Abstract
Mediated priming (e.g., from LION to STRIPES via TIGER), is predicted by spreading activation models but only by some integration models. The goal of the present research was to localize mediated priming by assessing two-step priming effects on N400 and reaction times [RT]. We propose that the N400 priming effect mainly reflects integration processes but, in contrast with RT, does not reflect spreading activation. The results show that RT- and N400-effects can dissociate. In a standard lexical decision task, we found mediated priming for N400, but not for RT. When a lexical decision to both the prime and the target was required, mediated priming was observed for both measures, but the RT effect was not influenced by list composition, whereas the N400 effect was. We conclude that two qualitatively different processes underlie the two types of mediated priming. A process of 'global integration' yields an N400 effect whereas an RT effect is evoked by spreading activation.
