Add custom metadata info about provenance of user data

[jcohenadad]

Currently, for data shared by users, we add the following info in the JSON. Example:

julien-macbook:~/data.neuro/sct-testing-large/sub-user0017/anat $ cat sub-user0017_T2star.json 
{   "Metadata":
    {
      "added_by": "Alexandru Foias",
      "added_on": "2021-05-27",
      "contact": "XXX",
      "URL": "email from 2021-05-26"
    }

This might not be BIDS-compliant, therefore we need to address this limitation and find a solution.

[alexfoias]

From my understanding, we have the freedom of using this custom field. Most field in the json sidecar files are recommended and very few are required. I think that the bids-validator doesn't check the contents of it.

https://bids-specification.readthedocs.io/en/stable/04-modality-specific-files/01-magnetic-resonance-imaging-data.html

[mariehbourget]

Thanks @alexfoias for the clarification.

I think you are right. I was under the impression that additional "non-standardized" metadata fields were prohibited but I did not find anything preventing it in the spec. Moreover, the BIDS-validator does not raise errors (nor warnings) for the unstandardized fields at the moment.

• For anat data : there are no required field in the JSON metadata, so having only the custom Metadata field is not an issue for validation.
• For micr data : we are planning to start archiving user data in a similar fashion than the sub-userXX method in sct-testing-large. In the case of Microscopy, to be BIDS-compliant, the JSON would need to have at least the PixelSize and PixelSizeUnits fields in addition of the custom Metadata field, as they are both always REQUIRED.
• For other data types, I recommend also making sure that all REQUIRED metadata field are also present in the JSON file.

As a reference, I found this about how to deal with "unstandardized" data and metadata in the spec. It does not name JSON metadata specifically but I think it applies:

• https://bids-specification.readthedocs.io/en/stable/02-common-principles.html#compulsory-optional-and-additional-data-and-metadata
• https://bids-specification.readthedocs.io/en/stable/02-common-principles.html#unspecified-data

Knowing all that, here are some thoughts, feel free to disregard if not appropriate:

• The name Metadata is pretty generic and I would recommend using a more specific name, ex: NeuropolyMetadata, so it is clear that it is not standardized.
• One of the way to make it entirely standardized would be to include the metadata entries from the custom Metadata field into columns of the participants.tsv file.

[jcohenadad]

The name Metadata is pretty generic and I would recommend using a more specific name, ex: NeuropolyMetadata, so it is clear that it is not standardized.

+1