No effect of targeted memory reactivation during slow‐wave sleep on emotional recognition memory
There are divergent lines of evidence with respect to the specific components of sleep responsible. Behavioural evidence links rapid eye movement (REM) sleep with emotional processing, although the neural mechanisms are unclear (Hutchinson and Rathore, 2015). A larger body of research implicates non‐REM sleep in memory consolidation, with a specific role for slow‐wave sleep (SWS) physiology. Active Systems accounts of sleep‐dependent consolidation propose that newly formed memories are reactivated during SWS, as indexed by spontaneous patterns of neural activity in the brain regions employed at learning (Diekelmann and Born, 2010). Reactivation is believed to reflect systems‐level communication between hippocampal networks and long‐term memory stores within the neocortex. This cross‐talk facilitates a shift in memory dependence from the hippocampus to the neocortex, leading to durable long‐term memory representations. Recent adaptions of this model propose that highly salient memories are preferentially reactivated in SWS, and thereby better stabilised in long‐term memory (Born and Wilhelm, 2012; Rasch and Born, 2013). Indeed, a selective benefit of SWS for the consolidation of emotionally negative memories has been observed in recent work (Cairney et al., 2015), suggesting that sleep‐specific emotional memory reactivations in SWS may be integral to affective memory processing.
Recently, a causal role of reactivation in memory consolidation has been demonstrated via a technique known as targeted memory reactivation (TMR). In a typical TMR experiment, individuals form associations between newly‐learned memories and semantically related sounds, a subset of which are then replayed in SWS to selectively cue memory reactivations (Oudiette and Paller, 2013). Memories are better remembered after cueing with TMR, indicating a mechanistic role of reactivation in sleep‐dependent consolidation (Cairney et al., 2016; Fuentemilla et al., 2013; Oudiette et al., 2013; Rudoy et al., 2009).
If emotional memories are preferentially reactivated in SWS, then the impacts of TMR may be stronger for emotional relative to neutral memory information. In keeping with this view, Cairney et al. (2014) observed a selective benefit of TMR in SWS for emotionally negative (versus neutral) memories. However, unlike previous work (Oudiette et al., 2013; Rudoy et al., 2009), the benefits of TMR were found in participant response times and not memory accuracy. How TMR influences emotional memory consolidation in sleep is therefore still very much unclear. The task employed by Cairney et al. (2014) included a strong spatial memory component, which may have dampened the emotional salience of the experimental stimuli at encoding. Placing greater emphasis on the explicit emotional properties of a novel memory trace may thus increase its responsiveness to TMR in sleep.
In the current study, we investigated the impact of TMR in SWS on the consolidation of highly salient emotional memory representations. To achieve this, we combined our TMR protocol with a recognition paradigm that promotes the emotional features of newly‐formed memories, and is compatible with previous studies of sleep and emotional memory (Hu et al., 2006; Sterpenich et al., 2007, 2009).