Using both drift models in nilearn

1

I’m removing confounds from my data using FirstLevelModel in nilearn, and I can either choose cosine or polynomial for my drift model.

However, I’d like to use both during GLM estimation, is this possible?

2

There is a possibility to add manual regressors, so you can supply regressors manually in such a way.
See e.g. add_regs parameter in nilearn.glm.first_level.make_first_level_design_matrix - Nilearn
HTH,
Bertrand