In this study we attempted to differentiate distinct components of the saccade network, namely cortical ocular motor centers and parieto-occipital brain regions, by means of a "minimal design" approach. Using a blocked design fMRI paradigm we evaluated the BOLD changes in a 2 x 2 factorial design experiment which was performed in complete darkness: while looking straight ahead with eyes open (OPEN) or closed (CLOSED) as well as during the execution of self-initiated horizontal to-and-fro saccades with the eyes open (SACCopen) or closed (SACCclosed). Eye movements were monitored outside the scanner via electro-oculography and during scanning using video-oculography. Unintentional eye-drifts did not differ during OPEN and CLOSED and saccade frequencies, and amplitudes did not vary significantly between the two saccade conditions. The main findings of the functional imaging study were as follows: (1) Saccades with eyes open or closed in complete darkness lead to distinct differences in brain activation patterns. (2) A parieto-occipital brain region including the precuneus, superior parietal lobule, posterior part of the intraparietal sulcus (IPS), and cuneus was relatively deactivated during saccades performed with eyes closed but not during saccades with eyes open or when looking straight ahead. This could indicate a preparatory state for updating spatial information, which is active during saccades with eyes open even without actual visual input. The preparatory state is suppressed when the eyes are closed during the saccades. (3) Selected ocular motor areas, not including the parietal eye field (PEF), show a stronger activation during SACCclosed than during SACCopen. The increased effort involved in performing saccades with eyes closed, perhaps due to the unusualness of the task, may be the cause of this increased activation.