Between- and within-scanner variability in the CaliBrain study n-back cognitive task

Gradin, V. et al. (2010) Between- and within-scanner variability in the CaliBrain study n-back cognitive task. Psychiatry Research: Neuroimaging, 184(2), pp. 86-95. (doi: 10.1016/j.pscychresns.2010.08.010)

Abstract

Structural and functional magnetic resonance imaging (sMRI and fMRI) show promise in their ability to facilitate the early diagnosis of disorders such as schizophrenia and dementia ( [Job et al., 2006] and [Whalley et al., 2006] ), and neuroimaging biomarkers of disease and clinical outcome data are increasingly regarded as useful in the development and evaluation of new treatments. However, psychiatric imaging studies tend to be small and based in one imaging centre. Multicentre studies, involving recruitment of subjects at several scanning centres, therefore have great appeal, but data pooling faces potential problems related to differences in scanner field strength and pulse sequences ( [Zou et al., 2005] and [Friedman et al., 2008] ). Before undertaking large psychiatric multi-centre fMRI studies, it is important to investigate the extent to which it is possible to obtain reproducible and reliable data, taking account of both between-scanner and within-scanner (over time) variability. If most of the variance in images were due to scanner-related differences, rather than to subject- and task-related factors of interest, imaging data would be largely scanner-dependent and of unclear generalisability (Costafreda et al., 2007). Such heterogeneity would raise questions about pooling fMRI data. Whilst a number of previous studies have focused either on within-scanner or between-scanner effects, to our knowledge, only two recent studies have examined both between- and within-scanner fMRI variability in the same subjects ( [Friedman et al., 2008] and [Suckling et al., 2008] ) as here. As will be discussed, a limited amount of work has been done on within-scanner reliability and less on between-scanner reproducibility. Here we report the results of an analysis of the n-back cognitive task (Owen et al., 2005) fMRI data, obtained as part of the Scottish ‘CaliBrain’ multi-centre's MRI and fMRI initiative. Our first hypothesis was that significant patterns of brain activation would be found with the three scanners. This was a prerequisite to testing our second hypothesis, which was no significant within- and between-scanner image differences with regard to these significant patterns of activity. Investigation of reproducibility and reliability of brain activation is a central focus of this study. Our third hypothesis was that after attention to ‘scanner harmonization’ (using the same field strength scanners and implementing the task in as similar a way as possible at each scanner), reproducibility values would be in the range reported for single scanner studies, and within-scanner reliability would be similar to between-scanner reliability.

• Item Type: Articles
• Status: Published
• Refereed: Yes
• Glasgow Author(s) Enlighten ID: Cavanagh, Professor Jonathan
• Authors: Gradin, V., Gountouna, V.E., Waiter, G., Ahearn, T.S., Brennan, D., Condon, B., Marshall, I., McGonigle, D.J., Murray, A.D., Whalley, H., Cavanagh, J., Hadley, D., Lymer, K., McIntosh, A., Moorhead, T.W., Job, D., Wardlaw, J., Lawrie, S.M., and Steele, J.D.
• Subjects: R Medicine > RC Internal medicine > RC0321 Neuroscience. Biological psychiatry. Neuropsychiatry
• College/School: College of Medical Veterinary and Life Sciences > School of Health & Wellbeing > Mental Health and Wellbeing; College of Medical Veterinary and Life Sciences > School of Medicine, Dentistry & Nursing
• Journal Name: Psychiatry Research: Neuroimaging
• ISSN: 0925-4927