Magnetic resonance imaging datasets with anatomical fiducials for quality control and registration

Taha, A. et al. Magnetic resonance imaging datasets with anatomical fiducials for quality control and registration. bioRxiv 2022.11.21.516173 (2022)

In this GitHub repo, you can find:

1. data: Markups Comma-separated values file (i.e., *.fcsv extension) files for both templates and datasets.
2. notebooks: Jupyter notebooks involved in creation of the JavaScript Object Notation (i.e., *.json) which is an adaptation from BIDS Appendix VIII on Coordinate systems.
3. other: Information about rater experience, afids protocol, and interactive glass brain

Compressed NIfTI-1 images for datasets involved in this data release are available for download:

1. 100 Unrelated Humman Connectome Project (AFIDs-HCP dataset; n = 30): https://github.com/afids/AFIDs-HCP
2. Open Access Series of Imaging Studies (AFIDs-OASIS dataset; n = 30): https://github.com/afids/AFIDs-OASIS
3. London Health Sciences Center Parkinson's Disease Dataset (LHSCPD ; n = 40): https://openneuro.org/datasets/ds004471
4. Stereotactic Neurosurgery 7-Tesla Control Dataset (SNSX; n = 32): https://openneuro.org/datasets/ds004470

*Please review and agree to the Data Usage Terms, outlined in links to the HCP and OASIS datasets, before downloading imaging data.

Accepting Data Usage Agreements (DUA):

For information about specific dataset DUA, please consult the individual directories linked above. To download HCP and OASIS data, the user must accept their specific DUAs. As for our local datasets (i.e., LHSCPD and SNSX), they are released under Attribution 4.0 International (CC BY 4.0) license and no further user intervention is needed before following instructions for download below. Users can either download those datasets directly from links above, or by cloning the super dataset here via instructions below.

Downloading the afids-data super dataset:

1. Visit the individual datasets linked above to appreciate licensing, accepting DUA, and acquiring any required user credentials

2. Install datalad: https://www.datalad.org/#install

3. Clone our super dataset (afids-data) on your machine: datalad install -r https://github.com/afids/afids-data.git run this where you want the data to live

4. Navigate to the cloned directory: cd afids-data

5. Retrieve file content: datalad get -r . run this to install all the data as part of our super dataset (. denotes “current directory” and the -r flag is used because afids-data is comprised of multiple DataLad datasets). You can also datalad get a full dataset or precise subset of to-be-retrieved files. To do this, you can supply paths or globbing expressions (such asdata/datasets/LHSCPD or sub-*/anat/* instead of .). Users will be asked to input username credentials or indicate they have accepted the DUA before this stage runs.

Feel free to consult the DataLad handbook for more information about working with DataLad: https://handbook.datalad.org/en/latest/

An important note on data licensing:

We release our anatomical landmark annotations under the Attribution 4.0 International (CC BY 4.0) license, available in DERIVATIVE_DATA_USE_AGREEMENT.txt at the dataset level. However, the HCP and OASIS-1 imaging data are protected by a data usage agreement (DUA). If users wish to access this imaging data, they must accept DUAs which we make availible in the IMAGING_DATA_USE_AGREEMENT.md file by following specific instructions at the level of each dataset.