| title | author | display |
|---|---|---|
Steering Group minutes |
true |
Date: Thursday, March 23rd, 2023
| Topic | Relevant Links |
|---|---|
| “Model” / “Resultset” directory proposal | https://github.com/bids-standard/bids-specification/pull/1280 |
| BIDS Derivatives retreat June 21st - 23rd | https://openneuropet.github.io/bidsderivativemeeting/ |
| Summary of the PET-BIDS derivatives meeting | BIDS Extension Proposal 23 (BEP023): PET Preprocessing Derivatives |
| Meeting with the DICOM team | Yaroslav sent a follow up email offering to join DICOM committee meeting |
|
Stefan: Here are three issues that could benefit from steering input (in particular those familiar with ephys experience) |
"Model" / "Resultset" directory proposal derivatives pull request
New proposal of models (e.g. dti models) where derivatives may be created from this model thus creating parameters and derivatives of a model that have overlapping meta-data.
Proposed Solution:
-
add a 'results' directory that contains the files of the model with metadata describing those files. Is it okay to add an additional nested directory like this?
-
MFP (model fit parameter) and MDP (model derived parameter) suffixes
Concern:
- this appears to be very dwi specific, should the suffixes be dwi specific?
- Why is datatype not under model label in the pull request example?
- Is the proposed organization of data aligned with user community? - should we request input from community to determine where these results should be stored
- bids-standard/bids-specification#751 (comment)
- How does this apply to more general data modalities? Like mEEG, SPECT - more use cases beyond dMRI would be helpful
- "BIDS-like derivatives in other modalities: https://data.donders.ru.nl/collections/di/dccn/DSC_3031000.00_191?0, see also https://github.com/Donders-Institute/infant-cluster-effectsize (for the scripts that generate the result)
- Other ideas: tarballs instead of directories, HDF5 files instead of directories.
BIDS Derivatives Retreat
- For general derivatives group (not just pet)
- See link for more information
Summary of the PET-BIDS derivatives meeting
- Meeting happened 2 weeks ago (March 10th)
- Mostly an onboarding meeting for the importance and relevance of PET-BIDS derivatives
- Raw PET data is now part of BIDS, now the second half of the white paper will address the derivatives aspect of PET
- The PET users are more clinically oriented, more diverse user groups, less agreement within the community in data standards
- The community is between two ideas: a more restrictive approach with data types allowed, or free datatypes with detailed provenance
Meeting with the DICOM team
- We need to formalize our current state of affairs, see which terms align/ dont align
- We need to follow up on email from Kim on Feb 14 asking to present at a dicom committee meeting.