we must create open datasets
Raamana, Pradeep Reddy
As I discuss niQC with various folks, a common, recurring, and urgent need connecting various QC tasks (auto QC, phantoms etc) is the need for open datasets sharing the relevant imaging and ratings data, so methods researchers (esp. those not at imaging centers) can target their brilliance towards. Hence, I encourage you all to encourage yourself and peers to share and publish any and all data that can be used to improve niQC.
Most of you already know datasets can be published in a few places now, but I would encourage everyone to mention and prefer the OHBM Aperture journal, for this purpose.
This is a necessary first step towards e.g. creating a central repository that Yarik mentioned.
Thanks,
Pradeep
JB Poline
--
You received this message because you are subscribed to the Google Groups "niQC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to niqc+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/niqc/DM6PR04MB67461F398CD8DCE1D5D63044B8E39%40DM6PR04MB6746.namprd04.prod.outlook.com.
Lei Wang
Agreed.
Pradeep – when is the Phantom meeting?
--Lei
--
Lei Wang, Ph.D., Lei....@osumc.edu
Psychiatry and Behavioral Health, Neuroscience
Ohio State University Wexner Medical Center
Psychiatry and Behavioral Sciences
Northwestern University Feinberg School of Medicine
To view this discussion on the web visit https://groups.google.com/d/msgid/niqc/CALMaa83MoCqi7pNF1Dh25cd5%3DrGQzJ-3MFt28wuEPGyr2Ddb7Q%40mail.gmail.com.
Raamana, Pradeep Reddy
Thanks Lei – the phantom meeting is next Tuesday 4-6pm EST (or longer for some if needed). I encourage everyone to subscribe to the niQC SIG calendar: https://calendar.google.com/calendar/u/0?cid=cHJyYjg0a2FqZDIxYmlla21tN3FmYzNpZmtAZ3JvdXAuY2FsZW5kYXIuZ29vZ2xlLmNvbQ
Raamana, Pradeep Reddy
Thanks JB! There is still one significant barrier towards the creation of open QC datasets (besides the effort to QC them before sharing etc): Data Usage Agreements (for ADNI, ABCD, etc) with a “Do Not Redistribute” (DNR) clause preventing us from openly sharing modified imaging data or associated QC ratings. Some folks are working around it (e.g. ABCD effort at some crowdsourcing by making the interested parties collaborators at UMN etc), but this is an unnecessary pain that is not justified. I haven’t seen a solid rationale for this beyond vague concerns for privacy etc, that can easily be addressed without carpet bombing our ability to redistribute privacy-safe versions of the datasets or their derivatives.
I did reach out to some NIH Data Sharing officers, but they didn’t really help (more like they shut me down). But with some help from here and right contacts, I would like discuss ways to remove or significantly modify the DNR clause for big datasets created with public funding.
PS: if needed, I am happy to organize a debate or brainstorming session for this – let me know if any of you are interested.
From:
JB Poline <jbpo...@gmail.com>
Date: Wednesday, July 21, 2021 at 12:33 PM
To: Raamana, Pradeep Reddy <RAAM...@pitt.edu>
Cc: niQC <ni...@googlegroups.com>
Subject: Re: [niQC] we must create open datasets
Cyril Pernet
following up on the last few weeks discussion and Yarik provocative results , Patrick Fisher (who joined the google group) and myself have decided to implement a weekly QC for fMRI for our 2 machines (a Prisma and a PET/MR). Currently the clinical people do a standard T1/T2 to check for obvious artefacts but that's about it - so we'll pump this up
Here is our plan, but I'd like the input from the group since we will also share those data and this can be used for some work we want to do
- 1/2 hour max including set-up
- standard phantom (need to find out exactly what we have - I'm new to the lab in Copenhagen ...) but we wouldn't buy anything new
- scan once a week but sample pseudorandomly (ie not just all Monday AM, or all Friday PM)
- T1/T2 with the same parameters as what is typically done here for our research protocols
- T2*EPI for BOLD; here I was thinking of 2 runs, one matching the 'standard' fmri done here and one with all gizmos on at maximum (in plane (grappa) and in volume (sms) accelerations with the fastest TR possible) ; how long those should be? longer better I guess, temporal SNR variance could matter for dynamic FC? no idea
- should we do a DWI? - do we actually expect changes here not captured otherwise
Maybe others here would want to set-up an equivalent / same QC so we would have the 'same' data from different places
Cheers
Patrick and Cyril
Raamana, Pradeep Reddy
Thank you Cyril – could you elaborate more on if Patrick and you have any specific goals with this dataset? Replicate Yarik’s results?
