When foreperiods (FPs) of different duration vary on a trial-by-trial basis equiprobably but randomly, the RT is faster as the FP increases (variable FP effect), and becomes slower as the FP on the preceding trial gets longer (sequential effects). It is unclear whether the two effects are due to a common mechanism or to two different ones. Patients with lesions on the right lateral prefrontal cortex do not show the typical FP effect, suggesting a deficit in monitoring the FP adequately [Stuss, D. T., Alexander, M. P., Shallice, T., Picton, T. W., Binns, M. A., Macdonald, R., et al. (2005). Multiple frontal systems controlling response speed. Neuropsychologia, 43, 396-417]. The aim of this study was two-fold: (1) to replicate this neuropsychological result testing cerebral tumor patients before and after surgical removal of the tumor located unilaterally in the prefrontal, premotor or parietal cortex, respectively and (2) to investigate whether the sequential effects would change together with the FP effect (supporting single-process accounts) or the two effects can be dissociated across tumor locations (suggesting dual-process views). The results of an experiment with a variable FP paradigm show a significant reduction of the FP effect selectively after excision of tumors on right prefrontal cortex. On the other hand, the sequential effects were reliably reduced especially after surgical removal of tumors located in the left premotor region, despite a normal FP effect. The latter dissociation between the two effects supports a dual-process account of the variable FP phenomena. This study demonstrates that testing acute cerebral tumor patients represents a viable neuropsychological approach for the fractionation and localisation of cognitive processes.