Nightmare sufferers show atypical emotional semantic associations and prolonged REM sleep-dependent emotional priming

Highlights

• Nightmare (NM) sufferers have uncommon emotional semantic associations.

• Rapid eye movement (REM) naps lead to restricted breadth of associations for primed negative cue words.

• REM naps lead to increased breadth of associations for primed positive cue words.

• After 1 week, the healthy control (CTL) group no longer showed altered semantic access for primed words.

• NM sufferers retain their priming effect at the 1-week retest.

Abstract

Study Objectives

The objective of this study was to investigate whether nightmare (NM) sufferers exhibit an abnormal network of emotional semantic associations as measured by a recently developed, rapid eye movement (REM) sleep-sensitive, associational breadth (AB) task.

Design

NM sufferers were compared to healthy controls (CTL) for their performance on an emotional AB task containing positive and negative cue words both before and after a nap with REM sleep. AB was assessed in both a priming condition, where cue words were explicitly memorized before sleep, and a non-priming condition, where cue words were not memorized. Performance was assessed again 1 week later.

Setting

The study was conducted in a sleep laboratory with polysomnographic recording at the Hôpital du Sacré-Coeur de Montréal

Participants

Twenty-eight participants between the ages of 18 and 35 years (M_age = 23.3 ± 3.4) were included in the study.

Measurements and Results

The NM group scored higher than the CTL group on both positive and negative AB, with group differences persisting at the 1-week retest. However, the two groups did not differ as expected in the AB priming effect following REM sleep. Both groups showed decreased REM sleep-related AB priming for negative cue words and increased AB priming for positive cue words. However, the NM group maintained these effects 1 week later, whereas the CTL group did not.

Conclusions

NM sufferers may access broader than normal emotional semantic networks in the wake state, a difference that may lead to this group being perceived as more creative. The fact that the AB priming effect is maintained at the 1-week retest for NM sufferers suggests that the presence of frequent NMs may alter REM sleep-dependent emotional processes over time.

Introduction

Nightmares (NMs) are powerful unpleasant dreams associated with feelings of threat, anxiety, fear, or other negative emotions that are clearly recalled upon awakening and that arise primarily during late-night rapid eye movement (REM) sleep [1]. Individuals suffering from frequent idiopathic NMs may have disturbed sleep patterns, both during the NM experience itself and potentially in its absence [2], [3], [4], [5], [6]. For example, during an ongoing NM, participants may experience increased heart rate (HR), eye movements, and shortened breath [3], [7]; REM periods not marked by NMs have also shown altered structure, including increased REM latency, or increased high alpha spectral power (10–14.5 Hz) [8], [9]. Such disturbances may be disruptive to emotional regulation mechanisms provided by sleep.

REM sleep in particular is theorized to regulate emotion via a reduction in amygdala activation, while simultaneously improving the cognitive processing of emotion through increased functional medial prefrontal cortex (mPFC) connectivity [10]. Accordingly, sleep loss leads to increased emotional reactivity as measured both physiologically and behaviorally, as well as to decreased outward expression and experience of emotion by sleep-deprived individuals [11], [12]. This pattern mirrors clinical findings that patients with frequent NMs suffer from both hyperarousal and alexithymia, a deficit in cognitive labeling of emotion [13]. Further, NM sufferers are prone to affective disorders, including anxiety and depression, implicating a relationship between NMs and dysfunctional emotion processing [14]. NM sufferers also show increased cognitive perseveration in a word fluency task and perform poorly in an emotional Stroop task [4], [15]. Thus, frequent NMs may contribute to deficits in both the affective and cognitive domains, although more research is warranted to assess emotional cognition specifically.

REM-dependent emotional memory consolidation in the NM population has been largely unexplored. One generally supported theory is that REM sleep functions to adaptively integrate emotional experiences into the vast autobiographical network. The unique neurophysiological state of REM sleep allows the activation of emotional memory traces within an environment of increased associative cortical connections, thus promoting emotional memory integration [16], [17], [18], [19], [20]. Substantial research supports the role of REM sleep in emotional memory consolidation, such as in the consolidation of fear and safety memories or of the negative component of complex pictures [2], [18]. Further, behavioral experiments support an associative function for REM sleep, showing increased associative capacity immediately upon awakening from REM sleep and improved performance on associative tasks that had been primed before a REM sleep nap [21], [22], [23]. Both of these findings support claims that REM sleep enables enhanced access to associative semantic content. Notwithstanding such findings – and theoretical speculations on an emotional integration function of REM sleep – the interaction between emotional memory and associative access has not been studied in depth.