Great idea to invite other imaging centers to join you as well, love it. I am checking if I can do it here at Pitt.
From:
ni...@googlegroups.com <ni...@googlegroups.com> on behalf of Cyril Pernet <wamc...@gmail.com>
Date: Wednesday, August 18, 2021 at 10:37 AM
To: niQC <ni...@googlegroups.com>
Subject: Re: [niQC] we must create open datasets
To view this discussion on the web visit https://groups.google.com/d/msgid/niqc/20c47721-2801-443a-bc39-c67c6014ae7en%40googlegroups.com.
Yaroslav Halchenko
On Wed, 18 Aug 2021, Cyril Pernet wrote:
> - should we do a DWI? - do we actually expect changes here not captured
> otherwise
me. Asked on specific purpose. Sample(s) can be found on
http://datasets.datalad.org/?dir=/dbic/QA/sub-qa/ses-20191202/dwi
BTW -- a new feature of datasets.datalad.org -- for files we host (like in this
case) just hover on the file, get ⠇ and click on it to get context menu, then
you can open those .nii.gz a viewer e.g.
https://bioimagesuiteweb.github.io/webapp/viewer.html?image=https://datasets.datalad.org/dbic/QA/sub-qa/ses-20191202/dwi/sub-qa_ses-20191202_acq-DTIX30Xp2_dwi.nii.gz
for such a DWI. It is purely browser-client-side so file will be first
downloaded, but then you could even get through dynamics/volumes.
--
Yaroslav O. Halchenko
Center for Open Neuroscience http://centerforopenneuroscience.org
Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
WWW: http://www.linkedin.com/in/yarik
Dr Cyril Pernet
Sent from my Huawei phone
Mary Beth Nebel
To view this discussion on the web visit https://groups.google.com/d/msgid/niqc/DM6PR04MB67462FC9B261785DD4564E61B8E59%40DM6PR04MB6746.namprd04.prod.outlook.com.
Lei Wang
If folks want, I can help find out. We are working with the ABCD folks to set up our own multisite project.
--Lei
--
Lei Wang, Ph.D., Lei....@osumc.edu
Psychiatry and Behavioral Health, Neuroscience
Ohio State University Wexner Medical Center
Psychiatry and Behavioral Sciences
Northwestern University Feinberg School of Medicine
To view this discussion on the web visit https://groups.google.com/d/msgid/niqc/CANAA4ruvmuR95N%2BPjW2e6cafRfd%3DeE7UU2d7y6Bq1rryS6ARow%40mail.gmail.com.
Raamana, Pradeep Reddy
That would be great, Lei. Thanks.
Wang, Lei
A few updates from ABCD:
- They are open to sharing non-human phantoms in principle, and will look into sharing them through the NDA (where the ABCD data are being shared). So more on that as it evolves.
- Anders Dale mentioned that Joshua Kuperman leads an ABCD regular meeting on phantom data. They suggested that someone from the niQC group makes a guest appearance to connect our efforts. We can learn about their processing and QA/QC metrics. I will find out about the dates and times and connect with whomever is interested or nominated.
--Lei
--
Lei Wang, Ph.D., Lei....@osumc.edu
Psychiatry and Behavioral Health, Neuroscience
Ohio State University Wexner Medical Center
Psychiatry and Behavioral Sciences
Northwestern University Feinberg School of Medicine
To view this discussion on the web visit https://groups.google.com/d/msgid/niqc/DM6PR04MB67464A45F4B5700200910DCBB8FF9%40DM6PR04MB6746.namprd04.prod.outlook.com.
