BIDS at MRC

To be able to co-register data from different imaging modalities in CIR, share data and analysis pipelines, it is important that data collected are saved in the standard Brain Imaging Data Structure (BIDS).

MRC raw data

Data collected at MRC should be pushed to FOU1/Alvik after each completed session. A unique subject-ID is created for each session and data is stored under that ID and the time of scanning, in folders numbered according to MR-sequence.

It is therefore important to note what ID gets created and assigned at each session.

If a subject was scanned with three sequences at 2024-12-01, 13:45:15 and given the ID: 12345, their raw DICOM data would be located on Alvik in:

/data/dicom/12345_20241201_134515/
├── 00000001
├── 00000002
└── 00000003

There are different tools available to make data collected at MRC compliant with BIDS, these are collected in the following repositories:

MRC-DICOM-to-BIDS

Written to run on the MRC compute server nick-named Alvik. Gather DICOM data collected at MRC and convert to BIDS compliant naming and format. Designed to extract all sequences in the CIR-protocol for up to three sessions per subject and rename subjects to study specific ID.

Require dcm2bids and GNU parallel in your environment.

Note that ASL and SWI sequences are currently not defined in BIDS. See BIDS extension proposal BEP004 & BEP005 for updates. For now, these are converted separately and copied to BIDS-derivatives.