Dreamlike Mentations During Sleepwalking and Sleep Terrors in Adults

Abstract

Background:

Sleep terrors and sleepwalking are described as arousals from slow wave sleep with no or poor mental recollection.

Objective:

To characterize the mental content retrospectively associated with sleep terrors or sleepwalking.

Setting:

University Hospital

Design:

Controlled prospective cohort

Participants:

Forty-three patients referred for severe sleepwalking/sleep terrors (age: 26 ± 7 y, 46% men, 5 with sleep terrors only, 8 with sleepwalking only, and 30 with both), matched with 25 healthy control subjects.

Intervention:

Thirty-eight of the 43 patients (88%) underwent an interview about the frequency, time, behaviors, and mental content associated with the episodes of sleepwalking and sleep terrors, whenever they occurred over a lifetime. The mental contents were classified for complexity (Orlinski score), and for characters, emotions, fortune/misfortune, and social interactions (Hall and Van de Castle categories). Patients and control subjects underwent an overnight video-polysomnogram.

Results:

Seventy-one percent of the patients reported at least 1 dreamlike mentation associated with the sleepwalking/sleep terrors episode. The dreamlike mentation action corresponded with the observed behavior. A total of 106 dreamlike mentations were collected (mean: 3 ± 3.4 dreamlike mentations/patient, range 0-17). Most (95%) dreamlike mentations consisted of a single visual scene. These dreamlike mentations were frequently unpleasant, with aggression in 24% (the dreamer being always the victim), misfortune in 54%, and apprehension in 84%. The patients with dream mentations reported more severe daytime sleepiness.

Conclusion:

Short, unpleasant dreamlike mentations may occur during sleepwalking/sleep terrors episodes, suggesting that a complex mental activity takes place during slow wave sleep. Sleepwalking may thus represent acting out of the corresponding dreamlike mentation.

Citation:

Oudiette D; Leu S; Pottier M; Buzare MA; Brion A; Arnulf I. Dreamlike mentations during sleepwalking and sleep terrors in adults. SLEEP 2009;32(12):1621-1627.

SLEEPWALKING AND SLEEP TERRORS CONSIST OF A SERIES OF COMPLEX BEHAVIORS THAT ARE INITIATED DURING SUDDEN AROUSALS FROM SLOW WAVE sleep (SWS). Episodes often begin with sitting up in bed and looking about in a confused manner. Sleepwalking and sleep terrors show considerable overlap and share many features. Indeed, at least 1 of the following is present in both of the disorders: (1) difficulty arousing the person, (2) mental confusion when awakened from an episode, (3) complete or partial amnesia for the episode, and (4) dangerous or potentially dangerous behaviors during the episode.¹ Both disorders arise from SWS, frequently cooccur in the same family or patient, and have a strong genetic background.²

In contrast with SWS, rapid eye movement (REM) sleep has, for a long time, been considered as the neurobiologic basis of dreaming. Nevertheless, complex mentations are reported in 5% to 74% of non-rapid eye movement (NREM) sleep awakenings.^3–8 The extent to which the reported mentation may legitimately be described as “dreaming” is still debated.^8,9 Indeed, there are qualitative and quantitative differences between NREM- and REM-sleep dreams. NREM-sleep dreams, especially when elicited during the first half of the night, are usually reported as being shorter, less complex, less vivid, and more “thought like” than REM-sleep dreams.^7,10–12 NREM sleep dreams can also be a recollection of the previous REM-sleep dreams.

Although there are sporadic descriptions of the mental content associated with sleepwalking or sleep terrors in the literature,^13–18 there is a general consensus against a complex dream activity associated with the sleepwalking/sleep terrors.¹⁹ Moreover, some authors have proposed that the sleepwalking/sleep terrors is triggered by the physiologic changes occurring during the arousal state (e.g., accelerated heart rate and respiratory changes) rather than by the mental activity preceding the arousal state.²⁰ In a cohort of patients with sleepwalking/sleep terrors, we investigate whether dreamlike mentations are frequent at the very moment of the episode and compare the sleep structure and daytime sleepiness of these patients with those of healthy matched control subjects.

