Add "Other" sex option

[hoechenberger]

As suggested in https://github.com/mne-tools/fiff-constants/issues/25#issuecomment-590406638

This also made me wonder why there's an "unknown" sex value, while we don't have that for any other personal information on the participant (e.g., handedness, height, weight, birthday… if we don't know them, they typically shouldn't be set, rigtht?)

Also, if anything, "Unknown gender" should be "Unknown sex" or just "Unknown", since we're talking about sex here and not gender… is it safe to change this?

[agramfort]

I am not convinced this change is necessary to the FIF Standard. Implications for all the software that use fif files is not trivial.

[hoechenberger]

@agramfort I do see the issue you mention, but without an option like Other, some participants cannot be properly classified… It would be like limiting "age" to 75 years max, but then what do you do when you measure an 80-year-old person, idk … :) Also there's currently a proposal to amend BIDS to include an Other option as well: https://github.com/bids-standard/bids-specification/issues/458 So that's maybe also something to consider

I'm open for discussion though (in case this wasn't clear)

[agramfort]

I would not be at the forefront here as again it's not such an easy fix. Let's wait to see how this evolves. In the analysis code I would treat Other as n/a so it's not a blocker to proceed with analysis pipeline.

[hoechenberger]

This is from a scientific survey we're currently preparing:

Now this is gender and not sex… but no-one's going to check their participants' crotches anyway, I suppose ;)

In Germany, all governmental institutions and businesses are required since 2018 to accept a third "gender/sex" option, "diverse" (there is only one word in German referring to both gender and sex). We have included all three on our consent forms in the past whenever we needed gender/sex information. (I even believe most ethics committees wouldn't approve studies otherwise anymore anyway.)

Just wanted to leave this here, and also ask @larsoner: how painful would it be to adopt the changes proposed here?

[larsoner]

I don't see it as being too difficult for MNE-Python, @agramfort would know better how challenging it would be for other packages

[agramfort]

my reaction after working on AP-HP data is that gender in a medical context if present will be used for models and visualisation. For covid we do an age pyramid by gender and we have 2 options and the rest is missing data. We need do discuss IRL.

[hoechenberger]

and we have 2 options and the rest is missing data.

Yes but this is in France and everything seems to be pretty binary in that regard around here ;) The forms and contracts I signed here would be illegal in Germany (and potentially other countries). Yes, let's discuss IRL over a beer / wine once the circumstances permit :)