Key points are not available for this paper at this time.
PURPOSE: To provide a between-site comparison of functional MRI (fMRI) signal reproducibility in two laboratories equipped with identical imaging hardware and software. Many studies have looked at within-subject reliability and more recent efforts have begun to calibrate responses across sites, magnetic field strengths, and software. By comparing identical imaging hardware and software, we provide a benchmark for future multisite comparisons. MATERIALS AND METHODS: We evaluated system compatibility based on noise and stability properties of phantom scans and contrast estimates from repeated runs of a blocked motor and visual task on the same four subjects at both sites. RESULTS: Analysis of variance (ANOVA) and region of interest (ROI) analysis confirmed that site did not play a significant role in explaining variance in our large fMRI dataset. Effect size analysis shows that between-subject differences account for nearly 10 times more variance than site effects. CONCLUSION: We show that quantitative comparisons of contrast estimates derived from cognitive experiments can reliably be compared across two sites. This allows us to establish an effective platform for comparing group differences between two sites using fMRI when group effects are potentially confounded with site, as in the study of neurocultural differences between countries or multicenter clinical trials.
Sutton et al. (Wed,) studied this question.