METHODS

Patients

All patients with sleepwalking/sleep terrors who were monitored with video-polysomnography (including an extensive 8-electroencephalographic channel montage) in our sleep disorder unit during the last 2 years were prospectively evaluated. This is an adult unit with occasional investigations of children younger than 15 years of age. The patients were referred because of severe nocturnal behaviors that were dangerous, disturbing, or frequent. Patients with parasomnia overlap disorders (when sleepwalking/sleep terrors is combined with REM sleep behavior disorder [RBD]), nocturnal epilepsy, sleep sex, and sleep-related eating disorder (as the exclusive symptoms) and patients with psychosis were excluded. There were 43 patients with sleepwalking/sleep terrors (sleepwalking only, n = 8; sleep terrors only, n = 5; and both, i.e. displaying either sleep terrors some nights, sleepwalking some other nights, or night terrors followed by sleepwalking, n = 30). Sleepwalking was defined as (1) a history of ambulation during sleep; (2) the persistence of sleep or impaired judgment during ambulation; and (3) the disturbance was not better explained by another sleep, medical, mental, neurological disorder, or medication/drug use.¹ Sleep terrors were defined as (1) a history of a sudden episode of terror occurring during sleep, usually initiated by a cry or loud scream with sympathetic and behavior manifestations of intense fear; (2) difficulty in arousing the person, or mental confusion when awakened from an episode, or complete or partial amnesia for the episode, or dangerous or potentially dangerous behaviors; (3) the disturbance was not better explained by another sleep, medical, mental, neurological disorder, or medication/drug use.¹ In addition to these clinical criteria, we observed in all patients at least 1 of the following features on the videopolysomnography (although they are neither totally sensitive nor specific, they are supportive in the context of a systematic study): (1) at least 1 arousal during NREM sleep stage 4 associated with an abnormal motor behavior suggesting surprise, confusion, or fear (startling, sitting in the bed, or looking around surprised); (2) numerous sudden arousals during NREM sleep stage 4; or (3) hypersynchronous delta waves before an arousal in NREM sleep stage 4. Because sleepwalking and sleep terror share many features, we pooled patients with sleepwalking with those with sleep terrors in the next part of this article. Their demographic, clinical (body mass index, Epworth Sleepiness Scale score), and sleep characteristics during polysomnography were compared with those from 25 age- and sex-matched healthy controls without sleep disorders. The control subjects were retrospectively selected from our database after they took part in another paid study.

Investigations

All patients underwent a semistructured interview administered by a sleep specialist that included assessments of the medical history, any family history of parasomnia, the age at sleepwalking/sleep terrors onset, the frequency and timing of the sleepwalking/sleep terrors episodes, and predisposing factors (such as alcohol intake, sleep deprivation, stress, or change of room). In addition, we asked the patients about their mental content during the episodes of sleepwalking or sleep terrors using the following question: ”What was going through your mind at the very moment of the episodes, whenever the night?” Our data had been gathered retrospectively, so that the dreamlike mentation collection covered a lifetime span for each patient. Most patients had the greatest difficulty precisely dating the time of the sleep terrors/sleepwalking-associated dreamlike mentation. After obtaining the spontaneous mentation report, we interviewed the patient more pointedly to elicit additional information, including characters, emotions, and bizarre elements. We classified the mental contents according to their complexity using the Orlinski score^21,22 (0: No dream recall; 1: Feeling of having dreamt but no memory of it; 2: One scene; 3: More than 1 scene; 4: Coherent dream; 5: Coherent and detailed dream with the dreamer as an actor; 6: Prolonged, image-filled scenario resembling real life; 7: Prolonged, image-filled scenario with bizarreness). The mental content was defined as a dreamlike mentation if the score ranged from 2 to 7. We used the nominal categories (characters; emotions; social interactions including aggression, friendliness, and sexuality; misfortune and good fortune; and success and failure) of the Hall and Van De Castle method to qualify each mental-content component.²³ We determined the presence or absence of any given category, in accordance with the “at least 1” procedure. Because most dreamlike mentation contents in sleepwalking/sleep terrors contained fewer than 50 words, we used only the categories of this method, but not the normative values, which have been established for 50- 350-word dream contents. Whenever possible, family members were also interviewed. A demonstrative case report of dreamlike mentation contents obtained during sleepwalking/sleep terrors is presented in the Appendix.

Sleep was monitored for 1 night in all subjects using video-electroencephalography (8 channels) in patients and 3 electroencephalographic channels in control subjects, and we used classic channels (eyes movements, chin and leg electromyography, nasal pressure, body position, thoracic and abdominal respiratory efforts, pulse oximetry, electrocardiography, and sounds) in all participants. The sleep stages, arousals, periodic leg movements, and respiratory events were scored using the international criteria, as previously described.²⁴

Statistics

Data are presented as mean ± SD, unless otherwise specified. The comparisons between the patients and the control subjects were performed using the student t test.

