Better Collaborative Infrastructure

295 views
Skip to first unread message

Horea Christian

unread,
Dec 22, 2017, 2:45:11 PM12/22/17
to bids-discussion
I often found myself quite frustrated by the current BIDS communication channels. I particularly find it hard to stay posted on mailing list threads of relevance to me without getting flooded by threads which are irrelevant to me. The Google Docs infrastructure is (becoming) very slow and clunky, and it's hard to keep track of and link against discussions and individual replies. The mixing of content and formatting on the document itself makes it challenging to disentangle contributions, and it's easy to make inadvertent formatting changes to the text of others while trying to contribute some new text yourself.

I commonly manage collaborative documents with my students and my supervisor via git-repository websites with issue tracking support (commonly BitBucket and GitHub). I find it's much cleaner to keep the document in a WYSIWYG format (e.g. Markdown) and have the discussions on a dedicated issue tracker where you can link and tag them individually and they're not constrained to a 50px width. GH in particular has excellent accessibility features for casual contributors with no grasp of git. I would really recommend moving to GH entirely.

I agree that mailing lists may have some broadcasting uses not adequately addressed by issue trackers; but in order to provide a good broadcasting channel, the mailing list should exclude any and all discussions which are better suited for an issue tracker.

So, what do you say, would you support packing up and moving to a new infrastructure?

Kirstie Whitaker

unread,
Dec 22, 2017, 2:57:49 PM12/22/17
to bids-di...@googlegroups.com
Hi everyone,

I had some conversations with Chris G a couple of months ago about the communications channels for BIDS.

I have very little energy for engaging on mailing lists in general, and in particular I think the mix of development and support for this list makes it difficult for people to ask beginner questions about BIDS.

Chris put in a hard sell for moving more questions to NeuroStars and I promised to find time to move some of the beginners resources to a more prominent position on the website. (Which I haven’t done, but still would be delighted to work on).

I would personally like to work in GitHub issues as Horea suggests, but I think moving the mailing list to GH might be tough as it’s the current work flow for most on the list.

I don’t have any strong suggestions because I think this is something the BIDS community should decide together.

So, my take home is: yes to different ways of communicating, and with a minor (lightly held) suggestion to move different *types* of communication in different stages.

Kirstie 

Sent from my iPhone, please excuse any typos or excessive brevity
--
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 post to this group, send email to bids-di...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/1c5120d1-ceed-4e03-8aa3-bb8b00cfb5fe%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Chris Gorgolewski

unread,
Dec 22, 2017, 4:41:16 PM12/22/17
to bids-discussion
Thanks for bringing this up. Let me sum up the problems that were brought up (here and via other exchanges):

1. The google doc is slow to load.
2. Comments on the doc are not well exposed and easy to miss (unless you explicitly sign up for them).
3. Comments on the doc are not great for having a longer conversation.
4. It's not possible to unsubscribe from one mailing list thread - you either receive all of them or none.
5. Since the mailing list is dominated by technical conversation it can be intimidating for beginners trying to ask basic questions.
6. The google doc includes both content as well as text formatting.

Proposed solutions (not all mutually exclusive):
1. Move everything to GitHub.
2. Create a separate mailing list for users/beginners.
3. Direct beginners to Neurostars.

Some stats:
1. Since March 24th here were 50 questions mentioning BIDS on NeuroStars, and around 120 email threads in the same period on the mailing list.
2. 41% of the 272 mailing list members have disabled email delivery.

The way we communicate is essential for the success of this community. I would love to hear what others think.

Best,
Chris

To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.

--
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-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/4E47CB0C-15E2-45DF-B07C-E030253BC579%40cam.ac.uk.

Dmitry Petrov

unread,
Dec 22, 2017, 5:03:38 PM12/22/17
to bids-discussion
I think it is an important discussion. My thoughts:

> 1. Move everything to GitHub.

I think there should be BIDS tutorials GitHub which covers essential BIDS infrastructure and has necessary links. For example, it may have organized notebooks with PyBIDS examples, data conversion examples + most useful BIDS apps tutorials. And beginners may ask specific questions there about specific tutorials in issues. 

> 2. Create a separate mailing list for users/beginners.

I think mailing list is important for general/conceptual discussion (like this one)

> 3. Direct beginners to Neurostars.

In my opinion, Neurostars is good for practitioners and advanced beginners with non-trivial questions which are not covered by tutorial GitHub.

To sum up, person involves with BIDS first through tutorials GitHub, then asks non-trivial questions at Neurostars and if this person has an idea how to make BIDS better he writes to the mailing list. 

I can help with part 1.

Dmitry 

Best,
Chris

To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussi...@googlegroups.com.
To post to this group, send email to bids-di...@googlegroups.com.

--
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 post to this group, send email to bids-di...@googlegroups.com.

Chris Gorgolewski

unread,
Dec 22, 2017, 5:08:50 PM12/22/17
to bids-discussion
Thank you for sharing your thoughts. There is in fact a tutorial github repository spearheaded by Dora Hermes. You can find it at  https://github.com/INCF/bids-starter-kit. It's still work in progress so your contributions would be very much welcomed.

We should probably advertise it better (see https://github.com/INCF/BIDS/issues/4).

Best,
Chris

To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.
 To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/b68e5e9c-5803-4323-ae29-2169e01d1fcc%40googlegroups.com.

