Michał Szczepanik
95e810d24b
The BIDS conversion page contained a mix of general statements and fine details (reflecting open discussions at the time of writing). Now, considering the "less is more" principle, the introduction is kept, but the other sections are changed to communicate the following: - selected tooling (heudiconv / dcm2niix), what id does - shared resources (public repos, reused across sites) - DataLad layout of a BIDS dataset (what subdatasets and why)
2.9 KiB
| title |
|---|
| BIDS conversion |
Converting the heterogeneous, site-specific raw MRI data acquisitions into a standardized dataset is an essential precondition for the collaborative work in TRR379. It readies the data for processing with established pipelines, and applies a pseudonymization as a safeguard for responsible use of this personal data.
TRR379 uses the Brain Imaging Data Structure (BIDS) as the standard for its datasets.
Conversion to BIDS
The conversion of raw MRI data in DICOM format to a BIDS-compliant dataset is a largely automated process. The recommended software to be used for conversion is heudiconv. Heudiconv uses dcm2niix as the actual DICOM→NIfTI converter. Heudiconv tutorials further illustrate how the software works.
Heudiconv performs the task of mapping DICOM series to BIDS entities (ie. determining BIDS-compliant file names). A key heudiconv concept is a heuristic: a Python program (function) which receives the DICOM series properties and matches them with a file naming pattern. A heuristic typically relies on DICOM series naming (set at the scanner console), but it can also use other properties such as number of images or acquisition parameters.
Heudiconv workflows have been implemented at each MRI acquisition site, and reuse shared components.
Shared Resources
- q02/heudiconv-container is a DataLad dataset providing a heudiconv container.
- q02/rdmtools contains Python scripts which further automate BIDS conversion, pushing created datasets to dedicated locations, and other data curation tasks.
- q01/phantom-mri-dicoms is a superdataset grouping phantom MRI scans from all sites, used to develop and test the conversion workflow; it also contains the reference TRR379 heudiconv heuristic.
Important
Each site maintains its own copy of the heuristic (as part of the site-specific DICOM superdataset) to account for site-specific differences.
Details
Modular datasets
A DataLad dataset containing a BIDS dataset will typically have these subdatasets. Subdatasets help in recording provenance while enabling different access scopes for different components:
code/heudiconv: heudiconv container (q02/heudiconv-container)code/rdmtools: optional, shared scripts for data curation tasks (q02/rdmtools)sourcedata/dicoms: a site-specific DICOM dataset, itself containing one subdataset per session.heudiconv: heudiconv cache / intermediate outputs are stored in a separate dataset to avoid including DICOM information in the BIDS dataset