Workflow strategy

[tsalo]

1. Collect preprocessed echo-wise data in boldref space.
   • The preprocessed echo-wise data are STC+MC+SDCed. @handwerkerd do you think it would be better to collect the raw echo-wise data and apply the STC+MC, but not SDC, before running tedana internally?
2. Collect transforms to anatomical and standard spaces.
3. Collect confounds and other useful derivatives.
   • We could include confounds from other BIDS derivatives (e.g., physio regressors) for use as the external confounds tedana now allows.
   • Tissue type masks for Group-level metric collection #2.
   • Are there other derivatives that might be useful?
4. Run tedana on the boldref-space multi-echo data.
   • Use the boldref-space mask from fMRIPrep here.
5. Warp BOLD data to standard space, regenerate standard-space component weight maps, and plot components overlaid on the standard-sapce T1w (Plot component weights on standard-space T1w #3)
6. Collect confounds and write them out.
   • Rejected component time series
   • Rejected component time series orthogonalized w.r.t. the accepted component time series
7. Warp target maps/timeseries to requested standard spaces
   • Optimally combined data, in case our optimal combination method differs from what the users requested with fMRIPrep
   • Denoised BOLD (or maybe denoise the optimally combined data in the standard space, per Denoise optimally combined data in target spaces #4)
   • T2* map
   • S0 map
   • Model fit map