Raamana, Pradeep Reddy
Sent: Wednesday, August 18, 2021 6:44:09 PM
To: Raamana, Pradeep Reddy <RAAM...@pitt.edu>; Mary Beth Nebel <mb.n...@gmail.com>
Cc: jbpo...@gmail.com <jbpo...@gmail.com>; niQC <ni...@googlegroups.com>
Raamana, Pradeep Reddy
Sent: Wednesday, August 18, 2021 7:11:18 PM
To: Wang, Lei <Lei....@osumc.edu>; Mary Beth Nebel <mb.n...@gmail.com>
Mary Beth Nebel
Yaroslav Halchenko
listener-mode to not derail the meeting ;) ) if there would be no
time overlaps with other commitments.
On Wed, 18 Aug 2021, Mary Beth Nebel wrote:
> I have been involved in efforts to implement ABCD-like sequences on the
> Philips scanners at our institute. As part of that effort, we've been
> discussing the attached figure on cross-talk saturation in SMS imaging. I
> think the ABCD fMRI sequence uses what the figure indicates is
> a suboptimal number of slice stacks (60 total slices/MB 6 = 10 stacks), so
> I'm curious if they see any evidence of cross-talk saturation in their
> phantom data.
it for such or similar "effects" (which might even not yet be
known/identified) if phantom data was available.
Raamana, Pradeep Reddy
Sent: Thursday, August 19, 2021 12:58:42 PM
To: niQC <ni...@googlegroups.com>
Subject: Re: [niQC] we must create open datasets
listener-mode to not derail the meeting ;) ) if there would be no
time overlaps with other commitments.
On Wed, 18 Aug 2021, Mary Beth Nebel wrote:
> I have been involved in efforts to implement ABCD-like sequences on the
> Philips scanners at our institute. As part of that effort, we've been
> discussing the attached figure on cross-talk saturation in SMS imaging. I
> think the ABCD fMRI sequence uses what the figure indicates is
> a suboptimal number of slice stacks (60 total slices/MB 6 = 10 stacks), so
> I'm curious if they see any evidence of cross-talk saturation in their
> phantom data.
indeed -- and should be possible to analyze to find support or lack of
it for such or similar "effects" (which might even not yet be
known/identified) if phantom data was available.
--
Yaroslav O. Halchenko
Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
--
You received this message because you are subscribed to the Google Groups "niQC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to niqc+uns...@googlegroups.com.
Dr Cyril, Pernet
Hi all,
update - after discussion with Yarik we'll include a short DWI, to compute simple FA and ADC maps
--> still welcoming commenst on how long we should scan for
fMRI / tSNR ?? @tom et al
+ we may as well prospectively register to replicate https://f1000research.com/articles/9-1131
thx
Cyril
- simple water phantom 1/2 hour scanning total
- T1/T2 with the same parameters as what is typically done here
for our research protocols
- T2*EPI for BOLD; here I was thinking of 2 runs, one matching the
'standard' fmri done here and one with all gizmos on at maximum
(in plane (grappa) and in volume (sms) accelerations with the
fastest TR possible) ; how long those should be? longer
better I guess, temporal SNR variance could matter for dynamic FC?
no idea
- DWI, at least simple 6 directions or more depending on time given the fMRI scans
-- Dr Cyril Pernet, PhD, OHBM fellow, SSI fellow Neurobiology Research Unit, Building 8057, Blegdamsvej 9 Copenhagen University Hospital, Rigshospitalet DK-2100 Copenhagen, Denmark wamc...@gmail.com https://cpernet.github.io/ https://orcid.org/0000-0003-4010-4632
Raamana, Pradeep Reddy
Sent: Friday, August 20, 2021 2:25 AM
To: Raamana, Pradeep Reddy; niQC; Patrick Fisher
Dr Cyril, Pernet
Cool! How long are you planing to acquire this dataset? 3, 6 or 12 months?
forever - QC must be taken seriously and we scan al the time ...
of course if we want to preregister to reproduce Yarik's paper, I
suggest 24 months to have 2 season cycles ? and work on the
sharing, BIDS, etc in between
anyone else want to scan water like us? two labs independent but
coordinated replications would be really good
cyril
Raamana, Pradeep Reddy
I meant for the purpose of the initial analyses and paper! We have a related important talk next week from the Gold Standard Phantoms and Phantom Lab companies scheduled on Aug 24th. I believe they will touch upon their work with few big datasets like ADNI, ABCD etc, and that might be a good time to discuss and finalize this.
