Introduction: Recent evidence suggests a critical role of sleep in consolidating emotional memories. Most studies have focused on REM sleep and targeted negative vs. neutral memories, reporting inconsistent results. Here, we further explore the role of REM, NREM and sleep spindles in the encoding and consolidation of pleasant, unpleasant and neutral memories. \n \nMethods: Forty-seven (33F) healthy university students were exposed to a first set of 40 pleasant, 40 unpleasant and 40 neutral pictures at 1:00PM and to a second equivalent set at 5:00PM. During a later recognition test (5.15PM), participants saw the previous 240 pictures intermixed with 120 novel stimuli. They had to decide whether they had seen the picture before. In the Nap group (N=31), participants took a 120-min nap after the first set presentation, while in the No-Nap group (N=16), participants remained awake. Based on polysomnographic recording, participants were segregated into REM (N=14) and NoREM groups (N=17). \n \nResults: Memory discrimination (d’), arousal and valence ratings did not differ between the REM and NoREM groups, therefore their data were collapsed into a Nap group. d’ was greater for sleep relative to wake. Although the Nap and No-Nap groups better discriminated recent (15m pre-retrieval) versus delayed (4h pre-retrieval) pictures, only the Nap group showed a significantly higher d’ for recent compared to delayed emotional pictures. Overall, neutral pictures showed a greater d’ than pleasant items, but no difference relative to unpleasant stimuli. An association between N2 sleep spindles and d’ for delayed negative stimuli was also observed. \n \nConclusion: Taken together, our results indicate that daytime nap, regardless of the presence of REM sleep, facilitates the consolidation of declarative memories independent of their valence. We suggest that sleep promotes the formation of new emotional memories and that sleep spindles may critically affect their subsequent consolidation.