MRG, DOC: add EEG referencing tutorial

[drammock]

This is ready for review but should probably wait until a solution to #6647 is merged in.

[codecov]

Codecov Report

Merging #6648 into master will decrease coverage by <.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #6648      +/-   ##
==========================================
- Coverage   89.48%   89.48%   -0.01%     
==========================================
  Files         420      420              
  Lines       75580    75582       +2     
  Branches    12381    12382       +1     
==========================================
  Hits        67635    67635              
- Misses       5138     5139       +1     
- Partials     2807     2808       +1

[drammock]

here's the revised rendering: https://14689-1301584-gh.circle-artifacts.com/0/dev/auto_tutorials/preprocessing/plot_55_setting_eeg_reference.html

@wmvanvliet thanks for the feedback, I've re-worked the introduction and I think it's much improved. @cbrnr please check whether I've explained add_reference_channels to your satisfaction.

[cbrnr]

Yes, very nice! There's a small formatting error in "signal without restoring the signal at EEG 999)" (the closing parenthesis shouldn't be there and EEG 999 should be in code font).

[drammock]

@cbrnr yep, that formatting error is already fixed in 3aea686

[drammock]

updated rendering: https://14699-1301584-gh.circle-artifacts.com/0/dev/auto_tutorials/preprocessing/plot_55_setting_eeg_reference.html

[agramfort]

@cbrnr feel free to merge if you're happy.

[drammock]

This should wait for a solution to #6647, which I hope to have today

[cbrnr]

OK, please ping me if it works.

[jona-sassenhagen]

Looks good so far.

[drammock]

NB: after #6670 is merged, this should get rebased and have the Circle rendering checked.

[drammock]

rebased to incorporate #6670, let's see if the docs look OK.

[drammock]

@cbrnr this should be good to merge now. Here's the raw plots with titles:
https://14963-1301584-gh.circle-artifacts.com/0/dev/auto_tutorials/preprocessing/plot_55_setting_eeg_reference.html#creating-the-average-reference-as-a-projector

[cbrnr]

I've pushed a few minor improvements. This tutorial is really nice, there's just one point I'd like to see addressed before merging. The last section talks about why average referenced data is important for source imaging, but it strongly recommends the "average-reference-as-projection approach". However, the reasons given are just reasons for the average reference in general. I really miss an explanation why users should prefer projection=True in this scenario.

[agramfort]

@cbrnr merge if you're happy

[drammock]

@cbrnr are you happy with the new wording?

[cbrnr]

Yes! Thanks @drammock!