Vince Calhoun

unread,
Dec 22, 2017, 5:23:38 PM12/22/17
to bids-di...@googlegroups.com

Dare I suggest slack? ;-)

 


Best,

Chris

 

To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussi...@googlegroups.com.
To post to this group, send email to bids-di...@googlegroups.com.

--
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 post to this group, send email to bids-di...@googlegroups.com.

 

--

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 post to this group, send email to bids-di...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/CAAQzouMehOj83jtumr-zFJ1sy0kVWktcuM%3D%3D6Wbi9oMG75K8LQ%40mail.gmail.com.

Chris Gorgolewski

unread,
Dec 22, 2017, 5:28:32 PM12/22/17
to bids-discussion
Of course, but it would be great if you could elaborate on how it addresses the existing problems and what potential drawbacks do you see.

BTW there are a couple of bids channels on the brainhack slack team. For example https://brainhack.slackarchive.io/bids_eeg/page-2

Best,
Chris


Best,

Chris

 

To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.

--
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-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.

--
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-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.

--
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-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.
 To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/07cc01d37b73%248125a5e0%248370f1a0%24%40mrn.org.

Vince Calhoun

unread,
Dec 22, 2017, 5:38:05 PM12/22/17
to bids-di...@googlegroups.com

I just think it would be a good way to communicate and enable us to categorize the topics better that email.

 

VDC

 

From: bids-di...@googlegroups.com [mailto:bids-di...@googlegroups.com] On Behalf Of Chris Gorgolewski
Sent: Friday, December 22, 2017 3:28 PM
To: bids-discussion <bids-di...@googlegroups.com>
Subject: Re: [bids-discussion] Better Collaborative Infrastructure

 

Of course, but it would be great if you could elaborate on how it addresses the existing problems and what potential drawbacks do you see.


Best,

Chris

 


Best,

Chris

 

To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussi...@googlegroups.com.
To post to this group, send email to bids-di...@googlegroups.com.

--
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 post to this group, send email to bids-di...@googlegroups.com.

--
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 post to this group, send email to bids-di...@googlegroups.com.

--
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 post to this group, send email to bids-di...@googlegroups.com.


For more options, visit https://groups.google.com/d/optout.

 

--

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 post to this group, send email to bids-di...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/CAAQzouNA8KE2vMzGofVbBoKBcDGE2%2B7XUid%2BLyfbE_QD_V%3DBOQ%40mail.gmail.com.

wamc...@gmail.com

unread,
Dec 23, 2017, 4:37:43 AM12/23/17
to bids-discussion
Hey Chris

I still think Google group mailing list is good but worth
- limiting what you want it for (beginners, expert, technical)
- slowly, a together, make sure the header includes the BIDS BEP)

Kirstie Whitaker

unread,
Dec 23, 2017, 5:07:04 AM12/23/17
to bids-di...@googlegroups.com
Hi everyone,

Following on from Vince’s slack suggestion here’s another weakly held opinion from me. Absolutely feel free to disregard!

In my opinion one of the problems with Slack is that there’s usually a barrier to entry - that is, you have to apply to join and be approved. That isn’t always the case - I think you can get around that requirement with a heroku app but it’s still a bit of a faff.

I really like Gitter. It would be easy to have rooms corresponding to the different BIDS repositories that already exist, and it’s also easy to build additional rooms as needed within one organisation. One can view public rooms without having to sign up to anything, and you only need a GitHub or twitter account to join public rooms. Gitter also links nicely to PRs & issues on GitHub so it’s quite a nice complement to conversations on GitHub.

The other challenge is that it means we’d have ANOTHER slack team to monitor. And this goes for Gitter rooms too! I personally find monitoring Gitter rooms a little easier than slack teams but I totally understand if there’s not much appetite for another app/website to check!

On a different note: thank you Chris for pointing out the BIDS starter kit!! I’m very exited to contribute there :)

Kx

Sent from my iPhone, please excuse any typos or excessive brevity

> --
> 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 post to this group, send email to bids-di...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/b638ab9b-3a14-4236-a925-a80a08870310%40googlegroups.com.

Michael Hanke

unread,
Dec 23, 2017, 5:30:54 AM12/23/17
to The Brain Imaging Data Structure (BIDS) discussion
Hey,

question re slack and all modern communication forms. Do they all have public archives (i.e. i do not have to be member of a club to read them)? Are these archives indexed by common search engines?

Mailing list archives can give people a sense of how a community behaves, what topics are and have been important previously. They are more accurate than post-hoc summaries and codes of conduct, at the cost of verbosity, of course.

Everytime a comment on the google doc is "resolved" the rational behind a decision is essentially lost to all but the most motivated forensic historians.

I do not know what is best for BIDS, but I think it is very important for the development of any standard that one is able to point to references of previous discussions, if only to pick up on a previously unresolved debate even months or years later.

Cheers,

Michael


On Dec 23, 2017 11:07, "Kirstie Whitaker" <kw...@cam.ac.uk> wrote:
Hi everyone,

Following on from Vince’s slack suggestion here’s another weakly held opinion from me. Absolutely feel free to disregard!

In my opinion one of the problems with Slack is that there’s usually a barrier to entry - that is, you have to apply to join and be approved. That isn’t always the case - I think you can get around that requirement with a heroku app but it’s still a bit of a faff.

I really like Gitter. It would be easy to have rooms corresponding to the different BIDS repositories that already exist, and it’s also easy to build additional rooms as needed within one organisation. One can view public rooms without having to sign up to anything, and you only need a GitHub or twitter account to join public rooms. Gitter also links nicely to PRs & issues on GitHub so it’s quite a nice complement to conversations on GitHub.

