Polysomnographic Sleep in Idiopathic Nightmares: A Systematic Review and Meta-Analysis

Abstract

Objectives

This study conducted a systematic evaluation of polysomnographic differences between individuals with nightmares and healthy controls.

Methods

Electronic searches of All EBM databases, EMBASE, Web of Science, and MEDLINE identified 14 eligible case-control studies; seven studies involving 112 individuals with nightmares and 112 controls were included in meta-analysis.

Results

The meta-analysis of sleep macro-architecture (such as total sleep time, wake after sleep onset, and sleep stage percentages) revealed no significant differences between individuals with nightmares and healthy controls. Furthermore, the limited available data regarding sleep micro-architecture (cyclic alternating patterns (CAP) and power spectral analysis (PSA)) and autonomic measures (heart rate and heart rate variability) in studies not subjected to meta-analysis makes it difficult to draw confident conclusions.

Conclusions

Our meta-analysis revealed no significant differences in sleep macro-architecture. This finding highlights the importance of investigating sleep micro-architecture (such as CAP and PSA), which may be relevant to the physiology of nightmares. However, these findings should be interpreted cautiously and validated in future studies because of the limited available data.

Introduction

Nightmare disorder, a type of parasomnia, is defined by the frequent occurrence of vivid and intensely distressing dreams that are typically recalled upon awakening. These dreams mainly occur during rapid eye movement (REM) sleep1 and typically arise in the second half of a sleep episode.2 Frequent nightmares result in impaired daytime functioning, such as occupational, social, or educational dysfunction. Nightmares are linked to the progression of psychiatric disorders, such as major depressive disorder (MDD),3–5 anxiety disorder,3 and posttraumatic stress disorder (PTSD).6,7

Polysomnography (PSG) is recognized as the gold standard for the objective evaluation of sleep, which may help elucidate the pathophysiology and etiology of nightmares. Although previous PSG studies have suggested alterations in sleep quantity and quality in individuals with nightmares, the findings (ie, total sleep time (TST) and N1 percentage) remain notably inconsistent.8–10 Mechanically, the developments of nightmares are associated with the disrupted fear extinction11 and emotional processing during REM sleep,12,13 suggesting the alterations in REM parameters may be considered as biomarkers of nightmares. However, the findings for REM parameters are also inconsistent. For instance, previous studies reported prolonged REM sleep percentage14 and REM latency15 in individuals with nightmares, but it was not observed in other studies.8,16 In addition, for power spectral analysis (PSA), Simor et al17 reported elevated high alpha power during REM sleep in individuals with nightmares compared with controls. By comparison, Marquis et al16 found altered slow-theta activity in individuals with nightmares during REM sleep compared with controls. These varied findings across different studies may be largely due to variations in demographic characteristics, clinical features, and experimental methodology, highlighting the need for a quantitative synthesis of PSG findings in individuals with nightmares.

Meta-analytic approaches can synthesize inconsistent findings, evaluate potential moderators, and determine whether routine PSG measures are clinically informative in idiopathic nightmares. Identifying objective sleep features may be helpful to improve clinical management of the disease. We hypothesized that individuals with nightmares would exhibit differences in PSG measured sleep parameters compared with healthy controls, particularly in REM sleep parameters. The null hypothesis was that PSG measured sleep parameters would not show significant differences between individuals with nightmares and healthy controls. Where possible, we investigate potential moderators that might account for heterogeneity among studies.

Methods

Protocol and Registration

The study protocol was registered on PROSPERO (ID: CRD42024582896; https://www.crd.york.ac.uk/PROSPERO/view/CRD42024582896) following the PRISMA statement, which outlines the preferred reporting items for systematic reviews and meta-analyses.18

Literature Searching

The databases MEDLINE, EMBASE, all EBM databases via OVID, and Web of Science were searched using the search terms (“Nightmare” OR “Nightmares”) AND (“Polysomnogra*” OR “PSG” OR “Sleep architect*” OR “Sleep monit*” OR “Sleep stage*”) in the Title/Abstract or Keywords section. The search was initially executed on September 9, 2024, and repeated between October 20 and December 13, 2025, to find recent publications and incorporate the latest data (Tables S1–S4). Two reviewers (J.X.H. and M.O.W.) conducted independent study selection. Any disagreements were resolved through consultation with the senior author (Y.Z.) to achieve consensus.

