ME-EPI + fieldmaps (will be) broken

[oesteban]

When I merge #2547, I'll be breaking the multiecho path when fieldmaps are present. This issue is just a note to remind myself to look at this after the single-echo pathways are finalized.

[Lestropie]

Was going to report an issue with 21.0.0rc0; suspect it might be the issue you're reminding yourself of here, but want to confirm, will list separate issue if it's not a duplicate. Problem arises here as for multi-echo data "bold_file" is a tuple.

[oesteban]

Well, I think the way I broke the pipeline is more profound (in the sense that, for the sake of making smaller steps on the overall integration, I thought that was the best way to proceed), but I would say yes, you are pointing at the earliest symptom of this feature failing.

I don't think the particular bug you have reported was created by #2547, so thanks for catching it - I'll address it together with this bug report.

Indeed, the difficulties to run ME-EPI + fieldmaps at this point are deeper: since the ME-EPI pipeline is probably going to undergo an overhaul, I think the fieldmap graph should be designed after the pipeline has been revised.

All things considered, thanks much for the bug report - I think you are right we can have it unified under this issue.

[Lestropie]

Unsurprising that it's a deeper issue to resolve; "earliest symptom" is apt. 👍

[effigies]

Should be resolved.