Sleep contributes to both memory consolidation and emotional regulation, but how non-rapid eye movement (NREM) and rapid eye movement (REM) sleep support these functions has been debated. This preprint reports within-subject EEG evidence that reactivation of emotional memories during nocturnal sleep is both stage-dependent and tied to specific sleep events, with dissociable relationships to memory retention and overnight changes in negative affect.
The authors used a within-subject crossover design in which participants experienced two experimental nights while high-density electroencephalography (EEG) was recorded. Across nights participants viewed either neutral or aversive film clips in a counterbalanced order. The design allowed comparison of how encoding of neutral versus aversive material related to subsequent nocturnal reactivation. The article is a preprint and the abstract was the source for this summary; additional methodological details such as sample size, participant characteristics, or exact timing were not reported in the provided source.
To identify memory reactivation during sleep, the study combined nocturnal EEG with daytime functional localizers that established neural patterns associated with aversive versus neutral emotional processing. These localizer-derived patterns were then used in a multivariate pattern analysis framework to detect whether, when, and in which sleep stage the emotional memory representations reemerged. The approach enabled detection of valence-specific neural reactivation rather than general replay alone.
In NREM sleep, the researchers observed that reactivation signals distinguishing aversive from neutral memories were time-locked to complexes of slow oscillations and spindles (SO–spindle complexes). Crucially, this valence-specific reactivation was not evident when slow oscillations or spindles were considered separately; it emerged specifically at the coupling of the two events. The authors further report that the percentage of NREM SO–spindle coupling was associated with consolidation of temporal memories, linking this event-specific reactivation to overnight memory stabilization.
In REM sleep, emotional memory reactivation showed a different profile: it occurred selectively during phasic REM periods that are marked by rapid eye movements. The analysis therefore implicates transient, event-defined epochs within REM rather than a uniform REM process. The strength of phasic REM reactivation was reported to correlate with overnight dissipation of negative affect, suggesting a specific REM-related mechanism that supports emotional processing and affective downregulation.
The dissociation observed in the neural data was paralleled by differential behavioral relationships. NREM SO–spindle coupling percentage correlated with improvement in temporal memory measures across sleep, supporting a role for NREM event coupling in memory consolidation. By contrast, the magnitude of reactivation during phasic REM predicted decreases in negative affect from evening to morning, consistent with a role for REM reprocessing in emotional dissipation. The abstract reports these associations; quantitative effect sizes, statistical details, and behavioral measures were not provided in the source extract.
Taken together, the findings provide direct evidence—within the limits of the reported data—that sleep reprocesses emotional experiences through dissociable, stage- and event-specific mechanisms. The authors frame these results as establishing a conceptual scaffold for targeted sleep-based interventions that might aim to enhance memory consolidation or modulate emotional responses by manipulating specific sleep events (for example, promoting SO–spindle coupling or modulating phasic REM epochs).
Important caveats: this manuscript is a preprint and has not undergone peer review. The summary above is based on the article abstract and front-matter information available in the source; many methodological and statistical details (sample size, demographics, exact EEG analysis pipeline, timing of stimuli and sleep recordings, and precise behavioral measures) are not reported in the abstract and therefore were not available for inclusion. Readers should consult the full preprint and subsequent peer-reviewed reports for complete methods, results, and replication.
High-density EEG combined with functional localizers and multivariate pattern analysis indicates that emotional memories are reactivated in sleep in a stage- and event-specific manner: NREM reactivation aligns with SO–spindle coupling and relates to memory consolidation, whereas REM reactivation occurs during phasic REM marked by rapid eye movements and relates to overnight negative affect dissipation. These dissociable mechanisms point to distinct roles for sleep microarchitecture in processing emotional experiences.