RESULTS

Clinical Characteristics of Patients with NREM Sleep Parasomnia

The 43 patients with sleepwalking/sleep terrors were aged 11 to 72 years at the time of the videopolysomnography (mean age: 26 ± 7 years). Only 1 patient was a child, the other patients being adult. Forty-six percent were male, with a mean body mass index of 22.7 ± 3.4 kg/m² (not different from controls). Patients were 4 to 61 years old at the onset of parasomnia (mean age: 12 ± 12). They experienced sleepwalking/sleep terrors for 20 ± 12 years. They had great difficulties estimating the frequency of their episodes. Most patients said the episodes came irregularly, with clusters for several days to weeks. The events occurred during the first 2 hours of sleep in 88% patients. When they occurred, there was a single event per night in 55% and more than 1 event per night in 45%. The precipitating factors were alcohol intake in 14%, sleep deprivation in 28%, stress or argument in 58%, a specific context from the preceding day (such as watching a horror film or an action-filled film or reading a thriller) in 16%, sleeping at someone else's house in 5%, and unknown in 16%.

Mental Content During Sleepwalking or Sleep Terrors

Thirty-eight out of 43 patients (88%, including 2/5 patients with sleep terrors only, 6/8 patients with sleepwalking only, and 30/30 patients with sleep terrors and sleepwalking) could reliably answer questions about their mental content during the parasomnia episodes. The mental content was classified using the Orlinski score (Table 1). Among them, 27 of 38 (71%) reported at least 1 incidence of mental content associated with an episode. Four additional patients had the feeling of having dreamt during at least 1 episode but had no clear memory about the dreamlike mentation. Most mental contents (95%) were brief, corresponding with the vision of a single scene. A single patient (a child, 11 years old) experienced a prolonged, image-filled, and complex dreamlike mentation, with several scenes and odd features. An example of dreamlike mentations reported by a patient with sleepwalking and sleep terrors is detailed in Appendix. According to the categories of the Hall and Van de Castle method, 39% of patients described the presence of at least 1 person. The people seen during sleepwalking/sleep terrors were mostly unknown by the dreamer (uncertain identity: 33%, strangers: 24%), whereas only 33% were relatives of the dreamer. Eleven percent of patients saw animals, generally aggressive or frightening, during sleepwalking/sleep terrors. A dreamlike mentation was classified as containing aggression when the situation involved a deliberate intentional act on the part of 1 character to harm or annoy some other character. The classification of misfortune was used to handle those situations in which injury, mishap, or adversity occurred to a character through chance or environmental circumstances over which it was impossible to exert personal control (“something bad happens”). During sleepwalking/sleep terrors, most mental contents (80%) were bad dreamlike mentations with aggression (26% of the total number of dreamlike mentations) and misfortune (54%). In the case of aggression dreamlike mentations, the patient was never the first aggressor, except for one patient who “was ordered to stab the flesh of somebody on pain of being killed.” In addition, 12% of the dreamlike mentation reports contained at least 1 act of friendliness. In all cases, the patient was a befriender who attempted to protect someone (generally a relative) from danger. In contrast, no patient in a dangerous situation was befriended or rescued. There were no sexual elements in the dreamlike mentation reports. The majority of dreamlike mentations (84%) were negatively perceived, with apprehension, fear, or terror at the heart of them. The 4 patients with no clear dreamlike mentation recall had only a feeling of pure fear, with no associated images (Table 2). Seventeen patients acted out nonaggressive behaviors such as laughing, singing a dirty song, or saying “how nice are the bees” with no associated dreamlike mentation recall. All patients had some additional, more usual, “good,” and classic dreams that they would recall the following morning.

Table 1.