Our recent study of healthy college students specifically assessed the effects of REM sleep on the associative integration of emotional semantic stimuli, as measured by an associational breadth (AB) task [24]. Following a morning nap, participants were asked to provide word associations to emotional cue words they had memorized before sleeping. Participants who had REM sleep during their nap gave less common word associations (scored relative to word associate norms) [25] than did participants who had only non-rapid eye movement (NREM) sleep during their nap or who stayed awake. This finding suggests that the emotional words studied before sleeping were consolidated within a broad semantic network exclusively during REM sleep. Further, this priming effect was particularly strong for positive, compared to negative, cue words. The positively valenced cue words were considered to stimulate an increased spread of activation in REM sleep, similar to an effect observed in waking-state priming studies. Specifically, induction of a positive mood facilitates spread of activation [26], improves success with Remote Associates Test solutions [27], [28], [29], and increases the uniqueness and diversity of word associations [30]. Further, positive stimuli promote faster speed of semantic access [31]. Thus, positive cue words in our study may have led to increased breadth and speed of spreading activation in REM sleep.

These findings support the validity of the emotional AB task as being sensitive to both REM sleep and the emotional valence of experimental stimuli. Thus, it is an appropriate task for assessing REM sleep-related emotional processes among frequent NM sufferers. However, the available literature points to two distinct hypotheses about whether NM sufferers are expected to show a restricted or an expanded breadth of emotional semantic associations.

The most important symptom of the NM sufferer, the NM experience itself, is composed of rather repetitive and perseverative content, suggesting that NM psychopathology is characterized by restricted emotional semantic access. The NM experience has vivid sensory imagery and intense emotional expression. The NM often depicts an unrelenting threat, whether being pursued by an aggressor, an out-of-control car, or an imminent tidal wave. The NM's main theme and imagery seem to grow more potent and imposing over time, with an increase in emotional arousal, and a resistance to the associative fluidity that normally permeates dreams [32]. This characterization of NMs as being associationally restricted is in line with findings in the literature on waking cognition, which shows that positive emotion increases associative access while negative emotion restricts and slows it [33], [34]. Thus, the NM may reflect a temporary failure of REM processes to integrate a dysphoric emotional memory. Given this possibility, we would expect the presence of NMs to correlate with reduced AB and longer associational reaction times (RTs) on the AB task and a blunting of REM sleep-dependent emotional priming.

A second possible explanation of NM pathology is that frequent NM sufferers are characterized by a broader than normal access to emotional semantic networks. NM sufferers report higher than average recall of positively toned dreams, e.g., non-NM dreams and other “intensified” dreams [35], [36] such as lucid dreams that contain self-reflective awareness and subtle kinesthetic imagery, and archetypal dreams that contain blissful emotion and, often, spiritual encounters. Even in the waking state, NM participants report more bizarreness in their daydreams – bizarreness has been likened to broad semantic access, as in the AB task [22], [24]. NM sufferers are also characterized by “thin boundaries,” a personality construct that includes creativity and artistic expression, both of which seem to draw more flexibly and frequently upon unusual associations [32]. Together, such findings support the expectation that NM sufferers may demonstrate an expansion of associational processing relative to control subjects.

In sum, while research supports the role of REM sleep in the associative integration of emotional memory, any potential effect of frequent NMs on this process is unknown. On the one hand, the NM experience seems to reflect associative restriction imposed by intense negative emotion; on the other hand, NM sufferers often report positive and bizarre dreams, and they are frequently characterized as being creative, artistic individuals. Assessment of their performance on an emotional AB task both before and after REM sleep will provide further insight into waking and REM sleep-dependent processes.

To determine whether the presence of persistent NMs is associated with either restriction or broadening of access to emotional semantic networks, we used a nap protocol and a recently developed task for assessing semantic AB in response to negative and positive emotional words [24]. Our objective was to assess baseline and REM sleep-dependent changes in emotional word associations among frequent NM sufferers and controls (CTL). We also assessed whether group differences were maintained at a 1-week follow-up.

Hypotheses: Current evidence is scarce, although it tends to suggest that NM sufferers will show restricted associative access, particularly in response to negative cue words, on an emotional AB task both without and with REM-sleep dependent priming: 1) lower AB scores, particularly for negative cue words; 2) slower RTs for associational responses; 3) lower scores on the REM sleep-dependent Priming Effect, particularly for negative cue words; and 4) slower RTs for Priming Effect responses.

All effects are expected to be maintained in NM sufferers after a 1-week delay.