Vivid dreams are associated with a high percentage of REM sleep: a prospective study in veterans

Abstract

Study Objectives:

Vivid dreams are dreams that feel real or are associated with dream enactment behavior. They are prevalent in veterans, especially in those with psychiatric disorders such as post-traumatic stress disorders. Such psychiatric disorders have known association with abnormalities in rapid eye movement (REM) sleep. Vivid dreams are also described in neurological conditions, such Lewy body dementias, which are also associated with REM sleep abnormality. Although vivid dreams occur in neuropsychiatric disorders that have REM sleep abnormalities, there are no studies that have directly investigated an association between vivid dreams and REM sleep. We sought to study vivid dreams and REM sleep in veterans.

Methods:

Veterans undergoing polysomnography at our hospital were invited to enroll. Participants completed a dream-related questionnaire the morning after their polysomnography.

Results:

We prospectively enrolled 505 veterans. After a night in the sleep laboratory, 196 of 504 (39%) reported experiencing a dream, and, of those, 117 of 190 (62%) described their dream as vivid. Discrepancies in patient totals are secondary to missing questionnaire data. Our novel finding is that participants with a high percentage of REM sleep (above 25%) were more than twice likely to report a vivid dream than participants with a lower percentage of REM sleep (P < .0001). Nonvivid dreams were not associated with a high percentage of REM sleep.

Conclusions:

Vivid dreams are associated with a high percentage of REM sleep. Further research into the role of REM sleep abnormalities in vivid dreams may help to advance understanding of neuropsychiatric disorders.

Citation:

Fattal D, Platti N, Hester S, Wendt L. Vivid dreams are associated with a high percentage of REM sleep: a prospective study in veterans. J Clin Sleep Med. 2023;19(9):1661–1668.

BRIEF SUMMARY

Current Knowledge/Study Rationale: Vivid dreams are described in various neuropsychiatric disorders including post-traumatic stress disorder, depression, and Lewy body dementias. Abnormalities in rapid eye movement (REM) sleep are established in these same neuropsychiatric disorders. To our knowledge, there are no studies that have directly assessed an association between vivid dreams and REM sleep. Our study aimed to address this knowledge gap in a veteran population.

Study Impact: We found that vivid dreams were associated with a high percentage of REM sleep (%REM), defined as more than 25% REM sleep during an overnight polysomnography, whereas nonvivid dreams did not show an association with %REM. The association between vivid dreams and high %REM may aid in advancing our understanding of REM abnormalities in neuropsychiatric disorders that involve vivid dreams.

INTRODUCTION

A vivid dream is a dream that feels as clear and realistic as normal vision.^1,2 Vivid dreams occur more frequently during rapid eye movement (REM) sleep and can be seen in healthy individuals.^2–6 In addition, vivid dreams can be seen in individuals with neuropsychiatric disorders. From a psychiatric perspective, vivid dreams and nightmares, which are a form of vivid dreams with negative connotations,⁷ are reported in depression, anxiety disorders, and post-traumatic stress disorder (PTSD), including in veterans.^8–12 Depression and PTSD can be associated with a high percentage of REM sleep (%REM), including in veterans.^13–15 Moreover, a reduction in REM sleep (for example, by waking up participants during periods of REM sleep in the sleep laboratory) is associated with improved mood.¹⁶ From the perspective of neurology, vivid dreams are seen in patients with dream enactment behavior (DEB),¹⁷ a condition in which observable behavior and movements of a sleeping person replicate the behavior and movements that the person is simultaneously executing in their dream.⁷ Vivid dreams are seen in more than three-quarters of individuals who execute violent behavior during sleep.¹⁸ Furthermore, DEB might be a sign of REM sleep behavior disorder (RBD), as seen in Lewy body dementias (LBD), and vivid dreams occur in more than half of people with RBD.¹⁹ Thus, vivid dreams are linked to REM sleep both in health and in neuropsychiatric disorders.

Although studies have investigated the association between vivid dreams and neuropsychiatric disorders, including in veterans,^20,21 and the association between neuropsychiatric disorders and a high %REM,^13–15 no study has directly investigated the association between vivid dreams and %REM. The present work sought to study the relationship between vivid dreams and REM sleep in veterans.

