Address correspondence to Paolo Tinuper, IRCCS Istituto delle Scienze Neurologiche, Department of Neurological Sciences, University of Bologna, Bellaria Hospital, Via Altura 3, Bologna, Italy. E-mail: firstname.lastname@example.org
Although parasomnias should be considered benign conditions without a deleterious impact on sleep quality and quantity, especially in children, it is important to recognize and properly diagnose these phenomena. Moreover, parasomnias may be misdiagnosed as epileptic seizures, in particular seizures with a predominant complex motor behavior as seizures occurring in nocturnal frontal lobe epilepsy (NFLE), leading to unnecessary and expensive investigations and prolonged and unsuccessful treatment. In this article we describe the clinical and neurophysiologic features of the most common parasomnias, giving the most reliable elements of differential diagnosis between parasomnias and epileptic nocturnal seizures, namely the typical seizures occurring in NFLE. The diagnostic value of history-taking, video-polysomnography, home video recording, and diagnostic scales is discussed. Next we describe the intriguing aspect of the frequent coexistence, in the same family and even in the same patients, of epileptic and parasomniac attacks, giving a common neurophysiologic interpretation. Finally some brief indications to the treatment of parasomnias are suggested.
According to the International Classification of Sleep Disorders (ICSD-II), the nonepileptic paroxysmal motor events during sleep include parasomnias, sleep-related movement disorders, isolated symptoms, apparently normal variants, and unresolved issues. Parasomnias are defined as “clinical disorders that are not abnormalities of the processes responsible for sleep and awake states per se but are undesirable physical phenomena that occur predominantly during sleep” (American Academy of Sleep Medicine (ASDA), 2005). The (“always or often”) prevalence in the population is 1–11% (Partinen & Hublin, 2000). Parasomnias tend to occur during childhood, often with episodic recurrence, but onset or persistence during adulthood is well recognized. The exact causes of parasomnias are unknown, but the postulated mechanism is an underlying impairment of arousal mechanisms that trigger dissociation between the motor component of the awake state and electroencephalography (EEG) electrical activity (i.e., dissociated state) (Mahowald & Schenck, 2005). Parasomnias include several clinical features, usually manifesting as central nervous system activation, with autonomic nervous system changes and skeletal muscle activity. The familial nature of parasomnias is well known: epidemiologic surveys found a clear familial recurrence (sleepwalking, sleep terrors, nightmares, and sleep enuresis) and a co-occurrence (sleepwalking with nightmares, for example) in some patients (Hublin & Kaprio, 2003) without a clear mode of transmission (Abe & Shimakawa, 1966; Kales et al., 1980). Twin studies show a higher concordance for sleepwalking in monozygotic than dizygotic twins (Bakwin, 1970; Hublin et al., 1997). There seems to be an important overlap in the genetic predisposition to sleepwalking and, to a lesser extent, night terrors. Lecendreux et al. (2003) found an association between sleepwalking and the subtypes of the human leukocyte antigen (HLA) DQB1*0501, especially in the familial forms, demonstrating the first genetic susceptibility factor for this parasomnia.
Although parasomnias should be considered benign, without a deleterious impact on sleep quality and quantity, especially in children, it is important to recognize and diagnose them properly for several reasons. They may represent an embarrassing condition for the person with the condition, especially if occurring outside the domestic environment. Sometimes they are interpreted as a psychological disturbance of the child. In some particular forms, especially in adults, the sufferer may act aggressively, potentially endangering the bed partner, with medicolegal implications. Parasomnias may be misdiagnosed as epileptic seizures, leading to unnecessary and expensive investigations and prolonged and unsuccessful treatment.
Herein we describe the clinical and neurophysiologic features of the more common parasomnias, and the most reliable elements that distinguish parasomnias from nocturnal frontal lobe epilepsy (NFLE). In addition, we discuss the intriguing and frequent coexistence, in the same family and even in the same patients, of epileptic and parasomniac attacks, suggesting a common neurophysiologic explanation.
Parasomnias are divided into non–rapid eye movement (NREM) parasomnias and REM parasomnias. NREM parasomnias, otherwise called disorders of arousal (i.e., episodes arising from NREM stages of sleep), include confusional arousals, sleep terrors, and sleepwalking. REM parasomnias (i.e., episodes arising from REM sleep) include REM sleep behavior disorders (RBDs), nightmares, and sleep paralysis. Others parasomnias include sleep-related dissociative disorders, sleep enuresis, sleep-related groaning (catathrenia), exploding head syndrome, sleep-related hallucinations, sleep-related eating disorder, parasomnia unspecified, parasomnia due to drug or substance, and parasomnia due to medical conditions (Table 1).
