Cubic first-level model

[Lestropie]

Your idea

I observed that with data from the Siemens Cima-X magnet, executing the same protocol as that previously acquired on Prisma magnets, the default eddy current correction was inadequate, and a cubic first-level model was required.

While technically it would be possible to change the model order depending on the source magnet by checking the corresponding BIDS field, I suspect it might be safer to just always impose a cubic first-level model, even if it's more computationally expensive.