Did Dreams Once Assist Survival? Investigating
Dream-Dependent Fear Memory Consolidation as
Part of the Threat Simulation Theory
Alex van Keeken
Student number: 11689919
Supervisor: Elsa Juan
Final thesis Psychobiology (BSc)
University of Amsterdam
19-06-2020
ABSTRACT
The role of dreaming in emotional memory consolidation remains unclear. Some evidence suggests that dreams promote emotional regulation. In this line, the threat simulation theory (TST) claims that dreams serve the evolutionary purpose of replaying threatening
experiences in a virtual environment, where threat-avoidance behaviour can be rehearsed safely. The current study aimed to test the theory’s implicit prediction that fear memory consolidation takes place as a result of threatening dream experiences, and whether these dream experiences are accompanied by emotional regulation. To assess this, memory performance and affective valence ratings of both fear-associated memory items and safety-associated memory items were tested before and after a night of sleep. A serial awakening paradigm was used during sleep to collect dream reports. Results show that there was no difference in performance improvement on the fear-associated memory items between participants who reported threatening dream experiences and participants who did not. Also, affective valence ratings became more positive after sleep for both groups. These findings conflict with the TST, because they suggest that threatening dream experiences do not reinforce fear memory consolidation and such dreams are not required for emotional regulation occur.
Keywords: dreaming, fear memory consolidation, emotional regulation, threat simulation
theory
1. Introduction
Dreams are conscious mental representations during sleep, consisting of self-generated cognitive, emotional and sensory experiences (Desseilles, Dang-Vu, Sterpenich, & Schwartz, 2011). While the occurrence of dreams was initially thought to be limited to Rapid Eye Movement (REM) sleep (Hobson, Stickgold, & Pace-Schott, 1998), it is now widely accepted that they also occur during non-REM (NREM) sleep (Foulkes, 1962; Antrobus, 1983;
Wamsley & Stickgold, 2019). Yet, a distinction is made between REM and NREM dreams, as the former are typically more vivid, bizarre and emotionally rich than the latter, which have a more thought-like nature (Foulkes, 1962; Merritt, Stickgold, Pace-Schott, Williams, &
Hobson, 1994). Interestingly, dream disturbances are implicated in a range of psychiatric disorders. For instance, recurring nightmares are one of the defining criteria of posttraumatic stress disorder (PTSD). Another primary characteristic of PTSD is the inability to extinguish trauma-related memories (American Psychiatric Association, 2013). Although no causal relation between these symptoms has yet been established, it is possible that trauma-related nightmares reinforce the memory of the traumatic event, contributing to intrusive
re-experiences in daily life. This idea, among many others, has inspiredinvestigations into the fundamental relationship between dream experience and memory processes.
Sleep facilitates memory consolidation processes, possibly through the reactivation of existing memory traces (Pavlides & Winson, 1989; Ego‐Stengel & Wilson, 2010; Schreiner & Rasch, 2015). Dreams are thought to be conscious reflections of this reactivation mechanism and therefore to contribute to the consolidation of newly formed memories (Payne & Nadel, 2004). This idea is supported by recent empirical evidence, demonstrating that dreaming of a spatial learning task (Wamsley & Stickgold, 2019) or a word-picture association learning task (Schoch, Cordi, Schredl, & Rasch, 2019) during following sleep improves performance on these tasks the next morning. Additionally, post-sleep memory performance improvement can be predicted by the extent to which elements of learning tasks are incorporated into dream content (Fogel, Ray, Sergeeva, De Koninck, & Owen, 2018). It is important to note that components of this word-picture association task were successfully incorporated into dreams when dream reports were collected over multiple awakenings at night, but not when dreams were only collected in the morning (Schoch et al., 2019), suggesting that a serial
awakening paradigm shows links between dream content and memory consolidation processes that a single awakening in the morning does not.
While more and more evidence suggests that dreaming facilitates declarative memory consolidation, its role in emotional memory consolidation remains unexplored. Numerous studies do point out that dreams generally include more negative than positive emotions, with fear as the most frequently reported emotion (Nielsen, Deslauriers, & Baylor, 1991; Merritt et al., 1994). In addition, elements of memories created during wake are often incorporated into dreams (Stickgold, Hobson, Fosse, & Fosse, 2001) and, independent of the type of emotion, the emotional intensity of these elements is lower when they are part of dream content compared with their original emotional intensity during wake (Vallat, Chatard, Blagrove, & Ruby, 2017). Also, people who report a predominance of fear-related emotions in dreams show smaller physiological and neurobiological fear responses to aversive faces during wake than individuals who report less fear in dream content (Sterpenich, Perogamvros, Tononi, & Schwartz, 2020). Together, this evidence supports the claim that dreaming serves a purpose of emotional regulation through re-exposure to, mostly negative, emotions (Scarpelli,
Bartolacci, D’Atri, Gorgoni, & Gennaro, 2019).
In that vein, Revonsuo (2000) proposed the threat simulation theory (TST),
hypothesizing that dreams have the adaptive function of replaying threatening life events in order to practice threat-avoidance in a safe environment. In prehistoric times, this would increase the organism's chance at survival and hereby reproduction when faced with future threats. The proposed mechanism behind this theory is that threatening experiences during wake prompt the activation of the “threat simulation system” (TSS), which in turn retrieves the memory of the experience for threat simulation during dreams. The findings above, pointing out that dream content typically contains wake-related elements and predominantly includes negative emotions, thus support this theory. One of the central predictions the TST makes is that threatening life events modulate following dream content. Posttraumatic nightmares are clear evidence of this. Not only do traumatized people report more dreams containing threatening events than non-traumatized people, but the type of threats they report are also more severe (Valli, Revonsuo, Pälkäs, & Punamäki, 2006). In addition, the TST proposes that the ability to select mnemonic traces of threatening life experiences for simulations in dreams is genetically encoded in order for it to be passed on to later
generations. The heritability of having a disposition to nightmares (Hublin, Kaprio, Partinen, & Koskenvuo, 1999) supports this proposal. Moreover, the TST predicts that dreams are biased towards including threatening events as they occurred in ancestral survival, such as being chased by a predator or by a member from a hostile tribe. A study among
undergraduate university students in Canada indeed showed that being chased or pursued was the most common theme within dream experiences, as it was reported by over 80% of the students (Germain, Nielsen, Zadra, & Montplaisir, 2000).
However, other empirical evidence conflicts with the TST. For instance, additional common themes in dream content, such as sexual experiences (76%) or flying (50%), involve positive emotions and do not directly equip the dreamer with an environment to rehearse threat-avoidance (Germain et al., 2000). Also, a disproportionate amount of negative emotions (Nielsen et al., 1991; Merritt et al., 1994) and wake-related elements (Stickgold et al., 2001) in dream reports does not necessarily imply that they are truly overrepresented in dream experience. Instead, they might simply be reported more
frequently due to better memory for these emotions and elements upon awakening (Voss & Klimke, 2018). Furthermore, people who regularly experience nightmares are at greater risk for developing PTSD following trauma exposure (Mellman, David, Kulick-Bell, Hebding, & Nolan, 1995) and, conversely, experiencing nightmares after trauma exposure generally relates to a more severe clinical picture of PTSD (Mellman, David, Bustamante, Torres, & Fins, 2001). So, in the case of PTSD, replaying threatening experiences during dreams does not appear to be functional nor adaptive in any sense.