Inclusion and Exclusion Criteria

Studies were eligible if they met the following inclusion criteria: (1) nightmare disorder was diagnosed according to the Diagnostic and Statistical Manual of Mental Disorders19 and the International Classification of Sleep Disorders20–22 (2) including a healthy control group; (3) comparing PSG variables between individuals with nightmares and healthy individuals; (4) written in English. If participants were included in multiple studies, only the dataset containing the most detailed information was selected. We excluded (1) studies that included individuals with neurological and psychiatric disorders (such as PTSD, MDD, and/or anxiety disorder); (2) case series/case reports, guidelines, review papers, editorials, statements, comments, and animal studies; (3) studies unrelated to nightmares; and (4) abstracts of studies that explicitly reported the absence of PSG data or the lack of comparisons of PSG data between individuals with nightmares and healthy controls.

Data Collection Process and Quality Assessment of Included Studies

We extracted sleep macro-architecture (TST, wake after sleep onset (WASO), sleep efficiency (SE), sleep latency (SL), REM density, REM efficiency, REM latency, non-rapid eye movement (NREM) awakenings, REM awakenings, arousal index, number of REM periods, REM/NREM cycle duration, and the percentage of stages N1, N2, N3, and REM sleep). Additionally, we extracted autonomic parameters during sleep, including heart rate (HR), heart rate variability (HRV), respiratory sinus arrhythmia (RSA), breathing cycle length, and sleep micro-architecture (cyclic alternating patterns (CAP) and PSA). The extracted demographic and clinical variables encompassed the number of participants, age, sex (percentage of male participants), comorbidity of other sleep disorders (obstructive sleep apnea (OSA) and periodic limb movement disorder (PLMD)) (Yes vs No), adaptation nights (Yes vs No), and PSG scoring criteria (American Academy of Sleep Medicine (AASM) or Rechtschaffen and Kales (R&K)) (Tables 1 and S5).30 The risk of bias in the included studies was evaluated by employing the Quality Appraisal Criteria - adapted from the Joanna Briggs Institute.31

Table 1.

Study Characteristics (for More Information Please See Supplementary Table S5)

Study	Sample Size	Percentage Male	Mean Age	Adaptation	PSG Scoring Methods	Excluding Drug Impacting Sleep
Germain et al, 20038	11 NM	45.5%	28.2 ± 5.3	YES	R&K	YES
	13 controls	53.8%	32.6 ± 11.2
Nielsen et al, 201015	16 NM	37.5%	26.1 ± 8.7	NO	NR	YES
	11 controls	36.4%	27.1 ± 5.6
Simor et al, 201214	17 NM	58.8%	20.65 ± 1.73	YES	R&K	YES
	23 controls	47.8%	21.35 ± 1.61
Simor et al, 2013a17	19 NM	52.6%	20.87 ± 1.57	YES	R&K	YES
	21 controls	52.4%	21.57 ± 1.47
Simor et al, 2013b23	17 NM	58.8%	20.65 ± 1.73	YES	R&K	YES
	23 controls	47.8%	21.35 ± 1.61
Simor et al, 201424	19 NM	52.6%	20.94 ± 1.58	YES	AASM	YES
	21 controls	52.4%	21.57 ± 1.47
Paul et al, 201510	17 NM	5.9%	25.18 ± 3.66	YES	AASM	YES
	17 controls	5.9%	24.31 ± 4.69
Paul et al, 201925	19 NM	5.3%	24.1 ± 4.5	YES	AASM	YES
	19 controls	5.3%	25.2 ± 3.7
Blaskovich et al, 2020a9	22 NM	36.4%	22.94 ± 3.54	YES	AASM	YES
	22 controls	27.3%	21.73 ± 2.73
van der Wijk et al, 202026	22 NM	36.4%	22.91 ± 3.1	YES	NR	YES
	21 controls	28.6%	21.46 ± 1.43
Blaskovich et al, 2020b27	22 NM	36.4%	22.91 ± 3.1	YES	AASM	YES
	23 controls	26.1%	21.83 ± 1.85
Tomacsek et al, 202428	24 NM	37.5%	23.125 ± 3.33	YES	AASM	YES
	30 controls	36.7%	22.1 ± 3.49
Perogamvros et al, 201929	11 NM	27.3%	34 ± 9.3	NR	AASM	YES
	11 controls	27.3%	33 ± 9.2
Marquis et al, 201716	18 NM	46.7%	23.17 ± 4.15	NO	AASM	YES
	15 controls	16.7%	23.67 ± 3.83

