Open tsalo opened 6 months ago
Ultimately, we may want to support three basic patterns (see below). I've included fMRIPost-AROMA derivatives in the examples as well.
xcp_d \
/derivatives/<pipeline> \ # path to pipeline with only standard-space derivatives (e.g., HCP)
/derivatives/xcp_d \
--output-spaces MNI152NLin6Asym:res-2 \ # will raise an error if it doesn't match the available spaces
--derivatives aroma=/derivatives/fmripost-aroma \
...
E.g., fMRIPrep with --level resampling
xcp_d \
/derivatives/<pipeline> \ # path to pipeline with native-space derivatives and warps
/derivatives/xcp_d \
--output-spaces MNI152NLin6Asym:res-2 \ # will raise an error if warps aren't available
--derivatives aroma=/derivatives/fmripost-aroma \
...
E.g., fMRIPrep with --level minimal
xcp_d \
/dset \ # path to raw dataset
/derivatives/xcp_d \
--output-spaces MNI152NLin6Asym:res-2 \ # will raise an error if warps aren't available
--derivatives \
preprocessed=/derivatives/<pipeline> \ # path to preprocessing derivatives with warps
aroma=/derivatives/fmripost-aroma \
...
Summary
NiPreps'
resampler
will be used in fMRIPost and sMRIPost workflows to warp native-space derivatives to target spaces. I would ultimately like to include this in XCP-D as well.Additional details
I need to double-check the fMRIPrep
--level
documentation, but I think we need to requireresampling
-level outputs from fMRIPrep.I'm not sure if
resampler
can handle slice timing correction yet.Next steps
--output-spaces
parameter, along with space-parsing.EDIT: I just want to track https://github.com/nipreps/fmripost-aroma/issues/34 in this issue because fMRIPost-AROMA will probably be the first functioning implementation of this general workflow.