Vivid dreams and nightmares are particularly prevalent in the veteran population due partly to the prevalence of neuropsychiatric disorders such as PTSD.¹² Moreover, vivid dreams and nightmares continue to have limited treatment options and thus continue to be a source of significant morbidity and increased risk of suicidality.^12,22 Nightmares can lead to avoidance of sleep, which leads to sleep deprivation, which in turns worsens nightmares.¹² In addition, sleep disorders among veterans have been increasing over the past two decades, making the morbidity and mortality related to sleep disorders a growing public health issue with their associated PTSD and other neuropsychiatric disorders and associated vivid dreams and nightmares.^22–24 Therefore, further understanding of vivid dreams and sleep is needed to better help veterans and the general population.

METHODS

We prospectively investigated the prevalence of vivid dreams and its relationship to %REM during polysomnography of several hundred veterans in the sleep laboratory at our Veterans Affairs (VA) Medical Center. Veterans undergoing polysomnography who agreed to participate in our study were asked whether they had vivid dreams during the night of monitoring. We hypothesized that vivid dreams are associated with a higher %REM.

Study participants

This study was conducted at our VA Medical Center, where we enrolled veterans. The term “veteran” means “a person who served in the active military, naval, or air service (of the United States of America), and who was discharged or released therefrom under conditions other than dishonorable”.²⁵ The study was conducted in two parts during two time periods: Part 1 (May–December 2021) and Part 2 (July–August 2022). The methods used for both parts were similar, with the exception that more detailed data was collected during Part 2.

During Part 1, we enrolled veterans when they came to our sleep laboratory for a polysomnography that was ordered by their providers, regardless of the reason for ordering the procedure. All veterans who consented were asked to complete a short questionnaire about dreams at the conclusion of their polysomnography the following morning. The only inclusion criterion was age (18–89 years). To keep our study simple and efficient, we excluded veterans 90 years of age or older; due to the small number of veterans in their 90 s, further research restrictions apply to that age group. No exclusion criteria were set. In Part 1 of our study, which was done under strict COVID restrictions, we aimed to get our veterans in and out of the lab as quickly as possible and to conduct our study efficiently, so we did not keep records of those who did not participate (we estimate that we enrolled at least 80% of veterans during that time). During Part 2 of our study, we kept an exact record of the number of veterans who consented and those who did not; our enrollment rate for Part 2 was 76% (74 out of 98 consented).

Because Part 1 of this study was performed during the COVID-19 pandemic, VA guidelines on personal protective equipment were followed by both staff and veterans. As such, our questionnaire was kept short to minimize the amount of time veterans needed to stay in the laboratory postpolysomnography. In Part 2, we were able to spend more time with the veterans based on the current VA guidelines, which allowed us to collect more data with an expanded questionnaire. Our study was approved by the Institutional Review Board at our VA Medical Center (Approval #202101393 for Part 1 of the study and #202206057 for Part 2).

The questionnaire

The morning after the polysomnography, each enrolled veteran was given a questionnaire. In Part 1 of the study, we used the following questionnaire (modified from Postuma et al.²⁶):

• Did you have a dream tonight? (Yes/No)

• If yes, is it a vivid dream, ie, looks real to you, or parts of the dream feel very close up, or in colors, or you think you acted it out with arm or leg movements or vocals? (Yes/No)

In Part 2 of the study, we used the following questions the morning after the sleep study:

• At home, have you ever been told, or suspected yourself, that you seem to “act out your dreams” while asleep (for example, punching, flailing your arms in the air, making running movements, etc.)? (Yes/No)

• This night in the sleep lab, did you have a dream? (Yes/No)

• If yes, do you suspect you acted it out? (Yes/No/Do not know)

During Part 2 of the study, we also collected data about the presence of PTSD, depression, or anxiety disorders, as listed in the problem list of each veteran in the electronic medical record on the sleep consult itself.

Polysomnography

Each veteran participated in an overnight polysomnography. All polysomnographies were performed in accordance with American Association of Sleep Medicine criteria^27,28 and included the following measurements: electroencephalography activity (electroencephalography leads at locations F3–A1, C3–A1, and O1–A1 on the scalp); electrooculographic activity; submental electromyographic activity; lower limb movement, monitored by video; electrocardiographic activity; chest and abdominal movements for monitoring airflow; and oxygen saturation (SaO₂; we measured the lowest SaO₂ recorded for that night). We also documented by video and electromyographic activity whether there was evidence of abnormal muscle movements during REM sleep, known as REM sleep without atonia (RWA). We use the “RWA” abbreviation, as proposed by the American Association of Sleep Medicine, although we are aware that other bodies use the abbreviation “RSWA.” If RWA was observed, we did not wake up our participants; instead, we relied on their self-reported report of DEB the following morning, as recorded in the questionnaire. All polysomnographies were scored and interpreted in accordance with standard criteria ^berry by an independent sleep physician who was blinded to our research study and questionnaire responses.