Notes: Simor et al, 2013a refers to the study published in Biological Psychology; Simor et al, 2013b refers to the study published in Sleep; Blaskovich et al, 2020a refers to the study published in Sleep; and Blaskovich et al, 2020b refers to the study published in Journal of Sleep Research.

Abbreviations: NM, nightmare group; %, percentage; NR, not reported; PSG, polysomnography; AASM, American Academy of Sleep Medicine; R&K, Rechtschaffen and Kales.

Statistical Analysis

The meta-analysis was performed using the Comprehensive Meta-Analysis software. To determine the combined effect size (standardized mean difference (SMD)) for the variations in PSG parameters between individuals with nightmares and healthy controls, the sample size, mean, and standard deviation for each group were entered for analysis. The Q statistic and I² were used to assess the presence and magnitude of between-study heterogeneity.32,33 Forest plots were also inspected to visualize the overlap of 95% confidence intervals across studies. Due to the methodological and clinical heterogeneity across studies, a random-effects model was employed to obtain relatively conservative estimates.34,35 The Egger regression method36 was employed to assess publication bias, with p-values of less than 0.05 indicating the existence of bias. Sensitivity analyses were conducted to investigate potential sources of heterogeneity (comorbidity of sleep disorders and adaptation nights). Additionally, a subgroup analysis was performed based on PSG scoring rules to investigate heterogeneity between studies. A meta-regression analysis was conducted to assess the effects of age, gender, and year of publication on the meta-analysis findings.

Results

Study Selection

Our initial search resulted in 1517 publications (Figure 1). Following the elimination of duplicates, we conducted a screening of the titles and abstracts of the 1033 remaining articles. An additional study was identified by searching the reference lists.16 Our systematic review encompassed fourteen studies (Tables 1 and S5), yet the meta-analysis incorporated only seven out of these fourteen studies. The reasons for excluding certain studies are detailed in Table S6.

Figure 1.

Flow chart used for the identification of eligible studies. A total of 1517 records were identified through database searches, and 1033 records remained after duplicate removal. After title and abstract screening, 33 full-text articles were assessed for eligibility. Fourteen studies were included in the qualitative synthesis, and seven studies were included in the quantitative meta-analysis.

Description of the Included Studies

The fourteen studies included sample sizes varying from 22 participants29 (11 individuals with nightmares and 11 healthy controls) to 54 participants28 (24 individuals with nightmares and 30 healthy controls). The average age of individuals with nightmares and healthy controls spanned from 20.7 to 34 years, as reported in fourteen studies. The proportion of males among individuals with nightmares and healthy controls ranged between 5.3% and 58.8%. AASM PSG criteria were employed in eight studies for sleep scoring,9,10,16,24,25,27–29 while the remaining studies utilized R&K criteria8,14,17,23 (the scoring criteria were not specified in two studies15,26). Eleven studies8–10,14,17,23–28 incorporated adaptation nights for PSG recordings, whereas two studies15,16 conducted only one night of PSG without an adaptation period, and another study29 did not explicitly mention whether there was an adaptation night. All included studies conducted PSG on medication-free nightmare patients. Table S7 contains the quality assessments for these studies.

