Slice time correction defaults

Hi,

It would be nice to have a feature for implementing slice time corrections for first level modeling. Or, at least to provide some documentation about how nilearn treats slice time corrections and how to adjust our data ,if slice time correction was done, to be compatible with nltools. Some (like me) may not realize that different softwares have different views and defaults on slice time correction (e.g. [default of fmriprep is to do it](https://reproducibility.stanford.edu/slice-timing-correction-in-fmriprep-and-linear-modeling/), nltools does not do it by default).

Thank you!

Start with nltools' first-level modeling entry points and the slice-time correction behavior described in the issue, then review the linked fMRIPrep documentation for comparison. Done would mean either a documented path for making already-corrected data compatible with nltools or a clearly specified slice-time correction feature and default.

Written by the indexing model from the issue text.