Sleep parameters

From the polysomnographies we collected age, sex, body mass index, apnea-hypopnea index, total sleep time, percentage sleep efficiency, percentage awake time, SaO₂, and %REM. The normal %REM in healthy adults is 19.0% (18.5–19.6%).²⁹

Statistical analysis

Categorical variables are described using counts and percentages, while continuous variables are described using median and interquartile ranges. Associations between categorical variables were evaluated using Fisher’s exact test. For continuous measures, participants were compared between strata using Wilcoxon rank-sum tests. Spearman rank correlations were used to assess the relationship between two continuous variables. Due to the exploratory nature of the analysis, no multiple comparison adjustments were implemented. After the first part of the study was conducted, an interim analysis was performed and deemed that more data should be collected. Thus, we did a second part of the study and then combined all samples into one analysis. SAS version 9.4 and R version 4.1.3 were used for all analyses. Data from Parts 1 and 2 of the study were combined whenever the questions asked in the two studies were compatible and were analyzed separately in other instances.

RESULTS

We enrolled a total of 505 veterans. During Parts 1 and 2 of our study, we enrolled 433 and 72 veterans, respectively. In Part 1, 17 of the participants who stated they did not have a dream during the night also responded to the follow-up question by stating that their dream was vivid. Since they stated that they had a vivid dream, we considered these participants to have had a dream. Data were missing for 7 individuals, which we described later with the relevant analyses.

The ages of our 505 enrolled participants ranged from 24 to 89 years (median 60; interquartile range 46–72). Thirty-one (6.1%) were female while 474 (93.9%) were male.

Prevalence of vivid dreams

All but one participant (504 participants) answered the question on whether they dreamt during the night. In the total cohort (504), 196 (39%) indicated experiencing a dream in the laboratory the night of the polysomnography. Of those who dreamt in the laboratory, 117 of 190 (61.6%) described their dream as vivid (6 had missing data). Specifically, in Part 1 of the study, 115 of 179 (64.2%) described their dream as vivid (ie, the dream felt real, or they thought they acted it out). In Part 2 of the study, 12 of 72 participants indicated they dreamt in the laboratory. Of those, 2 of 11(18%) thought they acted it out (1 had missing data). In Part 2 we also asked participants whether they thought they acted out their dreams at home; 15 of 72 (20.8%) responded in the affirmative.

Vivid dreams and high %REM

In the total sample of 505, we analyzed the association between %REM and the presence of vivid dreams. Of these, 7 participants had missing relevant data, making the total sample size 498 for this analysis. As mentioned, 117 of the 190 who reported having a dream described their dream as vivid. Within each of three ranges for %REM (0%–10%, 10%–20%, or 20%–25%), the proportion of participants who had vivid dreams was similar (15%, 20%, and 19% for each range, respectively). However, of the 93 veterans with %REM more than 25%, 43 (48%) had vivid dreams (P < .001; Table 1). Nonvivid dreams did not have a similar association with high %REM (Table 1).

Table 1.

Vivid and nonvivid dreams in laboratory/at home and %REM sleep.

Characteristic	0%–10% (n = 142)	10%–20% (n = 184)	20%–25% (n = 85)	>25% (n = 93)
Vivid dream, n (%)	21 (15)	36 (20)	16 (19)	43 (48)
Missing data	0	1	2	4
Nonvivid dream, n (%)	18 (13)	24 (13)	18 (22)	13 (15)
Missing data	0	1	2	4

%REM = percentage of REM sleep.

Vivid dreams, age, and sleep parameters

Participants who reported experiencing a vivid dream were generally younger (average age 60 years) than participants who reported experiencing a nonvivid dream (average age 64 years) (P = .035, Table 2). Otherwise, we found no associations between vivid or nonvivid dreams with any sleep parameter tested: the presence of apnea, severity of apnea, body mass index, sleep efficiency, total awake time, or SaO₂ (Table 2).

Table 2.

Comparing sleep variables between vivid dreams/dream enactment behavior (DEB) and nonvivid dreams.

