FBIRN Best Practices for Multi-site fMRI studies

FBIRN Best Practices for Multi-site fMRI studies

In which we delineate as best we can the accumulated wisdom of having performed both traveling subjects and multi-site clinical studies; and having developed tools for uploading, storing, curating and maintaining fMRI data on the SRB. (For more general information on BIRN, please see http://www.nbirn.net .)

This is a work in progress. Differing opinions may be expressed here. The answer to most of these questions is: "It depends." We try to flesh out here what it depends on.

*If you are logged-in as Birn_users, you should log out. This page is viewable to the public, and viewable/editable by FBIRN Wiki users only. Thanks!

FAQ

Calibration

• Why can't I just put functional or structural imaging data together from different sites without any worries?
• What functional imaging protocols should my multi-site study use?
• What structural imaging protocols should my multi-site study use?
• Can I do a multi-site study at all if some of the sites have different scanners?
• What pre-study measures should be performed at all sites?
• What is the agar phantom for and what should I do with it?
• My agar phantom has bubbles?!
• What is this idea of a traveling subjects study anyway?
• What difference does smoothing make in fMRI data?
• Should I care about head coil differences?
• Why should a senior engineer do a site inspection tour?

• I have a cognitive paradigm that I've used successfully and want to expand it to a multi-site study. What do I need to do?
• Do all sites have to have the same response devices, stimulus device, etc.?
• Where can I find examples of tasks used in multi-site studies?
• What is the FBIRN standard of IRB submission wording for multi-site data sharing?

• What imaging, clinical, and behavioral data QA is recommended, both before and during a study?
• What automated QA has been implemented and how can I use it in my study?
• What are the differences among imaging formats and which one is recommended
• What the heck are ontologies for and what role do they play in all this?
• Do I need to anonymize my data, and if so, how do I do it?
• If I use the SRB, how do I put my data on it and get my data off it?
• What's all this Scommands stuff?
• How can I make my data available to the rest of the scientific community?
• How does data replication work on the SRB? due out in Fall 2007
• What provenance information is required from my data and analyses? due out in Fall 2007

Manuals and tutorials are available for the Federated Informatics Research Environment (FIRE; the database formally known as HID); improved manuals for the updated upload and download interfaces are scheduled for fall 2007.

For more information, see also the FBIRN pages on the BIRN website.

• How do I get and install a FIRE? coming soon

• What analytical tools exist to handle multi-site fMRI data?
• Where can I find documentation of how the FBIRN multi-site data from 2005-2006 (Phase II) has been analyzed?
• What are the FBIRN recommendations for multi-site structural MRI data analysis using Freesurfer?
• How should I model the effect of site differences?
• How should I account for differences in Signal to Noise?
• Is it a good idea to share data that does not pass current QA measures?
• How do I include the BH or ASL measures for each subject as a correction factor? due out in Summer 2008

• What are your authorship policies?