Multifocal visual evoked potentials (VEP) allow one to assess whether stimulation at specific visual field locations elicits cortical activity; it might therefore enable us to conduct objective visual field perimetry. However, due to the cortical folding, which differs markedly between subjects, a particular electroencephaologram generator may fail to project signal on some recording electrodes. This may lead to false alarms for potential scotomata. Here we compare pattern-reversal and pattern-onset stimulation in their efficacy to activate the visual cortex and recorded mfVEPs to 60 locations comprising a visual field of 44° diameter. We report three main findings: (1) Pattern-onset compared to pattern-reversal enhances the amplitude by 30% for stimulation of the central visual field (<10° radius), while evoking 30% less response in the periphery (>15°). (2) Although pattern-onset and pattern-reversal responses differ markedly in their eccentricity dependence, they have a similar topographical distribution. (3) By combining both stimuli, the number of false positives was reduced to less than 1.5% of the visual field locations tested. We conclude that pattern-onset and pattern-reversal activate identical visual cortical areas but target different neural mechanisms within these areas. Furthermore, pattern-onset stimulation greatly increases the sensitivity of the mfVEP to assess the cortical representation of the central 10° of the visual field.