Group-level analysis in BIDS

[Mastaneh Torkamani]

Hi all,

I could not find the answer to this question on https://bids-specification.readthedocs.io/en/stable, so here is a question to the forum: 

Are there any suggestions for where to store the results of group-level analyses, say, for MRI or electrophysiological recordings- such as MEA? Shall they be added to derivatives or to a dedicated directory at the same level as derivatives? 

Thank you in advance for any suggestions,

Mastaneh 

mastaneh....@uef.fi 

[Remi Gau]

Maybe I am wrong but I do not think that anything in the BIDS specification addresses this... yet.

I did a quick check in some of the extensions for BIDS (work in progress) and BIDS apps to see if they were hints of this.

BEP17 (connectivity) mentions a group level folder

<dataset>/derivatives/<pipeline_name>/

   group/<modality>/

     <source-file>_[meas-<label>_desc-<label>_]NodeIndicesFile.tsv

<source-file>_[meas-<label>_desc-<label>_]NodeIndicesFile.json

<source-file>_[meas-<label>_desc-<label>_]relmat.[sparse.|dense.]<tsv,h5,zar>

<source-file>_[meas-<label>_desc-<label>_]relmat.json

BEP41 (stats output) also mentions a group level

derivatives/

univariate_glm/

group/

design.tsv

task-nback_contrast-2backMinus0back_stat-effect_statmap.nii.gz

task-nback_contrast-2backMinus0back_stat-effect_statmap.json

task-nback_contrast-2backMinus1back_stat-variance_statmap.nii.gz

task-nback_contrast-2backMinus1back_stat-variance_statmap.json

task-nback_contrast-2backMinus1back_stat-z_statmap.nii.gz

task-nback_contrast-2backMinus1back_stat-z_statmap.json

In terms of BIDS apps and what they do:

Fitlins I think save each level of the analysis as different derivatives datasets: https://fitlins.readthedocs.io/en/latest/

For Bidspm it saves the group level in a different derivative dataset: https://bidspm.readthedocs.io/en/latest/output_files.html#statistics

outputs/derivatives/bidspm-stats
├── CHANGES
├── dataset_description.json
├── derivatives
│   └── bidspm-groupStats
│       ├── CHANGES
│       ├── dataset_description.json
│       ├── jobs
│       ├── sub-ALL_task-<taskLabel>_space-<spaceLabel>_FWHM-6_node-dataset_contrast-<contrastLabel>
│       │   ├── beta_0001.nii
│       │   ├── con_0001.nii

Hope this helps.

Rémi

--
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/292f7f8b-08d1-4cd2-b9e2-981835e774b0n%40googlegroups.com.

--

Rémi Gau

[Mastaneh Torkamani]

Hi Remi,

Thank you! That was really helpful. I had not checked the extension proposals and they indeed provide helpful insights. 

We are collecting a large set of in vivo and ex vivo modalities, and currently we have the following structure within derivatives: 

 

derivatives/ sub-01/

            ses-exvivoMEA/

            ses-exvivoMRI/

            ses-invivoMRI/

...

Now I am thinking if we should re-order these and create separate directories for each modality to store the subject-level and group-level results. 

Best,

Mastaneh 

[Dr Cyril Pernet]

Dear Mastaneh

results, being subject or group level can go into your derivatives folder - possibly nesting another 'derivatives' for your group

- sub-01

- sub-02

-derivatives/sub-01

-derivatives/sub-02

-derivatives/derivatives

you can find some explanations here https://bids-specification.readthedocs.io/en/stable/derivatives/introduction.html for the principles involved, but otherwise these are in development in different BEPs -- for sure there is some progress on that for PET and electrophys, I am not sure there is much going on for fMRI

Cyril

--
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/292f7f8b-08d1-4cd2-b9e2-981835e774b0n%40googlegroups.com.

-- 
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