Saccades are fast eye movements that direct the point of regard to a target in the visual field. Repeated post-saccadic visual errors can induce modifications of the amplitude of these saccades, a process known as saccadic adaptation. Two experiments using the same paradigm were performed to study the involvement of the cerebrum and the cerebellum in the processing of saccadic errors using functional magnetic resonance imaging and in-scanner eye movement recordings. In the first active condition, saccadic adaptation was prevented using a condition in which the saccadic target was shifted to a variable position during the saccade towards it. This condition induced random saccadic errors as opposed to the second active condition in which the saccadic target was not shifted. In the baseline condition, subjects looked at a stationary dot. Both active conditions compared with baseline evoked activation in the expected saccade-related regions using a stringent statistical threshold [the frontal and parietal eye fields, primary visual area, MT/V5, and the precuneus (V6) in the cerebrum; vermis VI-VII; and lobule VI in the cerebellum, known as the oculomotor vermis). In the direct comparison between the two active conditions, significantly more cerebellar activation (vermis VIII, lobules VIII-X, left lobule VIIb) was observed with random saccadic errors (using a more relaxed statistical threshold). These results suggest a possible role for areas outside the oculomotor vermis of the cerebellum in the processing of saccadic errors. Future studies of these areas with, e.g., electrophysiological recordings, may reveal the nature of the error signals that drive the amplitude modification of saccadic eye movements.