Aaron, Richard, (if you are not on vacation this week), any comments? 😊
Raamana, Pradeep Reddy
Sent: Friday, August 20, 2021 8:30:24 AM
To: Dr Cyril, Pernet <wamc...@gmail.com>; niQC <ni...@googlegroups.com>; Patrick Fisher <patrick...@nru.dk>
Dr Cyril, Pernet
Also, do you expect to be able to share this dataset publicly, Cyril?
Yaroslav Halchenko
On Fri, 20 Aug 2021, Raamana, Pradeep Reddy wrote:
> Cool! How long are you planing to acquire this dataset? 3, 6 or 12 months?
> Yarik, do we need any other specific hypotheses / experiments on this
> about?
I guess there are two major aspects to reproduce in the original paper
(well -- one would be enough to start with I guess)
- ability to describe variance within phantom SNR based on scanning
intrinsic properties (SAR, IOPD -- position, etc) as in the paper
- note: for this you do not even need any new data, could just take
old phantom data and see what you get
- if human data available -- ability to describe some variance in
morphometry results; but could be extended to any other measure
Thank you Cyril in advance!
Cheers,
--
Yaroslav O. Halchenko
Center for Open Neuroscience http://centerforopenneuroscience.org
Lei Wang
Maybe we should start with asking Joshua Kuperman to give a talk. Do you have someone that helps with scheduling talks that I can connect them?
Also, once they let me know when their regular meetings are, we can nominate someone to make a guest appearance at their meeting.
--Lei
--
Lei Wang, Ph.D., Lei....@osumc.edu
Psychiatry and Behavioral Health, Neuroscience
Ohio State University Wexner Medical Center
Psychiatry and Behavioral Sciences
Northwestern University Feinberg School of Medicine
From: ni...@googlegroups.com <ni...@googlegroups.com>
On Behalf Of Raamana, Pradeep Reddy
To view this discussion on the web visit https://groups.google.com/d/msgid/niqc/DM6PR04MB67469FA33FF2BDBBAF0FCE33B8C09%40DM6PR04MB6746.namprd04.prod.outlook.com.
Raamana, Pradeep Reddy
That sounds good Lei - we have Malin Sandstrom from INCF helping but she is on vacation right now, so I’ve been coordinating them so far, and happy to take care of this.
Aaron Oliver-Taylor
Antoine Lutti
>> --> still welcoming commenst on how long we should scan for fMRI / tSNR ?? @tom et al
To view this discussion on the web visit https://groups.google.com/d/msgid/niqc/06fe14e9-d5f8-a87b-09a0-4016eb1500a1%40gmail.com.
Raamana, Pradeep Reddy
Thanks Aaron – I am quite looking fwd to tomorrow’s session with talks from you and Richard, and the subsequent discussion. As a reminder to everyone, it’s at 11am EDT Tuesday at this zoom room: https://pitt.zoom.us/my/raamana
Registered in England and Wales.
Company number: 9342804
VAT Reg: GB211807049
This email and any attachments to it may be confidential and are intended solely for the use of the individual to whom it is addressed. Any views or opinions expressed are solely those of the author and do not necessarily represent those of Gold Standard Phantoms.
If you are not the intended recipient of this email, you must neither take any action based upon its contents, nor copy or show it to anyone.
Please contact the sender if you believe you have received this email in error.
--
You received this message because you are subscribed to the Google Groups "niQC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
niqc+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/niqc/38550dd3-f91c-4797-867b-c0bf6f174979n%40googlegroups.com.
Raamana, Pradeep Reddy
Gentle reminder everyone for the talk happening NOW – join us : https://pitt.zoom.us/my/raamana
To view this discussion on the web visit https://groups.google.com/d/msgid/niqc/DM6PR04MB674632E733FDBE31FE5500CDB8C49%40DM6PR04MB6746.namprd04.prod.outlook.com.