The other challenge is that it means we’d have ANOTHER slack team to monitor. And this goes for Gitter rooms too! I personally find monitoring Gitter rooms a little easier than slack teams but I totally understand if there’s not much appetite for another app/website to check!

On a different note: thank you Chris for pointing out the BIDS starter kit!! I’m very exited to contribute there :)

Kx

Sent from my iPhone, please excuse any typos or excessive brevity

> On 23 Dec 2017, at 09:37, wamc...@gmail.com wrote:
>
> Hey Chris
>
> I still think Google group mailing list is good but worth
> - limiting what you want it for (beginners,  expert,  technical)
> - slowly,  a together,  make sure the header includes the BIDS BEP)
>
> --
> 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-discussion+unsubscribe@googlegroups.com.
> To post to this group, send email to bids-discussion@googlegroups.com.

> To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/b638ab9b-3a14-4236-a925-a80a08870310%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

--
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-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/282EAC17-0505-4732-92B6-14270844F956%40cam.ac.uk.

Kirstie Whitaker

unread,
Dec 23, 2017, 6:07:36 AM12/23/17
to bids-di...@googlegroups.com
Great question.

I don’t think Gitter gets deleted so I think conversations there would persist. I don’t think they’re super easily searchable.

The free slack account deletes messages after 10,000 (which is not that much as a community grows), and the archive is not public. (Again, it may be possible to make that archive and put it somewhere easy to access, but we’d have to do that work to ensure Michael’s point remains true for BIDS).

Conversations on GitHub issues definitely persist and are a really nice way to look back at how a project has developed. (But may be too much of a barrier to entry for researchers old and new who don’t use GitHub regularly.)

Kx


Sent from my iPhone, please excuse any typos or excessive brevity
To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussi...@googlegroups.com.
To post to this group, send email to bids-di...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/CAK-KoxSsYuW8i-zwX8dxtLeiFK0qasrWeKSEsU%3DinJYb7%2BzEwg%40mail.gmail.com.

Vince Calhoun

unread,
Dec 23, 2017, 9:23:47 AM12/23/17
to bids-di...@googlegroups.com

All great points. I don’t have a strong opinion about which option we use, but wanted to mention that slack does have a nonprofit program, which if you apply to and receive provides persistent archiving of all messages for no cost (we use this at MRN).  I’m not sure that option will work for this group though as there is no central organization everyone belongs to and the process requires a bit of paperwork.  As for a public archive, I think there are some public archive options for slack as well though I haven’t tried these. I agree that public searchable archiving is essential.

 

Best,

 

Vince

 

From: bids-di...@googlegroups.com [mailto:bids-di...@googlegroups.com] On Behalf Of Kirstie Whitaker
Sent: Saturday, December 23, 2017 4:08 AM
To: bids-di...@googlegroups.com
Subject: Re: [bids-discussion] Better Collaborative Infrastructure

 

Great question.

> To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussi...@googlegroups.com.
> To post to this group, send email to bids-di...@googlegroups.com.

> To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/b638ab9b-3a14-4236-a925-a80a08870310%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

--
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 post to this group, send email to bids-di...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/282EAC17-0505-4732-92B6-14270844F956%40cam.ac.uk.
For more options, visit https://groups.google.com/d/optout.

--
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 post to this group, send email to bids-di...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/CAK-KoxSsYuW8i-zwX8dxtLeiFK0qasrWeKSEsU%3DinJYb7%2BzEwg%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

--
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 post to this group, send email to bids-di...@googlegroups.com.

Yaroslav Halchenko

unread,
Dec 23, 2017, 12:02:25 PM12/23/17
to bids-discussion

On Fri, 22 Dec 2017, Horea Christian wrote:
> (commonly BitBucket and GitHub). I find it's much cleaner to keep the
> document in a WYSIWYG format (e.g. Markdown) and have the discussions on a
> dedicated issue tracker where you can link and tag them individually and

A one important IMHO aspect is to keep the issues and some pointed
discussions "linked" to the actual document in development. Github
works quite nice whenever code/docs are under git on github as well.
But also keep in mind that software code issues/fixes rarely have
a clear association with the particular block of code and most often
touch a discontinuous set of line blocks across multiple files.

ATM, BIDS specification is under google docs [1] and it worked
out, again in just my opinion, quite nicely: Specific
discussions/issues/suggestions are done at the point of needed
modification. They are "Resolved" one way (closed) or another
(accepted/edited) and comments disappear when that is done. I think
that making a dedicated github repo just for the sake of issues tracking
as related to the specification, disconnected from the actual document,
would be inferior.

In summary, whatever channel is chosen for providing "issues" tracking
should ideally be tightly linked to the standard document itself.

[1] I am saying that even though I have similar to your reservations
with google services.

--
Yaroslav O. Halchenko
Center for Open Neuroscience http://centerforopenneuroscience.org
Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
Phone: +1 (603) 646-9834 Fax: +1 (603) 646-1419
WWW: http://www.linkedin.com/in/yarik

Chris Madan

unread,
Dec 23, 2017, 12:38:11 PM12/23/17
to bids-discussion
Hi all,

I only saw this thread because Chris G shared it on Twitter.

My brief thoughts are that I'm interested in seeing BIDS evolve--and providing feedback, but can't keep up with yet another mailing list (or slack). I'm in favor of Github/Gitter. I agree with Yarik though, and a related consideration might be that there should be separate 'channels' for different types of discussions, e.g., technical questions about making a specific dataset BIDS compliant vs. discussing improvements to the standards and how to increase use of it vs. drafting of a specific collaborative document. 