Characteristic	Overall (n = 190)a	Nonvivid Dream (n = 73)a	Vivid Dream (n = 117)a	P b
Apnea-hypopnea index	10 (3, 23)	9 (3, 27)	10 (3, 21)	0.8
Age	62 (46, 72)	64 (50, 73)	60 (44, 71)	0.035
Body mass index	31 (28, 35)	31 (28, 36)	31 (28, 35)	0.5
Total sleeping time (hours)	5.68 (4.64, 6.42)	5.52 (4.46, 6.21)	5.85 (4.67, 6.53)	0.14
Sleep efficiency%	0.80 (0.66, 0.87)	0.79 (0.65, 0.87)	0.81 (0.67, 0.88)	0.6
Awake time%	0.15 (0.07, 0.25)	0.16 (0.07, 0.26)	0.14 (0.07, 0.25)	>0.9
Low SaO2	0.84 (0.77, 0.88)	0.84 (0.77, 0.88)	0.83 (0.77, 0.87)	0.6

^a Median (interquartile ratio). ^bWilcoxon rank sum test. SaO₂ = oxygen saturation.

Vivid dreams and associated psychiatric diagnoses

In Part 2 of our study, we collected psychiatric diagnoses data on the 72 participants (n = 72). Seventeen of 72 (24%) had PTSD, 13 (18%) had depression, and 10 (14%) had an anxiety disorder. Those with PTSD were more likely to report vivid dreams at home (46.7% vs 17.5%, P = .036). The combined measure of PTSD, depression, and anxiety disorders had a positive though nonsignificant association with %REM (P = .052) and no association with vivid dreams (P > .99).

Male and female distribution

The sex distribution of the veterans who participated in our study (6.1% females, 93.9% male) was consistent with that of the overall veteran population.³⁰ Because this was a predominantly male population, we did not do any further sex analyses.

DISCUSSION

Our novel finding is that veterans with a high %REM (defined as more than 25%) were more than twice as likely to report having a vivid dream than those with a lower %REM. We did not see a similar association between nonvivid dreams and high %REM. We did not find similar results in the literature. Therefore, we discuss our findings further using what is known about dreams, vivid dreams, and REM sleep.

Dreams are linked to REM sleep in both health and disease.³¹ The interrelation between dreams and REM sleep has been known since the 1950 s.³² In fact, it is the discovery of REM sleep in 1953 by Eugene Aserinsky and Nathanial Kleitman at the University of Chicago that “disrupted the supernatural zeitgeist (related to dreams) and placed dreams firmly into the realm of scientific investigation.”³¹ Due to the presence of hyperkinetic saccades, Aserinsky and Kleitman coined the term “rapid eye movement sleep.”³¹ However, many decades later the roles of sleep, REM sleep, and dreams remain largely unclear.³³

REM sleep is characterized by a loss of postural skeletal muscle tone (or atonia) due to processes in the pons and medulla that inhibit movement.⁷ In addition, REM sleep is characterized by several other components: electroencephalography activity akin to awake state, REM intermittent muscle twitches, a hippocampal theta rhythm, suppression of thermoregulation, autonomic and respiratory activation, and an elevated arousal threshold.³³ REM sleep exists in terrestrial mammals, birds, reptiles, and some aquatic invertebrates.³³ Its widespread conservation across species points to the pivotal role it plays, although its exact role is not fully understood.³³ For example, REM sleep is postulated to have a role in emotional memory consolidation; however, some highly intelligent animals, such as elephants, spend very little time, if any, in REM sleep.³³

Dreaming is defined as a subjective experience during sleep, which can include “complex multisensory content with temporal progression and narrative structure.”⁷ Although dreams can arise during non-REM sleep, most individuals have a higher recall of dreams that occur during REM sleep (up to 80%),^2,7,33 than during non-REM sleep (7%–50%).^2,5,7 Furthermore, dreams occurring out of REM sleep are often more complex, vivid, and bizarre than those occurring out of non-REM sleep.¹²

Prevalence of dreams and vivid dreams

In our study, 39% of participants indicated experiencing a dream after a night in the sleep laboratory. In population surveys, some people recall no dreams, while others recall a dream nightly, and on average people remember one dream a week.⁷ In another population survey, 76% recalled a dream at least once a week.³⁴

