To investigate the effects of repetitive transcranial magnetic stimulation (rTMS) on the central processing of motor preparation, we had subjects perform a precued-choice reaction time (RT) task. They had to press one of two buttons as quickly as possible after a go signal specifying both the hand to be used and the button to press. A precue preceding this signal conveyed full, partial, or no advance information (hand and/or button), such that RT shortened with increasing amount of information. We applied 1200 to 2400 pulses of 1-Hz rTMS over various cortical areas and compared the subjects' performances at various times before and after this intervention. rTMS delayed RT at two distinct phases after stimulation, one within 10 min and another with a peak at 20 to 30 min and lasting for 60 to 90 min, with no significant effects on error rates or movement time. The effect was significantly larger on left- than on right-hand responses. RT was prominently delayed over the premotor and motor cortices with similar effects across different conditions of advance information, suggesting that preparatory processes relatively close to the formation of motor output were influenced by rTMS. In contrast, the effect of rTMS over the supplementary motor area and the anterior parietal cortex varied with the amount of advance information, indicating specific roles played by these areas in integrating target and effector information. The primary motor area, especially of the left hemisphere, may take over this processing, implementing motor output based on the information processed in other areas.