Best,

Chris

elizabet...@gmail.com

unread,
Dec 23, 2017, 12:40:30 PM12/23/17
to bids-discussion
Hi all,

I'm excited to see this discussed, as I thinking improving our communication can only improve the health of the community that much more. My read is that there are two different issues at play here: communication as related to (1) people adopting the BIDS standard, and (2) improving the specification. The mailing list seems to be primarily used for issue (2) at this point, which might be because keeping up with the threads can be especially overwhelming to new BIDS users.
 
I think a lot of progress could be made by adopting Kirstie's idea of different channels for different stages of involvement, which to me would split along discussions for issue (1) and issue (2).

Regarding issue (1), I'm particularly excited to learn that a tutorial repository exists. I wonder if investing in those tutorials and creating something similar to what's available for nipype would provide the best introduction to interested beginners. Issues directly arising from the tutorials could then be brought either as a GitHub issue on that repository or to Neurostars.

For issue (2), I think a GitHub repository with the current standard, where new issues can be opened and discussed, would be ideal. Then issues could reference specific line numbers, and there would be an archive of how the discussion evolved.

My opinions for implementing (1) and (2) are weakly held, but I do think that different channels for different kinds of engagement are going to become more important as the community grows.

Elizabeth

JB Poline

unread,
Dec 23, 2017, 1:02:26 PM12/23/17
to The Brain Imaging Data Structure (BIDS) discussion
Hi,

I will investigate if INCF would have this option - this is to me our parent organization.
I found myself sometimes lost in the 50 channels X 10 workspaces in slack but -slowly- adapting :)

cheers
JB

> To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussion+unsubscribe@googlegroups.com.
> To post to this group, send email to bids-discussion@googlegroups.com.


> To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/b638ab9b-3a14-4236-a925-a80a08870310%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

--
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-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.

--
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-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.

--
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-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.

--
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-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/086601d37bf9%24a32c29d0%24e9847d70%24%40mrn.org.

Chris Gorgolewski

unread,
Dec 23, 2017, 1:46:29 PM12/23/17
to bids-discussion
Great to see such a vibrant discussion. At this stage I just wanted to add a couple of features of the current setup I learned to appreciate (both in context of standard development, not user support):

1) Ability to easily rope an outside expert into the conversation. In context of discussing changes to the standard it is very useful to get an opinion of an expert that might not necessarily be part of the community (aka member of the mailing list). Right now anyone can just CC any person on a mailing list thread (even if that person is not a list member). The list has been configure in a way that allows non-subscribers to post to it so if the expert presses "reply all" everyone will get the message. This works because everyone has an email address (in contrasts to say a GitHub account).

2) Ease of commenting and suggesting changes. Google docs makes it trivial for anyone (with or without Google account) to leave a comment or suggest a change. We once had a gentleman from an SF startup on a hackathon who read through the whole spec and left comments on it. This was easy because he could do it from withing the document itself. GitHub has gotten much better at editing and suggesting changes via the web UI, but it still requires multiple clicks, context switches and a GitHub account. There is also a lot of lingo that might be familiar to us, but confusing to others (such as fork or pull request).

Those are not deal breakers, but as the conversation continues it would be good to have them in mind. 

Best,
Chris

PS Don't get me wrong - I love GitHub and use it almost every day, it just might not be the best fit for all use cases.

Kirstie Whitaker

unread,
Jan 4, 2018, 4:52:36 AM1/4/18
to bids-di...@googlegroups.com
Happy new year everyone! I hope you all had a really wonderful holiday with lots of rest and relaxation :)

As this conversation happened mostly over the holidays I wanted to ping it again to see if anyone who missed it had any opinions. Or maybe someone wanted to try to summarise to help us (me!) get a consensus feeling?

Kx

To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/CAAQzouNC-vN_mxRMDj6HetrvrMqpD5SSYBz%2BSXSN7tAOZxLBtQ%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.



--
Kirstie Whitaker, PhD

Research Fellow, Alan Turing Institute
Research Associate, University of Cambridge

Mailing Address
Brain Mapping Unit
Department of Psychiatry
Sir William Hardy Building
Downing Street
Cambridge CB2 3EB

Phone: +44 7583 535 307
Website: https://whitakerlab.github.io
Twitter: @kirstie_j

Chris Gorgolewski

unread,
Feb 15, 2018, 7:43:24 PM2/15/18
to bids-discussion
This discussion died a little over the holidays, but my overarching feeling is that no consensus was found. There could be more ideas we should investigate.

Thinking about this made me realize that we should clarify where to find help and how to contribute to the project. I made a pull request to make such changes: https://github.com/INCF/BIDS/pull/17 Please have a look and let me know what do you think.

Here are the major proposed changes:
Inline image 1

Inline image 2


Let me know what do you think.

Best,
Chris

On Thu, Jan 4, 2018 at 1:52 AM, Kirstie Whitaker <kw...@cam.ac.uk> wrote:
Happy new year everyone! I hope you all had a really wonderful holiday with lots of rest and relaxation :)

As this conversation happened mostly over the holidays I wanted to ping it again to see if anyone who missed it had any opinions. Or maybe someone wanted to try to summarise to help us (me!) get a consensus feeling?

