Finalize fMRIPrep settings

[tsalo]

I'm planning to use version 23.1.5 (once it's released), since it has some nice CIFTI improvements.

Here are my planned settings:

--output-spaces func:res-native T1w:res-native MNI152NLin6Asym:res-2
--output-layout bids
--level full
--me-t2s-fit-method curvefit
--me-output-echos
--project-goodvoxels
--cifti-output
--slice-time-ref 0  # this should simplify the analysis for task data
--return-all-components
--random-seed 0

Is it worth it to include --use-aroma? Might be nice to have AROMA outputs to compare to tedana.

[handwerkerd]

I think AROMA would mostly be useful if we could use the same ICA mixing matrix for AROMA & tedana. If they use different ICAs, then it would be hard to do useful comparisions.

[tsalo]

It seems possible to use the same mixing matrix for both. It's not something anyone's done yet AFAIK.

[tsalo]

Actually I just remembered that AROMA was removed from fMRIPrep. I'll have to wait until fmripost-aroma is working, but that might be a useful additional derivative set once it's up and running.