Meta-Analyses

In the overall sample, meta-analysis indicated no significant differences between nightmare patients and controls in terms of TST, WASO, SE, SL, REM latency, REM efficiency, and percentage of stages N1, N2, N3, and REM sleep (p > 0.05) (Table 2 and Figure 2). Where possible, the Egger test revealed no evidence of publication bias in these results (Figures S1–S9).

Table 2.

Summary of Meta-Analysis Comparing Individuals with Nightmares and Controls

	No. of Comparisons	No. of Participants (NM/control)	Means of NM	Means of Controls	SMD (95% CI)	Q	I2
TST, min	5	78/72	416.814	424.136	−0.090 (−0.490 to 0.310)	5.918	32.411
WASO, min	5	84/84	33.626	28.038	0.238 (−0.067 to 0.544)	0.717	0.000
Sleep efficiency, %	7	112/112	91.530	92.399	−0.181 (−0.449 to 0.087)	6.104	1.698
Sleep latency, min	7	112/112	16.203	11.700	0.139 (−0.126 to 0.404)	3.824	0.000
N1 %(%TST)	7	112/112	6.679	7.076	0.206 (−0.060 to 0.472)	5.847	0.000
N2 %(%TST)	7	112/112	45.641	47.523	−0.020 (−0.323 to 0.283)	7.710	22.175
N3 %(%TST)	7	112/112	24.473	23.590	−0.274 (−0.739 to 0.192)	17.513	65.739
REM %(%TST)	7	112/112	20.936	20.311	0.223 (−0.069 to 0.515)	7.164	16.250
REM latency, min	6	101/101	80.906	86.197	−0.186 (−0.616 to 0.244)	11.400	56.142
REM efficiency, %	3	45/39	88.71	88.54	0.029 (−0.404 to 0.462)	1.154	0.000

Abbreviations: TST, total sleep time; WASO, wake after sleep onset; REM, rapid eye movement; NM, nightmare group; SMD, standardized mean difference; CI, confidence interval; Q, Cochran’s Q statistic; I², inconsistency statistic.

Figure 2.

Forest plots for meta-analyses of polysomnographic parameters in individuals with nightmares compared with healthy controls. Panels show pooled standardized mean differences (SMDs) with 95% confidence intervals (CIs) for (A) total sleep time, (B) wake after sleep onset, (C) sleep efficiency, (D) sleep latency, (E) N1 percentage, (F) N2 percentage, (G) N3 percentage, (H) REM sleep percentage, (I) REM latency, and (J) REM efficiency. Diamonds indicate pooled random-effects estimates. Values below zero indicate lower values in individuals with nightmares than in controls, whereas values above zero indicate higher values in individuals with nightmares than in controls.

Abbreviations: CI, confidence interval; PSG, polysomnography; REM, rapid eye movement; SMD, standardized mean difference.

Figure 2.

continued.

Source of Heterogeneity Across Studies

We performed sensitivity analyses by excluding studies without adaptation nights and those did not clearly report exclusion criteria for OSA or PLMD. The results still revealed no significant differences in sleep macro-architecture between the nightmare and control groups (Tables S8 and S9). A subgroup analysis revealed that different PSG scoring rules (AASM vs R&K) did not impact the meta-analysis findings (Table S10). A meta-regression analysis revealed that the proportion of male participants among individuals with nightmares was a significant source of heterogeneity across studies. Except for this finding, meta-regression analyses did not reveal significant associations of age and year of publication with the changes in PSG macro-architectural parameters.

Sleep Parameters That Were Not Amenable to Meta-Analysis

Regarding PSA, three studies reported PSA differences between nightmare patients and healthy controls during REM sleep;16,17,24 two of these studies also explored PSA differences during NREM sleep,16,17 and one specifically distinguished cortical activity during phasic and tonic REM.26

Additionally, between individuals with nightmares and controls, a study27 evaluated the number and duration of arousals prior to and following REM sleep in individuals with nightmares compared to controls; another investigated the REM/NREM cycle duration and REM fragmentation;15 a third examined the arousal index and the number of REM periods;10 and a fourth explored awakenings during NREM and REM sleep.9 REM density was not pooled because the available studies used different definitions or units for this parameter. Furthermore, one study examined CAP during NREM sleep.23 In addition, three studies15,24,25 assessed autonomic activity during sleep (including the HR, HRV, and RSA) to identify possible differences between groups. The main findings of these studies are presented in Table 3 (for more details, please see Figure S10).