Characteristics of Mental Content Associated with Sleepwalking or Sleep Terrors

	Patients
Patients with sleepwalking/sleep terrors, No.	43
Patients appropriately interviewed, No.	38
Patients with at least one mental content recall, %	71% (27 of 38)
Total number of dreamlike mentations	106
Number of dreamlike mentations per person (mean ± SD)	3.0 ± 3.6
Number of dreamlike mentations per person (range)	0-17
Type of mental content: Orlinski scale
No real mental content (% of 38 patients)
0. No dream recall	18% (7 of 38)
1. Feeling of having dreamt but no memory of it	11% (4 of 38)
Mental content recall (% of all dreams obtained)
2. Vision of one scene	95% (101 of 106)
3. Vision of more than one scene	0% (0 of 106)
4. Coherent dream	4% (4 of 106)
5. Coherent and detailed dream, with the dreamer as an actor	0%
6. Prolonged, image-filled scenario, resembling real life	0%
7. Prolonged, image-filled scenario, with bizarreness	1% (1 of 106)
Dreamlike mentations (% of all dreams obtained) with at least one
Character
People	40% (42 of 106)
Animals	11% (12 of 106)
Social interaction
Aggression	24% (25 of 106)
Dreamer as a victim, % of aggression dreams	100% (25 of 25)
Dreamer as an aggressor, % of aggression dreams	4% (1 of 25)
Friendliness	11% (12 of 106)
Befriender, % of friendliness dreams	100% (12 of 12)
Befriended, % of friendliness dreams	0% (0 of 12)
Sexuality	0% (0 of 106)
Fortune and misfortune
Fortune	0% (0 of 106)
Misfortune	54% (57 of 106)
Emotions
Anger	2% (2 of 106)
Apprehension	84% (89 of 106)
Sadness	0% (0 of 106)
Confusion	0% (0 of 106)
Happiness	0% (0 of 106)

Table 2.

Examples of the Mental Content During Sleepwalking/Sleep Terrors and Associated Behaviors

Patient	Mental content	Behaviors (observed by a cosleeper or by the dreamer)
Patient 4, F, 34 y	She was going to get run over by a truck.	She leapt out of bed and out of the mezzanine to avoid it.
Patient 4	There were spiders.	She spat in her bed to drown them.
Patient 5, F, 32 y	Her baby was jeopardized.	She grabbed her baby and ran out of the room with it.
Patient 7, F, 66 y	She plunged into death.	She yelled.
Patient 9, F, 27 y	She was locked in a box she could not escape. She felt herself suffocating.	She pushed back the walls around her.
Patient 14, F, 26 y	She broke a vial containing a lethal virus.	She shouted.
Patient 18, F, 32 y	She saw Death with its dark cape and an emaciated hand coming up close to her face and about to kill her.	She yelled and ran away.
Patient 18	She saw a crack above her bed containing dead flesh. Cockroaches crawled on the dead flesh, then fell in her bed and hair. She rubbed her hair to remove them and saw a cockroach drop and run on the floor	She ran out of her bed, rubbed her hair, and saw the cockroach and needed time to realize that she was in the dark without any insects.
Patient 19, M, 11 y	He went to the roof of the house because people were following him and his little brother. He forbade his brother to climb on the roof because of the danger. On the way, he picked up luminous balls. There were also people at the bottom of the house.	He climbed onto the roof of the house.
Patient 25, F, 72 y	The ceiling was collapsing.	She shouted, jumped out of the bed, and pushed the wall.
Patient 28, M, 30 y	He was skiing and hurtling down an icy steep slope, with a ravine at the end of the slope. He could not stop from falling in the abyss	He shouted and leaped out of the bed.
Patient 34, M, 28 y	His girlfriend was in danger.	He pulled his girlfriend out of the bed to save her.

We tried to distinguish patients with isolated sleep terrors (n = 5), with isolated sleepwalking (n = 8), and with both phenomena (n = 30). Unfortunately, only 2 of 5 patients with isolated sleep terrors were appropriately interviewed, and only 1 of these 2 patients remembered at least 1 mentation (2 frightening images) associated with the sleep terrors. As for isolated sleepwalking, 6 of 8 patients were appropriately interviewed and only 2 of the 6 remembered at least 1 mentation associated with the sleepwalking. These 2 patients reported 6 different mentations, with 2 of 6 containing a frightening image (Supplemental Table 1 available online at www.journalsleep.org).

Sleep Measures in Sleepwalkers and in Controls

With regard to the sleep architecture (Table 3), there was no difference between the patient and control groups for sleep duration and efficacy, sleep-onset latency, percentages of sleep stages, arousal index, apnea-hypopnea index, and periodic leg movements. The number of sudden arousals from SWS varied from 1 to 13 in patients and 0 to 4 in control subjects. It was 4 times higher in patients than in control subjects. The sudden arousals from SWS were not preceded by any respiratory event, including flow limitation. The patients had an Epworth Sleepiness Scale score of 10 ± 4, and 47% of them had an abnormal sleepiness score (greater than 10). Their mean sleepiness score was higher than in the control group, but the control subjects were selected for having no abnormal sleepiness.

