Previous studies have found that inhibition of a biologically dominant prepotent response tendency is required during the execution of a less familiar, non-prepotent response. However, the lasting impact of this inhibition and the cognitive mechanisms to flexibly switch between prepotent and non-prepotent responses are poorly understood. We examined the neurophysiological (ERP) correlates of switching between prosaccade and antisaccade responses in 22 healthy volunteers. The behavioural data showed significant switch costs in terms of response latency for the prosaccade task only. These costs occurred exclusively in trials when preparation for the switch was limited to 300 ms, suggesting that inhibition of the prepotent prosaccade task either passively dissipated or was actively overcome during the longer 1000 ms preparation interval. In the neurophysiological data, a late frontal negativity (LFN) was visible during preparation for a switch to the prosaccade task that was absent when switching to the antisaccade task, which may reflect the overcoming of persisting inhibition. During task implementation both saccade types were associated with a late parietal positivity (LPP) for switch relative to repetition trials, possibly indicating attentional reorienting to the switched-to task, and visible only with short preparation intervals. When the prosaccade and antisaccade task were contrasted directly during task implementation, the antisaccade task exhibited increased stimulus-locked N2 and decreased P3 amplitudes indicative of active inhibition. The present findings indicate that neurophysiological markers of persisting and current inhibition can be revealed using a prosaccade/antisaccade-switching task.