Table 3.

Summary of Findings That Were Not Subjected to Meta-Analysis

Study	Outcome	Sample size	Main findings
Simor et al, 2013b23	CAP in NREM sleep and arousal index in REM sleep	17 NM; 23 controls	The individuals with nightmares exhibited reduced amounts of CAP A1 subtype and increased CAP A2 and A3 subtypes, as well as longer duration of CAP A phases compared to controls. There were no significant differences in absolute number and frequency of REM arousals between groups.
Simor et al, 2013a17	PSA in NREM and REM sleep	19 NM; 21 controls	Compared to control participants, individuals with nightmares exhibited increased relative high alpha and fronto-central increases in high delta power during REM sleep, and a trend of increased fronto-central low alpha power in NREM sleep.
Nielsen et al, 201015	HRV and REM/NREM cycle duration	16 NM; 11 controls	In the autonomic measures, there were no significant differences in time domain (ie, heart rate) and absolute frequency domain (eg, LF power and HF power) between the individuals with nightmares and control groups. However, in sleep macro-architecture, the groups showed a difference in REM/NREM cycle duration.
Marquis et al, 201716	PSA in NREM and REM sleep	18 NM; 15 controls	Compared to control participants, individuals with nightmares are characterized by increased theta (3–4 Hz) activity within limbic-prefrontal circuits across all sleep-wake stages, especially in REM sleep.
Simor et al, 201424	PSA and HRV in pre- and post-REM periods	19 NM; 21 controls	During the transition from NREM to REM sleep, individuals with nightmares exhibited increased alpha power. There was a significant difference in HF power and RMSSD between the individuals with nightmares and control groups during both pre- and post-REM periods. Furthermore, individuals with nightmares exhibited a significant difference in LF power during the pre-REM period, but not during the post-REM period.
Paul et al, 201510	Arousal index and number of REM periods	17 NM; 17 controls	There was no significant difference in arousal index and number of REM periods between the individuals with nightmares and control groups.
Germain et al, 20038	REM density	11 NM; 12 controls	There was no significant difference in REM density between the individuals with nightmares and control groups.
Paul et al, 201925	REM density and autonomic activity in the last 5 minutes of REM sleep before awakening	19 NM; 19 controls (two PLMD participants in NM)	In the last 5 minutes of REM sleep before awakening, there were no general differences in autonomic activation in individuals with nightmares compared to controls. However, when reporting a nightmare, individuals with nightmares had significantly higher HR, lower HF, lower RSA, higher REM density, and shorter breathing cycle length in the last 5 minutes of REM sleep before awakening compared with their own non-nightmares.
Blaskovich et al, 2020b27	Arousal number in pre- and post-REM periods	22 NM; 23 controls	Compared with controls, individuals with nightmares experienced an increased number of arousals during pre-REM period, whereas this difference disappeared during the post-REM period.
van der Wijk et al, 202026	PSA in phasic and tonic REM sleep	21 NM; 22 controls	Individuals with nightmares and healthy controls did not show significant differences in either oscillatory power or functional synchronization during phasic and tonic REM microstates.
Tomacsek et al, 202428	HR and HRV	24 NM; 30 controls	Compared to control participants, individuals with nightmares had elevated HR across pre-REM, REM, post-REM and stable NREM sleep, but not during resting state before sleep onset. However, HRV values (specifically, the HF power and RMSSD) were not significantly different between the two groups.
Blaskovich et al, 2020a9	NREM awakenings and REM awakenings	22 NM; 22 controls	NREM awakenings and REM awakenings were not significantly different between individuals with nightmares and control participants.

