Temporal analysis of the multifocal cortical visual evoked potential (VEP) was studied using pseudo-random (m-sequence) achromatic stimulation. The effects of variation of luminance contrast on the first-order response were complex. At low to mid contrasts (< 60%), a wave doublet (P100-N115) predominated. A second wave complex (N100-P120-N160) dominated at high contrasts. The second-order responses, however, showed an extremely simple variation with luminance contrast. Intrinsic differences in the adaptation time of the generators of these two components caused a distinct separation in the slices of the second-order response. A rapidly adapting nonlinearity saturating at low contrasts was only observable when measuring the responses from two consecutive flashes. Its latency coincided with the contrast saturating first-order response component. By comparison, the nonlinearity derived from the responses to the stimuli with longer interstimulus intervals (second and third slices) yielded a much more linear contrast response function with lower contrast gain and latencies, which clearly corresponded to the longer latency component of the first-order response. Thus, the second-order responses show a first slice which is predominantly driven by neural elements that have a latency and contrast function that mimic those of the magnocellular neurons of the primate LGN and a second slice which is dominated by a generator whose properties resemble primate parvocellular function. This division into magno and parvocellular contribution to the VEP is based on function (interaction time) as distinct from other currently available analyses, with potential for neural analysis of visual disease.