Average T2* maps across runs

[tsalo]

I was thinking we could average the T2* maps across any runs before running tedana. It probably won't have much of an effect, but could still improve results (both optcom and ICA).

[tsalo]

Maybe we weight by the RMSE or length of the runs or something?

[handwerkerd]

This sounds like it would combined well with temporal concatenation of runs as discussed here ME-ICA/tedana#1133 (comment) It's already possible for someone to provide their own T2* map, but, if we allow users to input multiple runs, then the fit could be averaged across all voxels in all runs in a single step.

[tsalo]

Since this is a BIDS App, it will automatically process all runs from a given subject/session, so it shouldn't be an issue to aggregate T2* maps across runs.

[tsalo]

Random thought- We'll need to normalize to a shared space (anatomical space maybe?) before we average the maps across runs.