Clearly, some controversy exists about dream function in the way that it is explained in the TST. Interestingly, the memory aspect of this theory has neither been supported nor disproved by empirical evidence so far. An implicit prediction the TST makes is that
dream-dependent fear memory consolidation takes place during sleep, as memories that encode fearful events are reactivated and replayed for threat rehearsal in dreams. Whereas sleep in general is known to consolidate fear memories (Menz et al., 2013), this has never been linked to dreams specifically, let alone to dreams involving fear. Therefore, in an attempt to further define the relationship between dreaming and emotional memory consolidation and to simultaneously explore the validity of the TST, this study aims to investigate the influence of threatening dream experiences on fear memory consolidation and emotional regulation during sleep. Consistent with the TST, it is hypothesized that threatening dream experiences reinforce the consolidation of fear memories. Also, since a high prevalence of fear in dream experience is associated with reduced fear responses to fear-evoking stimuli during wake (Sterpenich et al., 2020), threatening dream experiences are expected to be accompanied by a form of emotional regulation across sleep.
In order to test these hypotheses, a feeling of fear or safety was assigned to specific memory items during a memory encoding task. From now on, these memory items will be referred to as negative and positive memory items, respectively. Participants performed a recognition task before and after a night of sleep. Pre-sleep and post-sleep affective valence ratings of all memory items were obtained. During sleep, a serial awakening paradigm was used to collect multiple dream reports from all participants. Overnight memory performance improvement on the negative memory items was compared between participants who reported having experienced a threat during one or more dreams and participants who did not report any threatening dream experiences. In line with the TST, it is expected that participants with threatening dream reports will show a greater overnight memory performance improvement for the negative memory items than participants without. Moreover, in participants with threatening dream experiences, emotional regulation as described by Sterpenich et al. (2020) is expected to be demonstrated by a larger increase in affective valence ratings of the negative memory items across sleep than in participants without.
2. Methods
2.1. Participants
An international sample of 40healthy first year psychology students from the University of Amsterdam (28 female, 11 male, 1 other), aged between 18 and 36 years (M ± SD: 20.55 ± 3.30), was used for this study. Participants were included if they met the inclusion criteria (Appendix A) based on self-report. After inclusion, the data of 5participants were excluded, because they did not complete the entire experiment (n=2) or because part of their data were outliers exceeding the limits of mean ± 3 SD (n=3). All participants were instructed to wake up at 08:00, to not drink any caffeine-containing beverages as of 14:00 or take naps on the day of the experiment. In addition, they were instructed to abstain from alcohol and drugs within 24 hours before the start of the experimental night and to sleep alone during the experimental night. Violation of these instructions resulted in a rescheduling of the experimental procedure. An online informed consent was obtained from all participants. Participants were rewarded research credits upon completion of the full experiment. This study was approved by the Ethics Review Board of the Department of Psychology, University of Amsterdam.
2.2. Computer tasks
All surveys and computer tasks were created and carried out on the online survey platform Qualtrics. As participation was home-based due to the COVID-19 outbreak, all tasks were carried out on participants’ personal computers with screen sizes ranging between 11 and 21.5 inches and screen resolutions between 1366x768 and 3200x1800 pixels.
The first computer task consisted of watching two movie scenes, one emotionally negative and one emotionally positive, presented in a counterbalanced order across
participants. The negative scene was an extract from “American History X” (Morrissey & Kaye, 1998), showing a neo-Nazi violently responding to a robbery and murdering an
African-American man by forcing him to place his mouth on a curb before smashing his head into it. This scene was chosen because a previous study showed that it induced the most negative affect among nine other distressing movie scenes from different movies (Schaefer, Nils, Sanchez, & Pierre, 2010). The positive scene was an extract from “March of the
Penguins” (Darondeau, Lioud, Priou, & Jacquet, 2005), showing how Emperor penguin chicks in an Antarctic environment take their first steps after being carried on their mothers’ feet. The scenes were 2:51 minutes and 2:49 minutes long, respectively. To ensure that participants payed attention to the movie scenes, they were asked to give a short, written description. Finally, participants rated each scene on affective valence and arousal.
The memory encoding task (Fig. 1) always followed this part. During encoding, 40 images of different butterfly sketches (Cramer, 1780) were presented, which participants were explicitly instructed to memorize. All butterfly images used in the current study were selected from a pilot study (Appendix B), where three groups with similar arousal and affective valence ratings were formed. During the encoding task, each butterfly image was preceded by a 2000-ms fixation cross, a 100-ms tone and a 3000-ms movie excerpt from one of the two movie scenes and followed by an affective valence rating of the image. There were 20 negative and 20 positive trials. In negative trials, tone A (1000 Hz) and an excerpt from the negative movie scene, showing the moment where the victim places his mouth on the curb while screaming, were presented. This specific excerpt was chosen, because it was thought to be the most distressing moment in anticipation of the brutal murder. Positive trials contained tone B (1500 Hz) and an excerpt from the positive movie scene. The positive excerpt was selected based on its pleasant view and calm audio, showing how three penguin chicks walk across the Antarctic landscape. The tones and excerpts were randomly
counterbalanced across participants.
The pre-sleep recognition task (Fig. 2) consisted of 80 trials, presented in a random order. All 40 butterfly images from the encoding task (“memory items”) were presented, along with 40 other butterfly images that had not been presented before (“foils”). Each image was presented for 3000 ms. After viewing each image, participants were asked whether they had seen it before in the memory encoding part, and to provide a confidence rating of their answer.
The post-sleep recognition task (Fig. 3) was similar, using 40 new foils to avoid any confusion with the previous foils. This time, an affective valence rating followed the
confidence rating. In this way, the overnight development of butterfly valence could be examined.
Figure 1. Schematic overview of the memory encoding task. During negative trials, tone A,
the negative movie excerpt and a memory item were shown consecutively, followed by an affective valence rating. Positive encoding trials were identical, except tone B was presented instead of tone A and the positive excerpt was shown instead of the negative.
Figure 2. Schematic overview of the pre-sleep recognition task. Each trial consisted of a
memory item, a recognition question and a confidence rating.
Figure 3. Schematic overview of the post-sleep recognition task. During each trial, a memory
item, a recognition question, a confidence rating and an affective valence rating of the memory item were presented.
