Dissociating the contributions of slow-wave sleep and rapid eye movement sleep to emotional item and source memory

Sleep benefits the consolidation of emotional memories, and this influence is commonly attributed to the rapid eye movement (REM) stage of sleep. However, the contributions of sleep stages to memory for an emotional episode may differ for the event per se (i.e., item memory), and the context in which it occurred (source memory). Here, we examined the effects of slow wave sleep (SWS) and REM sleep on the consolidation of emotionally negative and neutral item (picture recognition) and source memory (recall of picture-location and picture-frame color association) in humans. In Study 1, the participants (n= 18) learned 48 negative and 48 neutral pictures which were presented at specific locations and preceded by colored frames that had to be associated with the picture. In a within-subject design, learning was either followed by a 3-h early-night SWS-rich or by a late-night REM sleep-rich retention interval, then retrieval was tested. Only after REM-rich sleep, and not after SWS-rich sleep, was there a significant emotional enhancement, i.e., a significantly superior retention of emotional over neutral pictures. On the other hand, after SWS-rich sleep the retention of picture-frame color associations was better than after REM-rich sleep. However, this benefit was observed only for neutral pictures; and it was completely absent for the emotional pictures. To examine whether this absent benefit reflected a suppressive effect of emotionality on associations of minor task relevance, in Study 2 we manipulated the relevance of the picture-frame color association by combining it with information about monetary reward, following otherwise comparable procedures. Here, rewarded picture-frame color associations were equally well retained over SWS-rich early sleep no matter if the frames were associated with emotional or neutral pictures. Results are consistent with the view that REM sleep favors the emotional enhancement of item memory whereas SWS appears to contribute primarily to the consolidation of context-color information associated with the item.