Need to downsample one nifti - how should i edit the JSON file?

bids
neuroimaging
fmriprep

#1

hi all,

i have one functional scan which was taken at a higher resolution than the rest, which i would like to use in analysis. i have been able to downsample it to the right resolution using freesurfer mri_convert, but now i am wondering what I should do in terms of making this BIDS compatible.

i though i’d save the original volume as “sub-04_task-func_run-01_bold_2mm_orig.nii.gz” and the appropriately downsampled volume as “sub-04_task-func_run-01_bold.nii.gz”, to match the rest of the same-resolution runs. then, i’ll submit all of the runs to pre-processing in fmriprep, and use only the same-resolution runs in further analyses.

is this advisable, and if so, how should I go about editing the json file for this run?

thanks much,
nick


Retrospective tagging
#2

Hi Nick,

Thanks for using FMRIPREP! I assume that you are attempting to resample this one run to maintain the same resolution of your output files. This is not necessary and actually not recommended.

Instead you should keep the original data intact (which is more of a general rule) and only modify/resample the FMRIPREP outputs. Alternatively you can use the --output-grid-reference FMRIPREP argument to point to a nifti file that will provide reference resolution, field of view, and orientation information for your output files. This will enforce that all the outputs will have the same resolution (independent of the original data). More info http://fmriprep.readthedocs.io/en/stable/usage.html

Finally don’t forget to report the scanning protocol inconsistencies in your paper and check if it’s not driving your effects as an outlier.


#3

Hey Chris. That’s great. Thanks for the suggestions and for providing a great preprocessing framework in fmriprep.

-Nick