Changes in the level of unitization moderate the impact of unitization on associative memory and its underlying processing.

The viewpoint that unitization provides a possibility of increasing the contribution of familiarity to associative memory has been widely accepted, but its effects on associative memory and recollection remain controversial. The current study aims to explain these mixed results by considering a potential moderator: changes in the level of unitization from encoding to retrieval phases. During the encoding phase, participants learned the related and unrelated picture pairs (i.e., high vs. low levels of unitization). Subsequently, they needed to distinguish between the intact and rearranged pairs during retrieval, where, in these rearranged pairs, the level of unitization from encoding to retrieval phases may or may not change. Meanwhile, the scalp electroencephalographic activity (EEG) was recorded. The results showed a significant familiarity-related FN400 old/new effect for related picture pairs alone, which supported the above viewpoint. However, its impact on the associative memory and recollection-related LPC old/new effects varied with the level of unitization changes-specifically, under the unchanged conditions. Although related pairs elicited significant FN400 and LPC old/new effects, the differences in these old/new effects and associative memory between the related and unrelated picture pairs were not significant. Conversely, under the changed conditions, related picture pairs not only elicited significantly larger FN400 and LPC old/new effects but also improved associative memory more than unrelated picture pairs. These findings not only clarify some of the inconsistencies in the literature concerning the impact of unitization on associative memory but also suggest that unitization affects the contributions of familiarity and recollection to associative memory differently, its effectiveness varying with the level of unitization changes.