Neural correlates of explicit and implicit memory at encoding and retrieval: A unified framework and meta-analysis of...

The extent to which explicit memory (EM) and implicit memory (IM) involve similar or differential neural substrates remains unclear. To address this issue, this study provides a direct, meta-analytic comparison of functional neuroimaging studies involving EM and IM tasks. The meta-analysis comprised two separate meta-analytic comparisons. First, to compare EM and IM in terms of encoding activity, subsequent memory effects (remembered > forgotten) and repetition suppression effects (first > repeated) were directly compared. Second, to compare EM and IM in terms of retrieval activity, retrieval success effects (hit > correct rejection) and repetition suppression effects were directly compared. Based on the notion that reduced activity during repeated processing is a 'by-product' or direct consequence of the stimulus processing performed in the same regions at initial exposure, regions showing repetition suppression were thought to play an important role in both IM-encoding and IM-retrieval activities. The results indicated that subsequent memory and repetition suppression effects had extensive overlaps and no significant separations, suggesting that EM- and IM-encoding activities involve largely common regions. Retrieval success and repetition suppression effects had strong segregations and only modest overlaps, suggesting that EM- and IM-retrieval activities involve largely separate regions. Consistent with these results, Explicit/Implicit Memory Encoding and Retrieval (EIMER), a neurocognitive model of EM and IM that suggests a common-encoding, separate-retrieval hypothesis for EM and IM is proposed herein.