How does testing affect retrieval-related processes? An event-related potential (ERP) study on the short-term effects of repeated retrieval.

The testing effect is conceptualized as the benefit for remembering items that were studied and tested rather than just studied. Thus far, little is known about the underlying neurocognitive mechanisms. In an event-related potential (ERP) study, we investigated the immediate consequences of testing on recollection processes. During an initial study phase, participants encountered object names together with pictures of the denoted objects ("perceived items") or with the instruction to mentally visualize them ("imagined items"). Directly afterward, they had to differentiate between perceived, imagined, and new items, in two consecutive source memory tests. Half of the studied items were presented in the first run, and all items in the second. Behaviorally, repeated testing led to improved item and source memory, as well as faster reaction times, relative to items that had been tested once. In accordance with these behavioral changes, the left parietal old-new effect (500-900 ms) as the putative correlate of recollection was strongly enhanced by previous testing. An enhancement after testing was also observed for the early portion of the late right frontal old-new effect (700-900 ms). In contrast, old-new effects after 900 ms were not modulated by previous testing. The finding of a stronger left parietal old-new effect for previously tested items suggests that testing leads to an elaboration of memory traces, whereas the faster reaction times are more likely explained in terms of transfer-appropriate processing. The combination of more elaborated memory traces and transfer-appropriate processing provides a tentative explanation for the effectiveness of testing in enhancing retrieval performance.