fmriprep-ingress regressor naming issue

[sgiavasis]

With @e-kenneally 's idea, I'm picturing something like this:

1. Have a field you can place in the regressor selector (under each strategy) that is like "include_from_confounds:" or something and it's a list of column names that can be found in the ingressed confounds TSV that comes in from the output directory; if left blank or isn't present, the default is to include all (even if this crashes 3dTproject)
2. This functionality will filter the ingressed confounds TSV first - with a function node, maybe using Pandas ideally; drop the columns not included in the list
3. If there are other regressors in the selector for that strategy, now have it such that they are added to this filtered/reduced TSV
4. Send that merged TSV to the RPool as desc-confounds_timseries