There are many reasons for wanting to perform multi-site investigations employing MRI or fMRI, including reduced study duration with large study populations and access to a wider set of demographic characteristics. However, inter-site differences in scanner vendors and pulse sequences as well as calibration of scanners can cause difficulties when trying to combine anatomic images in multi-site studies utilizing MRI. These problems are exacerbated greatly in fMRI because of the additional need to monitor and maintain scanner stability and to carefully standardize fMRI scanning practices (e.g. subject preparation) and ancillary hardware such as response boxes and projection and auditory stimulus equipment. Uncontrolled differences in any of these aspects of study design and implementation can introduce significant inter-site variance in the data pool and render combination from multiple sites questionable. In this presentation, the use of scanner QA methods, standardization of pulse sequence parameters, inter-site and inter-subject calibration methods and the importance of controlling for scanning procedures across all sites are considered. A case study (the NIH NCRR-funded FIRST-BIRN: Functional Imaging Research in Schizophrenia Testbed Biomedical Imaging Research Network) will be discussed, including solutions that have been developed to deal with scanner? upgrades and on-going data QA.
Audio (MP3 File, Podcast Ready)
Copyright. All Rights Reserved.