Kx

Dianne Patterson

unread,
Feb 15, 2018, 8:32:43 PM2/15/18
to bids-di...@googlegroups.com
That looks potentially valuable. I'm a big fan of improved documentation. 

It would be nice if you could periodically review the documentation for individual apps and make sure they work.  For example, the docker invocation on github for the freesurfer bids app (https://github.com/BIDS-Apps/freesurfer) is wrong because it fails to mention that you have to create a bind for the license.txt (or that there is a license.txt inside the image).  I worry that for a lot of researchers, incorrect documentation is a deal breaker.  

This documentation is wrong:
docker run -ti --rm \
	-v /Users/filo/data/ds005:/bids_dataset:ro \
	-v /Users/filo/outputs:/outputs \
	bids/freesurfer \
	/bids_dataset /outputs participant --participant_label 01 \
	--license_file "license.txt"
Maybe just have a naive user try to follow the instructions and find out where those instructions fail. 

-Dianne

To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/CAAQzouPjvZ8krZcO89k7g9733tS_tycOyyTV5soA3Y0B%3DeQ64Q%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.



--
Dianne Patterson, Ph.D.
Research Scientist
d...@email.arizona.edu
or 
dian...@gmail.com
University of Arizona
Speech and Hearing Science 314
1131 E 2nd Street, Building #71
(Just East of Harvill)
==============
If you don't hear back from me (and you expected to), 
I blame the University's new SPAM filter.
Please write to my gmail account.
==============
Antipiphany: That moment when you realize how little you actually know
==============

Chris Gorgolewski

unread,
Feb 15, 2018, 8:38:54 PM2/15/18
to bids-discussion
Thanks for reporting. I have proposed a fix: https://github.com/BIDS-Apps/freesurfer/pull/40

I think the best place to report issues with individual BIDS apps is in there respective repositories (so for Freesurfer BIDS App this would be https://github.com/BIDS-Apps/freesurfer/issues)

Best,
Chris

On Thu, Feb 15, 2018 at 5:31 PM, Dianne Patterson <d...@email.arizona.edu> wrote:
That looks potentially valuable. I'm a big fan of improved documentation. 

It would be nice if you could periodically review the documentation for individual apps and make sure they work.  For example, the docker invocation on github for the freesurfer bids app (https://github.com/BIDS-Apps/freesurfer) is wrong because it fails to mention that you have to create a bind for the license.txt (or that there is a license.txt inside the image).  I worry that for a lot of researchers, incorrect documentation is a deal breaker.  

This documentation is wrong:
docker run -ti --rm \
	-v /Users/filo/data/ds005:/bids_dataset:ro \
	-v /Users/filo/outputs:/outputs \
	bids/freesurfer \
	/bids_dataset /outputs participant --participant_label 01 \
	--license_file "license.txt"
Maybe just have a naive user try to follow the instructions and find out where those instructions fail. 

-Dianne
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/CAGpV9FKMpLboiaofV1L6qnPZZho_oU%2Bw%2BaPPv%2BzonTdcurdVag%40mail.gmail.com.

Dianne Patterson

unread,
Feb 15, 2018, 9:20:24 PM2/15/18
to bids-di...@googlegroups.com
Thanks for fixing it.  I'll post issues to the github sites in the future.

-Dianne

To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/CAAQzouMDF%2BBfr9oiBM2CBhYMwOsUftQ4VyH%2BvZ8jFioBY7bhkQ%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

Melanie Ganz

unread,
Feb 16, 2018, 3:43:11 AM2/16/18
to bids-di...@googlegroups.com
Hi Chris,

I think this looks good and would really help.
Where do you want us to put full BIDS examples like I am preparing for PET? On openneuro.org or on OpenfMRI? I actually just got permission to put them on OpenfMRI (am waiting for the signed consent forms from earlier participants to come in), but am not sure how easy that is transferable to another site.

Cheers,
Mel
To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussi...@googlegroups.com.
To post to this group, send email to bids-di...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/CAAQzouPjvZ8krZcO89k7g9733tS_tycOyyTV5soA3Y0B%3DeQ64Q%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.


-- 
Melanie Ganz-Benjaminsen, MSc, Ph.D.
Neurobiology Research Unit
University of Copenhagen
Rigshospitalet
Rockefeller Center
Juliane Maries Vej 28/30, 3.
DK-2100 Copenhagen
Denmark

phone:  +45 3545 6718
e-mail: Melani...@nru.dk
web: http://melanie.clausundmelanie.de/

Melanie Ganz

unread,
Feb 16, 2018, 10:40:52 AM2/16/18
to bids-di...@googlegroups.com
Dear BIDS list,

while preparing data sets for the PET BIDS, I now stumbled upon the fact that for use with PET we actually gradient unwarp (correct for gradient non-linearities of) our anatomical MR images. I guess for pure MRI or fMRI studies this doesn't necessarily need to be done and isn't always done, but if you want to overlay a PET image you need to remove those distortions first.

But how can I specify if I am giving the unwarped or warped T1 or T2 image in the "anat" folder? I searched in the main BIDS spec and couldn't really find anything relating to this.