Our findings were that 64% of our veterans (both with and without PTSD) had a vivid dream. We did not find literature on vivid dreams and %REM, but there is literature about prevalence of nightmares. In the general population, the prevalence of nightmares is 5% to 27%,^5,35 and in a review of PTSD and nightmares, 50% to 70% of patients with PTSD had recurrent nightmares.¹¹ In veterans with PTSD, Neylan et al found that 52.4% of veterans with PTSD vs 5.7% of veterans without PTSD had nightmares.⁸ Woodward et al found that of 63 Vietnam veterans, 76% reported trauma-related nightmares.⁹ Furthermore, Vietnam veterans with PTSD reported on average 0.8 nightmares per night or had a nightmare 5 to 6 nights per week.¹⁰

Prevalence of DEB

We found that 20% of our participants have DEB at home and that this number was 40% in participants with PTSD vs 17% of those without PTSD. These findings are consistent with the literature. In a study of 78 veterans with PTSD, Jones et al found that 24.4% indicated experiencing DEB at least once per week at home, and this behavior occurred more frequently in veterans with more severe PTSD.³⁶ Furthermore, Lee et al found that veterans with PTSD (n = 20) reported more frequent DEB than controls without a history of trauma (n = 21) at 55% and 23.8%, respectively.³⁷

Vivid dreams and age

In our study, veterans who reported having a vivid dream were younger (average age 60 years for vivid dream vs 64 years for nonvivid dream, P = .035). In the literature, Haridi et al found nonstatistically significant differences in reported vivid dreams between young (average age of 30–39) (75.6%) and old (average age of 60–69 (63.6%) controls; %REM was not measured.³⁸ Other studies have found DEB to be more prevalent among older adults,³⁹ possibly because older people have a higher risk of LBD than younger individuals. More research into the association between age and vivid dreams, independent of the risk of LBD, is warranted.

Vivid dreams and sleep parameters

Our finding that vivid dreams did not correlate with sleep apnea, sleep apnea severity, or other sleep parameters aligns with the literature on nightmares. A study of nightmares in 35 Australian veterans in a hospital psychiatric ward did not find a significant relationship between the presence or absence of nightmares and the presence of sleep apnea or severity of apnea.⁴⁰ Woodward et al did not find any association between nightmares and total sleep time in 63 veterans with PTSD.⁹ Finally, in another veterans study, the lowest SaO₂ was the same in veterans with PTSD compared to trauma-exposed individuals without PTSD and controls.³⁷

Vivid dreams and elevated %REM

In this study, veterans with %REM of more than 25% were more than twice as likely to report having a vivid dream than those with a lower %REM. We know that vivid dreams are associated with REM sleep itself. Dreams that occur during REM sleep tend to be emotional, presumably due to elevated limbic system activity during REM sleep.⁴¹ In conditions of low stress, dreams might be crucial for processing emotions, and they might reflect a sense of well-being; however, during times of high stress, the level of dream recall can increase, specifically of vivid dreams with negative valiance like nightmares.⁶ It is possible that an individual’s dream experience may play a pivotal role in emotional regulation, although our understanding of the neural basis for and the function of REM-based dreams remains limited.⁵ One study that evaluated emotional dreams and REM sleep found that when an individual’s recent (defined as within 2 days prior to an emotional dream) waking-life experiences were incorporated in REM dreams, the dreams were “positively correlated with frontal theta activity in the REM sleep period.”⁴² In this study, no such correlation was observed for older memories (defined as events in daily logs collected 3–10 days prior to an emotional dream), nor was there an electroencephalography correlate during slow-wave sleep dreams.⁴² These results agree with theories that “dreaming reflects emotional memory processing taking place in REM sleep.”⁴² Furthermore, in individuals who have nightmares, theta changes (2.0–5.0 Hz) are most clearly reported during REM sleep.⁴³ Of interest is that the left hippocampus is a specific brain region that is implicated in the vividness of a dream: this region’s mean diffusivity (as identified via diffusion tensor imaging and impacted by the number of cellular barriers that prevent free water diffusion into tissue) is negatively associated with the visual vividness of a recalled dream.¹ According to the stress acceleration hypothesis of nightmares, fear-processing networks such as the amygdala, hippocampus, and medial prefrontal cortex extinguish fear during REM sleep, and thus REM sleep disorders can lead to nightmares.⁴⁴ Therefore, vivid dreams are associated with REM sleep.

