Improve documentation and config validation of `loose` and `depth` parameters; drop support for `loose=None`

[hoechenberger]

@larsoner @SophieHerbst WDYT?

We currently don't support volumetric or mixed source spaces anyway, so I thought we could simplify things a little.

I was even considering entirely removing the loose and depth settings but then wasn't sure if some of you might be using them sometimes? Personally I never touch those.

Before merging …

• [x] Changelog has been updated (docs/source/changes.md)

[larsoner]

Eventually we could pretty easily add support for volumetric source estimates I think. So I'm not sure it's worth getting rid of 'auto' just to potentially add it back. And I think loose and depth are both useful from time to time, especially when working with template / non-individualized MRIs (you often want loose=1.)

[SophieHerbst]

I agree about keeping the loose and depth parameters, we do use them sometimes. No opinion about the 'auto' option, but if mixed source spaces can be implemented easily, that would be useful.

[hoechenberger]

@larsoner @SophieHerbst I now changed things to basically only improve the documentation and config validation, except for one change: I removed support for loose=None, as this is equivalent to looe=0, and I always disliked we had two ways to achieve the same outcome.

This, of course, introduces an inconsistency with MNE, which allows for loose=None, but I can live with that…

Thoughts?

[SophieHerbst]

I have always found the options for those parameters a bit confusing, so I am fine with removing the None.

[larsoner]

Okay with me to remove None

[larsoner]

Whoops wrong PR comment, thanks @hoechenberger !