Only operate on masked data in array format and keep masks as images

[tsalo]

Summary

Per discussion at the hackathon, it is confusing to have our data variously in XxYxZxT-, SxExT-, and MxExT-shaped arrays. Since passing around strings (to load/write files in each function) or images (per #251) are both too slow, we have committed to operating on arrays, but also want to limit our functions to MxExT-shaped arrays. Masks should then be images, and we will need custom functions to mask and unmask multi-echo data (as nilearn automatically concatenates lists of files/imgs in the time dimension).

Additional Details

My hope is that this change will make it easier for advanced users to load multi-echo data and apply tedana's individual functions outside of the workflow. It will also make it easier for us to interact with other packages (e.g., AROMA) or to port some of our general-purpose methods (e.g., GODEC maybe?) to another package without relying too much on highly-specific IO.

Next Steps

1. In all cases where we pass around SxExT data, we should change it to just use the MxExT (i.e., masked) data.
2. Add new functions to apply masks and unmask data that works elegantly with lists of files.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions to tedana:tada: !

[jbteves]

We're doing the work, bot.

[stale[bot]]

This issue has been automatically marked as stale because it has not had any activity in 90 days. It will be closed in 600 days if no further activity occurs. Thank you for your contributions to tedana:tada: !