These authors found that isolates presented at the start of a list were remembered better than standard words, even though there was not yet a context to make them distinctive. This finding, and ours, suggests that the von Restorff effect is not produced by novelty processing, but by other mechanisms. However, there are also studies with findings that contradict ours. Several studies have found that recalled isolates elicited
larger P3s at study than nonrecalled isolates (Fabiani et al. 1990; Fabiani and Donchin 1995; Otten and Donchin 2000). This was only the case when color was used to make words distinctive and not when this was done with a surrounding frame (Otten and Donchin 2000), and only when participants were Inhibitors,research,lifescience,medical instructed to use rote rehearsal as their learning strategy, not Inhibitors,research,lifescience,medical when elaboration was used (Fabiani et al. 1990). Wiswede et al. (2006) found a larger P3 for both recalled isolates and recalled standard words as compared with not-recalled words. In all of these studies, immediate free recall was used to test memory. We tested memory with delayed cued recall and recognition. Our contradictory findings suggest that P3 amplitude indexes a process
that helps in free recall, but not cued recall or recognition. One candidate for such a process is attention to the unique, novel feature itself. In free recall, but not cued recall or recognition, features such as color Inhibitors,research,lifescience,medical or font size can be used as cue to retrieve the word. For
example, in our study only one of 80 words was presented in green font. This would Inhibitors,research,lifescience,medical GDC-0449 probably make “green” a good cue to retrieve this word in free recall. Thus, in free recall, the “green” word would often be recalled (see McDaniel et al. 2005 for a discussion of similar retrieval-based accounts). By contrast, during Inhibitors,research,lifescience,medical cued recall, participants cannot search their memory for a green word, as they would not know for which word stem cue “greenness” would be of any help in retrieval.2 If the P3a indeed indexes attention to the novel feature, this would thus aid free recall of isolates (Karis et al. 1984; Fabiani et al. 1990; Fabiani and Donchin 1995; Wiswede et al. 2006), but not cued recall. It would also, presumably, aid more in designs in which there was just a single isolate Mephenoxalone per list, as opposed to more than one as in our design. Whether this is the case, it is clear that a von Restorff effect can be found for isolates that do not elicit larger P3s, for example, when isolates are made distinctive with a surrounding frame (Otten and Donchin 2000) or when elaboration is used (Fabiani et al. 1990; our results). Moreover, Fabiani and Donchin (1995) reported larger P3s for recalled as compared to not-recalled isolates in conditions where no von Restorff effect was found: for semantic isolates when participants were asked to focus on physical appearance, and for a recognition task that followed lexical decision.