Licensing & legal

This package is a convenience tool for accessing publicly available data. It does not host, relicense, or redistribute any dataset. When you use it to download data, you are downloading directly from the provider, and you are responsible for reviewing and complying with that provider's terms before using, redistributing, or publishing results derived from the data.

mridata.org

Datasets on mridata.org are contributed by many groups and each dataset may carry its own license and citation/attribution requirements. Check the dataset's page on mridata.org for its specific terms and the publication it should be attributed to. Do not assume a uniform license across datasets.

OCMR

The OCMR cardiac k-space repository is distributed under its own data-use terms. In particular, use of OCMR data requires citing the OCMR publication:

Chen C, Liu Y, Schniter P, Tong M, Zareba K, Simonetti O, Potter L, Ahmad R. OCMR (v1.0)—Open-Access Multi-Coil k-Space Dataset for Cardiovascular Magnetic Resonance Imaging. arXiv:2008.03410, 2020.

Review the terms on the OCMR website and in the dataset repository, and comply with any restrictions on redistribution and use.

CMRxRecon2024

The CMRxRecon2024 challenge dataset is hosted on Synapse. Access is gated:

1. Register for a free Synapse account.
2. Apply to join the challenge and complete the external team-information form — this is mandatory. Until the challenge registration is finalized, a generated Personal Access Token (PAT) will not have permission to download the data.
3. Create a Synapse PAT with view and download scopes and provide it to this package via set_synapse_token! (or the SYNAPSE_AUTH_TOKEN environment variable).

See the Task 2 page for technical context on the random-sampling CMR reconstruction task, and the FAQ for terms and citation requirements. Cite the CMRxRecon publications as required by the organizers.

CMRxRecon-300

The CMRxRecon-300 dataset (the revised CMRxRecon-2023 dataset) is hosted on Synapse under the CC-BY license. Unlike the 2024 challenge data, access is not gated by a challenge application — a free Synapse account suffices:

1. Register for a free Synapse account.
2. Create a Synapse PAT with view and download scopes and provide it to this package via set_synapse_token! (or the SYNAPSE_AUTH_TOKEN environment variable).

CC-BY permits redistribution and reuse with attribution: cite the CMRxRecon Scientific Data paper (Wang C, Lyu J, Wang S, et al. CMRxRecon: A publicly available k-space dataset and benchmark to advance deep learning for cardiac MRI. Scientific Data, 2024, 11(1): 687) when you use the data.

USC Speech (SPAN 75-speaker)

The USC SPAN 75-speaker real-time speech production MRI dataset is hosted on figshare (13725546) under the Creative Commons Attribution 4.0 (CC BY 4.0) license. No account or registration is required to download it.

CC BY 4.0 permits redistribution and reuse — including commercially — provided you give appropriate credit. Cite the Scientific Data data descriptor when you use the data:

Lim Y, Toutios A, Bliesener Y, et al. A multispeaker dataset of raw and reconstructed speech production real-time MRI video and 3D volumetric images. Scientific Data, 2021, 8(1): 187.

and acknowledge the figshare dataset DOI (10.6084/m9.figshare.13725546).

M4Raw

The M4Raw low-field brain MRI dataset is hosted on Zenodo (record 8056074) under the Creative Commons Attribution 4.0 (CC BY 4.0) license. No account or registration is required to download it.

CC BY 4.0 permits redistribution and reuse — including commercially — provided you give appropriate credit. Cite the Scientific Data data descriptor when you use the data:

Lyu M, Mei L, Huang S, et al. M4Raw: A multi-contrast, multi-repetition, multi-channel MRI k-space dataset for low-field MRI research. Scientific Data, 2023, 10(1): 264.

and acknowledge the Zenodo record DOI (10.5281/zenodo.8056074).

fastMRI

The fastMRI dataset is a collaborative research project from NYU and Facebook AI Research (FAIR), providing knee, brain, prostate, and breast MRI k-space data. Access requires completing the fastMRI Dataset Agreement form at fastmri.med.nyu.edu — submission is required before any download. The confirmation email contains time-limited (90-day) pre-signed AWS S3 URLs for the dataset archives.

Key obligations:

• You must complete the Dataset Agreement form for each dataset (knee, brain, prostate, breast) you wish to use. Approval is typically automatic but requires an institutional affiliation.

• Re-request download links after 90 days. Supply new links to MRITestData with set_fastmri_urls!.

• Cite the relevant fastMRI publications for the anatomy or task you use:

  • fastMRI benchmark dataset (knee + brain):

    Zbontar J, Knoll F, Sriram A, et al. fastMRI: An Open Dataset and Benchmarks for Accelerated MRI. arXiv:1811.08839, 2018.

  • fastMRI+ annotations (knee):

    Zhao R, Yaman B, Zhang Y, et al. fastMRI+: Clinical Pathology Annotations for Knee and Brain Fully Sampled Multi-Coil MRI Data. Scientific Data, 2022, 9: 717.

  • fastMRI prostate:

    Tibrewala R, Tamir JI, Sriram A, et al. fastMRI Prostate: A Publicly Available, Multimodal Dataset for Prostate MRI. arXiv:2304.09254, 2023.

  • fastMRI breast:

    Steinhardt J, Fabian MS, Liu S, et al. The fastMRI Breast Dataset. ISMRM, 2024. (Check fastmri.med.nyu.edu for the current citation.)

Compliance with the Dataset Agreement is your responsibility. The MRITestData.jl maintainers cannot grant rights to the fastMRI data.

Your responsibilities

• Verify the license/terms of each dataset you download.
• Provide the attribution/citation each provider requires in any publication or derived work.
• Comply with restrictions on redistribution — do not re-host data unless the provider's license permits it.
• Note that some datasets may be for non-commercial or research use only.

If you are unsure about the terms for a particular dataset, consult the provider directly. The maintainers of MRITestData.jl cannot grant rights to third-party data.