derivatives/fmriprep/sub-01/func/sub-01_task-mixedgamblestask_run-02_space-MNI152NLin6Asym_desc-preproc_bold.nii.gz   float32 [ 91, 109,  91,  16] 2.00x2.00x2.00x2.00

input BOLD files:

~/code/fmriprep-testing$ nib-ls ds005/sub-01/func/*.nii.gz
ds005/sub-01/func/sub-01_task-mixedgamblestask_run-01_bold.nii.gz float32 [ 50,  50,  34,  16] 4.00x4.00x4.00x2.00
ds005/sub-01/func/sub-01_task-mixedgamblestask_run-02_bold.nii.gz float32 [ 50,  50,  34,  16] 4.00x4.00x4.00x2.00

[mgxd]

@snastase - specifying your outputs explicitly with :res-native should do the trick in previous versions, but this has been patched in the latest pre-release (20.1.0rc2).

[snastase]

Excellent—thanks!