Notes: Simor et al, 2013a refers to the study published in Biological Psychology; Simor et al, 2013b refers to the study published in Sleep; Blaskovich et al, 2020a refers to the study published in Sleep; and Blaskovich et al, 2020b refers to the study published in Journal of Sleep Research.

Abbreviations: CAP, cyclic alternating pattern; NREM, non-rapid eye movement; REM, rapid eye movement; PSA, power spectral analysis; HRV, heart rate variability; HR, heart rate; NM, nightmare group; PLMD, periodic limb movement disorder; LF, low-frequency power; HF, high-frequency power; RMSSD, square root of the mean of the squares of differences between adjacent normal sinus to normal sinus intervals.

Discussion

This systematic review and meta-analysis synthesized PSG findings in individuals with idiopathic nightmares. The pooled results did not show significant differences in sleep macro-architecture (ie, TST, SE, REM latency, and REM sleep percentage) between individuals with nightmares and healthy controls. The results failed to reject the null hypothesis of no significant group differences in PSG measured sleep parameters between groups. Furthermore, the limited available data regarding autonomic measures (HR and HRV) and sleep micro-architecture (CAP and PSA) in studies not subjected to meta-analysis makes it difficult to draw confident conclusions.

Several studies utilizing self-report measures, including questionnaires or diaries, reported impaired subjective sleep quality in relation to nightmares.3,37,38 Approximately 83% of adults report that they experience at least one nightmare annually.39,40 The clinical relevance of nightmares grows with their frequency. Wong et al reported that frequent nightmares were significantly associated with depressive and anxiety symptoms, as well as with a 30-day major depressive episode or generalized anxiety disorder.3 Individuals experiencing frequent nightmares exhibited a 5.74 times greater risk of psychiatric disorders and demonstrated a significantly higher likelihood of mood disorders compared to those without frequent nightmares.37 Although the subjective assessments of nightmares are clinically important, sleep profiles through subjective methodologies may be insufficient to elucidate the etiology and neurobiology of nightmares. PSG provides objective sleep measures that may help identify candidate biomarkers of nightmares.

Our meta-analysis revealed no significant differences in sleep macro-architecture metrics (such as N3 and REM sleep percentage) between groups. However, in our included studies not subjected to meta-analysis, Nielsen et al15 observed a significant increase in REM/NREM cycle duration in nightmares compared to controls. Paul et al25 found that REM density was significantly increased in the last 5 minutes before awakenings from nightmares, in contrast to awakenings from non-nightmare sleep among individuals with nightmares. Nightmares are associated with elevated alpha power during the transition from the NREM to the REM sleep state.24 In addition, Marquis et al16 proposed that nightmares are characterized by increased theta activity within limbic-prefrontal circuits across all sleep-wake stages, with particular prominence during REM sleep, rather than the previously proposed elevation of relative REM “high alpha” activity.17 These findings suggest that individuals with nightmares may exhibit alterations in REM sleep parameters. The primary mechanisms linking REM sleep disturbances to nightmares are hypothesized to involve a limbic-prefrontal network that is essential for emotion regulation and fear processing.41–43 In particular, REM sleep involves activation of neural circuits, including the amygdala, medial prefrontal cortex, hippocampus, and anterior cingulate cortex.11,44 During REM sleep, this limbic-prefrontal network may facilitate the fear extinction function of dreaming. However, dysfunction in this network may impair the process of fear extinction.11 This makes it difficult to process fear memories, potentially resulting in nightmares where these memories are triggered but not properly contextualized or dismissed.43 However, due to the lack of neuroimaging or intracranial electroencephalography (EEG) data from the included studies, this interpretation should be verified in the future. Future studies exploring the associations of PSG measured sleep with neuroimaging changes are warranted to understand the neurophysiological mechanisms of nightmares.