Table 3.

Sleep Measures in Patients with NREM Sleep Parasomnia and Controls

	Patients	Controls
Number	43	25
Nighttime sleep
Total sleep period, min	522 ± 82	543 ± 73
Total sleep time, min	460 ± 76	491 ± 78
Sleep efficiency, %	88 ± 8	90 ± 6
Latency to sleep onset, min	28 ± 21	30 ± 23
Sleep duration, min
Stage 1	21 ± 12 (4.6%)	19 ± 12 (3.9%)
Stage 2	221 ± 56 (47.9%)	242 ± 55 (47.5%)
Stage 3-4	117 ± 33 (25.7%)	123 ± 28 (25.3%)
REM sleep	107 ± 54 (21.7%)	105 ± 37 (21.0%)
Sleep fragmentation
Periodic leg movements index	4.6 ± 5.1	3.4 ± 11.7
Apnea-hypopnea index	3.4 ± 6.9	2.0 ± 2.7
Arousals index	14.0 ± 9.2	15.5 ± 5.1
Number of arousals from SWS	4.8 ± 3.0*	1.2 ± 1.0
Epworth Sleepiness Scale score	10 ± 4
Percentage of patients with a score > than 10	47	0

NREM refers to non-rapid eye movement sleep; REM, rapid eye movement sleep; SWS, slow wave sleep.

Thirty-five of the 43 (81%) patients exhibited at least 1 arousal during non-REM sleep stage 4 associated with an abnormal motor behavior suggesting surprise, confusion, or fear (startling, sitting in the bed, or looking around surprised). Most of the episodes were minor and much less important than those described at home. Because these patients were first referred to our sleep unit for a clinical (and not a research) purpose, they were not interviewed by the nurse just after the brief parasomnia to collect their mentations. No patient reported any dreamlike mentation when interviewed in the morning after the videopolysomnography night. A video of a patient with sleep terror (opening the eyes and looking scared before screaming) is displayed as supplemental video available online at www.journalsleep.org. On the next morning, she did not remember any mental content. During similar events at home, she had the vision of being locked in a box and suffocating or being attacked by something or somebody unknown.

DISCUSSION

In this group of 43 adults with sleepwalking/sleep terrors, as many as 71% of patients reported at least 1 dreamlike mentation associated with a nocturnal motor episode. The dreamlike mentations were mostly short and frequently unpleasant. There was no major difference regarding the frequency and nature of the dreamlike mentation content in sleepwalkers compared with patients with sleep terrors. Compared with healthy controls, the patients with sleepwalking/sleep terrors had more frequent arousals from SWS (despite a similar percentage of SWS) and reported more severe daytime sleepiness.

This study highlights the presence of dreamlike mentation content during an episode of sleepwalking/sleep terrors in adults. These results are concordant with those from a study of a group of 12 patients with sleep terrors who slept with electroencephalographic and heart rate monitoring during 4 to 23 nights.¹⁸ In this group, a mean of 58% recall of mental content was obtained after spontaneous arousals from NREM sleep associated with heart rate acceleration (but not necessarily with a scream). Apart from this case series reported in 1974, dreamlike mentations associated with a sleepwalking/sleep terror episode have been described incidentally.

A young patient rushed into the room where his parents were sitting and threw the butter dish out of the window, believing it to be a bomb.¹⁷

A sleepwalker threw his wife on the floor, ran to his 2 children, took them into his arms, and ran outside. He afterwards said he believed that the house was on fire.¹⁴ A father took his baby girl from her cradle, brought her into the attic and tied her, while dreaming that officials from child welfare wanted to deprive him of the custody of the baby.¹⁶

These cases illustrate that a complex mental content can be synchronous with elaborate motor behavior erupting from SWS, suggesting a mind-behavior isomorphism. One may notice, however, that the patients reported at least 1 dreamlike mentation associated with a sleepwalking/sleep terrors episode. They could not be precise enough to determine how many sleepwalking/sleep terrors episodes were associated with dreamlike mentations, as they were frequently amnesic of numerous episodes. In this study, we did not systematically interview the patient spouse, who could possibly recall many more episodes than the patient. All in all, dreamlike mentation recalls were occasionally associated with sleepwalking/sleep terrors, but the exact frequency of this association is still unknown.