Since we did not find a similar association between vivid dreams and high %REM in the literature, we explored the literature on nightmares. Woodward et al noted that in 63 unmedicated, nonapneic, Vietnam combat veterans, %REM was “mildly elevated” at 26% in veterans with nightmares and 27.7% in veterans with trauma-related nightmares.⁹ Kis et al studied %REM in those with nightmares vs a control group over 2 nights. Although their aim was not to compare the two groups specifically for %REM, they found that %REM for those with nightmares vs controls was 25.9% vs 22.2%, respectively, during the first night of the sleep study, and 29.1% vs 25.9%, respectively, during the second night.⁴⁵ In this study, an interaction model with group and night was constructed, and only marginal significance was detected between those with nightmares and controls. In another study, Nielsen et al predicted that participants with frequent nightmares would have a high %REM.⁴⁶ However, among 14 participants with frequent nightmares vs 11 controls, they did not find increased %REM, which surprised the authors. They postulated that since no nightmares were reported during polysomnography recordings, the samples of sleep they collected in the laboratory “may not be representative of these subjects’ typical sleep.”⁴⁶ Furthermore, the authors stated, “This lack of representativeness may mean that, for still unexplained reasons, the suspected intensification of REM propensity (as measured by %REM) was temporarily absent during our polysomnography recordings.”⁴⁶ Although we did not ask specifically whether our participants’ vivid dream was a nightmare, our finding of high %REM in those with a self-reported vivid dream is similar to that reported and predicted in those with nightmares.

Putting it all together: vivid dreams and nightmares, neuropsychiatric disorders, and high %REM

There seems to be a triangular relationship among vivid dreams/nightmares, certain neuropsychiatric disorders, and high %REM. First, vivid dreams, including nightmares, are associated with neuropsychiatric disorders.^8–12,47 Individuals who attempt suicide have higher rates of nightmares than those who are depressed and without suicide attempts (96% vs 79%, respectively).⁴⁸ Depression is associated with more negative dream content.⁴⁹ PTSD patients report vivid dreams related to the traumatic events they have experienced, which may be secondary to both the amplification and change in function of the amygdala during REM sleep in these patients.⁵

Second, neuropsychiatric disorders seem to be associated with high %REM. High %REM has been reported in depression, and a reduction in %REM is associated with improved mood.¹⁶ In addition, antidepressant medications, many of which suppress REM, possibly exert their effect by normalizing %REM.^50,51 Furthermore, a systematic review that included more than 100 participants with seasonal affective disorder found that %REM was increased during acute and remitted phases of seasonal affective disorder.⁵¹ Furthermore, increased REM density (another quantitative measure of %REM) was seen in healthy participants who self-reported having a first-degree relative with an affective disorder.⁵² However, such findings are not universal. For example, in a review by Benca et al, most but not studies on affective disorder vs control found elevated %REM.⁵³ In our study, our finding of a trend of association (P = .052) in our small sample of veterans with PTSD, depression, or anxiety disorders and %REM is consistent with high %REM in psychiatric disorders.^49,50,52 Percentage of REM also seems to be higher in those with PTSD. In a PTSD study of 11 veterans with PTSD vs 8 veterans without PTSD, the average %REM was 26.1% vs 19.4%, respectively.¹³ In another study, 17 veterans with PTSD vs 11 controls (including 2 veterans) without PTSD had an average %REM of 19.1% vs 13.0%, respectively.¹⁴ Additionally, a large community sample of older adult male veterans (n = 59) who had been exposed to war trauma 28 to 50 years earlier and who did not take sleep-affecting medications found that the 30 veterans with PTSD had a higher %REM than the 29 non-PTSD veterans (23% vs 19%).¹⁵ However, high %REM has not been consistently documented in all PTSD studies. In a study of 24 veterans with PTSD, with or without mild traumatic brain injury, Wallace et al did not find an increase in %REM in PTSD participants vs healthy veteran controls.⁵⁴ In another study of 20 participants with a history of trauma and PTSD vs 23 with a history of trauma but without PTSD vs 21 controls, the average %REM was 20.48%, 20.57%, and 18.14%, respectively.³⁷ The lower %REM in PTSD veterans found in some studies could be due to antidepressants that suppress REM.^55,56 Therefore, there seems to be a connection between PTSD and high %REM sleep disturbances,¹¹ and this association is worth noting and exploring further.

