Knowing the connections between different brain regions and the direction of the flow of information along these connections is of paramount importance for understanding the functioning of the visual system both in its healthy and diseased states. By adapting a body of theoretical and empirical results from basic research in low-level visual psychophysics, we propose that functional magnetic resonance imaging (fMRI) can be used to image the flow of information without the need of a very high temporal resolution. With reasonable assumptions, we can show that when noise is added to a visual stimulus, the impact of this noise on neural activity will depend on the amount of nonlinear processing that occurs between the stimulus and the brain region of interest. This general effect leads to a number of fMRI-measurable quantities that we can use to determine the input-output ordering between adjacent cortical areas and subregions. The goal of this study is to develop and test this novel method, based on clear-cut predictions from the underlying theory, in those cortical regions of the human visual system where large-scale connectivity is reasonably well known. ? ?