Table 1. Classification of parasomnias
REM sleep behavior disorders (RBDs)
Sleep-related dissociative disorders
Sleep-related groaning (catathrenia )
Exploding head syndrome
Sleep-related eating disorder
Parasomnia due to drug or substance
Parasomnia due to medical conditions
NREM parasomnias (disorders of arousal)
The disorders of arousal are due to impairment in arising rapidly and fully from the deepest stage of NREM sleep (stages 3), typically in the first third of the night. Table 2 summarizes their common features. Polysomnographic recordings in NREM parasomnias show that the episodes arise from slow-wave sleep, with a regular, rhythmic, sometimes hypersynchronous delta activity pattern, associated with a marked increase in muscle tone and heart rate, and change in respiratory rate.
Table 2. Characteristic features of NREM and REM (RBD) parasomnias
REM sleep behavior disorder
Age at onset and disease course
Childhood; the episodes cease in adolescence
Usually after 50 years
Yes: sleep deprivation, febrile illness
Time of occurrence during the night
In the first third of the night
In the latter part of the night
No same-night episode occurrence
Usually once per night, up to 4 times per night
Arousal from slow-wave sleep
REM without atonia
Episode motor behavior
Sudden arousals from sleep with various degrees of motor activity: screaming, incoherent verbalizations, inconsolable crying and often fear
Violent behaviors mimicking the content of the dream: defence against attack by people or animals, punching the arms, kicking, gesturing, grabbing and crawling
Autonomic activation during the episode
Minimal or none
Recall of the event
Patients report a complex “dream-tale”
Disorders of arousal include confusional arousals, sleep terrors, and sleepwalking, each reflecting different degrees of motor, emotional, and autonomic activation.
Confusional arousals are “partial awakenings in which the state of consciousness remains impaired for several minutes without any accompanying major behavioral disorders or severe autonomic responses” (ASDA, 2005). The subject may present with more or less complex movements in bed but without walking or terror behavior because of only minimal motor and central nervous system activation (Broughton, 1968) (Fig. 1). Confusional arousals normally occur at awakening in the morning or, more rarely, during an arousal from deep sleep. In children, they are frequent physiologic events, whereas in adults, they may represent a nonphysiologic condition related to sleep deprivation, medication, or altered circadian sleep–wake pattern. Some authors call confusional arousals in adults “sleep drunkenness” (Stores, 2010) or “sleep inertia,” comprising a state of confusion and automatic behavior after awakening.
Sleep terrors are dramatic events consisting of sudden arousals from sleep with various degrees of motor activity, such as screaming, incoherent and intense verbalizations, inconsolable crying, and often fear and extreme emotional and autonomic activation (tachycardia, tachypnea, flushing, sweating, and mydriasis). The children, even if apparently alert, do not recognize their parents. Attempts at consolation are fruitless and may only prolong or even intensify the confusional state. Patients do not usually leave the bed and the episodes generally last 1–5 min. If awakened after the episode, the child does not recall a structured dream but refers to vague feelings of fear or danger.
Sleepwalking is characterized by “a series of complex behaviors arising from slow-wave sleep, culminating in walking around with an altered state of consciousness and impaired judgment” (ASDA, 2005). The episode lasts up to 15–30 min, during which the patient sometimes answers questions or follows suggestions. If restrained, the patient may react aggressively, although injuries are uncommon as normally the patient avoids obstacles or dangerous situations. After the episode, the patient may return to bed and in the morning has little or no memory of the episode, but may recall fragments, flashes, or vague impressions of the event (Oudiette et al., 2009).
Disorders of arousal may be triggered by various factors in susceptible individuals, such as febrile illness, physical activity, emotional stress, sleep deprivation, alcohol, or medications. Forced awakenings from slow-wave sleep in the early part of the night may also precipitate an episode.
Confusional arousals are ubiquitous in healthy young children until the age of 5–7 years (Broughton, 1968; Stores, 2010). Sleep terrors affect between 1% and 6% of prepubertal children, with a peak incidence at age 5–7 years; sleepwalking peaks some years later at age 8–12 years (Ohayon et al., 1999). Disorders of arousal become less common with increasing age and eventually cease (Ohayon et al., 1999). The prevalence of NREM parasomnias in adults is unknown, but mostly represent a continuation of episodes after adolescence, sometimes after having been symptom free for several years.