2.3. Procedure
Participants started with the online surveys and computer tasks around 21:30 on their personal computers, in a quiet room with a strong internet connection. First, participants completed a Positive and Negative Affect Schedule (PANAS; Watson, Clark & Tellegen, 1988) to obtain a baseline of their mood. Then, they performed the movie watching task. Participants were instructed to watch the movie scenes on full screen mode, without looking away or closing their eyes and to immerse themselves into the scenes. In this way, the scenes were thought to induce the most emotion as possible. Next, participants continued with the memory encoding task, where they were explicitly instructed to remember the butterflies and informed that their memory would be tested at a later stage. They were also instructed to pay attention to the tone-excerpt association. To check whether participants had learned the association, they were asked to indicate which excerpt followed which tone after the encoding task. This was followed by a break of up to 10 minutes. Participants then completed a Pittsburgh Sleep Quality Index (PSQI; Buysse, Reynolds, Monk, Berman, & Kupfer, 1989) to assess their general sleep quality, a general dream characteristics survey (based on Schredl, Berres, Klingauf, Schellhaas, & Göritz, 2014) a Beck Depression Inventory-II (BDI-II; Beck, Steer, & Brown, 1996) and a State Trait Anxiety Inventory – Trait scale (STAI-T; Spielberger, Gorsuch, & Lushene, 1970). After the STAI-T, participants performed the pre-sleep recognition task. The final step of the evening program was completing a second PANAS to assess any change of mood after completing the tasks. Altogether, this evening procedure took approximately an hour.
Participants were instructed to go to sleep at 23:00, providing them with a 9-hour sleep opportunity window. Starting at 00:00, they were woken up every hour by a telephone call until 08:00, resulting in a total of nine awakenings. During each awakening, participants answered questions from the “Dream report questionnaire” (Appendix C) to collect all the dreams, thoughts or feelings or they might have had before the telephone call. Participants’ answers to the “Dream report questionnaire” were recorded. Instead of collecting dream reports during a single awakening in the morning, a serial awakening paradigm was used to increase the amount and the accuracy of dream reports. A pleasant ringtone was installed on participants’ phones in preparation for the experimental night to ensure a smooth awakening.
After the final awakening at 08:00, participants had 30 min to recover from sleep inertia. At 08:30, they started with the morning tasks. Participants first completed the PANAS survey again – to detect any change of mood after sleep – followed by a survey assessing their sleep quality of the previous night. Next, they performed the post-sleep recognition task and the movie watching task. Finally, they completed an exit survey about their overall experiences with the experiment, where they were also able to report dreams of the previous night that they only remembered after the last awakening. The morning procedure took about 30 minutes to complete.
2.4. Data analysis
Only the dream reports, the answers to the recognition questions of the recognition tasks and the pre-sleep and post-sleep valence ratings of the negative memory items were analyzed to test the current hypotheses.
Participants were divided into two groups, based on their answer to the question “Did you experience a threat or violence to you or to others?” in the “Dream report questionnaire”. Participants who reported at least one threatening experience as part of a dream during any of the awakenings were assigned to the “Threat” group (T) and participants who did not formed the “Non-Threat” group (NT).
Memory performance was assessed by calculating the d-prime (d’), a measure for sensitivity that is commonly used to accurately assess memory performance in a recognition task (Bayley, Wixted, Hopkins & Squire, 2008; Hamilton & Gotlib, 2008; Wynn, Kessels & Schutter, 2020). The d’ was quantified by subtracting the Z-score corresponding to the false alarm rate from the Z-score corresponding to hit rat, where the false alarm rate is defined as the amount of wrongly identified foils divided by the total amount of foils and the hit rate as the amount of correctly identified memory items divided by the total amount of memory items (Macmillan, 1993).
The overnight difference in memory performance of the negative memory items was compared between T and NT by first subtracting participants’ d’-scores of the pre-sleep recognition task from the d’-scores of the post-sleep recognition task for both groups. Next, the assumptions of normality within the groups and equal variances between the groups were tested with Shapiro-Wilk tests and an F-test, respectively. A non-parametric Welch Two Samples t-test was carried to directly compare the overnight difference scores of the
negative memory items between the two groups.
A similar approach was used when comparing valence ratings between T and NT. After testing the assumptions above with the mentioned tests, a parametric Two Sample t-test was used to compare the overnight difference in valence ratings of the negative memory items between the two groups. All statistical tests were performed in R-Studio version
1.1.383 and a significance level of α = 0.05 was applied.
3. Results
Out of all 315 planned awakenings (35 participants x 9 awakenings), 7 were not carried out because participants slept through the telephone call (n=2), the experimenter decided to not call because participants were awake during the three preceding awakenings (n=1) or because of procedural errors (n=4). Among the remaining 308 awakenings, 161 dream reports were collected (see Appendix D for dream report examples). Seven additional dream reports were obtained from the exit survey. Participants reported 4.60 ± 2.24 dreams on average (Fig. 4A). Out of all dream reports, 21 reports across 14 participants contained an experienced threat (Fig. 4B). Groups “Threat” (T; n=14) and “Non-Threat” (NT; n=21) were formed accordingly.
A.
B.
Figure 4. (A). General dream report frequency
(n=35). The number of dream reports per participant ranged from 1 to 10 during the experimental night. Most participants reported 5 dreams. (B). Threat dream report frequency (n=35). Seven participants reported one threatening dream experience and 7 reported two. These 14 participants formed the T group. The remaining 21 participants (NT) did not report any threatening dream experiences. 3.2. Memory performance
Calculating the d’-scores required correcting for several cases with 0 false alarms before quantifying the hit and false alarm rates. The correction entailed adding 0.5 to the number of hits and false alarms for all participants and adding 1 to the number of memory item and foil trials (Hautus, 1995).
Overall, performance on the post-sleep recognition task (d’ = 1.21 ± 0.52) was considerably higher than on the pre-sleep task (d’ = 0.86 ± 0.36). This was tested with a Paired Student’s t-test (n = 35, t(34) = - 4.106, p < 0.001) after confirming that the assumption
of normality was met (W = 0.939, p = 0.05).
Overnight memory performance improvement on the negative memory items was compared between T and NT. The assumption of normality was met for T (W = 0.964, p = 0.79) and NT (W = 0.947, p = 0.30), but the assumption of equal variances of the two groups was not (F(13,20) = 2.854, p = 0.034). Therefore, a non-parametric Welch Two Sample t-test
was used, revealing no significant differences between T (median[IQR]: 0.23[-0.21 0.63]) and NT (0.29[-0.09 0.59]; t(19.105) = 0.347, p = 0.73; Fig. 5).
In order to explain this absence of difference, additional statistical tests were performed. First, pre-sleep and post-sleep memory performance on the negative memory items were compared within both groups to investigate how fear memory evolved overnight. After confirming that the assumption of normality was met in both cases ([T: W = 0.964, p = 0.79]; [NT: W = 0.947, p = 0.30]), Paired Student’s t-tests were done for T ([pre-sleep: d’ = 0.95 ± 0.44]; [post-sleep: d’ = 1.16 ± 0.68]) and NT ([pre-sleep: d’ = 0.77 ± 0.39]; [post-sleep d’ = 1.05 ± 0.44]; Fig. 6). The results showed no significant differences between pre-sleep and post-sleep performance for T (t(13) = -1.262, p = 0.23), whereas post-sleep performance
was significantly higher than pre-sleep performance for NT (t(20) = -3.408, p = 0.003).