The mental content associated with sleepwalking/sleep terrors in our patients was mostly short and contained 1 single scene, except for 5 patients (Patients 4, 15, 21, 18, and 1 child climbing on the roof as he followed a long and movie-like inner scenario). This scene was visual. Similar to our patients, a patient with sleepwalking/sleep terrors in another series attempted to flee or fight in self-defense during the night because he felt threatened or attacked by individuals with “blank eyes” and by a cow with a knife hidden in its mouth as a tongue.¹³ One may discuss whether these brief scenes are closer to a hypnagogic hallucination from SWS than a more classic dream. Patient 9 (video) opened her eyes and turned her head, while still in quiet SWS and then dreadfully screamed. She behaved as somebody discovering a frightening scene. These brief, visual, mental contents during sleepwalking/sleep terrors suggest that the episodes could be either the terminal part of a longer dream forgotten at time of arousal or a phasic, short, mental creation elicited before or just at the time of arousal. One may wonder if these phasic visions could be triggered by pontogeniculooccipital waves that occur as rare isolated events during SWS in cats, whereas they are numerous and grouped in bursts during REM sleep.²⁵ Some authors have proposed that sleepwalking/sleep terrors episodes are triggered by accelerated heart and respiratory rates occurring during the arousal state rather than by the mental activity preceding the arousal state.²⁰ Patient 9 here, however, does not fit the model, as she quietly opened her eyes with a normal heart rate before screaming.

The subject of scenes associated with sleepwalking/sleep terrors is mostly frightening, to the point that many patients call them nightmares in our clinical experience. As many as 54% of the reports contained elements of misfortune (whereas none contained fortune), 24% were aggressive, and 84% were apprehensive. In addition, 4 patients had “blank” dreams (the feeling of having dreamt before arousal but with no recall) associated with a pure intense fear. The strongly negative emotions associated with sleepwalking/sleep terrors suggest that the emotional network, and especially the amygdala, would be activated during the event. A single functional imaging session in a sleepwalker showed that the posterior cingulate cortex (which is part of the emotional processing network) and the anterior cerebellum are activated during sleepwalking, whereas the frontoparietal cortices are deactivated.²⁶ Whether patients with sleepwalking/sleep terrors exhibit an abnormal processing of emotion, even during the daytime, as was recently shown in patients with narcolepsy/cataplexy, should be determined.²⁷

Some sleepwalkers in our series, however, experienced not only misfortune, but also sang dirty or childish songs and quietly spoke while asleep. Similar to the previous report of nonviolent behaviors during RBD,²⁸ one may imagine that people with nonviolent sleepwalking/sleep terrors are less likely to seek medical advice. There could be a bias in our series towards violent, unpleasant, or more disturbing cases. The association of a dreamlike mentation (or a nightmare) recall with violent frightened behavior in sleepwalkers suggests that the distinction between sleepwalking and RBD can be difficult when it is based only on a sleep interview. Confusion during an interview could lead to the misclassification of nocturnal violence during epidemiologic surveys²⁹ or case-reports.³⁰ Some elements in the interview may however direct the clinician toward a diagnosis of sleepwalking rather than RBD. They include the age (young in NREM parasomnias, middle-aged or old in patients with RBD), the occurrence of walking (frequent in NREM parasomnias, exceptional in RBD),³¹ the frequency (low in patients with NREM parasomnias, high in patients with RBD), and nature (a single brief scene in NREM parasomnias vs a long detailed dream in RBD) of the dreamlike mentation associated with the behavior and, to some extent, its content. In our study, sleepwalkers report misfortune, whereas patients with RBD report aggression.³² Misfortune is defined by any injury, mishap, or adversity occurring to a character through chance or environmental circumstances (e.g., collapsing ceiling) over which it was impossible to exert personal control.²³ In most dreamlike mentations reported by patients with RBD and pseudo-RBD associated with severe obstructive sleep apnea syndrome,³³ the dreamers (or their loved ones) are a victim of an aggression (attack, rape, theft) by a human or an animal and fight back.^32,34 There is, however, possible an overlap between the contents of dreamlike mentations in NREM parasomnia and RBD, so that a videopolysomnography will be helpful in doubtful cases.

This study confirms previous observations that sudden arousals from SWS are frequent in patients with sleepwalking/sleep terrors and shows that the arousals are 4 times more common than in healthy control subjects. Half of the patients with sleepwalking/sleep terrors reported abnormal levels of daytime sleepiness. Sleepiness as a symptom of sleepwalking/sleep terrors has not been previously specifically highlighted. Because there is no difference other than the arousals from SWS between the patients' and control subjects' sleep structures (especially no more respiratory events during stage 4 sleep), we wonder whether long confused arousals from stage 4 sleep would not affect the restorative properties of sleep.