Recent epidemiologic studies in randomly selected young adults (18 years and older) showed a lifetime prevalence of confusional arousals of 18.5% and actual prevalence (in the previous months) of 6.9%. For sleepwalking these prevalences were 22.4% and 1.7% and for sleep terrors they were 10.4% and 2.7% (Bjorvatn et al., 2010).
Many patients have a family history of either sleep terrors (Kales et al., 1980) or sleepwalking, or both, supporting the hypothesis that NREM parasomnias share a common genetic predisposition (Hublin et al., 2003). The first genetic locus for sleepwalking was recently found at chromosome 20q12-q13.12 (Licis et al., 2011).
Disorders of arousal begin in stage 3 NREM sleep and are associated with marked tachycardia, tachypnea, and increased muscle tone. Bursts of hypersynchronous rhythmic delta waves immediately precede the parasomniac episode. This hypersynchronous delta activity before an arousal from slow-wave sleep had long been implicated as a possible diagnostic sign of an NREM parasomnia. More recent intracerebral EEG recordings show that during confusional arousals, the motor and cingulate cortices are activated and show the same activity as seen during wakefulness, whereas the frontoparietal associative cortices show the enhanced delta activity characteristic of sleep (Terzaghi et al., 2009). These data suggest that confusional arousal is the expression, not of a global cerebral phenomenon, but rather of the coexistence of different local states of being.
REM sleep behavior disorders (RBDs) were first reported by Schenck et al. (1986). They are characterized by episodes of motor agitation of varying intensity during REM sleep, associated with absence of the physiologic muscle atonia of REM sleep, thereby permitting the “acting out” of dreams (Schenck et al., 1986). The dreams are usually vivid, accompanied by vigorous, often violent, sleep behaviors enacting attack or defense reactions. Patients and bed partners are frequently injured; attacks can cause bruising, lacerations, and fractures, and the violence of the sleep-related behavior is often discordant with the awake personality (Tinuper et al., 2007). During the episodes, there is usually absent or mild autonomic activation. RBD patients have between one and multiple events per night; the episodes are of short duration and on awakening there is a rapid return to alertness and orientation, allowing patients to report dream mentation appropriate to the observed behavior. The events tend to occur in the middle of the night or early in the morning.
RBD is more common in people older than 50 years of age and predominates in male patients (male-to-female ratio 9:1) (Olson et al., 2000). RBD has both acute and chronic forms. The acute form is almost always induced by medications, most commonly antidepressants (tricyclic antidepressants, particularly the serotonin-specific reuptake inhibitors) (Mahowald & Schenck, 2005), or associated with medication withdrawal (alcohol, barbiturate, or meprobamate) (Silber, 1996; Plazzi et al., 2002). The chronic form is usually idiopathic or associated with central nervous system disorders, particularly the synucleinopathies, that is, Parkinson’s disease, multiple system atrophy, and dementia with Lewy bodies (Schenck et al., 1996; Plazzi et al., 1997; Boeve et al., 2001; Vetrugno et al., 2004). RBD may be the first manifestation of these conditions, preceding other symptoms by many years.
The polysomnographic features of RBD include excessive augmentation of chin electromyography (EMG) tone or excessive chin or limb EMG twitching (Fig. 2). A single night of recording is generally sufficient to establish the diagnosis, because there are usually REM sleep polysomnographic abnormalities, even when a behavioral episode does not occur.
Nightmares are “disturbing mental experiences that generally occur during REM sleep and that often result in awakening” (ASDA, 2005). They are characterized by a sudden awakening from sleep, usually in the middle or latter half of the night, with intense fear, anxiety, embarrassment, disgust, and other negative feelings, accompanied by minimal motor activity, vocalization, and mild autonomic activity. The sleeper is usually easily aroused and, when awakened, shows full alertness and can immediately recall the frightening dream content (Tinuper et al., 2007). Nightmares are common in young children, peaking at aged 3–6 years, and subsiding thereafter (Leung & Robson, 1993).
Polysomnography is indicated if there is suspicion of concurrent sleep apnea or RBD. Patients with nightmares show long, frequently interrupted REM-sleep periods with the recollection of a long frightening dream and clear orientation on awakening. There is no difficulty with respiration and there may be no, or only minor, autonomic symptoms, such as palpitations.
Sleep paralysis is an inability to perform voluntary movements on awakening from sleep, with episodes lasting between a few seconds and several minutes. The sensation of being paralyzed causes intense anxiety. Sleep paralysis may occur as an isolated event affecting 15–40% of the population, especially persons younger than 30 years of age, and may recur in members of the same family. Episodes often arise in the course of narcolepsy.
