BIDS youtube channel and podcast

55 views
Skip to first unread message

Remi Gau

unread,
Jun 25, 2021, 7:06:12 AM6/25/21
to bids-discussion

Hey everyone,


Short version:




Now for the longer version


As mentioned at the last BIDS town hall there is now a BIDS youtube channel: https://www.youtube.com/channel/UCxZUcYfd_nvIVWAbzRB1tlw


Its main goals are to try to:

  • centralize and curate the videos that already out here into a series of thematic playlist (tutorials, BIDS related tools, presentations of BIDS extensions),

  • host new original content on many aspects of BIDS and the vibrant community around it.


At the moment, you’ll find 3 “BEP debrief” videos where I had a chat with the leads of the most recently merged BIDS extension (ASL, qMRI, PET) to get their view on the whole process of creating extension to BIDS. Big thank you to Melanie, Martin, Agah, Gilles, Patricia and Henk who readily decided to play along for those.


There should soon be a video “BEP briefing” video with Sylvain Takerkart and Julia Sprenger to introduce the BEP 32 on animal electrophysiology where it’s going and what input they are now looking for from the community.


Gilles and Melanie independently suggested that those videos could be turned into podcast episodes. So now there is also a BIDS podcast: https://anchor.fm/bids-maintenance.


I have some other ideas using the same format, for example chats with:

  • BIDS apps developers

  • Users telling us about their experience of using BIDS and how they did managed before

  • Feedback from PIs about the cost and benefits of switching to BIDS


But I really would like to get ideas about where this channel could and should go. So don’t hesitate to get in touch (bids.mainten...@gmail.com) if you:

  • know of some videos out there that are not listed on the channel, 

  • have ideas for some other videos (like if the BEP you are working on needs a bit of publicity to get more community feedback or if you want to make a series of quick video tutorials on BIDS),

  • know the basics of video shooting and editing and would like to help

  • want to help in general with content creation 

  • want to help “host the show”, (would love a co-host who is less dude and less white than me

  • are interested in creating a mirror of the channel on bilibili to allow access to the content from countries that can’t connect to youtube 

  • want to create content in other languages or help with sub-titles


@+


Rémi


--

Rémi Gau

Dr Cyril Pernet

unread,
Sep 2, 2021, 4:58:29 AM9/2/21
to bids-di...@googlegroups.com

As part of Open NeuroPET, I am creating many datasets - those will be shared under GDPR (ie user registration)

Datasets do not have licenses but have a DUA

-------- in the BIDS spec ---------

A LICENSE file MAY be provided in addition to the short specification of the used license in the dataset_description.json "License" field. The "License" field and LICENSE file MUST correspond. The LICENSE file MUST be either in ASCII or UTF-8 encoding.

what about allowing DUA? we all know this is coming ...


-- 
Dr Cyril Pernet, PhD, OHBM fellow, SSI fellow
Neurobiology Research Unit
Copenhagen University Hospital, Rigshospitalet
Building 8057, Blegdamsvej 9
DK-2100 Copenhagen, Denmark

wamc...@gmail.com
https://cpernet.github.io/
https://orcid.org/0000-0003-4010-4632

Robert Oostenveld

unread,
Sep 2, 2021, 5:55:23 AM9/2/21
to bids-di...@googlegroups.com
Hi Cyril

For the web interface of https://data.donders.ru.nl we use the term "Data Use Agreement” (DUA). On other places the agreement is called a Data Transfer Agreement (DTA, e.g. https://www.uu.nl/sites/default/files/data_transfer_agreement_youth_template.pdf ), or Data Use Terms  https://www.humanconnectome.org/study/hcp-young-adult/data-use-terms, or Material Transfer Agreement (see https://autm.net/surveys-and-tools/agreements/material-transfer-agreements), or a Data UseR Agreement (also DUA, see https://onlinelibrary.wiley.com/doi/full/10.1002/hbm.25351)

When considering code, it is common to have a file COPYING in the top level directory with for example the GPLv3 license in it (which is autodetected and formatted nicely by github, e.g. https://github.com/fieldtrip/fieldtrip/blob/master/COPYING), but if you look at https://github.com/fieldtrip/playground/blob/master/LICENSE you see it is also detected when the file is named LICENSE, or LICENSE.md, or LICENSE.rst. See also https://docs.github.com/en/github/creating-cloning-and-archiving-repositories/creating-a-repository-on-github/licensing-a-repository 

In the actual file collections in the Donders Repository we automatically add a LICENSE.txt file with the DUA upon publication as explained here https://data.donders.ru.nl/doc/help/faq/publish-data.html?5#faq-files-added-to-dsc. See https://public.data.donders.ru.nl for examples that allow redistribution (i.e. not the ones with potentially identifying data, which are under the RU-DI-HD agreement). When logged in, you can also browse https://webdav.data.donders.ru.nl/  for all public collections, including the ones with a restricted DUA that you have requested and been granted access to. 

