- To switch from one cognitive task to another is thought to rely on additional control effort being indicated by performance costs relative to repeating the same task. This switch cost can be reduced by advance task preparation. In the present experiment the nature of advance preparation was investigated by comparing a situation where an explicit task cue was presented 2000 ms in advance of the target stimulus (CTI-2000) with a situation where cue and target were presented in close succession (CTI-100). We mapped the blood-oxygenation-leveldependent (BOLD) activation correlates of switch-related control effort and advance task preparation to test alternative explanations why advance preparation is reducing switch costs. A previously reported control-related cortical network of frontal and parietal brain areas emerged that was more strongly activated for switching between tasks. However, this was true exclusively for CTI-100 where no advance task preparation was possible. At CTI-2000 these same brain areas were equally engaged in both switch and repeat trials. For some of these areas, this common activation was time-locked to the presentation of both the cue as well as the target. Other areas were exclusively associated with target processing. The overall pattern of results suggests that advance task preparation is a common process of pre-activating (cue-locked activation) the currently relevant task set which does not face interference from a persisting N − 1 task set. During target processing the same brain areas are re-engaged (subsequent target-locked activation) to apply the pre-activated task set. Though being common to repeat and switch trials, advance preparation has a differential benefit for switch trials. This is because the instructed task set has time to settle into a stable state, thus becoming resistant against disruption from the previous task set, which is retrieved by the current target stimulus. © 2004 Elsevier Ltd. All rights reserved.