Sleep-related groaning (catathrenia) is characterized by an unusual expiratory groaning noise during sleep, without observed respiratory distress or associated motor phenomena; arterial oxygen saturation remains normal. The disorder appears almost nightly, predominantly or exclusively occurring during REM sleep. Groaning usually lasts 2–49 s, and often is repeated in clusters and recurs many times per night. The duration of the noise and the absence of any concomitant abnormal motor phenomena distinguish this expiratory sound from the moaning that occurs during epileptic seizures. The pathophysiology of groaning and the long-term consequences of this chronic disorder are unknown (Vetrugno et al., 2001).
Catathrenia is characterized by a deep inspiration without any sound followed by a prolonged expiration with groaning, with no evidence of respiratory muscular effort or oxygen desaturation (Fig. 3).
Differential Diagnosis between Epileptic Seizures and Parasomnias
There is, as yet, no single, valid and reliable procedure (including video-polysomnography, VPSG) to distinguish parasomnias from NFLE. The diagnostic difficulty results from, first, the similarity of the behavioral patterns in non-REM and REM parasomnias and NFLS; second, the possible coexistence of these sleep disturbances in the same person (Provini et al., 1999; Bisulli et al., 2010); and third, the lack of a “gold standard” objective test for both parasomnias and NFLE. The response to therapy is not a proper ex juvantibus diagnostic tool, as even the response to antiepileptic drugs does not necessarily support a diagnosis of NFLE. Although carbamazepine helps most patients at low doses, the therapy did not change seizure frequency at all in 32% of them (Provini et al., 1999). Therefore, the diagnosis remains electroclinical, based on the semiology of the ictal behavioral and polygraphic features, and on the clinical context.
Many experts consider video-EEG or VPSG the simultaneous recording of EEG, EMG, and cardiorespiratory variables under video control—to be the gold standard for diagnosing paroxysmal sleep-related events. Nevertheless, the sensitivity of VPSG is below 100% (especially in patients with rare seizures), and it is not 100% specific because of unsatisfactory interobserver agreement (Vignatelli et al., 2007) and being nondiscriminating for some complex fearful behaviors (Nobili, 2009). Hence, VPSG is not the “gold” standard but just one of several diagnostic tools.
When VPSG is not available, home video recording may help in diagnosis, particularly as family members often record attacks on home videos or mobile phone cameras (Tinuper et al., 2007).
The diagnosis is, in most cases, based more on the ictal semiology (i.e., the ictal symptoms, signs, and behaviors) from history taking or recorded events, rather than on EEG correlates (Manni et al., 2008). History taking helps to distinguish disorders of arousal from epileptic seizures, particularly the hypermotor seizures of NFLE. In arousal disorders, the age of onset is earlier, episodes are infrequent and have a long duration, rarely appear more than once per night, and tend to disappear in adulthood. In contrast, NFLE start between the ages of 10 and 20 years, with seizures persisting into adulthood and occurring very frequently, sometimes every night and many times during the same night. Arousal disorders do not show the dystonic or dyskinetic postures typical of NFLE seizures, and the complex motor sequence is not stereotyped in an individual patient as it is in epileptic seizures (Provini et al., 1999; Tinuper et al., 2007).
RBD patients are usually older than those with NFLS. RBD episodes are short, and the timing of the episodes during sleep reflects the REM sleep time, often occurring in the second half of the night. The memory of dream mentation appropriate to the behavior, together with complete alertness and orientation on awakening, help to distinguish RBD from NFLS (Tinuper et al., 2007). Although rare cases of idiopathic RBD show interictal epileptiform abnormalities on routine EEG (Manni et al., 2006), the diagnosis is usually straightforward. In doubtful cases, however, the diagnosis is readily established by video-EEG monitoring during nighttime sleep together with concomitant recording of sustained EMG activity, intermittent loss of REM atonia, or excessive phasic muscle twitching of the submental or limb EMG during REM sleep.
Nightmares typically arise in the second part of the night when REM propensity is high, whereas NFLS arise from any night period. The poor motor behavior during the episodes of nightmares, the absence of confusion on awakening, together with detailed dream reports, if available, help to distinguish nightmares from nocturnal seizures. Polysomnography, indicated in doubtful cases, shows abrupt awakenings from REM sleep and desynchronized tracings during episodes (Tinuper et al., 2007).
