How to share your data responsibly
The Open Data Working Group has built a searchable, user friendly XNAT database to store MRI, EEG and MEG scan data directly from the scanners. The database also has the capability to store other research data alongside the scans to create a research dataset. Image conversion tools are be integrated into the database to convert raw image files to standard formats and the community standard Brain Imaging Data Structure (BIDS) file structures.
Data will only ever be shared when the participant has given explicit consent to open sharing. All access protocols have been developed to ensure the highest levels of security to protect against accidental or malicious data breaches.
The database has the capability to share data at between specified individuals, openly to all WIN members, or externally based on the requirements of the research lead.
Before data is shared externally, it will be checked against a list of defined criteria to ensure the data are appropriately de-identified and any risk of identification of individual participants is negligible. The criteria for de-identification will include removal of identifying facial features (defacing), the removal of personal data from raw dicom images, and the removal of any linkage with consent or experimental participant identification numbers.
Individuals accessing shared data will additionally be required to agree to a Data Usage Policy, where they explicitly confirm that they will not attempt to re-identify participants, nor share the data with any third party who has not signed the same agreement.
The Data Usage Agreement and de-identification process is being developed with the support of Departmental and University level Information Security and Compliance teams.
WIN members will also be encouraged to run and share the results of predefined quality control algorithms, so anyone accessing the data can have a ready measure of image quality.
THIS TOOL IS CURRENTLY IN DEVELOPMENT. PLEASE REFER TO THE INFORMATION BELOW TO UNDERSTAND THE AIM AND AMBITION OF THIS PROJECT. THE “HOW TO” GUIDE WILL BE BUILT BY THE COMMUNITY AND TOOL DEVELOPERS IN THE COMING MONTHS.
Note while the XNAT system is in development, we suggest you use the below wording in your data availability statement in journal submissions:
“Data will be made available on the WIN Open Data server. This is currently in development. Register here to find out when materials are available for download: https://web.maillist.ox.ac.uk/ox/subscribe/win-open-data”
WIN XNAT (accessible from the university network or VPN) is currently being built and will be used to share imaging data internally within WIN.
You can log into WIN XNAT using your WIN IT account (currently requires a connection to the university network or VPN).
The XNAT website has useful background information about the XNAT platform.
For the current overview of how BIDS works in XNAT, see the BIDS page.
There are several python libraries](data/python.md) that can be used to write scripts against the XNAT API. See python libraries for more info on pyxnat, xnatpy and dax.
To see how Docker works in XNAT, see the Docker page.
The OPDC project uploaded DICOM data from jalapeno using python and dcmtk.
TODO: More case studies!
External users will be able to search the database for resources which individual teams have chosen to make openly available. These may be deposited to support publications as supplementary methods material, or they may form the main body of research in data papers.
Sign up to the win-open-data mailing list to receive updates on new data shared to the WIN XNAT server or new features. This mailing list will be “broadcast only” with messages sent from WIN XNAT administrators. You can unsubscribe at any point. Enquiries about external access to WIN XNAT should be directed to email@example.com.
We are grateful to the following WIN members for their contributions to developing the Open Data server.