To examine whether overnight memory performance improvement of the NT group was specific to the negative memory items or of it was a general improvement, pre-sleep performance on all memory items was compared with post-sleep performance for NT. The assumption of normality was not met (W = 0.904, p = 0.041), so a Wilcoxon Signed Rank test was used, showing that post-sleep performance (d’ = 1.13 ± 0.37) was significantly higher than pre-sleep performance (d’ = 0.80 ± 0.28; V = 25, p < 0.001). For T, there was no
0 1 2 0 5 10 15 20 25 21 7 7
Number of dream reports containing a threat
N um be r o f p a rt ic ip a nt s 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 3 3 4 6 8 3 3 2 2 1
Number of dream reports
N um be r o f p a rt ic ip a nt s
significant difference between pre-sleep (d’ = 0.94 ± 0.46) and post-sleep (d’ = 1.24 ± 0.68) performance on all items (W = 0.889, p = 0.08; t(13) = -1.746, p = 0.10).
Finally, T was divided into T1 (n = 7) and T2 (n = 7). T1 contained the participants that reported one threat dream and T2 those that reported two. This distinction was made to investigate whether fear memory consolidation differed between NT and T2, since the difference in fear memory consolidation was thought to be the largest between these groups. Overnight difference scores in performance on the negative memory items were compared between NT (Δd’ = 0.28 ± 0.37) and T2 (Δd’ = 0.20 ± 0.60). The assumption of normality was met for both groups ([NT: W = 0.947, p = 0.30]; [T2: W = 0.919, p = 0.47]) and the variances were equal (F(20, 6) = 0.390, p = 0.10). A Two Sample t-test then revealed no significant
differences between NT and T2 (t(26) = 0.42, p = 0.68).
Figure 5. Boxplot of the overnight memory performance improvement on the negative
memory items for T (n=14) and NT (n=21). Performance improvement was calculated by subtracting the pre-sleep d’ scores from post-sleep d’ scores. Negative Δd’ scores indicate performance decline across sleep. The whiskers (dotted vertical lines) show the variance of the data. There was no significant difference in performance improvement between the two groups (N.S.).
Figure 6. Bar graph of the average memory performance on the negative memory items for
T and NT. Error bars represent the standard deviations. Post-sleep performance was significantly higher than pre-sleep performance for NT (**), but not for T (N.S.). 3.3. Affective valence ratings of the memory items
The pre-sleep affective valence ratings of the negative and positive memory items were compared among all participants (n = 35) to confirm that the emotional manipulation of the memory items was successful. The assumption of normality was violated (W = 0.87, p < 0.001), so a Wilcoxon Signed Rank test was used. It revealed that the negative memory items (37.91 ± 12.70) had a significantly lower pre-sleep affective valence rating than the positive memory items (55.55 ± 10.54; V = 620, p < 0.001), demonstrating that the emotional manipulation worked as intended.
The overnight difference scores of the affective valence ratings on the negative memory items were compared between the two groups. The difference scores of T (W = 0.909, p = 0.15) and NT (W = 0.938, p = 0.20) followed a normal distribution and their variances were equal (F(13, 20) = 2.207, p = 0.11). A Two Sample t-test then revealed that
there were no significant differences between T (12.3[8.48 21.49]) and NT (8.65[2.20 13.40]; [t(33) = 1.408, p = 0.17]; Fig 7.).
Next, pre-sleep ratings of the negative memory items were compared with post-sleep ratings for T and NT to examine whether sleep-dependent emotional regulation occurred at all. For both groups, the assumption of normality was met ([T: W = 0.909, p = 0.15]; [NT: W = 0.938, p = 0.19]) and Paired Student’s t-tests showed that post-sleep valence ratings ([T: 49.39 ± 8.97]; [NT: 50.16 ± 6.26]) were significantly higher than pre-sleep ratings ([T: 33.48 ± 15.44; t(13) = -3.574, p = 0.003]; [NT: 40.87 ± 9.82; t(20) = -3.799, p = 0.001]; Fig. 8).
The same division as previously described was made within T to compare the overnight difference scores on the negative memory items between NT and T2. The assumption of normality was met for the difference scores of NT (W = 0.938, p = 0.19), but not for those of T2 (W = 0.805, p = 0.046). The variances between the two groups were equal (F(20, 6) = 0.485, p = 0.21). A Mann-Whitney U test then showed no significant
differences between NT (9.30 ± 11.22) and T2 (17.64 ± 16.10; W = 46, p = 0.16).
Figure 7. Boxplot of the overnight difference in affective valence ratings of the negative
pre-sleep ratings from post-pre-sleep ratings. The dots represent outliers. No significant difference was found between T and NT overnight difference ratings (N.S.).
Figure 8. Bar graph of the average affective valence ratings of the negative memory items
for T and NT. A rating of 0 indicates extremely negative, 50 indicates neutral and 100
extremely positive. Post-sleep ratings were significantly higher than pre-sleep ratings for both T and NT (**).
3.4. Affective valence and arousal ratings of the negative movie scene.
Apart from examining the affective valence ratings of the negative memory items, some extra analyses were done on the affective valence and arousal ratings of the negative movie scene to acquire additional information about emotional regulation in T and NT (Appendix E). The main findings were that post-sleep affective valence ratings (11.24 ± 12.06) were significantly higher than pre-sleep ratings (7.33 ± 8.87) for NT (t(20) = 2.087, p = 0.049), but not for T
sleep: 3.79 ± 6.53]; [post-sleep: 8.79 ± 11.14]; V = 26, p = 0.05) and that neither T ([pre-sleep: 73.14 ± 25.34]; [post-([pre-sleep: 74.79 ± 15.72]; t(13) = 0.230, p = 0.82) nor NT ([pre-sleep:
76.19 ± 14.15]; [post-sleep: 73.48 ± 13.54]; t(20) = -0.791, p = 0.44) showed an overnight
change in arousal ratings.
4. Discussion
The aim of this study was to investigate the relation between dream content and emotional memory consolidation by providing evidence for the TST. We hypothesized that threatening dream experiences reinforce fear memory consolidation during sleep and that these dreams are accompanied by a form of emotional regulation. Contrary to our expectations, there was no difference in memory performance improvement on the negative memory items between participants who had threatening dreams and participants who did not. Further investigations even showed that performance on these and all items improved after sleep for participants without threatening dreams, but not for participants with. Affective valence ratings of the negative memory items became equally more positive after sleep for both groups. Together, these results suggest that threatening dream experiences do not reinforce fear memory consolidation and that emotional regulation occurs independently of these dreams.
The unexpected overnight performance improvement on the negative memory items for participants who did not experience a threat in their dreams can be explained by the general performance improvement of this group. However, why an absence of threatening dream experiences is related to overall memory performance improvement remains unclear.
A possible explanation is that the dreams of this group were characterized by lower levels of distress, which may have resulted in a higher sleep quality. Sleep of high quality is indeed known to facilitate memory consolidation processes in general (Vyazovskiy, 2015) and promote recognition memory specifically (Monaghan, Shaw, Ashworth-Lord, & Newbury, 2017). So, a difference in sleep quality between the two groups could account for the difference in overall memory performance improvement. Unfortunately, the scope of this study did not allow for an electroencephalography (EEG) spectral analysis, which would have been the most accurate way to measure and compare sleep quality.
A methodological explanation for the absence of fear memory consolidation in participants with threatening dream experiences could be that groups T and NT were not formed with full accuracy with regard to the TST. This has several causes. First of all, no distinction was made between objective and subjective threats. For instance, some
participants reported that one of their dreams was threatening, while their description of the dream did not directly imply that it included any form of danger (Appendix D, “Threat dream 2”). A more objective classification process could have resulted in a cleaner T group,
exclusively formed by participants who experienced a direct danger of being harmed, ideally accompanied by a feeling of fear and/or anxiety. Nevertheless, we chose to assign
participants who reported a subjective threat to T, because participants were the only observers of their dreams, so they were in the best position to recognize a threat. Second, classification was based on participants’ answer to the question whether they experienced a threat in their dream, either to themselves or to others. Revonsuo (2000) does not mention anything about dreams being of adaptive value when the threats in it are directed at others, most likely because it does not offer the dream-self to rehearse threat avoidance.Adapting this question in a way that it only concerns threats to the dream-selves could therefore test the TST more directly. Participants who experienced threats to others (Appendix D, “Threat dream 3”) were nonetheless included in T, because we hypothesized that fear memory consolidation takes place regardless of whom the threat in the dream was directed at, since the memory of a threatening event is reactivated either way.Third, and most importantly, dream collection fully depended on participants’ memory recall and introspective abilities upon awakening. It is reasonable to assume that (parts of) dreams were forgotten or
unknowingly reconstructed when prompted to report them. This is a significant challenge that empirical research involving dreaming faces in general (Desseilles et al., 2011). In this study specifically, some participants might have forgotten a threatening dream experience upon awakening, while the consolidation of the corresponding fear memory did take place. We addressed this concern by collecting dream reports within an hour after dreams occurred, making the chances of distorting them upon report slim.
Another reason why threatening dream experiences did not benefit memory
performance on the negative memory items could have been that the negative memory items were not as emotionally charged as intended. However, pre-sleep affective valence ratings indicated that the emotional manipulation was completed successfully. Alternatively, the experimentally induced emotion might not have been fear, specifically. Whereas a previous study by Schaefer et al. (2010) and the obtained affective valence ratings indeed
demonstrate that the used negative movie scene induces strong negative affect, Schaefer et al. (2010) also highlight that this negative affect is mostly characterized by feelings of anger and disgust rather than fear.
Based on the overnight development of valence ratings of the negative memory items, the presence of a threat in dreams did not influence emotional regulation. These findings conflict with the formerly introduced claim that a function of dreaming is to regulate emotions through re-exposure (Scarpelli et al., 2019) and are hereby inconsistent with previous research demonstrating that fear-related emotions in dreams are correlated with lower fear responses during wake (Sterpenich et al., 2020). This inconsistency could be due to the different measures used to define emotional responses. Sterpenich et al. (2020) used pupillary responses as an index of arousal and measured insula activity, both of which are significantly less vulnerable to biases than self-reported affective valence ratings.
points in time are not directly comparable, due to the different natures of the tasks they were embedded in. On the one hand, pre-sleep ratings were obtained during the encoding task, where memory items were directly preceded by the movie excerpts for the purpose of
inducing emotions. On the other hand, post-sleep ratings were collected during a recognition task, where the movie excerpts were not presented again. So, rather than attributing the less emotionally loaded post-sleep valence ratings to sleep-related emotional regulation, they could be attributed to a lower emotional intensity of the post-sleep task.
The fact that threatening dream experiences did not facilitate memory consolidation of the negative memory items seems to contradict the TST’s prediction regarding
dream-dependent fear memory consolidation. Yet, it cannot be rejected without deeper inspection. While the TST merely suggests that threatening dreams strengthen the memories of
threatening experiences through reactivation, we assumed – perhaps wrongfully – that these dreams would at the same time promote the consolidation of other non-related fear
memories. Consequently, our experimental setup was suitable for testing associative fear memory in general, but not for testing the memory of specific threatening events.
So, based on this study, no predictions can be made about the influence of
posttraumatic nightmares on the memory of a traumatic experience. This is also because the vast majority of the reported threat dreams were not nightmares, which makes comparing their effects with those of posttraumatic nightmares tricky. Taking the limitations of the current study into account, future studies could investigate the same prediction of the TST in a more accurate way and further examine the influence of posttraumatic nightmares on intrusive trauma memories in PTSD patients. This could be done by correlating patients’ physiological fear responses to objective threats directed at themselves during nightmares with the amount of intrusive trauma memories and the intensity of physiological emotional responses to these memories during the following days.
Overall, our findings suggest that threatening dream experiences do not reinforce fear memory consolidation, nor do they promote emotional regulation more than non-threatening dream experiences do. These findings conflict with the TST and therefore challenge its validity. However, this theory, stating that dreams have an adaptive threat-rehearsal function, cannot be rejected based on this evidence. Future research is needed to further explore the TST and eventually the overall relation between dreaming and emotional memory.
References
American Psychiatric Association (2013). Diagnostic and Statistical Manual of Mental Disorders DMS V. Washington, DC: American Psychiatric Association.
Antrobus, J. (1983). REM and NREM sleep reports: Comparison of word frequencies by cognitive classes. Psychophysiology, 20(5), 562-568.
Bayley, P. J., Wixted, J. T., Hopkins, R. O., & Squire, L. R. (2008). Yes/no recognition, forced-choice recognition, and the human hippocampus. Journal of Cognitive Neuroscience, 20(3), 505-512.
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for Beck Depression Inventory-II. San Antonio, TX: Psychological Corporation.
Buysse, D. J., Reynolds, C. F., Monk, T. H., Berman, S. R., & Kupfer, D. J. (1989). The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Research, 28(2), 193-213.
Cramer, P. (1779). De uitlandsche kapellen voorkomende in de drie waereld-deelen Asia, Africa en America. Amsterdam: S.J. Baalde.
Darondeau, Y., Lioud, C., & Priou, E. (Producers), & Jacquet, L. (Director). (2005). March of the Penguins [Motion Picture]. United States: Warner Independent Pictures.
Desseilles, M., Dang-Vu, T. T., Sterpenich, V., & Schwartz, S. (2011). Cognitive and emotional processes during dreaming: a neuroimaging view. Consciousness and Cognition, 20(4), 998-1008.
Fogel, S. M., Ray, L. B., Sergeeva, V., De Koninck, J., & Owen, A. M. (2018). A novel approach to dream content analysis reveals links between learning-related dream incorporation and cognitive abilities. Frontiers in Psychology, 9, 1398.
Foulkes, W. D. (1962). Dream reports from different stages of sleep. The Journal of Abnormal and Social Psychology, 65(1), 14.
Germain, A., Nielsen, T. A., Zadra, A., & Montplaisir, J. (2000). The prevalence of typical dream themes challenges the specificity of the threat simulation theory. Behavioral and Brain Sciences, 23(6), 940-941.
Hamilton, J. P., & Gotlib, I. H. (2008). Neural substrates of increased memory sensitivity for negative stimuli in major depression. Biological Psychiatry, 63(12), 1155-1162.
Hautus, M. J. (1995). Corrections for extreme proportions and their biasing effects on
estimated values of d′. Behavior Research Methods, Instruments, & Computers, 27(1), 46-51.
Hobson, J. A., Stickgold, R., & Pace-Schott, E. F. (1998). The neuropsychology of REM sleep dreaming. Neuroreport, 9(3), R1-R14.
Hublin, C., Kaprio, J., Partinen, M., & Koskenvuo, M. (1999). Nightmares: familial aggregation and association with psychiatric disorders in a nationwide twin cohort. American Journal of Medical Genetics, 88(4), 329-336.
Macmillan, N.A. (1993). Signal detection theory as data analysis method and psychological decision model. In G. Keren & C. Lewis (Eds.), A handbook for data analysis in the behavioural sciences: Methodological issues (pp. 21-57). Hillsdale, NJ: Erlbaum. Mellman, T. A., David, D., Bustamante, V., Torres, J., & Fins, A. (2001). Dreams in the acute
aftermath of trauma and their relationship to PTSD. Journal of Traumatic Stress, 14(1), 241-247.
Mellman, T. A., David, D., Kulick-Bell, R., Hebding, J., & Nolan, B. (1995). Sleep disturbance and its relationship to psychiatric morbidity after Hurricane Andrew. The American Journal of Psychiatry. 152(11), 1659-63.
Menz, M. M., Rihm, J. S., Salari, N., Born, J., Kalisch, R., Pape, H. C., Marshall, L., & Büchel, C. (2013). The role of sleep and sleep deprivation in consolidating fear memories. Neuroimage, 75, 87-96.
Merritt, J. M., Stickgold, R., Pace-Schott, E., Williams, J., & Hobson, J. A. (1994). Emotion profiles in the dreams of men and women. Consciousness and Cognition, 3(1), 46-60. Monaghan, P., Shaw, J. J., Ashworth-Lord, A., & Newbury, C. R. (2017). Hemispheric
processing of memory is affected by sleep. Brain and language, 167, 36-43.
Morrissey, J. (Producer), & Kaye, T. (Director). (1998). American History X [Motion Picture]. United States: New Line Cinema.
Nielsen, T. A., Deslauriers, D., & Baylor, G. W. (1991). Emotions in dream and waking event reports. Dreaming, 1(4), 287.
Payne, J. D., & Nadel, L. (2004). Sleep, dreams, and memory consolidation: the role of the stress hormone cortisol. Learning & Memory, 11(6), 671-678.
Revonsuo, A. (2000). The reinterpretation of dreams: An evolutionary hypothesis of the function of dreaming. Behavioral and Brain Sciences, 23(6), 877-901.
Scarpelli, S., Bartolacci, C., D'Atri, A., Gorgoni, M., & De Gennaro, L. (2019). The functional role of dreaming in emotional processes. Frontiers in Psychology, 10.
Schaefer, A., Nils, F., Sanchez, X., & Philippot, P. (2010). Assessing the effectiveness of a large database of emotion-eliciting films: A new tool for emotion researchers. Cognition and Emotion, 24(7), 1153-1172.
Schoch, S. F., Cordi, M. J., Schredl, M., & Rasch, B. (2019). The effect of dream report collection and dream incorporation on memory consolidation during sleep. Journal of Sleep Research, 28, e12754.
Schredl, M., Berres, S., Klingauf, A., Schellhaas, S., & Göritz, A. (2014). The Mannheim Dream questionnaire (MADRE): Retest reliability, age and gender effects. International Journal Of Dream Research, 7(2). 141-147.
Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press.
Sterpenich, V., Perogamvros, L., Tononi, G., & Schwartz, S. (2020). Fear in dreams and in wakefulness: Evidence for day/night affective homeostasis. Human Brain
Mapping, 41(3), 840-850.
Stickgold, R., Hobson, J. A., Fosse, R., & Fosse, M. (2001). Sleep, learning, and dreams: off-line memory reprocessing. Science, 294(5544), 1052-1057.
Vallat, R., Chatard, B., Blagrove, M., & Ruby, P. (2017). Characteristics of the memory sources of dreams: A new version of the content-matching paradigm to take mundane and remote memories into account. PloS one, 12(10).
Valli, K., Revonsuo, A., Pälkäs, O., & Punamäki, R. L. (2006). The effect of trauma on dream content - A field study of Palestinian children. Dreaming, 16(2), 63.
Voss, U., & Klimke, A. (2018). Dreaming during REM sleep: autobiographically meaningful or a simple reflection of a Hebb'ian-based memory consolidation process. Archives Italiennes de Biologie, 156(3), 99-111.
Vyazovskiy, V. V. (2015). Sleep, recovery, and metaregulation: explaining the benefits of sleep. Nature and Science of Sleep, 7, 171-184.
Wamsley, E. J., & Stickgold, R. (2019). Dreaming of a learning task is associated with enhanced memory consolidation: Replication in an overnight sleep study. Journal of Sleep Research, 28(1), e12749.
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: the PANAS scales. Journal of Personality and Social Psychology, 54(6), 1063.
Wynn, S. C., Kessels, R. P., & Schutter, D. J. (2020). Electrocortical indices of subjectively perceived confidence in episodic memory. International Journal of
Appendix A: Inclusion and exclusion criteria
Inclusion criteria
The participant…
Rates his/her sleep quality of the past week as fair, good or very good.
Usually went to sleep between 22:00-23:30 and woke up between 06:00-10:30 in the past week.
o Participants who go to sleep between 00:00 and 00:30 are included if they agree to alter their sleep schedule by going to sleep at 23:00 on the three nights preceding the experiment and wake up no later than 09:00 on these days.
Is able to fall asleep within 45 minutes.
Does not have clear trouble sleeping in a different place. Is easily woken up by an alarm clock.
Has the option of sleeping alone.
Is staying in an EU country and has a telephone number with an EU country code. Has an English proficiency of intermediate level or higher.
Has normal or corrected-to-normal vision and hearing. Agrees to watch a movie fragment involving violence.
Exclusion criteria:
The participant…
Works night shifts or works past midnight.
Uses medication that influences cognition (e.g. anti-depressants). Has a neurological, psychiatric, psychological or sleep disorder. Has a heart condition.
Appendix B: Pilot study to validate butterfly images
Purpose of the pilot
This pilot was carried out to test the amount of memory items that would be suitable for a recognition task and select a set of butterfly images for the current study accordingly. Another purpose of this pilot study was to validate these butterfly images by testing them on affective valence and arousal dimensions.
Participants
A total of 11 participants (8 female, 3 male), aged between 20 and 34 years (M ± SD: 25 ± 4.84) took part in this pilot.
Procedure
The study was created and carried out in Qualtrics. It contained three parts, using a total of 180 different butterfly images varying in shape, color and wing pattern (Cramer, 1780).
The first part was a memory encoding task, consisting of randomized 60 trials. In each trial, a different butterfly image was presented for 3000 ms, followed by an affective valence rating and an arousal rating of the image. Participants received explicit instructions to memorize the butterfly images and they were aware that their memory would be tested next.
The second part was a recognition task with 120 trials presented in a random order. Here, the 60 butterfly images from the encoding task (“memory items”) were presented again, as well as 60 new butterfly images (“foils”). All images were presented for 3000 ms, after which participants indicated whether they recognized the image from the encoding task, gave a confidence rating of that answer and provided an affective valence and arousal rating of the image.
In the third part, participants viewed 60 new foils for 3000 ms. They did not need to memorize them, but only rate them on affective valence and arousal dimensions. The scales of all rating questions ranged from 0 to 100.
Data analysis
Similar to the current study, the d-prime (d’) was calculated to measure memory performance, using the hit rates and false alarm rates.
Average valence and arousal ratings were calculated for the individual memory items and foils. After placing all 180 butterfly images together in one group, the images that looked very similar to another image (“lures”) were removed. Based on the memory performance outcome, 120 images with the most similar affective valence and arousal ratings were selected and randomly divided into three even groups for the current study: a memory item group, a foil group for the pre-sleep recognition task (“foils 1”) and a foil group for the post-sleep recognition task (“foils 2”). Two-Sample t-tests were done between the affective valence and arousal ratings of the three groups to examine whether they had similar means. Results
The average hit rate on the recognition task was 0.52 ± 0.17 and the average false alarm rate 0.21 ± 0.12. This led to an average d’ score of 0.90 ± 0.49.
The entire set of butterfly images had an average affective valence rating of 51.42 ± 5.07 and an average arousal rating of 41.53 ± 4.27. A total of 4 lures were removed. There were no differences between the affective valence and arousal ratings (Table 1) of the newly formed groups (Table 2).
Affective valence Arousal
Memory items 51.27 ± 3.92 42.59 ± 3.35
Foils 1 49.93 ± 3.69 42.12 ± 3.71
Foils 2 51.13 ± 3.50 42.05 ± 3.22
Table 1. Average affective valence and arousal ratings (M ± SD) of the memory items, foils 1
and foils 2 groups on a scale from 0 to 100.
Affective valence Arousal
Memory items – foils 1 t(78) = -1.57, p = 0.12 t(78) = -0.60, p = 0.55 Memory items – foils 2 t(78) = -0.17, p = 0.87 t(78) = -0.74, p = 0.46 Foils 1 – foils 2 t(78) = -1.49, p = 0.14 t(78) = 0.09, p = 0.93
Table 2. Outcomes of Two-Sample t-tests performed between the affective valence and
arousal ratings of the memory items, the foils 1 and foils 2 groups. No significant differences were found between any of the groups.
Conclusions
Based on the obtained hit rate, which was only slightly above chance level, and the average d’ score of under 1.0, it can be concluded that it was very difficult to distinguish the memory items and foils from each other. Therefore, a selection of 120 butterfly images was made for the current study, resulting in 40 memory items, 40 pre-sleep foils and 40 post-sleep foils, rather than 60 in each group. There was no significant difference between affective valence and arousal ratings of the three new groups, which makes them valid to use in the current study.
Appendix C: Dream report questionnaire
Participant number: _____Awakening number: _____ Time: _____
Instructor: “Please take your time to recollect what was going through your mind before the phone call. When you are ready, just tell me everything (that was going through your mind)”. [If the participant does not start talking spontaneously within 10-15 seconds]:
1. Can you tell me everything that was going through your mind before the phone call?
__________________________________________________________________________ __________________________________________________________________________ __________________________________________________________________________ __________________________________________________________________________ __________________________________________________________________________ __________________________________________________________________________ __________________________________________________________________________
2. Is there anything else that you can recall? / Can you report any impressions or
general feelings? __________________________________________________________________________ __________________________________________________________________________ __________________________________________________________________________ __________________________________________________________________________ __________________________________________________________________________ __________________________________________________________________________
Sleep report (asleep)
→ Please describe everything that was going through your mind “Something, but I do not remember” (asleep)
→ Please indicate any impressions or general feelings Nothing at all (asleep)
Wake report (awake)
3. Are you sure (about what you just reported)?
Sure Not sure
4. On a scale from 0 (extremely negative) to 100 (extremely positive), how negative or
positive were you feeling?
0 100
5. On a scale from 0 (totally calm) to 100 (totally excited), how calm or excited were you
feeling?
6. Did you hear any sounds (before the phone call)?
Yes
→ What did you hear? __________________ No
7. Were you asleep or awake (before the phone call?)
Asleep Awake I am not sure
8. Were you aware of being in your room (before the phone call)?
Yes No
I am not sure
Do you now remember anything else (about your experience?) → If yes, add answer to free recall in question 1.
---- STOP HERE IF NOTHING AT ALL / SOMETHING BUT I DO NOT REMEMBER / WAKE REPORT
----9. On a scale from 0 (short snapshot) to 100 (long story), how long and detailed was
your experience?
10. On a scale from 0 (not at all) to 100 (completely), to what degree were you perceiving (e.g. seeing, hearing) something during your experience?
11. Which sense was dominant in your experience?
Seeing Hearing Tasting Smelling Touching
Not applicable (if answer to previous question was 0)
12. On a scale from 0 (not at all) to 100 (completely), to what degree were you thinking
about something during your experience?
13. Which emotions or feelings best describe your experience?
14. Did you experience a threat or violence to you or others?
Yes
→ Namely _______________________________________________________ No
15. Did you feel in control?
Yes No
16. Did anything in your experience remind you of the online tasks that you did this
evening? Yes
→ What? __________________
→ And how did it make you feel? ___________________ No
17. Did anything in your experience remind you of a recent personal event?
Yes
→ Please describe ____________________________________________________ No
18. Would you say that you were dreaming?
Yes No
I am not sure
19. Would you say that you were having a nightmare?
Yes No
I am not sure
20. Were you aware of the fact that you were dreaming while you were dreaming?
Yes No
I am not sure
Appendix D: Dream reports
Below are three examples of dream reports that contained a threat and three examples of dream reports that did not. Only the answers to the three questions from the “Dream report questionnaire” that were most important for this study are shown.
Threat dream 1 (objective threat to self):
Can you tell me everything that was going through your mind before the phone
call? “I was running on the street where I live, but there were people who were trying
to talk to me, but it was really quite late, so I was quite scared and I felt unsafe, that they were going to rob me or anything. While I was running, I saw someone I know in a car. There was something before that. I was in some sort of restaurant school maybe and we had dinner with a group. There were two tables and the waitress gave us money before we sat down. I think it was some sort of school thing, but I don't know. That's all I can remember now.”
Is there anything else that you can recall? “There was the guy who participated in Temptation Island, he was sitting on the other table.”
Did you experience threat or violence to you or others? “Yes, I was afraid that the people in the street would do something to me.”
Threat dream 2 (subjective threat to self):
Can you tell me everything that was going through your mind before the phone
call? “I remember... Between the previous phone call and this phone call, I remember
it was up in the mountains, I was skiing. We were doing, like, a cross country ski, I think. And it was, like, small scenes from that. I don't remember much about it. Suddenly, I was in a cabin and I was telling this stranger, an American, about some biology sh*t. Oh sorry, not sh*t. Biology stuff, sorry. And I think it was related to.... I do have a “Brain and Cognition” test for Psychology coming up soon. It was kind of exactly… It was content from that. And my sister and my brother were also in the same room. But they were discussing something else, like, something *inaudible* to me, but I can't remember now.”
Is there anything else that you can recall? “He was comparing my answers to the content of my best friend. And she is also in Amsterdam, actually.”
Did you experience threat or violence to you or others? “Yes, I think a
psychological threat. I think maybe when he was asking me one of the questions and I didn't know the answer, maybe I might have felt a bit of threat in that sense. A psychological threat in the sense that I didn't know the answer.”
Threat dream 3 (subjective threat to other):
Can you tell me everything that was going through your mind before the phone
call? “I was dreaming that I had different parents and I was dress shopping with
someone. Like, fancy dress shopping. And, also, there was this orphanage with kids that don't really age, and they all have different powers. And there were some people from my old high school there. We were biking back from the orphanage and I ran into some people from my high school, not really people I talked to. In a different scene, there were a bunch of guys with old clothes on, like, from a different century. And they tried to marry a queen off to someone else. And she has problems and they want to take the power. They were quite mean.”
Is there anything else that you can recall? “The people who were not my parents but were my parents. First off all, did not really care about me. But the woman also wanted me to get surgery to remove belly fat. She was mean about it.”
Did you experience threat or violence to you or others? “Yes, the guys who tried to marry off the queen seemed mean.”
Non-Threat dream 1:
Can you tell me everything that was going through your mind before the phone
call? “I remember that I was walking through one park in Prague and then I got to my
old golf club, in which there was a swimming pool. And for some reason the
swimming pool was filled up with golf balls, which I found funny and surprising at the same time. So, I swam in it and it was uncomfortable. And then I just went to the main building where I got something to eat and I remember… Yeah, I got something to eat, that's it.”
Is there anything else that you can recall? “No.”
Did you experience threat or violence to you or others? “No.”
Non-Threat dream 2:
Can you tell me everything that was going through your mind before the phone
call? “I was playing chess. Not sure what was happening, but I was playing with the
horse.”.
Is there anything else that you can recall? “Something about my opponent, but I can't really remember what that is. I think someone I know, someone I recognize.” Did you experience threat or violence to you or others? “No.”
Non-Threat dream 3:
Can you tell me everything that was going through your mind before the phone
call? “Okay, so now I... This time it is slightly clear. It was less literal. I... I remember
it was about, like, an imaginary scenario where people are… I don't know, there was probably a curse or something. People do not have their bodies anymore, but there were only their hearts, so you didn't see people. You only saw hearts, like, walking around, like, smaller and bigger. And then at some point, you had to be on a ship and you'd… You'd have to do a jump from an... Like... What is the name for those wooden things you have, like, in pirate ships? Like, when someone was… When they wanted to, like, kill someone, they made them jump from that. It's like a wooden plank. Okay, so you had to do that and maybe your heart would be you, because you didn't have, like, another physical form with, like, grow out wings. Or otherwise you would, like, just get into the water and you were supposed to, like, go there to heal yourself, although you were not really sure what would happen if you'd do that. And people, like, of course, really wanted to have their bodies and the only way to do that in the end is if you, like, if you replace your heart with another person, because then, instead of carrying your heart, you carry that other person and in its own physical form. So, if everyone chooses one person, then in the end everyone will, like, become a person again.
Is there anything else that you can recall? “No.”
Appendix E: Results of extra analyses
Affective valence ratings of the negative movie scene
The overnight difference scores in the valence ratings of the negative movie scene were compared between T (5.00 ± 9.22) and NT (3.90 ± 8.57). A Mann-Whitney U test revealed no significant differences between the groups (W = 154.5, p = 0.81). Although the difference scores of NT met the assumption of normality (W = 0.919, p = 0.08) and the variances between the groups were equal (F(13, 20) = 1.156, p = 0.75), this non-parametric test
was used because the difference scores of T did not meet the assumption of normality (W = 0.744, p = 0.001).
Then, pre-sleep ratings were compared with post-sleep ratings for T and NT to investigate whether the groups showed emotional regulation towards the negative movie scene. The assumption of normality was violated for the difference scores of T (W = 0.744, p = 0.001), after which a Wilcoxon Signed Rank test showed no significant overnight
differences in affective valence ratings for T ([pre-sleep: 3.79 ± 6.53]; [post-sleep: 8.79 ± 11.14]; V = 26, p = 0.05). However, the difference scores of NT were normally distributed (W = 0.919, p = 0.08) and a Paired Student’s t-test showed significantly higher post-sleep ratings (11.24 ± 12.06) compared to pre-sleep ratings (7.33 ± 8.87) for this group (t(20) = 2.087, p =
0.049; Fig. 9).
Figure 9. Bar graph of the average affective valence ratings of the negative movie scene for
T and NT. The same scale was used as for the affective valence ratings of the memory items. Post-sleep ratings were significantly higher than pre-sleep ratings for NT (*), but not for T (N.S.).
Arousal ratings of the negative movie scene
No differences were found between the overnight difference scores of the two groups ([T: 1.64 ± 26.71]; [NT: -2.71 ± 15.73]; t(19.044) = 0.550, p = 0.59). This was tested with a Welch
Two Sample t-test, after finding that the assumption of normality was met for T (W = 0.930, p = 0.30) and NT (W = 0.938, p = 0.20), but the variances of the two groups were not equal
Finally, to examine whether the ratings changed after a night of sleep, pre-sleep and post-sleep arousal ratings of the negative movie scene were compared for T and NT. As mentioned above, the overnight differences in ratings were normally distributed for both groups. Paired Student’s t-tests revealed that there was no significant overnight difference in ratings for T ([pre-sleep: 73.14 ± 25.34]; [post-sleep: 74.79 ± 15.72]; t(13) = 0.230, p = 0.82) or
NT ([pre-sleep: 76.19 ± 14.15]; [post-sleep: 73.48 ± 13.54]; t(20) = -0.791, p = 0.44; Fig. 10).
Figure 10. Bar graph of the average arousal ratings of the negative movie scene for T and
NT. A rating of 0 indicates extremely calm, 50 indicates neither calm nor excited and 100 extremely excited. For both groups, there was no significant difference between pre-sleep and post-sleep ratings (N.S.).