Thanks,

Melanie

Melanie Ganz

unread,
Feb 16, 2018, 10:46:01 AM2/16/18
to bids-di...@googlegroups.com
P.S.: And sorry, of course I can look into the sequence name/descriptions and see if it is called ND or GD (that's a at least standard for the Siemens scanner sequences we use), but is this general enough? Shouldn't there be a much more general flag for this?
--
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 post to this group, send email to bids-di...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Chris Gorgolewski

unread,
Feb 16, 2018, 12:14:24 PM2/16/18
to bids-discussion
Thanks for bringing this up. Couple of suggestions:

1) In terms of file naming I would suggest using rec- keyword to distinguish warped and unwarped images.
2) I think most people save and share unwarped images, because gradient information is proprietary and thus unwarping is impossible without a legal agreement with scanner manufacturers.
3) I don't think we have a field for this - I looked for a DICOM one, but could not find anything. Might be worth thinking about adding something to the BIDS spec,

Best,
Chris

On Fri, Feb 16, 2018 at 7:45 AM, Melanie Ganz <Melani...@nru.dk> wrote:
P.S.: And sorry, of course I can look into the sequence name/descriptions and see if it is called ND or GD (that's a at least standard for the Siemens scanner sequences we use), but is this general enough? Shouldn't there be a much more general flag for this?


On 16-02-2018 4:40, Melanie Ganz wrote:
Dear BIDS list,

while preparing data sets for the PET BIDS, I now stumbled upon the fact that for use with PET we actually gradient unwarp (correct for gradient non-linearities of) our anatomical MR images. I guess for pure MRI or fMRI studies this doesn't necessarily need to be done and isn't always done, but if you want to overlay a PET image you need to remove those distortions first.

But how can I specify if I am giving the unwarped or warped T1 or T2 image in the "anat" folder? I searched in the main BIDS spec and couldn't really find anything relating to this.

Thanks,

Melanie

-- 
Melanie Ganz-Benjaminsen, MSc, Ph.D.
Neurobiology Research Unit
University of Copenhagen
Rigshospitalet
Rockefeller Center
Juliane Maries Vej 28/30, 3.
DK-2100 Copenhagen
Denmark

phone:  +45 3545 6718
e-mail: Melani...@nru.dk
web: http://melanie.clausundmelanie.de/

--
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-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.


-- 
Melanie Ganz-Benjaminsen, MSc, Ph.D.
Neurobiology Research Unit
University of Copenhagen
Rigshospitalet
Rockefeller Center
Juliane Maries Vej 28/30, 3.
DK-2100 Copenhagen
Denmark

phone:  +45 3545 6718
e-mail: Melani...@nru.dk
web: http://melanie.clausundmelanie.de/

--
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-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/6268b6d0-f4b6-e76c-a627-98d5bca9b355%40nru.dk.

Melanie Ganz

unread,
Feb 19, 2018, 3:38:23 AM2/19/18
to bids-di...@googlegroups.com

Hi Chris,

 

1)   Great, then I will propose naming accordingly in the main BIDS. I’ll make a comment in that Google doc.

2)   Regarding sharing unwarped images, this came up because I exactly got an unwarped image from collaborators that were sharing PET data in a trial PET BIDS format with me. And yes, the gradient table and tools are proprietary. Some used to be distributed through e.g. FreeSurfer (https://surfer.nmr.mgh.harvard.edu/fswiki/GradientUnwarping), but they are not freely accessible anymore.

3)   I think it might be a good idea to add a field – it should be easy and is an easy check instead of going into sequence names, etc.

 

Kind regards,

Mel


Best,

Chris

 

To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussi...@googlegroups.com.
To post to this group, send email to bids-di...@googlegroups.com.

 

-- 
Melanie Ganz-Benjaminsen, MSc, Ph.D.
Neurobiology Research Unit
University of Copenhagen
Rigshospitalet
Rockefeller Center
Juliane Maries Vej 28/30, 3.
DK-2100 Copenhagen
Denmark
 
phone:  +45 3545 6718
e-mail: Melani...@nru.dk
web: http://melanie.clausundmelanie.de/

--
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 post to this group, send email to bids-di...@googlegroups.com.

 

--

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 post to this group, send email to bids-di...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/CAAQzouOF3s9SCej7Ep3E445jrWfup4QCBUjGK1QR29G84ozeUQ%40mail.gmail.com.

Chris Gorgolewski

unread,
Feb 19, 2018, 1:41:00 PM2/19/18
to bids-discussion
Hi,
Sorry for the late reply. OpenfMRI has officially been rebranded as OpenNeuro so all submissions need to go through openneuro.org. It would be great to host PET examples there. 

There is one thing you might run into - OpenNeuro uploader performs validation and since the validator does not include PET extension (because it's still under development) you will get errors about unrecognized files. This can be worked around using .bidsignore file.

Best,
Chris


-- 
Melanie Ganz-Benjaminsen, MSc, Ph.D.
Neurobiology Research Unit
University of Copenhagen
Rigshospitalet
Rockefeller Center
Juliane Maries Vej 28/30, 3.
DK-2100 Copenhagen
Denmark

phone:  +45 3545 6718
e-mail: Melani...@nru.dk
web: http://melanie.clausundmelanie.de/

--
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-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.

Melanie Ganz

unread,
Feb 19, 2018, 9:45:21 PM2/19/18
to bids-di...@googlegroups.com
Dear BIDS list,

while preparing datasets for the PET BIDS, I now stumbled upon the fact
that for use with PET we actually gradient unwarp our anatomical MR
images. For pure MRI or fMRI studies this doesn't necessarily need to be
done and isn't always done, but if you want to overlay a PET image you
need to remove those distortions first.

But how can I specify if I am giving the unwarped or warped T1 or T2
image in the "anat" folder? I searched in the main BIDS spec and
couldn't really find anything relating to this.