The content of the agreement for data (which is factual and measured) is different than that for code (which is written and hence authored), as such it is important to distinguish copyright “contracts” from data transfer “contracts”. Hence I would personally not use COPYING for a DUA as it suggests rights related to an author. But for me either DUA, MTA, DTA, DUT, or LICENSE are all fine. 

In BIDS dataset_description.json I have so far been using the License field and added a LICENSE.txt file (with txt extension, although upon careful reading the spec it should not have the extension). If we all agree that the field is to be used for this, then I don’t think there is a reason to rename License into DUA or something else.

best
Robert



--
We are all colleagues working together to shape brain imaging for tomorrow, please be respectful, gracious, and patient with your fellow group members.
---
You received this message because you are subscribed to the Google Groups "bids-discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/e4ae8c2a-2a85-62b9-bd5d-7e72daa7bef1%40gmail.com.

Dr Cyril, Pernet

unread,
Mar 3, 2022, 5:13:32 AMMar 3
to bids-di...@googlegroups.com

Hi all,

As many of us are moving onto derivatives, I realize that different choices are being made regarding how to describe how a derivative set is being made.

Regardless of that choice, I am thinking we should have principles guiding the json derivatives information - here are a few ideas for the community

- how to describe successive steps that lead to the shared derivative data (e.g. must have 'step-1', 'step-2', must be in a {} hierarchy, or whatever else we can agree on)

- how to refer to a specific code located into 'code' which generated the data (1) when located in raw when raw are present (raw and derivative shared); (2 when located in raw which is itself another dataset (derivatives shared separately from raw - although I guess they become the raw folder) (3) when shared/available somewhere else

- if using a 'standard pipeline' (say fmriprep) and wanting to indicate the arguments in should this be a 'code' (a one liner .sh for instance) or we could have a 'arguments' key (my prefered option) like

"GeneratedBy": [ { "Name": "fmriprep", "Version": "1.4.1", "Container": { "Type": "docker", "Tag": "poldracklab/fmriprep:1.4.1", }

"Arguments": "fmriprep --xxx --xx"

},


With these 3 things, I believe we can encode 'any' workflow, each BEP specifying unique key-values pairs as needed and we can read those in a consistent manner

Cyril


-- 
Dr Cyril Pernet, PhD, OHBM fellow, SSI fellow
Neurobiology Research Unit, 
Building 8057, Blegdamsvej 9
Copenhagen University Hospital, Rigshospitalet
DK-2100 Copenhagen, Denmark

tibor...@gmail.com

unread,
Mar 3, 2022, 5:18:28 AMMar 3
to bids-di...@googlegroups.com

Hi Cyril,

 

There is already a BEP in the pipeline on BIDS provenance:

bids-standard/BEP028_BIDSprov: Organizing and coordinating BIDS extension proposal 28 : BIDS Provenance (github.com)

 

BIDS Extension Proposal (BEP028): Provenance - Google Doc

 

Kind regards,

Tibor

 

Auer, Tibor M.D. Ph.D.

Research Fellow

School of Psychology, Faculty of Health and Medical Sciences

University of Surrey, Guildford GU2 7XH

T.A...@surrey.ac.uk

@TiborAuer

--

We are all colleagues working together to shape brain imaging for tomorrow, please be respectful, gracious, and patient with your fellow group members.
---
You received this message because you are subscribed to the Google Groups "bids-discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussi...@googlegroups.com.

Dr Cyril, Pernet

unread,
Mar 3, 2022, 5:26:56 AMMar 3
to bids-di...@googlegroups.com, tibor...@gmail.com

yes I know but while it may answer my 'steps' issue (though this is not what other derivatives are doing and that's a concern), it doesn't addressed other IMO

Dr Cyril, Pernet

unread,
Mar 3, 2022, 5:29:16 AMMar 3
to bids-di...@googlegroups.com, tibor...@gmail.com

of of course, NIDM is great but BIDS covers a little more than MRI :-)

On 03/03/2022 11:18, tibor...@gmail.com wrote:

tibor...@gmail.com

unread,
Mar 3, 2022, 5:38:57 AMMar 3
to Dr Cyril, Pernet, bids-di...@googlegroups.com

I am not sure I can follow…

 

Although, BIDS-PROV also covers the particular use case of DICOM conversion (specific for MRI) (1.2.a in the Google Doc), it proposes a framework for provenance for BIDS derivatives in general, regardless of the modality (1.2.b in the Google Doc).