These data should be considered with a number of important limitations. Sleep terrors and sleepwalking are usually described as 2 distinct parasomnias. They are pooled here, which can be considered as a limitation. They indeed show considerable overlap, as 70% patients here have sleep terrors and sleepwalking at different periods of the night or the year or as a terror first (scream) followed by an attempt to flee out of the bed. Frightening images are not specific to sleep terrors (or sleep terrors/sleepwalking), however, as they occurred here in 2 patients with isolated sleepwalking. This result is observed in a small subsample, but it further suggests that the boundary between sleepwalking and sleep terrors is very thin. Another limitation in our study is that the dreamlike mentations have been gathered retrospectively, over a lifetime span, so that recall bias is possible. All studies on dreams are by definition retrospective, as the subject cannot at the same time experience the dream and report it to someone else. It is probable, however, that the time elapsed between the dreamlike mentation experience and the report alters the memory of the dreamlike mentation, so that an immediate recall would further enhance the percentage and the length of dreamlike mentations reported by our patients, as it is the case in normal subjects. This bias would underestimate (and not overestimate) the frequency of dreamlike mentations in our study. Also, healthy subjects tend to remember more easily the dreamlike mentations with striking rather than ordinary content.³⁵ To avoid this bias, one should perform a systematic, nightly based interview after each NREM parasomnia in a series of patients, which would be a costly but fascinating future study. Notably, because none of our 43 patients had any dreamlike mentation recall the next morning in the sleep-unit settings, such a large study may require either home monitoring, staying a week in the sleep unit, or using provocative methods to increase the number of parasomnia events.

In conclusion, dreamlike mentations (mostly brief, frightening visual images) may occasionally exist during sleepwalking and sleep terrors, suggesting that a complex mental activity takes place during SWS. Sleepwalking may thus represent acting out of the corresponding dreamlike mentations. A previous study proposed that awakening serves as an indirect measure of nightmare intensity, the emotions during nightmares being more intense than during bad dreams.³⁶ Similarly, the emotional character of the frightening image could be 1 of the proximal trigger of the sleepwalking and sleep terror.

DISCLOSURE STATEMENT

This was not an industry supported study. Dr. Arnulf has received research support from Actelion Pharmaceuticals Ltd. and has participated in speaking engagements for UCB. The other authors have indicated no financial conflicts of interest.

APPENDIX

Case Report

Patient 21 was a 23-year-old employee in a school who had experienced severe sleepwalking and sleep terrors for 16 years. Her grandfather was a sleepwalker (he jumped through the window into a rosebush when he was young), and her mother still talks and walks while asleep. The patient had no history of depression or psychological trauma. She had frequent nightmares, several times per night, but they occurred in clusters for several days or weeks at a time and then disappeared for several weeks. Fatigue, stress, changes in altitude, and horror films precipitated the episodes. She had fewer episodes during the holidays and when she did not sleep at home. She complained of daytime sleepiness, with an Epworth Sleepiness Scale score of 14. She hurt herself several times during the sleepwalking episodes, including a broken clavicle, a twisted ankle, a head injury, and an ear wound. During the nighttime videopolysomnography, the sleep architecture was normal (without epileptiform activity), except for 10 sudden awakenings from non-rapid eye movement sleep stage 4 (Figure 1). During these awakenings, she opened the eyes and looked, with a scared expression, around her. She sat twice on the bed and spoke shortly. At home during the sleepwalking episodes, she was agile, opening doors, and avoiding routine obstacles. Only unforeseeable obstacles, such as a bag left unattended on the floor the previous evening, would cause her to fall. If someone awoke her at that time, she would have superhuman strength and be very aggressive toward the intruder. She reported vivid mental imagery during the sleepwalking episodes. She experienced recurrent nightmares with short scenarios including floods, collapsing buildings, fires, and 2 walls getting closer to the point of crushing her. In these cases, she had to escape and felt extreme apprehension. She shouted loudly, ran away from the bed, and hid in the bathroom or under the living room table. She was able to dress within a few seconds to get out of the apartment and hurtle down the stairs before being aware that it was just a nightmare. She also experienced other types of mental content, that she herself called “the mission dreams.” In these dreamlike mentations, she had to do something to survive. For example, she had to stay under the duvet because a superior power ordered it, she had to wash her feet in the bathroom, and she had to plug the bathroom door with clothes. Once, she saw her aggressor as a Chinese dragon. In a recurrent dreamlike mentation, she was in a large white room with many Smarties (colored chocolate candies) on the floor. She had to sort and pile them up according to their colors, but the stack always fell and she had to start again. She required several minutes after awakening to admit that nothing bad had happened. When she resumed sleep, she experienced the same dreamlike mentation again. She also experienced dreamlike mentations resembling horror films (“the living-dead type”). She saw dead heads falling, people hanged in her room, acid falling from the roof, thieves, killers, and children from her school slaughtered by a madman. When she had these types of dreamlike mentations, she screamed (waking up her neighborhood), ran to the bathroom, and locked the door. In addition, she had a recurrent dreamlike mentation in which a gorilla came into her house, grabbed her by the hair, and dragged her down the stairs. During this dreamlike mentation, she was partly aware but could not refrain from screaming and taking part in her dreamlike mentation, while she simultaneously thought “Oh no, not the gorilla again'” Her dreamlike mentation could also be influenced by daytime context. During the rugby world cup 2007 in Paris, she dreamt that an army of rugbymen wanted to kill her. Of note, she also had good dreams but did not remember a sleepwalking episode associated with good dreams, nor did her boyfriend report such occurrences.