Thanks,

Melanie

Jolinda Smith

unread,
Feb 20, 2018, 12:48:11 PM2/20/18
to bids-discussion
Here's where to get the information in a Siemens dicom file. For Siemens scanners, this information is included the "Image Type" tag (0008,0008). Non-distortion corrected images will include "ND", and distortion corrected images will include "DIS2D" and/or "DIS3D". If a sequence was run with distortion correction and was set to save both corrected and uncorrected images, the uncorrected series will append "ND" to the series description, but sequence with no distortion correction applied will not do this, so this is not a reliable place to get the information. As for the sequence name, this is more accurately termed the "protocol name" and can be anything at all, and distortion correction can be turned off or on within the protocol regardless of the name, so that's not reliable either.

Siemens seems to have begun including this information some time between VA13 and VA25, so unless you're looking at 15-year-old data, there should always be either "ND" or "DISxD" in the image type field.

Jolinda

Best,
Chris

To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussi...@googlegroups.com.
To post to this group, send email to bids-di...@googlegroups.com.


-- 
Melanie Ganz-Benjaminsen, MSc, Ph.D.
Neurobiology Research Unit
University of Copenhagen
Rigshospitalet
Rockefeller Center
Juliane Maries Vej 28/30, 3.
DK-2100 Copenhagen
Denmark

phone:  +45 3545 6718
e-mail: Melani...@nru.dk
web: http://melanie.clausundmelanie.de/

--
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 post to this group, send email to bids-di...@googlegroups.com.

Chris Gorgolewski

unread,
Feb 20, 2018, 8:25:17 PM2/20/18
to bids-discussion, Jolinda
Thank you for clarification. I did some digging and could not find those codes (ND, DIS3D) in DICOM spec. I think they are Siemens specific.

Melanie suggested adding a new field in BIDS

NonlinearGradientCorrection: Boolean stating if the image saved  has been corrected for gradient nonlinearities by the scanner sequence. 

Based on your findings we could populate it at least for Siemens scanners. What do you think about such solution?


Best,
Chris

To unsubscribe from this group and stop receiving emails from it, send an email to bids-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.
 To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/779ae25d-73a7-4cdf-8999-8d1d7513a052%40googlegroups.com.

Ali Khan

unread,
Feb 20, 2018, 9:30:40 PM2/20/18
to bids-discussion
I think a Boolean would not describe all the situations, Siemens has both 2D and 3D variants, then there could also be data corrected offline (like HCP does). Perhaps a string?

Ali

Chris Gorgolewski

unread,
Feb 20, 2018, 11:26:46 PM2/20/18
to bids-discussion, Jolinda
Good point. Should we adopt the siemens codes? But then we might as well stick to ImageType field as it is.

Do any other scanners represent this information in DICOMs?

Best,
Chris

On Tue, Feb 20, 2018 at 6:30 PM, Ali Khan <al...@robarts.ca> wrote:
I think a Boolean would not describe all the situations, Siemens has both 2D and 3D variants, then there could also be data corrected offline (like HCP does). Perhaps a string?

Ali
--
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-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to bids-discussion@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/bdb829c0-90ff-4479-a91e-b385b3eb0f08%40googlegroups.com.

Melanie Ganz

unread,
Feb 21, 2018, 3:40:58 AM2/21/18
to bids-di...@googlegroups.com, Jolinda

Hi again,

 

I also did some digging for the other scanner types and it seems the ND and DISxD/GD definitions are Siemens specific (thanks Jolinda for pointing out the image type field).

At least our Philips 7T does not even offer to save the uncorrected (ND) image.

In one of the Phillips 7T dicom headers there’s also no nomenclature regarding this. The "ImageType"field is ["ORIGINAL", "PRIMARY", "PROJECTION", "IMAGE", "M", "FFE"].

 

So should we add string values in the form of “correctedGradientNonlinearitiesWithXXX” with details how it was done? Or isn’t that too much? For PET I just need to have an easy way of checking in the json if I got the right scan (the corrected one).

 

Cheers,

Melanie

 

 

Fra: bids-di...@googlegroups.com [mailto:bids-di...@googlegroups.com] På vegne af Chris Gorgolewski
Sendt: 21. februar 2018 05:26
Til: bids-discussion <bids-di...@googlegroups.com>; Jolinda <jol...@uoregon.edu>
Emne: Re: [bids-discussion] Unwarped vs. non-unwarped anatomical MR

 

Good point. Should we adopt the siemens codes? But then we might as well stick to ImageType field as it is.

 

Do any other scanners represent this information in DICOMs?


Best,

Chris

 

On Tue, Feb 20, 2018 at 6:30 PM, Ali Khan <al...@robarts.ca> wrote:

I think a Boolean would not describe all the situations, Siemens has both 2D and 3D variants, then there could also be data corrected offline (like HCP does). Perhaps a string?

Ali

--
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 post to this group, send email to bids-di...@googlegroups.com.


To view this discussion on the web visit

For more options, visit https://groups.google.com/d/optout.

 

--

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 post to this group, send email to bids-di...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bids-discussion/CAAQzouMRA2q6R7J93z1r2L5P9ezogN3iLmSLN%3DSk1KMcSFzKDw%40mail.gmail.com.

Jolinda

unread,
Feb 21, 2018, 12:15:29 PM2/21/18
to bids-di...@googlegroups.com

I think that BIDS datasets should always be saved along with the original DICOMs, with a way to tell which DICOMS were converted to BIDS, so that anything not included in the BIDS specification can be found by interrogating the original DICOMs. Then you don’t have this constant problem of finding out there was some useful information in a dicom file that wasn’t included in the BIDS spec.