Third and finally, our work showing that vivid dreams are associated with high %REM represents the third link, closing the triangle of these associations. Our work also supports Scarpelli’s observation that “abundant dreams or nightmares are related to REM sleep abnormalities and psychiatric disorders.”⁵ In neurological conditions, vivid dreams are seen in more than half of individuals with RBD,¹⁹ which serves as a surrogate for REM disturbances.¹¹ In the literature we did not specifically find that %REM sleep is higher in RBD or in LBD patients, but this is worth exploring.

Limitations

Our study had limitations. (1) We did not divide our sample into those with PTSD vs without PTSD but recommend exploring these groups further. (2) We did not account for participants who were taking antidepressants. Given the association between affective disorders such as depression and vivid dreaming, one can speculate that participants who reported vivid dreams may have been more likely to be depressed and taking antidepressants. Because certain antidepressants suppress %REM, it is possible that the %REM difference we found was an underestimate for participants using selective serotonin reuptake inhibitors, serotonin-norepinephrine reuptake inhibitors, tricyclic antidepressants, or monoamine oxidase inhibitors. The relationship may be overestimated in those using bupropion for depression, as this medication enhances REM.⁵⁰ (3) This study was done during the COVID-19 pandemic, thus we were unable to do a 2-night study. First-night effect on %REM is a well-known phenomenon, causing a reduction in %REM (by an average of 2%)⁵⁷; any differential effect of this phenomenon on persons with vivid vs nonvivid dreams, if any, will need further exploration. (4) Finally, our sample predominantly comprised a male population of veterans, thus limiting the generalizability of our results. Though there is no a priori reason or current evidence that the status of %REM would be different in a veteran vs matched control, further exploration of our results in the nonveteran population may be warranted.

Future directions

Another angle to this work is to research the presence of vivid dreams and %REM in individuals with psychiatric disorders who are on antidepressants (selective serotonin reuptake inhibitors, serotonin-norepinephrine reuptake inhibitors, tricyclic antidepressants, or monoamine oxidase inhibitors) vs those who are not medicated and excluding those taking bupropion alone or in combination. We can also study vivid dreams and %REM in the different types of neuropsychiatric disorders, such as in different anxiety disorders, and in PTSD vs no-PTSD.¹² Furthermore, since patients with LBD and RBD have abnormalities in REM sleep (namely RWA), and since they also commonly have vivid dreams, we recommend further exploration of vivid dreams and %REM in patients with RBD and LBD. In our study, we captured RWA in only 2 veterans, and thus we could not shed light on any association between RWA and %REM. In addition, we can study the number of REM epochs during nights with vivid dreams vs those with nonvivid dreams; preliminary work has shown that those with nightmares have a higher number of REM epochs.⁴⁶ Moreover, we recommend repeating this work with 2-night studies, adding more detailed questionnaires about sleep habits, exploring the REM index, and further studying REM sleep around the time of the vivid dream. Finally, we recommend studying lucid dreaming and %REM.

CONCLUSIONS

Veterans with a high %REM (more than 25%) were twice as likely to have vivid dreams than those with a lower %REM. There was no such association between %REM and nonvivid dreams. The relationship between vivid dreams and REM sleep needs further exploration in various psychiatric disorders such as PTSD, depression, and anxiety disorders.⁵⁸ There is also a need to further explore the relationship between vivid dreams and REM sleep in neurological conditions such as RBD and LBD, in which it is known that patients commonly have vivid dreams and have a known REM sleep abnormality (namely RWA). This study adds a new angle to the field. Further work is needed to elucidate the significance of the association between vivid dreams and REM sleep and their interaction and function in health and in neuropsychiatric disorders. This issue will remain relevant since sleep disorders, and their associated vivid dreams and nightmares, are becoming more prevalent over time, at least in the veteran population.²⁴

DISCLOSURE STATEMENT

All authors have seen and approved the manuscript. Institution where work was performed: Iowa City VA Medical Center. The authors report no conflicts of interest. Research reported in this publication was supported by the National Center for Advancing Translational Sciences of the National Institutes of Health under Award Number UL1TR002537. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

At the time of the study, Dr. Platti was affiliated with the University of Iowa Carver College of Medicine. She is currently affiliated with University of South Florida, Neurology Department.

ABBREVIATIONS

DEB

dream enactment behavior

LBD

Lewy body dementias

PTSD

post-traumatic stress disorder

RBD

REM sleep behavior disorder

%REM

percentage of REM sleep

REM

rapid eye movement

RWA/RSWA

REM sleep without atonia

SaO₂

oxygen saturation

VA

Veterans Affairs