Background Functional magnetic resonance imaging (fMRI) is one of the most popular methods to probe and understand the human brain, offering a noninvasive way for the in vivo investigation of brain function (1). fMRI accounts for the growth of neuroscience research with some 40,000 peerreviewed publications in the last two decades (2,3). The results of fMRI studies are largely determined by fMRI systems and informatics tools which process complex data generated from fMRI scan. fMRI quality assurance (QA) plays a critical role to guarantee high reliability of fMRI studies (4-11), since fMRI QA programs and methods can be used for calibrating fMRI scanner (10-13), testing the stability of fMRI system (14-17), assessing fMRI data quality (15,16,18-22) and evaluating informatics tools (4-9). Moreover, many QA-related metrics such as test-retest reliability and family-wise error rate (FWER) are applied to evaluate the reliability of fMRI studies (3,23,24). Many QA programs can test basic MRI system performances, such as resolution, signal contrast, geometric distortion, intensity uniformity, and ghosting artifacts (25-27). These programs are helpful but inadequate for fMRI study. The reasons are as follows: Scanner temporal stability The blood oxygenation level dependent (BOLD) signal