Derry et al. (2006b) developed the frontal lobe epilepsy and parasomnias (FLEP) scale in order to establish how reliably features from the history might distinguish NFLS from parasomnias. Through validating this scale in patients with established diagnoses, they showed its value for distinguishing various sleep-related disorders; they derived cutoff scores to stratify the likelihood of a nocturnal paroxysmal episode being either epileptic or parasomniac. However, Manni et al. (2008) challenged the usefulness of the scale after studying a tertiary sleep center population. They found that the FLEP scale risked misdiagnosing some patients, especially NFLE subjects presenting episodes of nocturnal wandering—misinterpreted as arousal parasomnias—reflecting the low sensitivity. Furthermore, the scale misleadingly gave an epileptic diagnosis in one third ofthe parasomnia patients—mainly from the RBD group—reflecting its low specificity. Differences in the composition of patient series, namely the high frequency of RBD in the Italian cohort, may partly explain the differences between the two studies. Moreover, the authors highlighted the inadequacy of some of the scale items (namely those investigating wandering) and the need to analyze the reliability of the FLEP scale items (Manni et al., 2008). Bearing these limitations in mind, the FLEP scale could be useful in such an algorithm, helping the physician to better address simple clinical impressions, suggesting the need for a video or polysomnographic recording, or avoiding these investigations (Manni et al., 2008).
Recently, we measured the accuracy of anamnestic features collected during clinical history taking for the diagnosis of NFLS. We identified two major anamnestic patterns (i.e., dystonic posturing or hyperkinetic behavior) for NFLE diagnosis, with a high specificity but unsatisfactory sensitivity, and four minor features that may increase the specificity of these clinical items when associated with one of the two major patterns (Bisulli et al., 2012). These clinical features give a standardized core of anamnestic information that we called Structured Interview for Nocturnal Frontal Lobe Epilepsy (SINFLE), helping physicians to differentiate NFLS from parasomnias or other sleep disorders during the clinical interview. Moreover, SINFLE disclosed possible elements for developing future shared diagnostic criteria, specific for NFLE, that are absent even from the new ICSD-II. This study confirmed the weakness of the clinical history in differentiating NFLE from parasomnias, a flaw intrinsically related to the peculiarity of these sleep disorders and therefore difficult to overcome. So far, there is no single valid and reliable diagnostic procedure (including VPSG) that can define parasomnias and NFLE. In this scenario, we agree with Manni et al. (2008) that future efforts must focus on developing a reliable algorithm (rather than a single reliable method) to help physicians diagnose paroxysmal motor sleep disorders. We clearly need an international consensus on these issues (Bisulli et al., 2012).
Treatment of Parasomnias
Parasomnic attacks in healthy children and adolescents normally require no treatment. Some cases respond to a psychological approach. Parents should be taught to avoid waking the children; instead, they should gently accompany them to bed. Several conditions or substances that increase stage 3, NREM sleep may trigger arousal parasomnias (Grigg-Damberger & Ralls, 2011). Avoiding or treating these conditions is usually sufficient (Table 3), with pharmacotherapy considered only when the episodes are frequent or dangerous to the patient or others. In the cases of nocturnal wandering or violent parasomnias, parents should make the environment as safe as possible (securing windows, removing obstacles, installing alarms) to avoid serious injuries (Kotagal, 2009). Clinicians should consider therapy if the episodes cause undesirable secondary consequences, such as excessive daytime sleepiness, or cause distress to the patient or family. The most widely prescribed therapies in adults with arousal disorders are clonazepam (0.5–2 mg at bedtime) and imipramine. The effectiveness of clonazepam for suppressing these disorders relates to inhibition of arousals or of locomotor activity, rather than to pharmacologic suppression of stage 3 NREM sleep. The treatment of RBD comprises primarily the prevention of potentially dangerous situations that can expose the patient to injuries. The first-line pharmacologic treatment is clonazepam at low doses at bedtime. This is effective in most patients, but the motor behavior often relapses after it is stopped. Melatonin (3–12 mg at bedtime) (Mahowald & Schenck, 2005), pramipexole (0.18–0.72 mg) (Fantini et al., 2003), or levodopa are also effective.
Maintain regular sleep–wake cycle with adequate amounts of sleep, avoiding sleep deprivation
Avoid all stimuli that may contribute to partial arousal and trigger an episode, for example, decreasing noise, light, pain, nocturia, which may contribute to partial arousal
Avoid extreme exercise, fatigue, and emotional or situational stress
Search for and treat, if present, other sleep disorders such as sleep apnea, restless legs, and gastroesophageal reflux
Avoid alcohol, antipsychotics, antidepressants, antihistamines, sedative hypnotics, and benzodiazepines
The authors have no conflicts of interest to disclose. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.