The redundant target effect (RTE) consists in the speeding of reaction time with single versus multiple targets and can be explained either by a neural coactivation or by a race model. To try to understand the role of the magnocellular and parvocellular systems in the determination of the RTE we carried out three experiments using onset or feature singletons. The former are likely to be mainly processed by the magnocellular system while the latter are mainly processed by the parvocellular system. In experiment 1 we found an RTE both when the target (red disk) was presented in isolation and when it was surrounded by equiluminant green distractors. Thus, the RTE occurred both with onset and feature singletons. However, with the former, the RTE could be accounted for by neural coactivation while with the latter it could be accounted for by a probabilistic explanation. In experiment 2 we tried to ascertain the role of distractors in yielding a probabilistic RTE: we used either targets in isolation or surrounded by distractors of lower luminance and found an RTE that could be explained by neural coactivation for both kinds of targets. This ruled out an effect of distractors per se in determining a probabilistic RTE. Finally, in experiment 3 we used targets of lower luminance than either the background or the distractors. We found that the RTE could be accounted for by neural coactivation with targets alone while it was probabilistic with distractors. Overall, these results show that stimuli presumably processed by the magnocellular system yield redundancy gains that result from a neural coactivation mechanism. In contrast, stimuli presumably processed by the parvocellular system are compatible with a probabilistic redundancy gain.