[ENH] Make explicit that "task" metadata applies to "beh" modality

[Remi-Gau]

Fixes #795

Adds task metadata that exist in other modalities to:

• behavioral
• PET

All metadata is RECOMMENDED.
Making TaskName REQUIRED (like it is for other modalities) could break backward compatibility, no?

I used a pretty global approach for beh assuming that these metadata could go in any of JSON flles on can find in this folder: should we have a more fine grain approach to this?

---

• update validator according to this PR (see this comment) ... see Behavior schema and validation bids-validator#1311

[Remi-Gau]

@melanieganz and @mnoergaard I am adding you as reviewers as this also relates to PET. 😄

[sappelhoff]

LGTM, probably something to remind ourselves of in #774

[Remi-Gau]

LGTM, probably something to remind ourselves of in #774

Good point.

Once this is merged I will update #774 (and also the validator ?)

[sappelhoff]

Mmh yes ... about updating the validator: Would you add it to the events.json schema, or the beh.json schema? The latter doesn't exist yet, but I think it'd be a better choice, and also something to clarify in this PR. WDYT?

Otherwise we'd have to specify it in the events.json schema: https://github.com/bids-standard/bids-validator/blob/master/bids-validator/validators/json/schemas/events.json

[Remi-Gau]

I would be more in favor of creating a beh.json schema.

This will need to happen at some point anyway.

[effigies]

LGTM. Suggested switching to Stroop instead of rest, since rest produces no behavioral data.

[sappelhoff]

This PR will have to be updated with respect to #774 --> metadata tables are now rendered via macros directly from the schema.

Other than that technical change request, I still approve of this PR

[Remi-Gau]

This PR will have to be updated with respect to #774 --> metadata tables are now rendered via macros directly from the schema.

Updated the PR to use the macros and the schema definitions.

The problem is that now the definitions for TaskName and Instructions both make reference to resting state which does not apply to behavior (see @effigies comments above).

Suggestions: shorten the base definitions and add the "resting state" mentions where necessary (func, eeg, meg, ieeg).

@tsalo does that make sense to you?

[tsalo]

👍 That works for me!

[Remi-Gau]

Tempted to add some info in the contributing.md on how to formulate the call to the macros to use the metadata description contained in the schema.
Just so that people are aware that those base description can be extended.

[sappelhoff]

Tempted to add some info in the contributing.md on how to formulate the call to the macros to use the metadata description contained in the schema.
Just so that people are aware that those base description can be extended.

I think that's a good idea 👍 we'll have to refer people to such docs quite some times in the future

[sappelhoff]

Thanks @Remi-Gau 🎉 let's get the validator PR fixed and merged next

[Remi-Gau]

yup.
in the meantime I will update #839