Figure 1.

Hypnogram of Patient 21 (case report) with 10 sudden awakenings from slow wave sleep (shown by arrows) with corresponding behavior of surprise, sitting, and talking evoking a “status parasomnicus.”

Supplemental Table 1.

Characteristics of Mental Content Associated with Isolated Sleepwalking, Isolated Sleep Terrors, and Both Sleepwalking and Sleep Terrors

	Sleepwalking	Sleep terrors	Both sleepwalking and sleep terrors
Patients with sleepwalking/sleep terrors, No.	8	5	30
Patients appropriately interviewed, No.	6	2	30
Patients with at least 1 mental-content recall, % (No.)	33 (2/6)	50 (1/2)	80 (24/30)
Total number of dreams	6	2	98
Number of dreams per person (mean ± SD)	1 ± 1.5	1 ± 1.4	3.6 ± 3.8
Number of dreams per person (range)	0-4	0-2	0-17
Type of mental content: Orlinski scale
- No real mental content, % of patients (No.)
0. No dream recall	67 (4/6)	0 (0/2)	10 (3/30)
1. Feeling of having dreamt but no memory of it	0 (0/6)	50 (1/2)	10 (3/30)
- Mental content recall, % of all dreams obtained (No.)
2. Vision of 1 scene	83 (5/6)	100 (2/2)	96 (94/98)
3. Vision of more than 1 scene	0 (0/6)	0 (0/2)	0 (0/98)
4. Coherent dream	0 (0/6)	0 (0/2)	4 (4/98)
5. Coherent and detailed dream, with the dreamer as an actor	0 (0/6)	0 (0/2)	0 (0/98)
6. Prolonged, image-filled scenario, resembling real life	0 (0/6)	0 (0/2)	0 (0/98)
7. Prolonged, image-filled scenario, with bizarreness	17 (1/6)	0 (0/2)	0 (0/98)
Dreams, % of all dreams obtained (No.) with at least 1
- Character
People	83 (5/6)	0 (0/2)	38 (37/98)
Animals	0 (0/6)	0 (0/2)	12 (12/98)
- Social interaction
Aggression	17 (1/6)	0 (0/2)	25 (24/98)
Dreamer as a victim, % of aggression dreams (No.)	100 (1/1)		25 (24/98)
Dreamer as an aggressor, % of aggression dreams (No.)	0 (0/1)		1 (1/98)
Friendliness	33 (3/6)	0 (0/2)	9 (9/98)
Befriender, % of friendliness dreams	100 (3/3)		100 (9/9)
Befriended, % of friendliness dreams	0 (0/3)		0 (0/9)
Sexuality	100 (0/6)	0 (0/2)	0 (0/98)
- Fortune and misfortune
Fortune
Misfortune	33 (2/6)	100 (2/2)	54 (53/98)
- Emotions
Anger	17 (1/6)	0 (0/2)	1 (1/98)
Apprehension	33 (2/6)	100 (2/2)	87 (85/98)
Sadness	0 (0/6)	0 (0/2)	0 (0/98)
Confusion	0 (0/6)	0 (0/2)	0 (0/98)
Happiness	0 (0/6)	0 (0/2)	0 (0/98)