Steering Group minutes 2023/05/25

Date: Thursday, May 25th, 2023

Topic Relevant Links
NSF POSE Workshop https://bidsstandard.slack.com/files/U02Q9B87D9P/F057GT2JDEU/data-metadata-standards-draft_v6.docx
BIDS workshop prep questions

1. What is your vision for derivatives? (a vision is the non technical expression of what you want it to do, e.g. I can see output X being reused across labs for Y or output X can be easily reused by ML experts)

2. There is 2 views on derivatives: A. it should store the outcome of a pipeline + share code B. it should document every steps, and thus store every step.

--> what is your view?

--> to help developing BEP derivatives, would it be useful to have a framework/guidelines considering the focus of derivatives i.e. reusage?

3. Given your involvement in BIDS and derivatives, what is/are the current roadblocks to advance your project?


Present: Cyril Pernet, Ariel Rokem, Guiomar Niso, Yaroslav Halchenko, Stefan Appelhoff, Kimberly Ray

Guest: Guy Jones

NOTES

Guest Guy Jones: discuss adding badges to the BIDS compliant datasets on Scientific Data Journal website.

Guy’s Feedback:

May not have a badge in a classic sense, but possible a notification on a set of collections

All this requires is a list of compliant dois that could be added to a group

This would allow for changes to be made at any time, even retrospectively

Collections can be searchable on the journal website.

Cyril: What are the tools that Scientific Data uses to check that BIDS data are compliant?

Reviewers seem to be a first point of identification for BIDS on a less official level.

It does not appear that there is an official way that Scientific Data confirms for BIDS validity. However the use of BIDS tools can facilitate this.

Guy: Some fields have tools that produce a “validation receipt” that demonstrates the validity of the dataset. A printout of the validator could do that.

BIDS validator could be used, however the validator outputs are not consistent. Scientific data could provide different levels (A,B,C grading) to indicate if a dataset has certain BIDS errors or is missing necessary metadata or files. For example datasets on OpenNeuro may be in BIDS but may still contain errors.

If OpenNeuro mandated that all datasets be BIDS compliant, then that could be one route for ensuring validity.

Scientific Data could consider a trial where the authors required a certificate of compliance. This could be an option (check box) during the manuscript submission period. This could be the easiest, but may not be a perfect solution because it's based on the user.

For other data modalities (e.g. genetics, etc), typically the journal just encourages it to be shared on a repository. There are not many efforts like BIDS, for standards themselves.

Nature Neuro may be another target journal that Guy could connect us with to expand this mechanism.

ACTION ITEM: add a button that produces a PDF of the BIDS validation results.

Should this PDF include a file listing to make sure that the dataset has not been altered.

Add a list of BIDS ignore files:

[https://github.com/bids-standard/bids-validator/issues/1676]{.underline}

[https://github.com/bids-standard/bids-validator/issues/1677]{.underline}

Should we reach out to the new Imaging Neuroscience journal while they are still developing their platform?

Potentially reach out to someone on their [editorial board]{.underline}

NSF POSE WORKSHOP

A funding mechanism for conferences and workshops. (note: Proposals under $100k are reviewed internally)

They intend to do a workshop on open source models for data and metadata standards.

Timing: First quarter of 2024 - Ariel is planning a submission that includes interdisciplinary colleagues at UW, proposal due mid June

BIDS Steering Committee may be included on a list of invited speakers (not required to attend).

Could this be a line of funding for BEPS? - the goal is not neuro-focused, BIDS has good practices to share, but a proposal would need to be interdisciplinary.

Are there cross sectional topics that could be considered? E.g. other modalities, or repositories, provenance. what other standards are out there? Physics, astronomy, CERN

BIDS workshop prep questions

Goal of the questions is to help focus the upcoming workshop, identify areas of agreement.

Suggested Meeting discussion point: what level of analysis steps should be included and how should those steps/data be reported and/or stored. This would allow reproducibility of results

Other questions to address:

OHBM BIDS Townhall:

BIDS should be considered as a part of the Best practices, Cyril will contact the Program committee about establishing a yearly townhall spot guaranteed

Add to the next agenda website updates - who has done it and how do we update in the future?