Modality specific contributions to true and false recognition: an event related potential study

Abstract

Retrieval in recognition memory is generally thought to operate based on two distinct processes, recollection and familiarity. The contribution of modality specific information to these processes is unclear, however. The Deese/Roediger and McDermott (D/RM) paradigm was used to create a situation in which test items differed according to perceptual experience, with old items associated with a previous sensory experience, but lure items not associated with such an experience. Additionally the modality of study lists was manipulated such that participants either saw or heard items at study, then were tested visually, such that there was a match or mismatch in study-test modality. Event-related potentials (ERPs) were used to provide a precise measurement of recognition memory processes as they unfolded. Significant differences were found for several ERP components. The P300 was larger for participants in the visual study condition and for false alarms to lure items. Analysis of the FN400 component revealed higher amplitudes for false alarms compared to hits in the visual study condition, but lower amplitudes for false alarms than hits in the auditory condition. A parietal old/new effect was found in the auditory but not the visual condition, with hits and false alarms producing ERPs consistent with recollection and familiarity processes, respectively. Finally, late frontal effects were found which indicated that participants in the visual condition were engaging in greater post-retrieval evaluation than those in the auditory condition. This effect was driven by a large late frontal positivity for false alarms in the visual study condition. These results are interpreted as indicating a possible retrieval strategy difference between the study modality conditions, consistent with an explanation based on distinctiveness.