A potential application of studying eye movements with functional MRI (fMRI) is to examine patient populations with known eye movement dysfunction, but the reliability with which normal subjects demonstrate activity in specific brain regions has not been established. To date, fMRI studies of smooth-pursuit eye movements have used relatively small numbers of subjects and have been restricted to fixed-effects analyses. We extend these studies to whole brain imaging at 1.5 T, properly accounting for intersubject variation using random effects analysis. Smooth-pursuit eye movements elicited activation consistently in dorsal cortical eye fields and cerebellum. Subcortical activation was greatly attenuated, but not eliminated, with the random-effects second-level analysis. In addition, session-dependent changes in activation were greater in some regions than others and may indicate areas of brain, such as the supplementary eye fields, that are sensitive to attentional modulation of eye movements.