A growing body of evidence indicates that sleep promotes memory consolidation; the protracted process by which fresh, labile memories are reorganized into stable memories [1]. Although the first evidence for this effect was collected more than a century ago [2], the potential underlying neural mechanisms are only just beginning to be conceptualized and characterized experimentally.Memory consolidation is a hackneyed expression that may take different meanings. At the cellular level, consolidation is often framed in terms of long-term experience-dependent changes in neural responsiveness, usually resulting from modifications in synaptic strength (Figure 1) [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19]. At the systems level, consolidation relates to the progressive reorganization of memories within brain circuits [20][21]. From a behavioral perspective, consolidation refers to an improvement of performance between practice sessions or a reduction of vulnerability of a memory trace after the acquisition of a novel skill [22][23], although other phenomena such as automatization [24][25] or generalization [26] might also be considered as behavioral markers of memory consolidation (Figure 1).It is therefore not surprising that the effects of sleep on memory consolidation were reported at various levels of description. The early studies assessed the effects of sleep on different memory systems at the behavioral level. They contrasted the influence of non-rapid eye movement sleep (NREMS) with rapid eye movement sleep (REMS) on declarative and procedural memories, respectively. As the understanding of the effects of sleep on memory consolidation progressed, the hypotheses were increasingly framed in terms of the neural processes occurring during NREMS or REMS, and specifically, those associated with phasic events such as slow waves, spindles or phasic REMS. This paper is organized into five sections. The first section summarizes the key findings of behavioral studies that suggest an effect of sleep on memory. Earlier reviews have already covered these results in detail [27][28][29][30][31]. The present article highlights recent data that call for a reappraisal of these effects that, although indisputable, seem quantitatively modest. The article also emphasizes the higher sensitivity of neuroimaging and neurophysiological measures in characterizing the effects of sleep on memory. The following sections review the two main hypotheses that are currently being tested, namely the homeostatic synaptic downscaling and the sleep-dependent systems-level consolidation. The final two sections emphasize the intervention of post-training REMs and wakefulness in memory processing.
Behavioral effectsThe recognition of different memory systems helped to characterize the effects of sleep on memory consolidation (Figure 2). On the procedural side, spontaneous gains in performance in a motor sequence learning task were observed between initial training and later testing, specifically if the interval included sleep rather ...