We examined the slow potential shifts preceding interictal spikes in the scalp EEGs of the children with benign rolandic epilepsy of childhood (BREC). In contrast to our expectations and the observations reported by Stodieck and Wieser (1987), we were not able to detect focal negative slow shifts preceding interictal discharges. Possible reasons for this result are as follows. The cerebral event to which we triggered might have been too 'weak'. Another problem is that not every single spike generated in the deep reaches the surface. Thus, the moment, a spike occurs in the scalp-recordings does not necessarily represent the moment the brain changes from one state to another. The reference electrode might also play a role in obscuring any slow shift.