Comorbidity of Sleep Disorder

The effects of comorbidity of other sleep disorders on our meta-analysis findings should be considered. Severe OSA often suppresses REM sleep, and individuals with nightmares and comorbid with OSA may experience more sleep interruptions.45 Germain et al8 found that nightmares exhibited increased periodic leg movement indices compared with controls. These comorbidities may also cause alterations in PSG parameters among individuals with nightmares. Thus, we conducted sensitivity analyses excluding studies that did not explicitly state their exclusion criteria concerning the comorbidity of OSA or PLMD. The findings are similar to those seen in the whole sample. However, this finding requires verification in future studies with larger samples.

Sleep Micro-Architecture and Autonomic Parameters

Regarding CAP, Simor et al23 suggested that there are similar overall CAP A phase rates between nightmares and controls, but nightmare patients show reduced A1 and increased A2 and A3 rates. These findings reflect the imbalance between sleep-promoting and arousing mechanisms during NREM sleep and dysfunctional sleep-wake regulation, possibly resulting in sleep instability.

Individuals with nightmares may exhibit autonomic dysregulation, characterized by elevated HR and changes in HRV. However, different studies have yielded conflicting findings. For HR, Tomacsek et al28 reported that frequent nightmare recallers showed elevated HR across several sleep stages, including pre-REM, REM, post-REM, and stable NREM sleep, but not during resting wakefulness before sleep onset. By comparison, Paul et al25 found that there were no general differences in autonomic activation across REM periods but reported increased autonomic activation in the last 5 minutes of REM sleep before awakening compared with participants’ own non-nightmare awakenings. Nielsen et al15 also observed no significant difference in HR throughout the night. These discrepancies reflect differences in study design, clinical characteristics, and the timing of autonomic assessment relative to nightmare occurrence. For HRV, similar inconsistencies exist in HRV analysis. Some findings indicate that during pre- and post-REM periods, nightmare sufferers had significantly lower high-frequency (HF) power and a lower square root of the mean of the squares of differences between adjacent normal sinus to normal sinus intervals,24 reflecting attenuated parasympathetic activity. By comparison, Tomacsek et al found no significant differences in HRV values (specifically HF power) between the two groups during the same sleep periods.28 The limited number of studies and methodological heterogeneity make it difficult to draw confident conclusions about autonomic nervous system alterations. Previous studies have suggested that autonomic dysfunction is associated with cardiometabolic diseases (such as type 2 diabetes, obesity, heart failure, and hypertension),46,47 which highlights the importance of investigating the autonomic nervous system in individuals with nightmares. Thus, the exact changes of HR and HRV in idiopathic nightmares should be further studied in the future.

Clinical Implications and Future Directions

Our findings have implications for nightmare assessment and indicate important directions for future studies. Clinically, the assessment of nightmare symptoms mainly depends on the use of valid questionnaires. Our findings suggest that PSG should be considered for the assessment of nightmares, although these findings should be verified in future studies. Importantly, it is warranted to determine whether alterations in REM sleep parameters can be considered as biomarkers and used diagnostically for nightmares in clinical practice. Methodologically, future studies exploring PSG changes in nightmares should consider the effects of nightmare frequency on PSG findings. It should also apply standardized PSG protocols, such as defining consistent EEG frequency bands, clearly excluding participants with OSA or PLMD, applying standardized adaptation-night procedures and consistent PSG scoring methods, which are helpful for cross-study synthesis. Moreover, sharing raw PSG data across different sleep laboratories and applying portable, ambulatory, and wearable systems with optimization algorithms may improve the scalability of sophisticated analyses (ie, PSA and CAP).48,49 These methodological considerations are helpful to conduct large-scale, multicenter studies with standardized protocols to further explore sleep micro-architecture and its cost-effectiveness. In addition, integrating multiple dimensional PSG data, such as sleep macro-architecture, micro-architecture (also including phasic and tonic REM sleep), autonomic measures (ie, HRV), respiratory parameters, electroencephalography-electrocardiogram coupling, and their associations with limbic-prefrontal connectivity, is important for underlying the mechanisms of nightmares.

