We investigate the effects of variations in response requirements on BOLD activation in a visual oddball task and consider implications for fMRI task designs. Sixteen healthy subjects completed 3 runs of a visual oddball task: passive, count and respond. Besides expected activation patterns during passive viewing, we identified joint activations, but more importantly crucial differences between the count and respond versions of the task. Middle frontal gyrus activation was seen in the respond but not the count condition suggesting that this region is associated with action execution rather than the decision-making aspect of the task. In addition, activation observed in the central opercular cortex and parietal operculum in the respond (but not count) condition is likely to reflect integration of the sensory, decision and response processes. We also observed activation in the supplementary motor area (SMA) during count as well as respond. Since the count condition requires no motor planning or response our data provide evidence for an SMA involvement in decision-making. Our study clearly shows that the count and respond versions of the visual oddball task result in different patterns of BOLD activation that could both be attributed to ‘target detection’ if information on the respective other condition was not available. We also show that considering the elements of a complex task is crucial when transferring it from one imaging modality to another and that a motor response is not always necessary in fMRI studies when the task has been set up appropriately. •Task design influences the attribution of cognitive function to activation patterns.•Overt and covert responses result in different interpretations of BOLD activation.•Provides evidence for the functional role of brain regions in the oddball task.