Bidirectional interactions between REM sleep and stress responses through the lens of the resilience framework

Rapid eye movement (REM) sleep and stress responses are tightly intertwined, yet the nature of their relationship remains contested. REM sleep has been linked to both attenuation and potentiation of stress responses across studies, and the sources for such inconsistency are poorly understood. Here, we synthesize evidence on the bidirectional interactions between REM sleep and stress responses across three levels of analysis and propose a unified framework for resolving some of these inconsistencies. First, we highlight intrusive memory formation and threat and safety learning as two core cognitive-emotional processes through which REM sleep may influence stress responses. Second, we review evidence that REM-associated emotional- and threat-related memory processing engages limbic-prefrontal circuitry that substantially overlaps with neural systems supporting stress responses, offering a plausible mechanistic substrate for REM-stress interactions. Third, we address the temporal dynamics of REM-stress interactions, reviewing evidence that stress alters subsequent REM architecture but also that pre-stress REM predicts later stress outcomes. Further, we acknowledge that these mutual interactions may involve both short-term state-like effects and longer-term trait-like processes. After reviewing these findings, we integrate them into a resilience-based framework. To this end, we propose that the critical variable in determining stress outcome is not absolute REM levels at any single time point but rather the within-person trajectory of REM perturbation and recovery following stress. In other words, an adaptive stress response may involve moderate changes in REM sleep architecture following stress exposure that resolve over time. Through four illustrative cases we demonstrate how this framework can reconcile existing contradictory findings in the literature. Finally, we outline potential translational and clinical implications, particularly in relation to posttraumatic stress disorder (PTSD), and highlight future longitudinal research designs that could further improve our understanding of the bidirectional interactions between REM sleep and stress responses.