It should also be noted that no psychiatric disorders were reported among participants with nightmares in the included studies. However, it is known that frequent nightmares increase the risk of mood disorders,3,37 and the distress of nightmares is associated with the severity of the psychiatric disorders.50 Besides, the sleep architecture alterations were observed in MDD and anxiety disorders.51,52 Accordingly, it is possible that comorbid MDD or anxiety disorders may have an impact on the PSG alterations observed in patients with nightmares. Therefore, future studies should evaluate the psychiatric comorbidities and explore their effects on PSG measured sleep in individuals with nightmares.

Strengths and Limitations

The use of medications that impact sleep, such as anticonvulsants, antidepressants, and antipsychotics, is associated with alterations in PSG.53,54 Meanwhile, taking antidepressants or non-benzodiazepine hypnotics was also strongly linked to nightmares.55,56 In our meta-analysis, it is worth noting that none of the participants in the included studies were taking medications that affect sleep. This suggests that our findings are not influenced by medications known to affect sleep, which is a strength of our meta-analysis. However, the current review has notable limitations. First, the effects of several factors on the meta-analysis cannot be explored in our meta-analysis, such as variability in bedtime routines across sleep laboratories and discomfort associated with wearing PSG devices. Second, methodological heterogeneity across the included studies, such as inconsistencies in nightmare frequency criteria, PSG scoring criteria, and whether there was an adaptation night, may impact our findings. Third, the small number of studies on micro-architecture and autonomic parameters warrants cautious interpretation of these findings.

Conclusions

To our knowledge, this is the first meta-analysis to characterize polysomnographic sleep in individuals with idiopathic nightmares. The meta-analysis showed no significant differences in sleep macro-architecture between the nightmare and control groups. Thus, the proposed hypothesis was not supported, and we failed to reject the null hypothesis that there were no significant differences in PSG measured sleep parameters between groups. However, our systematic review highlights the importance of investigating sleep micro-architecture (such as CAP and PSA), suggesting PSG should be used for the assessments of nightmares. In the future, larger multicenter PSG studies with standardized protocols are needed to identify and validate candidate objective biomarkers, particularly in sleep micro-architecture and autonomic parameters.

Funding Statement

Our work was supported by the National Natural Science Foundation of China (82341249, 82470095, 82170099, 82120108002) and the Ministry of Science and Technology of the People’s Republic of China (2025ZD0216400).

Abbreviations

REM, rapid eye movement; MDD, major depressive disorder; PTSD, post-traumatic stress disorder; PSG, polysomnography; TST, total sleep time; PSA, power spectral analysis; WASO, wake after sleep onset; SE, sleep efficiency; SL, sleep latency; NREM, non-rapid eye movement; HR, heart rate; HRV, heart rate variability; RSA, respiratory sinus arrhythmia; CAP, cyclic alternating patterns; OSA, obstructive sleep apnea; PLMD, periodic limb movement disorder; AASM, American Academy of Sleep Medicine; R&K, Rechtschaffen and Kales; SMD, standardized mean difference; HF, high-frequency; EEG, electroencephalography.

Data Sharing Statement

The original contributions presented in the study are included in the article files, further inquiries can be directed to the corresponding author.

Author Contributions

Jiaxue Huang: Conceptualization, Writing - original draft, Methodology, Investigation, Formal analysis, Data curation. Meiou Wang: Writing - review and editing, Data curation, Validation, Supervision. Wenxin Bao: Writing - review and editing, Validation, Supervision. Wenqi Lyu: Writing - review and editing, Validation, Supervision. Yongming Wang: Writing - review and editing, Validation, Supervision. Shuyao Wu: Writing - review and editing, Validation, Supervision. Xiangdong Tang: Writing - review and editing, Validation, Supervision, Funding acquisition. Rong Ren: Writing - review and editing, Validation, Supervision, Funding acquisition. Ye Zhang: Conceptualization, Writing - review and editing, Validation, Supervision, Funding acquisition. All authors contributed to database preparation and accuracy checks, provided important suggestions for improving the article, commented on and approved drafts and the final article, gave final approval of the version to be published, have agreed on the journal to which the article has been submitted, and agree to be accountable for all aspects of the work.

Disclosure

The authors have no conflict of interest to declare.