 

But failing that I would just include the “Image type” dicom field, in its entirety, in the BIDS spec, because it’s useful in lots of other cases too, and it’s much easier to just include the DICOM field and let the user parse it than to try to keep up with every vendor’s notation for this and that.

 

Jolinda


Best,

Chris

 


For more options, visit https://groups.google.com/d/optout.

 

--
You received this message because you are subscribed to a topic in the Google Groups "bids-discussion" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/bids-discussion/P6bwUza0alw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to bids-discussi...@googlegroups.com.


To post to this group, send email to bids-di...@googlegroups.com.

Harms, Michael

unread,
Feb 21, 2018, 1:24:31 PM2/21/18
to bids-di...@googlegroups.com

 

The contents of the ImageType field are already extracted in their entirety into the json as part of recent versions of ‘dcm2niix’.

 

Many times though, it isn’t feasible, or concerns about “hidden” PHI preclude making the DICOMs concurrently available.

 

Cheers,

-MH

 

-- 

Michael Harms, Ph.D.

-----------------------------------------------------------

Conte Center for the Neuroscience of Mental Disorders

Washington University School of Medicine

Department of Psychiatry, Box 8134

660 South Euclid Ave.                        Tel: 314-747-6173

St. Louis, MO  63110                                          Email: mha...@wustl.edu


For more options, visit https://groups.google.com/d/optout.

 

The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail.

Harms, Michael

unread,
Feb 21, 2018, 1:33:51 PM2/21/18
to bids-di...@googlegroups.com

 

@Chris: If you did add a distinct field to BIDS itself, I think a Boolean is too simplistic.  E.g., if you going to capture the relevant specifics of a nonlinear gradient correction in their entirety, here are the following issues that I can think of:

 

1) Was the correction 2D or 3D?

2) Was the correction done by the scanner, or as part of post-processing?  If the latter, what software (and settings) were used?

3) For dMRI data, is there any expectation that the correction is also applied to the bvals and bvecs?  (And if so, what format do you want to adopt for conveying that information?)

 

I’m sure there are other considerations as well.

 

cheers,

-MH

 

-- 

Michael Harms, Ph.D.

-----------------------------------------------------------

Conte Center for the Neuroscience of Mental Disorders

Washington University School of Medicine

Department of Psychiatry, Box 8134

660 South Euclid Ave.                        Tel: 314-747-6173

St. Louis, MO  63110                                          Email: mha...@wustl.edu

 

From: <bids-di...@googlegroups.com> on behalf of Jolinda Smith <jol...@uoregon.edu>


Reply-To: "bids-di...@googlegroups.com" <bids-di...@googlegroups.com>
Date: Wednesday, February 21, 2018 at 11:16 AM
To: "bids-di...@googlegroups.com" <bids-di...@googlegroups.com>


For more options, visit https://groups.google.com/d/optout.

Granville Matheson

unread,
Feb 22, 2018, 8:11:14 AM2/22/18
to bids-discussion
In my group, we work with data from both a Siemens system and a GE MR system which are coregistered to our PET images.

Siemens
I have been in touch with Siemens in Sweden.  The physicist there sent me the following, corroborating Jolinda's information above, and adding another field where this information is stored:

There are two DICOM tags where you can see if distortion correction has been used. If the tags below says DIS2D (or DIS3D) distortion correction has been used.
 
(0008,0008)

ORIGINAL\PRIMARY\M\NORM\DIS2D
(0051,1016)

p2 M/NORM/DIS2D

 
If you cannot find DIS2D in those tags (or if you see ND) no distortion correction has been used. See below.
 
(0008,0008)

ORIGINAL\PRIMARY\M\ND\NORM
(0051,1016)

M/ND/NORM

He also told me that images could (probably) be unwarped retrospectively using the scanner computer (given the uncorrected DICOM images), or that the coefficients file with the spherical harmonics can just be extracted from the scanner computer (from the following location: C:\MedCom\MriSiteData\GradientCoil\coeff.grad) to do the unwarping afterwards (using various tools e.g. https://surfer.nmr.mgh.harvard.edu/fswiki/GradientUnwarping).  I'm not sure if these tools would actually affect the DICOM tags though: I'm waiting to be sent these files to be able to try this.

GE
I spoke to the operator of our GE system here.  He told me that the camera can output the image with or without distortion correction - there is a checkbox in the interface to decide this.  However, within the ImageType DICOM header (0008,0008), our corrected images just display the following: ORIGINAL\PRIMARY\OTHER.   And we looked through the rest of the header and it doesn't seem to be anywhere.


Is this information of importance for others outside of PET though?  As Melanie says, for PET researchers, we are primarily concerned with either having the correct MR file, or with knowing that we should perform unwarping before coregistration with our PET images.  From my very basic understanding of this issue, the details of which procedure was performed are probably not of great importance for PET researchers specifically due to the spatial resolution of PET, unless something has gone wrong in the unwarping procedure.  Since this information seems not to be reliably extractable from the DICOM headers, it might even mean that it could be hard to determine what the exact correction procedure was even performed on some systems. 

Might it be possible then to include both a Boolean as Chris suggested, and an additional (optional) string field as Melanie suggested, to describe any further more detailed information about what was performed like the details that Michael suggested, in case there is an application for which this information is essential?

All the best,
Granville
Reply all
Reply to author
Forward
0 new messages