 

It also specifies file structure, format (JSON-LD), and terms allowing linking data and for a more in-depth provenance for less ‘standard’ workflows.

 

Tibor

 

Auer, Tibor M.D. Ph.D.

Research Fellow

School of Psychology, Faculty of Health and Medical Sciences

University of Surrey, Guildford GU2 7XH

T.A...@surrey.ac.uk

@TiborAuer

 

Dr Cyril, Pernet

unread,
Mar 3, 2022, 5:43:26 AMMar 3
to tibor...@gmail.com, bids-di...@googlegroups.com

is not not misleading prov vs common derivatives? i'll try that way in the next few weeks but seems that others been working on things separate from that model .. (look at any derivative coming up with their way to do thing) -- seems like a prblm beyond my specific questions

tibor...@gmail.com

unread,
Mar 3, 2022, 6:01:05 AMMar 3
to Dr Cyril, Pernet, bids-di...@googlegroups.com

As far as I can see, it is a generic problem of capturing provenance. Ideally, the same model could be used for any provenance, regardless of the processing stage (preprocessing, modelling, inferencing, sharing) or whether it is within the BIDS ecosystem or not. This would ensure a more efficient linkage and query.

For BIDS derivatives, however, it can potentially be simpler when using ‘standard’ workflow (e.g. a BIDS App).

 

When I read your mail first, I thought it was BIDS-PROV! The only difference is that BIDS-PROV has a specified set of terms.

Paola Di Maio

unread,
Mar 3, 2022, 6:06:47 AMMar 3
to bids-di...@googlegroups.com, Dr Cyril, Pernet
Thank you Tibor et al
Yes, I am working on developing a meta-model-object
which can be used by any standard , I am glad to see this is being
done in BIDS- can be used as an implementation example
PDM

tibor...@gmail.com

unread,
Mar 3, 2022, 6:10:34 AMMar 3
to bids-di...@googlegroups.com, Dr Cyril, Pernet

Hi Paloa,

 

There is also a more generic effort to build a data model for neuroimaging.

Please, check out http://nidm.nidash.org and feel free to comment and contribute.

 

Kind regards,

Tibor

 

Auer, Tibor M.D. Ph.D.

Research Fellow

School of Psychology, Faculty of Health and Medical Sciences

University of Surrey, Guildford GU2 7XH

T.A...@surrey.ac.uk

@TiborAuer

 

Paola Di Maio

unread,
Mar 3, 2022, 6:16:59 AMMar 3
to bids-di...@googlegroups.com, Dr Cyril, Pernet
Thank you Tibor
I will join!
My work is even more abstract,  is a general model for brain data-
 neuroimaging is subset of brain data, ie the image refers to a visual way of
representing the brain data, which can be equally expressed numerically and/or textually
It s a system level (SL) model-
Nonetheless, to see a neuroimaging model is a very good applied use case for the future SL metamodel, ie, will it fit the neuroimaging use case-  so, I ll try to learn and contribute if I can
see you there
pdm

yarikoptic

unread,
Mar 3, 2022, 10:22:01 AMMar 3
to bids-discussion

worth announcing on twitter (didn't spot it) and promoting.  Whenever reaching 100 subscribers, get a dedicated short name like we got for https://www.youtube.com/c/DataLad .

Remi Gau

unread,
Mar 7, 2022, 2:06:35 PMMar 7
to bids-di...@googlegroups.com, yarikoptic


On 03/03/2022 16:22, yarikoptic wrote:
There is BIDS youtube channel: https://www.youtube.com/channel/UCxZUcYfd_nvIVWAbzRB1tlw

worth announcing on twitter (didn't spot it) and promoting.  Whenever reaching 100 subscribers, get a dedicated short name like we got for https://www.youtube.com/c/DataLad .

Actually trying to make it a little more visible from the Starter Kit angle via this pull request.

https://github.com/bids-standard/bids-starter-kit/pull/263

Also don't hesitate  to let us know if there are videos / slides that are not listed.




--
We are all colleagues working together to shape brain imaging for tomorrow, please be respectful, gracious, and patient with your fellow group members.
---
You received this message because you are subscribed to the Google Groups "bids-discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussi...@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages