Anxiety, Sleep and Dreaming: an exploratory study using Trait Anxiety as a Predictor
for Sleep Quality and Dream Affect
Eva M. K. Beunk
Student number: 11707054
Deadline date: January 22, 2021
Supervisor: Dr. Elsa Juan
Abstract
Anxiety disorders have the highest prevalence of all mental disorders. A substantial body of literature shows that high levels of trait anxiety are related to significant impairments in quality of life,
including sleep quality. In addition, high trait anxiety is predictive of negative dream affect. But despite this seemingly clear relationship between trait anxiety, sleep and dreaming, very few studies have been conducted on this matter, and not a single study has been conducted that gathered data using serial awakenings throughout an experimental night. The current study examines the
relationship between trait anxiety and emotional processing occurring during wakefulness, sleep and dreaming. In addition, relationships between trait anxiety and dream content, dream affect and seep quality are examined. Thirty-eight first-year college students filled out several questionnaires before sleep and were subsequently awakened and asked about their dreams several times throughout the experimental night. Results from MANOVAs show that high trait anxiety is related to low emotional valence and high arousal when measured in a questionnaire that assesses general dream content. However, this same relationship is not found in Mixed Effect Models with measures of emotional valence and arousal conducted during serial awakenings. This contradicts the continuity hypothesis, which states that dream states are related to waking states. Results also suggest that more anxious individuals experience more sad dreams, and, in accordance with previous research, high trait anxiety scores are related to lower sleep quality. These results provide interesting directions for future
research that can be further substantiated with polysomnographic recordings.
Introduction
Out of all classified mental disorders in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV), anxiety disorders have the highest lifetime prevalence with 28.8%, and have a median age of onset as early as 11 years (Kessler et al., 2005). Given this high prevalence, it is
important to gain insights into the characteristics of anxiety and to improve our understanding of the underlying mechanisms.
Anxiety is an emotion that is characterized by psychological symptoms such as worried thoughts or feelings of tension, as well as physical changes, such as increased blood pressure or higher levels of cortisol (Kazdin, 2000; Vreeburg et al., 2010). In general, anxiety is linked to a high level of negative affect, but without perturbation of positive affect (Watson, Clark, & Carey, 1988). It can be subdivided into state anxiety and trait anxiety, with state anxiety being a transitory response, and trait anxiety being a constant feature, driven by the individual’s personality (Horváth et al., 2016). Far fewer studies have focused on trait anxiety, whilst Vasey and Dadds (2001) suggest in their book that trait anxiety is related to the development of Generalized Anxiety Disorder (GAD), one of five major types of anxiety disorder. In fact, evidence has shown that high trait anxiety is predictive of the development of stress-induced anxiety disorders (Saviola et al., 2020). According to the U.S.
Department of Health & Human Services (HHS; 2014), major types of anxiety disorder besides GAD include social anxiety disorder (SAD), obsessive-compulsive disorder (OCD), panic disorder and post-traumatic stress disorder (PTSD). Irrespective of the type of anxiety disorder , all anxiety disorders impair quality of life and psychosocial functioning. Interestingly, these impairments can also be found in individuals with subthreshold forms of anxiety disorders, although to a lesser extent (Mendlowicz & Stein, 2000). For example, high levels of trait anxiety have been related to issues in executive functioning tasks (Pacheco-Unguetti, Acosta, Callejas, & Lupiáñez, 2010).
An important aspect of life that is impaired in anxious individuals is sleep. Sleep is of vital importance for individuals’ physical and psychological well-being (Talamini, 2017). It can be divided
into two main stages: rapid eye movement sleep (REM sleep) and non-REM (NREM) sleep. NREM sleep can be further subdivided into lighter sleep stages and a deep sleep stage, also known as slow wave sleep (SWS) (Talamini, 2017). Research has shown that anxiety is related to decreased length of REM sleep and SWS, and increased light sleep. (András et al., 2015; Fuller, Waters, Binks, & Anderson, 1997). In addition, anxiety disorders are significantly associated with impaired sleep quality
(Ramsawh, Stein, Belik, Jacobi, & Sareen, 2009). The mentioned anxiety-related differences in the macrostructure of sleep may cause this poorer sleep quality.
Somewhat 30 years ago, sleep was thought to be a natural state of unconsciousness (Hediger, 1980). It has since become clear that in fact, conscious processes occur during sleep that drive several functions, including memory consolidation (Windt, 2020; Windt, Nielsen, & Thompson, 2016). For example, during REM sleep, vivid dreams occur which are thought to play a role in processing emotional information. However, the functional role of different sleep stages and of dreaming in relation to anxiety remains unclear. Some studies suggest that the lack of REM sleep in anxious individuals prevents the occurrence of emotional attenuation, leading to a small habituation response to emotional stimuli as compared to healthy individuals (Scarpelli et al., 2019). Yet another study has shown that higher REM sleep deprivation is related to greater emotional adaptation, which suggests precisely the opposite (Lara-Carrasco, Nielsen, Solomonova, Levrier, & Popova, 2009).
Nevertheless, anxiety seems to be related to differences in dream content and dream affect. For example, individuals with anxiety disorders have more non-nightmare distressed awakenings (NNDA) (Uhde, Cortese, & Vedeniapin, 2009). NNDA are awakenings that are similar to awakenings caused by nightmares, meaning individuals experience low emotional valence and high feelings of arousal, but without dream recall. In addition, research has shown that higher levels of anxiety, measured with Beck’s Anxiety Inventory (BAI), are related to an increase in dreams with negative and harmful emotional load (Komasi, Soroush, Khazaie, Zakiei, & Saeidi, 2018). However, the BAI
will apply for trait anxiety. In addition, results from this study were gathered using dream questionnaires that assess dreaming in general. Measures taken directly or very shortly after awakening have proven to be better related to dream content and dream affect than participants’ general self-ratings (Sikka, Pesonen, & Revonsuo, 2018). Such research with measures conducted shortly after awakening with healthy participants has found that individuals with higher scores on trait anxiety express more negative affect in their dream reports (Pesant & Zadra, 2006). Another recent study has shown that peace of mind was related to positive dream affect, whereas symptoms of anxiety were related to negative dream affect (Sikka et al., 2018).
These findings are consistent with the continuity hypothesis, which states that dreams and wakefulness share similar mechanisms and that waking states and concerns are reflected in dreams (Pesant & Zadra, 2006; Scarpelli et al., 2019). Since anxiety is linked to a high level of negative affect in a waking state (Watson et al., 1988), the continuity hypothesis seems to explain why more anxious individuals also experience more distress and negative affect in their dreams. However, research on the relationship between anxiety, dream content and dream affect with a non-clinical population is scarce. Specifically, there is a striking lack of research that examines this relationship using a serial awakening paradigm, in which participants are repeatedly awoken throughout the night and asked about their dreams. Instead, most research on the relationship between anxiety, sleep and dreaming is conducted using questionnaires that assess dreaming in general and just a few studies use dream questionnaires conducted directly or shortly after awakening in the morning. The questionnaires that examine dreaming in general are often an unreliable measure of dream affect, as the time between waking and filling in the questionnaire is usually long (i.e. hours to days), causing memories of the dream(s) and associated emotions to change or to get forgotten. In addition, dream reports from shortly after awakening have so far only been collected upon awakening in the morning, in home-based studies with no control over the time interval between awakening and filling in the dream questionnaire. Using a serial awakening paradigm is effective for the accurate collection of dreams from different sleep stages, without impairing sleep quality (Noreika, Valli, Lahtela, & Revonsuo,
2009). This allows for the collection of a large and representative sample of individuals’ sleep experience across an experimental night and avoids the confounding effect of memory that comes with questionnaires that assess dreaming in general (Siclari, LaRocque, Postle, & Tononi, 2013). Since many studies have shown that dreaming does not only occur during REM sleep but also during NREM sleep, it is interesting to investigate dream reports across all sleep stages using the serial awakening paradigm. With multiple measures after awakening from sleep per participant instead of solely one measurement after awakening in the morning, a more complete picture of dream content and dream affect will emerge. This can help establish a better view on the relationship between trait anxiety, dream content and dream affect.
Therefore, the aim of this study is to examine whether trait anxiety can be used to predict sleep quality, dream content and dream affect, using a serial awakening paradigm, with a non-clinical population of first year college students. Several studies have shown that levels of anxiety are higher and anxiety disorders are more prevalent in adolescents aged 13-18 years and in college students (Conley, Shapiro, Huguenel, & Kirsch, 2020; Merikangas et al., 2010). This makes the group of first year Psychology students in the current study an interesting subject pool. In addition, an important reason for studying the effects of anxiety is the current threat of anxiety symptom development. In their paper, Ornell and colleagues (2020) proposed that there is a fear pandemic that accompanies the current COVID-19 pandemic, which causes an increase of anxiety levels in healthy individuals and an aggravation of symptoms of people with pre-existing psychiatric disorders, as previously found in a study on the Ebola virus pandemic.
One of the expectations from the current study is that more anxious individuals will experience sleep disturbance. This can be clarified by previous research, which has shown that individuals with subclinical anxiety symptoms also suffer from impaired quality of life, and several studies have shown the existence of a close relationship between sleep quality and quality of life (Luyster & Dunbar-Jacob, 2011; Marques, Meia-Via, da Silva, & Gomes, 2017; Nunes et al., 2009;
Ramsawh et al., 2009; Tel, 2013; Zeitlhofer et al., 2000). In addition, and in line with the continuity hypothesis, high anxiety individuals are expected to be more aroused upon each awakening and experience lower emotional valence. High anxiety is also hypothesized to be related to a higher level of detail of dreams, reflected in a higher score on level of thinking, perceiving, and detail. To test these hypotheses, participants filled in an anxiety questionnaire and a sleep quality questionnaire. During the night, participants were woken up and asked about the content of their dreams and any associated emotions. Based on above-mentioned theories and studies, it is hypothesized that the trait anxiety score in this non-clinical population will be negatively related to sleep quality and positively related to experienced negative emotions during the night, which will be reflected in lower valence scores and higher arousal scores.
It is important to study the links between anxiety, sleep, and dreaming, as they all seem to be closely intertwined. Many anxiety disorders are paired with comorbid sleep disorders, and it is unclear which of the two precedes the other. Doing more research on sleep and dreaming in relation to anxiety can provide valuable insights that might be used for the treatment of anxiety disorders in the future. When comorbid sleep disorders are treated, quality of life of individuals with anxiety disorders significantly improves (Ramsawh et al., 2009).
Methods
The data used in this study has been previously obtained for studies on task-related dream content and emotional memory consolidation (Tuinman, 2020; van Keeken, 2020).
Participants
Forty-two first-year Psychology students were eligible for participation after filling in a screening questionnaire, in which they were checked for being healthy, proficient in English, and classified as good sleepers. Some participants were excluded from the study because of cancellation (n
= 1), failed attempts to contact (n = 1), or because they did not complete the entire experiment (n = 2). This leaves 38 participants (26 female, 11 male, 1 other), aged between 18 and 36 years old (M ± SD = 20.55 ± 3.37) who completed the entire study and were included for data analysis.
Participants were instructed to abstain from alcohol and drugs within 24 hours before the start of the experimental night, to wake up at 8:00 a.m. on the day of the experiment, to not drink
caffeinated drinks after 2:00 p.m. on the day of the experiment, and to sleep alone during the experimental night. They were rewarded with research credits for their participation in the study.
Procedure
Participants started with online surveys and computer tasks around 9:00 p.m. on their personal computers. First, participants completed a Positive and Negative Affect Scale (PANAS; Watson et al., 1988) to obtain mood at baseline. Next, they watched three-minute excerpts of a neutral movie and negative movie and rated these videos on valence and arousal. Afterwards, participants performed an encoding task in which three-second clips from the excerpts seen before were used to induce negative and neutral emotions. Following this, participants filled in multiple questionnaires, including the Pittsburgh Sleep Quality Index (PSQI; Buysse, Reynolds, Monk, Berman, & Kupfer, 1989), a general dream characteristics questionnaire (GDC; Appendix A; based on Schredl, Berres, Klingauf, Schellhaas, & Göritz, 2014), Beck’s Depression Inventory (BDI-II; Beck, Steer, & Brown, 1996) and the trait section of the state trait anxiety inventory (STAI-T; Spielberger, Gorsuch, & Lushene, 1970). Following this, participants completed a pre-sleep memory test on the previously encoded information, followed by completion of another PANAS to obtain mood after emotion induction.
At 11:00 pm, participants were asked to go to sleep. Starting at midnight, participants were awoken every hour by a phone call from the researcher and asked to “report everything that was going through their mind before the phone call”, a method used in previous research (Noreika et al., 2009). Additionally, participants were asked to score their valence and arousal ratings at every awakening. For each awakening that resulted in a dream experience (DE), a structured Dream Report
Questionnaire was conducted (Appendix B). From this, scores of the degree of thinking and perceiving in the dream, plus how detailed the experience was, were derived. Participants were awakened a total of nine times throughout the night. However, due to procedural adjustments and errors, the number of awakenings varies between 5-9 awakenings per participant.
After the final awakening at 8:00 a.m., participants filled in another PANAS and a sleep quality questionnaire about the experimental night (Appendix C). After this, they did a post-sleep memory test, followed by viewing the movie excerpts again and rating them on valence and arousal. Finally, participants filled out an exit questionnaire. Figure 1 provides a visual overview of the procedures during the experimental night.
Figure 1
Schematic overview of the experimental night. Participants start at 9:00 pm with several questionnaires and tasks. They go to sleep at 11 pm, after which they are awakened every hour to conduct the Dream Report Questionnaire. After final awakening at 8:00 am, final questionnaires and tasks are conducted.
Note. Questionnaires include: Positive and Negative Affect Scale (PANAS), Pittsburgh Sleep Quality
Index (PSQI), General Dream Characteristics Questionnaire (GDC), Beck’s Depression Inventory (BDI-II), the trait section from the State-Trait Anxiety Inventory (STAI-T). a post-sleep sleep quality survey and an exit questionnaire.
Main Measures
Several measurements from the described procedure were included in this study. Besides STAI-T score, main measures include the PANAS score measured at three different times, the valence and arousal scores of the videos measured before and after sleep, the PSQI score, GDC scores, scores from the Dream Report Questionnaire and the sleep quality survey score.
Multiple scores are derived from the GDC questionnaire, which can be found in Appendix A. Measurements of interest for the current study include dream recall frequency, dreams’ effect on mood, nightmare frequency, valence, arousal and attitude towards dreams. The measure of the general attitude towards dreams consists of a cumulative score of several statements. All other scores from the questionnaire are obtained through a single response value.
Furthermore, several measures were obtained from the Dream Report Questionnaire, which can be found in Appendix B. Main measures include valence and arousal scores for all awakenings, along with counts of task-related dreams, dreams about personal events and different experienced emotions in dreams, and scores for the degree of thinking, perceiving and amount of detail in dreams.
Data Processing
Scores for each participant on the various questionnaires were calculated and merged into a CSV file for data analysis. From the three PANAS questionnaires, two difference scores were derived: a film response score and an attenuation score. The film response score was calculated by subtracting the PANAS scores measured before watching the videos from the PANAS scores measured after watching the videos, both pre-sleep. The attenuation score was calculated by subtracting the post-sleep PANAS scores from the pre-post-sleep PANAS score measured after watching the videos. This resulted in a total of four scores: a positive affect (PA) film response score, a negative affect (NA) film response score, a PA attenuation score and an NA attenuation score. Besides this, awakenings from participants were categorized as wake report (W), dream experience (DE), dream experience without
recall (DEWR), or no experiences (NE), in line with previous research from Siclari and colleagues (2017). Following this, ratios for each awakening type were determined for participants by dividing the number of a certain awakening type by the total number of awakenings.
Other variables that were measured during awakenings, including the valence, arousal, perceiving, thinking and detail scores, were not averaged or added together but treated as separate entities. In addition, a variable named report length was made, consisting of the word count of the initial report of every awakening (i.e. the answer to the question 1: “report everything that was going through your mind before the phone call”). Before conducting analyses, the dataset was split into two subsets: a single-level variable dataset containing all variables that consisted of a single variable (e.g. PSQI score, PA film response score, or DE-ratio) and a multilevel variable dataset with the variables that were measured at each awakening (i.e. the mentioned subset of responses to the Dream Report Questionnaire).
Statistical Analysis
Single-level variables were analysed using multiple multivariate analyses of variance
(MANOVAs). Multilevel variables were analysed using multiple mixed effects models. For the mixed effects models, participants were included in the model as a random effect, to account for the fact that multiple awakenings belong to the same participant, which is needed to account for within-subject variance and prevent it from occluding the results. All analyses, both for single-level data and for multilevel data, used the trait anxiety score from the STAI (STAI-T) as a continuous predictor.
Exploration of the data before conducting analyses was done using JASP, version 0.13.1.0 (JASP Team, 2020). All analyses were conducted using RStudio, version 4.0.3 (RStudio Team, 2020). Firstly, analyses were conducted on the dataset with the single level variables. Since there were in total thirty-eight dependent variables to be tested, several MANOVAs were conducted to help protect against inflating Type 1 error rate in multiple univariate tests (Cramer & Bock, 1966). Prior to
dependent variables in order to test the MANOVA linearity assumption, that states that the dependent variables should be moderately correlated to each other (.20 - .60; Meyers, Gamst, & Guarino, 2006). The correlation matrix can be found in Appendix D. Based on these correlations and theoretical expectations, several groups of correlating variables were formed to be tested together in MANOVAs. This was deemed necessary, as not all dependent variables correlated moderately to each other, and the total number of dependent variables exceeds the advised limit of 10 for a MANOVA (Field, 2018). Dependent variables that did not meet the linearity assumption were not included in a MANOVA and were tested using a linear model analysis. The variables included in the different MANOVAs and the variables that were tested seperately, can be found in Table 1 and Table 3, respectively. In accordance with the central limit theorem, normality can be assumed in sample sizes above 30 (Field, 2018). In addition, MANOVA is robust to a few outliers, so these were not removed from the data. However, two of the variables had less than five observations that were not zero, including surprise and disgust. These were therefore not included in data analysis, as this number of observations is too small to produce reliable results. When a significant MANOVA result was obtained, univariate analyses were conducted as a follow-up test. To correct for multiple testing, the False Discovery Rate (FDR) method was used, as research has shown that this is a more powerful and more suitable correction to use than the Bonferroni correction (Benjamini, 2010). Besides this
adjustment of p-values after follow-up analyses, p-values from all seperatly tested variables were also adjusted using the FDR method.
Results
Single level variables
An overview of the conducted MANOVAs, the included variables and the results can be found in Table 1. Results from conducted follow-up analyses are presented in Table 2. Table 3 consists of results from univariate linear models of dependent variables that were not included in MANOVAs.
Trait anxiety might be predictive of sad DEs
There was no significant MANOVA effect found for negative emotions (p = .085; Table 1). However, the assumption of multivariate normality was not met because only few observations from the tested dependent variables deviated from zero. For increased robustness, and in order to fully explore available data, follow-up univariate linear models were conducted. These showed that for sadness in DEs, a significant positive relation with STAI-T score was found, meaning that more anxious individuals experience more sad dreams (p = .042; Table 2; Figure 2). There was no such relationship found between trait anxiety and anxious dreams (p = .092) or angry dreams (p = .55).
Figure 2
Trait anxiety is predictive of poor sleep quality
The sleep quality MANOVA yielded a significant effect (p = .038; Table 1). Follow-up linear models showed that STAI-T positively predits PSQI score (p = .022; Table 2; Figure 3). As a higher PSQI score indicates poorer sleep quality, a higher STAI-T score can be said to be related to poorer sleep quality. This same effect of trait anxiety on sleep quality was not observed for the sleep quality survey that was conducted the morning after the experimental night (p = .18; Table 2).
Figure 3
Trait anxiety (STAI-T) score as a predictor for sleep quality measured with the Pittsburgh Sleep Quality Index (PSQI).
Trait anxiety is predictive of nightmare frequency, dream recall frequency and the effect dreams have on mood
Scores of the GDC that were included in the MANOVA showed a signifcant predictive effect of STAI-T score (p = .014; Table 1). From follow-up linear models, it becomes evident that STAI-T score has a significant positive predictive effect on the frequency of nightmares (p = .028), dream recall frequency (p = .031) and the effect that dreams generally have on mood (p = .028; Table 2; Figure 4). This indicates that more anxious individuals experience more frequent nightmares, recall their dreams more often and that their mood is more affected by their dreams. Such a predictive effect of STAI-T score was not found for individuals’ attitude towards dreams (p = .25; Table 2).
Figure 4
Trait anxiety (STAI-T) score as a predictor for dream recall frequency (blue line), the effect of dreams on mood (red line), and nightmare frequency (green line), as obtained from the General Dream Characteristics (GDC) questionnaire.
Trait anxiety seems to be predictive of GDC valence and arousal ratings
A significant predictive effect of STAI-T score was found for the valence and arousal scores of dreams, as obtained by the GDC. Here, high anxiety is related to low dream valence and high arousal. However, all dependent variables that did not moderately correlate with other variables, were not included in MANOVAs but tested univariately using linear models. The included variables and the results can be found in Table 3. Because of the multiple conducted linear models, p-values were corrected using the FDR-method. After p-value correction, the effect of STAI-T score on valence and arousal measured in the GDC was no longer significant (p = .068; Figure 5).
Figure 5
Trait anxiety (STAI-T) score as a predictor for valence and arousal scores, as obtained from the General Dream Characteristics (GDC) questionnaire.
Trait anxiety does not predict sleep ratio, dream content, PANAS difference scores, happy DEs, fearful DEs, or post-pre sleep differences in valence and arousal ratings of emotional and neutral videos
For sleep ratio scores, dream content and PANAS difference scores, no significant MANOVA effects were obtained (Table 1; p = .96, p = .36 and p = .73, respectively). Follow-up univariate analyses were not conducted in this case. In addition, for seperately tested dependent variables, STAI-T score was not found to be significantly predictive of the number of task features present in dreams, as scored by the experimenter (p = .38), nor of happy DEs (p = .93) or fearful DEs (p = .25; Table 3). There were also no significant effects found for the arousal difference scores of the emotional and neutral videos (p = .39 and p = .34, respectively), nor for the valence difference scores of the emotional and neutral videos (p = .81 and p = .55, respectively; Table 3).
Table 1
Results from performed MANOVAs with a description of included dependent variables, Pillai’s trace, F-value with degrees of freedom and corresponding p-values.
MANOVA model Dependent variables included Pillai’s trace F(1,36) p-value
Sleep ratio scores W-ratio, DE-ratio, DEWR-ratio & NE-ratio 0.017 0.14 .96 Dream content Task features, threats, personal life 0.090 1.12 .36
Negative emotions Angry, sad, anxious 0.17 2.40 .085
PANAS difference scores
PA attenuation, NA attenuation, PA film response, NA film response
0.058 0.50 .73
Sleep quality PSQI, sleep quality survey 0.17 3.59 .038 GDC scores Attitude, effect on mood, recall frequency,
nightmare frequency
Table 2
Results from follow-up univariate models of dependent variables included in significant MANOVAs.
Dependent variable F(1,36) p-value p-value (FDR corrected)
Emotion: anxiety 3.73 .061 .092
Emotion: sadness 6.70 .014 .042
Emotion: anger 0.36 .55 .55
PSQI score 7.22 .011 .022
Sleep quality score 1.85 .18 .18
GDC: effect on mood 7.01 .012 .028
GDC: recall frequency 5.65 .023 .031
GDC: nightmare frequency 6.64 .014 .028
GDC: attitude 1.35 .25 .25
Note. FDR = False Discovery Rate, PSQI = Pittsburgh Sleep Quality Index, GDC = General Dream
Characteristics
Table 3
Results from univariate linear models, performed on the single level variables excluded from MANOVAs.
Dependent variable F(1,36) p-value p-value (FDR corrected)
Scored task features 0.80 .38 .59
Emotion: happiness 0.0078 .93 .93
Emotion: fear 1.39 .25 .59
Valence difference score: emotional video 0.74 .39 .59 Valence difference score: neutral video 0.93 .34 .59 Arousal difference score: emotional video 0.061 .81 .91 Arousal difference score: neutral video 0.36 .55 .71
GDC: valence 7.31 .010 .068
Multilevel variables
A total of 340 awakenings were conducted, of which 175 DEs. Assumptions were checked after fitting the mixed effect models for the different variables by examining several plots: residuals against fitted values for normality, STAI-T score against residuals for linearity and QQ-plots for the normality of the model residuals.
Trait anxiety is not predictive of valence, arousal or report length of all awakenings
First, analyses were conducted on variables from every awakening: valence, arousal and report length. Results can be found in Table 4. Valence score, arousal score and report length were not significantly predicted by STAI-T score (p = .27, p = .25 and p = .34, respectively). Inspection of the different plots showed that assumptions were met for valence and arousal. Because of extreme outliers, report length did not meet the linearity assumption. Twenty-six outliers were removed using the interquartile range (IQR) method, a method that makes use of percentiles to determine outliers, which is a more reliable method than one that is dependent on mean and standard deviation (Sharma, 2018). This resulted in a model with a better fit, as determined by plot-inspection and the notion that the standard error had become smaller than the t-value, instead of vice versa (Table 4). However, also in the corrected model, STAI-T score was not a significant predictor of report length (p = .19).
Table 4.
Results from Mixed Effects Models on data gathered from all awakenings
Dependent variable t(df) Standard error p-value
Valence t(35.57) = -1.13 0.12 .27
Arousal t(35.87) = 1.15 0.24 .25
Report length t(32.34) = .96 1.34 .34
Report length after outlier removal
Trait anxiety is not predictive of valence, arousal or report length of DEs
Following data from all awakenings, tests were conducted on the data that was collected during awakenings that resulted in dream reports. Tested variables and results can be found in Table 5. No significant effect was found for valence and arousal from DEs (p = 039 and p = .43, respectively). Both models met assumptions as determined by inspection of the plots. The thinking score, perceiving score and the detail score of the different DEs were also not significantly predicted by STAI-T score (p = .52, p = .83 and p = .36, respectively). The model of report length of DEs also did not yield a
significant effect (p = .40). However, assumptions were again not met due to outliers. After removal of 12 outliers using the IQR-method, the model fit upon examination of the residuals plots had
substantially improved, as was the standard error value. In the corrected model, STAI-T score was again not found to be predictive of dream report length (p = .064; Table 5).
Table 5
Results from Mixed Effects Models on data gathered from just awakenings resulting in dream experiences (DEs).
Dependent variable t(df) Standard error p-value
Valence t(39.38) = -0.87 0.17 .39
Arousal t(40.17) = .80 0.27 .43
Report length t(26.39) = .86 2.76 .40
Report length after outlier removal
t(32.31) = 1.92 0.64 .064
Thinking score t(34.86) = .66 0.43 .52
Perceiving score t(39.91) = -0.221 0.33 .83
Detail score t(38.35) = .93 0.33 .36
Discussion
Previous research on the relationship between anxiety and dream affect has been mainly conducted using questionnaires that assess dreaming in general, or upon awakening in the morning. This method does not provide a complete view of dream affect, because dreams may have been forgotten or reinterpreted at the time the questionnaires are conducted. Therefore, in the current study, serial awakenings were conducted throughout the night to form a comprehensive view of dream content and affect. Several factors have been explored that contribute to emotional experience during wakefulness, sleep, and dreaming, as well as dream content and general sleep quality. From the analyses, several interesting results present themselves.
To begin with, higher trait anxiety is related to lower sleep quality, as illustrated by a positive relationship between STAI-T score and PSQI score. This replicates results from Ramsawh et al. (2009), and corroborates the notion that trait anxiety is related to sleep problems. Considering the substantial body of literature that relates anxiety to sleep disorders, this relationship between trait anxiety and sleep quality is not surprising and in line with our hypotheses. However, response on the sleep quality survey that was conducted in the morning was not predicted by trait anxiety. A reason for this might be that the used survey is not a valid tool to measure sleep quality, possibly because it consists only of ‘Yes’ or ‘No’ statements, or because some included statements to do not make sense given the methodology of the current study, such as the statement “I woke up several times last night.” (Appendix C). In addition, due to intrapersonal variation and the unusual circumstance of being woken up several times during the night, the sleep quality survey score might not give a clear nor complete view of an individual’s overall sleep quality. Unfortunately, no polysomnographic (PSG) recordings were conducted for an objective assessment of sleep quality. In future research, another self-report measure of sleep quality should be conducted. This score can then be compared to sleep quality assessed from PSG recordings to provide valuable insights into a possible relationship between anxiety, changes in sleep macrostructure and sleep quality.
Besides these results on sleep quality, trait anxiety is predictive of several scores from the GDC questionnaire. Specifically, higher trait anxiety is related to higher dream recall frequency, higher nightmare frequency and an increased effect of dreams on mood. These findings are in line with the continuity hypothesis. Possible mechanisms contributing to these effects are higher vigilance and an attentional bias for threatening information concerning the self in anxious individuals
(Berggren & Eimer, 2020; Harris & Menzies, 1998) These aspects might cause more anxious individuals to be more aware of their dreams and possible threatening content, relating to higher dream recall and more frequent nightmares. In addition, sensitivity and susceptibility to stress might cause more anxious individuals to experience a greater influence of their dreams on their daily life and functioning.
In addition to these scores, valence and arousal as measured with the GDC provide interesting results, despite of relationships between trait anxiety and valence/arousal no longer being significant after p-value adjustment. Since the current study is exploratory and analyses were conducted on data that has previously been gathered, without prior expectations or pre-planned hypotheses, the
correction for multiple testing is not a critical part of the analysis process (Althouse, 2016). Therefore, results can be interpreted with the uncorrected p-value, meaning trait anxiety is related to lower scores on valence and higher scores on arousal of dreams in general. This finding again is in line with the continuity hypothesis. In addition, previous research has shown that anxiety relates to more negative dream affect and more frequent NNDA, both which relate to valence and arousal. These results on the relationship between trait anxiety and valence and arousal are thus in accordance with our hypotheses and previous research.
Interesting is, however, that this same finding has not been replicated for the valence and arousal ratings of DEs measured during the experimental night. Pointedly, both for all awakenings taken together and when just examining awakenings resulting in DEs, trait anxiety was not predictive of the valence nor the arousal score. This is not in line with our hypotheses, as it was expected to find
the same result as for valence and arousal scores measured by the GDC. Several methodological and theoretical explanations might account for this finding. For instance, the 100-point scale used to measure valence and arousal might be too broad for accurate collection because of participants’ drowsy state after being woken up from sleep. A narrower scale might be more useful to determine valence and arousal during awakenings, such as the 9-point valence scale used by Lara-Carrasco et al. (2009), which also made use of a serial awakening paradigm. The difference between measurements at awakenings and the measurement using the GDC questionnaire might also be attributable to trait anxiety itself. Previous research has shown that individuals high in trait anxiety are more likely to retain threat-related information and often show a negative cognitive appraisal of situations (Berggren & Eimer, 2020). Individuals that score high on trait anxiety might thus be biased when filling in the GDC questionnaire, due to a selection bias of threat-related information that causes them to have stronger memories for dreams with negative or threatening content. This might explain why valence and arousal scores from a general questionnaire show a different relationship with trait anxiety than valence and arousal scores obtained during the night.
Thirdly, concerning other variables measured during sleep, trait anxiety seems to predict the experience of sad emotions in DEs, with a higher trait score being related to the experience of more sad emotions. However, this finding should be interpreted with caution, because it is based on a small number of observations and the initial MANOVA did not provide a significant result. Of the 38 participants in this study, only 6 individuals experienced sad emotions in their dreams. This makes the result difficult to interpret, and it has to be validated in a more extensive study that focuses on obtaining more emotions from participants. Out of all of the six basic emotions tested in this study, fear was experienced by most participants (n = 13), followed by happiness (n = 9), sadness (n = 6), anger (n = 4), surprise (n = 3) and disgust (n = 1). Perhaps this small number of observations is attributable to the fact that an open-ended question (i.e. “Which emotions or feelings best describe your experience?”) was used to obtain experienced emotions. Perhaps participants had trouble finding words to describe their emotions, given the small number (n = 5) of native English speakers in the
current sample. An alternative is to ask about the experience and intensity of certain emotions with a pre-established emotions list, as in previous research by Lara-Carrasco et al. (2009). Using this type of questioning will likely generate more useful data with more observations. A disadvantage of this method might be that it takes longer to obtain the information, thus making the awakening time longer and therefore perhaps causing higher sleep disruption in participants. Even though the result from the current study is not conclusive and should be interpreted with caution, it can be stated that it points in an interesting direction for future research. In addition, this is the first result that seems to suggest that more anxious individuals experience more negative affect in their dreams, without perturbation of positive affect, as has been hypothesized before (Pesant & Zadra, 2006; Sikka et al., 2018; Watson et al., 1988).
Several limitations present itself in this study. To begin with, the conducted study was a home-based study, which prevented good environmental control. Many factors might have influenced sleep quality, such as for example environmental noise or use of mobile phone throughout the night. On the other hand, this home-based study decreases the “first night effect” (FNE), the phenomenon in PSG recordings characterized by decreased total sleep time, lower sleep efficiency, reduction in REM sleep, and longer REM latency (Agnew, Webb, & Williams, 1966). Besides limited control of
environmental factors, a limitation of the current study is the lack of PSG recordings. Without these recordings, no distinction can be made between dreams from NREM sleep and dreams from REM sleep. In addition, since nearly all data has been gathered through self-report measures, it is subject to self-report biases. All participants were first year Psychology students, which provides risk of possible precognition that affects the way participants answer in the different questionnaires. In future
research, it might be useful to check for this self-report bias by making use of the Lie scale scores from the Eysenck Personality Inventory (Lara-Carrasco et al., 2009). However, data and results from the current study are still interpretable and relevant. Because of the lack of PSG recordings, it was decided to perform a large exploratory study on the data, in such a way that results present interesting
original experiment was centred around a memory task that was performed before and after sleep, and those results were not examined in the present study, the effect of self-report bias is therefore likely negligible.
Another factor which might have influenced results, is the COVID-19 pandemic. Data was gathered during the first wave of the virus, in which global measures were being taken to prevent virus spread. Fear of the pandemic might have acted as a stressor. It is known that higher levels of stress relate to lower resilience and higher levels of anxiety (Anyan & Hjemdal, 2016). Besides this, social isolation or social distancing, the control measures that have been globally implemented by many governments, have been associated with higher levels of anxiety in adolescents (Duan et al., 2020). In fact, the number of subjects in the present study that meet clinical threshold for anxiety disorder (n = 16) is a lot higher than expected from a pool of healthy participants, with over 42% of the sample meeting the clinical threshold (Ercan et al., 2015). Perhaps future research conducted in times when nation-wide or global measures are being taken that have great impact on individual’s daily lives, an extra questionnaire can be conducted to control for pandemic fear.
In conclusion, the current study has provided valuable insights in the relationship between trait anxiety, sleep and dreaming. Higher trait anxiety is related to lower sleep quality measured with the PSQI, higher nightmare frequency, higher dream recall frequency and a larger effect of dreams on mood. Furthermore, it seems that high trait anxiety predicts low valence and high arousal when assessed in a general questionnaire, though this has not been found for data collected during the awakenings. This difference between general reports and awakening reports presents an interesting field of further research. Future research could be conducted using the 9-point scale previously mentioned to measure valence and arousal during awakenings, and using point-scales to obtain experienced emotions instead of an open-ended question. In addition, another measure of sleep quality for the experimental night can be used, either objectively using PSG recordings, using a more
accurate self-report measure, or both. Most importantly, a confirmatory study with pre-planned hypotheses should be conducted to validate the findings from this study.
Given the current threat of high anxiety symptom development due to the COVID-19 pandemic, it is important to study the characteristics of anxiety and to improve our understanding of the underlying mechanisms. If future research directions are followed and insights from the present study are validated and substantiated with PSG recordings, significant steps towards sleep
References
Agnew, H. W., Webb, W. B., & Williams, R. L. (1966). THE FIRST NIGHT EFFECT: AN EEG STUDYOF SLEEP. Psychophysiology, 2(3), 263–266. https://doi.org/10.1111/j.1469-8986.1966.tb02650.x
Althouse, A. D. (2016). Adjust for Multiple Comparisons? It’s Not That Simple. Annals of Thoracic
Surgery, 101(5), 1644–1645. https://doi.org/10.1016/j.athoracsur.2015.11.024
András, H., Anna, S., Montana, X., Lanquart, J. P., Hubain, P., Flamand, M., Linkowski, P., & Loas, G. (2015). Individual differences in sleep macrostructure: effects of anxiety, depression, aging and gender. Neuropsychopharmacologia Hungarica, 17(3), 146–156.
Anyan, F., & Hjemdal, O. (2016). Adolescent stress and symptoms of anxiety and depression:
Resilience explains and differentiates the relationships. Journal of Affective Disorders, 203, 213–220. https://doi.org/10.1016/j.jad.2016.05.031
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck Depression Inventory-II. Psychological Corporation.
Berggren, N., & Eimer, M. (2020). The role of trait anxiety in attention and memory‐related biases to threat: An event‐related potential study. Psychophysiology. https://doi.org/10.1111/psyp.13742
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.
Conley, C. S., Shapiro, J. B., Huguenel, B. M., & Kirsch, A. C. (2020). Navigating the College Years: Developmental Trajectories and Gender Differences in Psychological Functioning, Cognitive-Affective Strategies, and Social Well-Being. Emerging Adulthood, 8(2), 103–117.
Cramer, E. M., & Bock, R. D. (1966). Multivariate Analysis. Review of Educational Research, 36(5), 604. https://doi.org/10.2307/1169484
Duan, L., Shao, X., Wang, Y., Huang, Y., Miao, J., Yang, X., & Zhu, G. (2020). An investigation of mental health status of children and adolescents in china during the outbreak of COVID-19.
Journal of Affective Disorders, 275(April), 112–118. https://doi.org/10.1016/j.jad.2020.06.029
Ercan, I., Hafizoglu, S., Ozkaya, G., Kirli, S., Yalcintas, E., & Akaya, C. (2015). Examining cut-off values for the state-trait anxiety inventory. Revista Argentina de Clinica Psicologica, 24(2), 143–148.
Field, A. (2018). Discovering statistics using IBM SPSS statistics (5th edition.). SAGE Publications.
Fuller, K. H., Waters, W. F., Binks, P. G., & Anderson, T. (1997). Generalized anxiety and sleep architecture a polysomnographic investigation. Sleep, 20(5), 370–376.
https://doi.org/10.1093/sleep/20.5.370
Harris, L. M., & Menzies, R. G. (1998). Changing attentional bias: Can it effect self-reported anxiety?
Anxiety, Stress & Coping, 11(2), 167–179. https://doi.org/10.1080/10615809808248310
Hediger, H. (1980). The biology of natural sleep in animals. Experientia, 36(1), 13–16. https://doi.org/10.1007/BF02003943
Horváth, A., Montana, X., Lanquart, J. P., Hubain, P., Szűcs, A., Linkowski, P., & Loas, G. (2016). Effects of state and trait anxiety on sleep structure: A polysomnographic study in 1083 subjects.
Psychiatry Research, 244, 279–283. https://doi.org/10.1016/j.psychres.2016.03.001
JASP Team. (2020). JASP (Version 0.13.1) [Computer software].
Kazdin, A. E. (2000). Encyclopedia of psychology (Volume 3). Washington, D.C: American Psychological Association.
Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Walters, E. E. (2005). Lifetime Prevalence and Age-of-Onset Distributions of. Arch Gen Psychiatry, 62(June), 593–602.
Komasi, S., Soroush, A., Khazaie, H., Zakiei, A., & Saeidi, M. (2018). Dreams content and emotional load in cardiac rehabilitation patients and their relation to anxiety and depression. Annals of
Cardiac Anaesthesia, 21(4), 388–392. https://doi.org/10.4103/aca.ACA_210_17
Lara-Carrasco, J., Nielsen, T. A., Solomonova, E., Levrier, K., & Popova, A. (2009). Overnight emotional adaptation to negative stimuli is altered by REM sleep deprivation and is correlated with intervening dream emotions. Journal of Sleep Research, 18(2), 178–187.
https://doi.org/10.1111/j.1365-2869.2008.00709.x
Luyster, F. S., & Dunbar-Jacob, J. (2011). Sleep quality and quality of life in adults with type 2 diabetes.
Diabetes Educator, 37(3), 347–355. https://doi.org/10.1177/0145721711400663
Marques, D. R., Meia-Via, A. M. S., da Silva, C. F., & Gomes, A. A. (2017). Associations between sleep quality and domains of quality of life in a non-clinical sample: results from higher education students. Sleep Health, 3(5), 348–356. https://doi.org/10.1016/j.sleh.2017.07.004
Mendlowicz, M. V., & Stein, M. B. (2000). Quality of life in individuals with anxiety disorders.
American Journal of Psychiatry, 157(5), 669–682. https://doi.org/10.1176/appi.ajp.157.5.669
Merikangas, K. R., He, J., Burstein, M., Swanson, S. A., Avenevoli, S., Cui, L., Benjet, C., Georgiades, K., & Swendsen, J. (2010). Lifetime Prevalence of Mental Disorders in US Adolescents: Results from the National Comorbidity Study-Adolescent Supplement (NCS-A). J Am Acad Child Adolesc
Psychiatry, 49(10), 980–989. https://doi.org/10.1016/j.jaac.2010.05.017
Meyers, L. S., Gamst, G., & Guarino, A. J. (2006). Applied Multivariate Research: Design and Interpretation. SAGE.
Noreika, V., Valli, K., Lahtela, H., & Revonsuo, A. (2009). Early-night serial awakenings as a new paradigm for studies on NREM dreaming. International Journal of Psychophysiology, 74(1), 14–18. https://doi.org/10.1016/j.ijpsycho.2009.06.002
F. C. (2009). Impaired sleep reduces quality of life in chronic obstructive pulmonary disease.
Lung, 187(3), 159–163. https://doi.org/10.1007/s00408-009-9147-5
Ornell, F., Schuch, J. B., Sordi, A. O., & Kessler, F. H. P. (2020). ‘“Pandemic fear”’ and COVID-19: Mental health burden and strategies. Brazilian Journal of Psychiatry, 42(3), 232–235.
https://doi.org/10.1590/1516-4446-2020-0008
Pacheco-Unguetti, A. P., Acosta, A., Callejas, A., & Lupiáñez, J. (2010). Attention and anxiety: Different attentional functioning under state and trait anxiety. Psychological Science, 21(2), 298– 304. https://doi.org/10.1177/0956797609359624
Pesant, N., & Zadra, A. (2006). Dream Content and Psychological Well-Being: A Longitudinal Study of the Continuity Hypothesis. Journal of Clinical Psychology, 62(1), 111–121.
https://doi.org/10.1002/jclp.20212
Ramsawh, H. J., Stein, M. B., Belik, S. L., Jacobi, F., & Sareen, J. (2009). Relationship of anxiety
disorders, sleep quality, and functional impairment in a community sample. Journal of Psychiatric
Research, 43(10), 926–933. https://doi.org/10.1016/j.jpsychires.2009.01.009
RStudio Team. (2020). RStudio version 4.0.3.
Saviola, F., Pappaianni, E., Monti, A., Grecucci, A., Jovicich, J., & De Pisapia, N. (2020). Trait and state anxiety are mapped differently in the human brain. Scientific Reports, 10(1), 1–11.
https://doi.org/10.1038/s41598-020-68008-z
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(MAR), 1–16.
https://doi.org/10.3389/fpsyg.2019.00459
Schredl, M., Berres, S., Klingauf, A., Schellhaas, S., & Göritz, A. S. (2014). The Mannheim dream questionnaire (MADRE): Retest reliability, age and gender effects. International Journal of Dream
Sharma, N. (2018). Ways to Detect and Remove the Outliers. Towards Data Science.
https://towardsdatascience.com/ways-to-detect-and-remove-the-outliers-404d16608dba
Siclari, F., Baird, B., Perogamvros, L., Bernardi, G., LaRocque, J. J., Riedner, B., Boly, M., Postle, B. R., & Tononi, G. (2017). The neural correlates of dreaming. Nature Neuroscience, 20(6), 872–878.
https://doi.org/10.1038/nn.4545
Siclari, F., LaRocque, J. J., Postle, B. R., & Tononi, G. (2013). Assessing sleep consciousness within subjects using a serial awakening paradigm. Frontiers in Psychology, 4(AUG), 1–9.
https://doi.org/10.3389/fpsyg.2013.00542
Sikka, P., Pesonen, H., & Revonsuo, A. (2018). Peace of mind and anxiety in the waking state are related to the affective content of dreams. Scientific Reports, 8(1), 1–13.
https://doi.org/10.1038/s41598-018-30721-1
Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). Manual for the State-Trait Anxiety Inventory. Consulting Psychologists Press.
Talamini, L. M. (2017). Memory Manipulation During Sleep: Fundamental Advances and Possibilities for
Application. 313–334. https://doi.org/10.1007/978-3-319-45066-7_19
Tel, H. (2013). Sleep quality and quality of life among the elderly people. Neurology Psychiatry and
Brain Research, 19(1), 48–52. https://doi.org/10.1016/j.npbr.2012.10.002
Tuinman, Z. (2020). Sleep ’ s influence on negative emotional memory : loss of emotional tone and lower
beneficial influence on negative emotional memory than on neutral memory . February.
U.S. Department of Health & Human Services (HHS). (2014, February 12). What are the five major types
of anxiety disorders? | HHS.gov.
https://www.hhs.gov/answers/mental-health-and-substance-abuse/what-are-the-five-major-types-of-anxiety-disorders/index.html
and theoretical treatment implications. Current Psychiatry Reports, 11(4), 269–276. https://doi.org/10.1007/s11920-009-0039-4
van Keeken, A. (2020). Did Dreams Once Assist Survival ? Testing Dream- Dependent Fear Memory
Consolidation as Part of the Threat Simulation Theory.
Vasey, M. W., & Dadds, M. R. (2001). The Developmental of Generalized Anxiety. In The Developmental
Psychopathology of Anxiety (1st ed., Vol. 1, pp. 481–503). Oxford University Press.
Vreeburg, S. A., Zitman, F. G., Van Pelt, J., Derijk, R. H., Verhagen, J. C. M., Van Dyck, R., Hoogendijk, W. J. G., Smit, J. H., & Penninx, B. W. J. H. (2010). Salivary cortisol levels in persons with and without different anxiety disorders. Psychosomatic Medicine, 72(4), 340–347.
https://doi.org/10.1097/PSY.0b013e3181d2f0c8
Watson, D., Clark, L. A., & Carey, G. (1988). Positive and Negative Affectivity and Their Relation to Anxiety and Depressive Disorders. Journal of Abnormal Psychology, 97(3), 346–353.
https://doi.org/10.1037/0021-843X.97.3.346
Windt, J. M. (2020). Consciousness in sleep: How findings from sleep and dream research challenge our understanding of sleep, waking, and consciousness. Philosophy Compass, 15(4).
https://doi.org/10.1111/phc3.12661
Windt, J. M., Nielsen, T., & Thompson, E. (2016). Does Consciousness Disappear in Dreamless Sleep?
Trends in Cognitive Sciences, 20(12), 871–882. https://doi.org/10.1016/j.tics.2016.09.006
Zeitlhofer, J., Schmeiser-Rieder, A., Tribl, G., Rosenberger, A., Bolitschek, J., Kapfhammer, G., Saletu, B., Katschnig, H., Holzinger, B., Popovic, R., & Kunze, M. (2000). Sleep and quality of life in the Austrian population. Acta Neurologica Scandinavica, 102(4), 249–257. https://doi.org/10.1034/j.1600-0404.2000.102004249.x
Appendix A: General Dream Characteristics Questionnaire
How often have you recalled your dreams recently (in the past several months)?
6 = Almost every morning
5 = Several times a week
4 = About once a week
3 = Two or three times a month
2 = About once a month
1 = Less than once a month
0 = Never
How negative (0) or positive (100) are your dreams on average?
negative
positive
How often have you experienced nightmares recently (in the past several months)?
Definition: Nightmares are dreams with strong negative emotions that result in awakening
from the dreams. The dream plot can be recalled very vividly upon awakening.
0 = never,
1 = less than once a year,
2 = about once a year,
3 = about two to four times a year
4 = about once a month
5 = two to three times a month
6 = about once a week
7 = several times a week
If you currently experience nightmares, how distressing are they to you in your everyday life?
0 = Not at all distressing
1 = Not that distressing
2 = Somewhat distressing
3 = Quite distressing
4 = Very distressing
How calm or exciting are your dreams usually?
How often do your dreams affect your mood during the day?
7 = Several times a week
6 = About two to four times a year
5 = About once a week
4 = About once a year
3 = Two to three times a month
2 = Less than once a year
1 = About once a month
0 = never
How often have you experienced so-called lucid dreams (in the past several months)?
Definition: In a lucid dream, one is aware that one is dreaming during the dream. Thus it is
possible to wake up deliberately, or to influence the action of the dream actively, or to
observe the course of the dream passively.
7 = Several times a week
6 = About two to four times a year
5 = About once a week
4 = About once a year
3 = Two to three times a month
2 = Less than once a year
1 = About once a month
0 = never
Appendix B: 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’re ready, just tell me everything (that was going through your mind)”
[If participant doesn’t 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?
__________________________________________________________________________
__________________________________________________________________________
Categorize participant’s answer:
Sleep report (asleep)
→ Please describe everything that was going through your mind
“Something but I don’t 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?
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’m not sure
8. Were you aware of being in your room (before the phone call)?
Yes
No
I’m 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 DON’T 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
I’m not sure / Not applicable
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’m not sure
19. Would you say that you were having a nightmare?
Yes
No
I’m not sure
20. Were you aware of the fact that you were dreaming while you were dreaming?
Yes
No
I’m not sure
Appendix C: Sleep quality survey
PPNR:
DATE:
The following questions are related to your sleep quality of last night. To answer each question, please
circle "Agree" or "Disagree", even in those cases where you can hardly decide. Do not think about your
answer too long, circle according to your first impression. There is no alternative beside "Agree" and
"Disagree".
I think I slept very badly last night.
Agree / Disagree
I lay awake for more than half an hour last night before I fell asleep.
Agree / Disagree
I woke up several times last night.
Agree / Disagree
This morning I had a tired feeling after waking up.
Agree / Disagree
I feel like I lacked sleep last night.
Agree / Disagree
I got up last night.
Agree / Disagree
This morning, after I got up, I felt well rested.
Agree / Disagree
I feel like I have only slept a few hours last night.
Agree / Disagree
I think I slept well last night.
Agree / Disagree
I didn't sleep tonight.
Agree / Disagree
I fell asleep easily last night.
Agree / Disagree
Last night, after waking up, I had trouble falling asleep again.
Agree / Disagree
I was very restless last night.
Agree / Disagree
Appendix D: correlation matrix 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 1. W ratio 1 2. DE ratio -0.5 1 3. DEWR ratio -0.3 -0.2 1 4. NE rratio -0 -0.6 -0.3 1 5. taskfeature scored -0.1 0.15 -0.1 -0.1 1 6. perceived threats -0.4 0.57 0.11 -0.4 0.18 1 7. dreams about task -0.2 0.41 -0.2 -0.3 0.75 0.32 1 8. dreams about personal life -0.3 0.56 0.04 -0.4 0.3 0.42 0.49 1 9. anxious feelings -0.3 0.23 0.15 -0.2 -0.1 0.59 0.24 0.27 1 10. happy emotion -0.3 0.28 0 -0.1 0.24 0.06 0.34 0.44 -0.1 1 11. sad emotion -0.3 0.04 0.21 0.02 0.07 0.37 0.07 0.25 0.48 0.14 1 12. angry emotion -0 -0.1 0.09 0.03 -0 0.1 0.16 0.07 0.41 -0.1 0.19 1 13. fear emotion 0.06 0.26 -0.1 -0.3 0.4 0.4 0.34 0.08 0.08 0.1 0.07 -0 1 14. surprise emotion -0.1 0.28 0.07 -0.3 0.23 0.29 0.19 -0.1 0.11 -0 0.1 -0.1 0.38 1 15. disgust emotion -0 0.19 -0.2 -0.1 0.59 -0.1 0.59 0.27 -0.1 0.36 -0.1 -0.1 0.13 -0.1 1 16. PSQI 0.1 -0 0.09 -0.2 0.28 0.12 0.31 0.09 0.14 0.12 0.01 0.11 -0 0.16 0.18 1 17. sleep quality -0.3 0.45 -0.1 -0.2 -0.1 0.23 -0 0.23 -0 0.23 0.11 0.06 0.08 -0 0.03 -0.4 1 18. PA film response 0 0.09 -0.3 0.17 0.03 -0.1 -0 -0.1 -0.5 -0.1 -0.3 -0.2 0.13 0.03 0.04 0.13 0.11 1 19. NA film response -0.1 -0.1 0.14 0.14 0.21 0.13 -0.1 -0 0.02 -0.1 0.15 -0.1 0.02 0 -0 0.15 -0.2 0.07 1 20. PA attenuation response -0.1 0.21 0.13 -0.2 0.12 0.15 0.15 0.42 0.06 0.19 0.04 0.21 -0.2 -0 0.22 0.26 0.25 0 0.24 1 21. NA attenuation response -0 0.03 -0.1 0.03 -0 -0.1 0.07 -0.2 -0.2 0.11 -0 -0.1 0.03 0.27 -0 0.11 -0 0.34 -0.5 -0.5 1 22. GDC valence -0.1 -0 -0 0.08 0.09 0.05 0.12 -0 0.02 0.16 -0.2 -0 -0 -0.2 0.13 -0.4 0.21 -0 -0.2 -0.1 0.18 1 23. GDC arousal 0 0.32 0.11 -0.5 -0.1 0.32 0.05 0.13 0.28 -0.2 0.11 0.11 0.1 0.16 -0.3 0.11 0 -0.2 0.01 -0.1 0.03 -0.1 1 24. GDC attitude -0.2 0.33 0.13 -0.3 0.16 0.51 0.25 0.38 0.24 -0 0.09 0.1 0.08 0.24 -0.1 0.31 -0.1 0.01 0.26 0.26 0.07 0.01 0.25 1 25. GDC recall 0.05 0.13 0.05 -0.2 0.16 0.05 -0 0 0.01 -0.2 0.11 0 0.08 -0 -0.1 0.05 0.06 -0.2 0.01 -0.1 0.09 -0.1 0.35 0.15 1 26. GDC nightmare -0.2 0.3 -0.1 -0.2 0.14 0.26 0.17 0.16 0.19 0.36 0.38 0.15 0.2 0.19 -0.2 0.3 0.15 -0 0.03 0.03 0.17 -0.2 0.15 0.3 0.33 1 27. GDC effect on mood -0.1 0.13 -0 -0.2 0.34 -0 0.31 0.1 0.05 -0 0.09 -0.1 -0.1 0.19 0.11 0.26 -0.3 -0.1 0.31 0.18 -0.2 -0.1 0.21 0.27 0.3 0.14 1 28. GDC effect nightmares -0 -0.1 0.06 0.05 0.2 0.16 0.07 0.04 0.1 0.05 0.02 -0.1 0.32 0.04 -0.1 0.15 -0.4 -0.1 0.35 0.08 -0.1 -0.1 -0.1 0.4 0 0.16 0.2 1 29. GDC lucid dreaming 0.1 -0.1 0.2 -0.1 0.16 0.36 0.03 0.12 0.2 -0.1 0.32 0.11 0.23 -0.1 -0.2 -0.1 -0.1 -0.1 0.05 -0.1 0.03 0.29 0.28 0.08 0.33 -0 0.02 0.07 1 30. Valence change emotional video -0.1 -0.3 -0 0.42 0.06 -0.2 0 0.08 -0.1 0.17 0.23 -0 -0.3 -0.2 0.02 0.31 -0.1 0.06 -0 0.09 0.07 -0.2 -0.3 -0.1 0.09 0.23 -0.1 -0 -0.1 1 31. Arousal change emotional video 0.16 -0 0.11 -0.2 0.03 0.09 -0.1 0.02 -0 -0.1 0.08 0.16 -0.2 -0.2 0.01 0.31 0.06 0.31 0.19 0.32 0.02 -0.1 -0 0.11 0.18 0.11 0.04 -0.1 0.11 0.11 1 32. Valence change neutral video -0.1 -0.1 0.16 0.09 0.01 0.03 -0.1 -0.2 0.06 0.01 0.13 0.15 -0 0.16 0.01 -0.1 0.18 -0.1 -0.1 0.2 -0.1 0.35 -0.1 -0.1 0.04 -0.1 -0.1 -0.1 0.12 0.08 0 1 33. Arousal change neutral video -0.3 0.23 0.21 -0.1 0.35 0.26 0.42 0.26 0.04 0.06 -0.2 0.05 0.2 0.21 0.080.36 -0.1 0.3 0.22 0.19 0.03 0.04 0.21 0.46 -0.1 -0 0.28 0.15 0.09 -0.1 0.04 -0.1 1
Appendix E: Analysis figures
Single level data
Note. All plots display the confidence interval in grey scale, and all presented p-values noted are
corrected for multiple testing with the False Discovery Rate (FDR) method.
Figure E
STAI-T (trait anxiety) score and awakening type ratios: wake (W) ratio (A), dream experience (DE) ratio (B), dream experience without recall (DEWR) ratio (C) and no experience (NE) ratio (D).
Figure F
STAI-T (trait anxiety) score and dream content: dreams about tasks (self-report; A and scored; B), dreams about personal life (C) and dreams containing threats (D).
Figure G
STAI-T (trait anxiety) score and PANAS difference scores: positive affect (PA) film response (A) and PA attenuation (B), negative affect (NA) film response (C) and NA attenuation (D).
Figure H
STAI-T (trait anxiety) score and valence score pre-post sleep difference for emotional video (A), arousal difference for the emotional video (B), and valence and arousal difference for the neutral video (C and D, respectively).
Figure I
Figure J
STAI-T (trait anxiety) score and General Dream Characteristics (GDC) questionnaire score for attitude towards dreams.
Figure K
STAI-T (trait anxiety) score and emotions in dream experiences (DEs): angry (A), anxious (B), happy (C) and fear (D).
