The potentially important effect of gradient switching sound on brain function during functional magnetic resonance imaging (fMRI) was studied by comparing experiments with low and high scanner sound levels. To provide a low sound level experiment, a sparse scanning method was used, characterized by long, 9 sec, periods of scanner silence interspersed with 1 sec echoplanar imaging (EPI) bursts. For the condition with high sound levels, extra EPI gradient modules were inserted in the 9 sec inter-image intervals. Visual, motor, or auditory stimuli were presented in the interval between imaging. It was found that with the addition of gradient sounds, auditory activation was significantly decreased while motor and visual activation were not significantly altered. Other general factors relating to fMRI were also examined, such as experimental duration and fatigue. For example, motion of the subjects during the experiments was found to be related to the time spent in the scanner, rather than to the ambient sound level.