Visual Evoked Potentials (VEP) have been considered an important complementary diagnostic method for the study of the physiology of the visual pathways and cortex. It is suggested that topographic maps of Visual Evoked Potentials (VEP) generate information related to the anatomy and physiology of the visual cortex. The source derivation technique applied to electroencephalographic signals has been shown to improve the sensitivity and spatial selectivity of the results. Mapping VEP after source derivation was implemented in order to facilitate the interpretation of the VEP signals and as an aid in identifying the distribution of the sources generating them. The signals were obtained from normal individuals stimulated with full-field and half-field pattern reversal checker-boards. The electrode system used (16 channels) was that proposed by Bodis-Wollner; this system concentrates the electrodes in the scalp areas related to the visual cortex. Signals were sampled simultaneously with a 10-bit A/D converter, saved and, thereafter, processed in an IBM/PC compatible microcomputer. After coherently averaging the responses, source derivation considering inter-electrode distances was applied. It was observed that the spatial spreading of the signals was reduced, leading to an improved spatial selectivity, locating the VEP sources in the expected brain regions and suggesting the existence of electrical dipoles.