Abnormal behaviors during sleep from the viewpoint of sleep epileptology: current and future perspectives on diagnosis

Abnormal behaviors during sleep (ABDS) exhibit a myriad of symptoms. Their underlying diseases are also diverse, which include NREM/REM-related parasomnias, epilepsy and mental disorders. Since ABDS may severely affect a patient’s quality of life, giving an early and accurate diagnosis of the underlying disease (by analyzing video-polysomnographic data during the manifestation of ABDS) is of great importance. However, accurate diagnosis of ABDS is rather difficult. Recently it has been suggested that the pathology of (NREM/REM-related) parasomnias and epilepsy are closely related. In order to unravel the pathophysiological substrate of ABDS, it is essential to develop a novel approach based on sleep epileptology, a field which targets the interface between sleep medicine and epileptology.


Background
Abnormal behaviors during sleep (ABDS) can range across a wide spectrum, from simple and minor motor activities (e.g., paroxysmal arousals, limb myoclonus) to complex and intense behaviors (e.g., wandering, talking, screaming, hyperactivity, violence).
Diagnosis of ABDS is difficult for the following reasons (Breen et al. 2018;Ingravallo et al. 2014). On the patient side, i) since ABDS are rarely observed (or video recorded), patients have little or no subjective information about the symptoms (it is also often the case that nobody else notices them); ii) patients do not have a medical examination until they realize disadvantages in social life, or experience trauma caused by ABDS. On the doctor side, i) although video-polysomnography (V-PSG) using full-montage electroencephalography (EEG) is a highly regarded test, it requires a lot of time and money. Only a few patients take V-PSG; ii) the number of medical specialists who can accurately diagnose ABDS is still low.
Patients suffer from ABDS not only at night. ABDS cause inadequate sleep quantity and quality at night, which in turn causes excessive daytime sleepiness. ABDS tend to result in functional disorders in various situations that include a patient's family life, social life, career plan, and school life. Due to ABDS, patients and their bed partners can even get hurt, which sometimes requires forensic psychiatric evidence/decisions (Ingravallo et al. 2014). In short, ABDS may severely affect the quality of life (QOL) of patients and their families (Breen et al. 2018;Ingravallo et al. 2014).
In this paper, I propose a clinical-practice based classification of the underlying diseases of ABDS and present perspectives from some of the best research on pathophysiological relations between parasomnias and epilepsy, both representing typical ABDS. I also argue that it is imperative to develop a novel approach based on sleep epileptology (Chiba 2019), a field which targets the interface between sleep medicine and epileptology to uncover the pathophysiology behind ABDS. classification of Breen et al. (Breen et al. 2018) who incorporated the latest insights into the International Classification of Sleep Disorders (ICSD), third edition (ICSD-3, 2014) (American Academy of Sleep Medicine 2014). In the other category, from my clinical experience, of particular importance are delirium, panic disorder, posttraumatic stress disorder (PTSD), psychogenic nonepileptic seizures (PNES) and dissociative (conversion) disorder, which are frequently observed in medical practice. Hence we always bear them uppermost in mind when we make a differential diagnosis of ABDS.
Delirium has transient consciousness disturbance as its cardinal symptom, frequently observed in inpatients at general hospitals (10-82%) (Inouye et al. 2014). Hyperactive delirium requires prompt clinical treatment. Possible hyperactive delirium could sometimes turn out to be ictal/postictal delirium caused by epileptic seizures. It is frequently observed that elderly patients manifest nocturnal insomnias, sleep-wake cycle disturbance, and nightmares before delirium (Hatta et al. 2017). It is reported that early treatment of these sleep disorders could prevent development of delirium (Hatta et al. 2017).
Most patients with panic disorders have experienced panic attacks not only during the daytime but also at night (Staner 2003). Panic attacks are sometimes wrongly diagnosed as sleep terrors, nightmare disorders, or epilepsy. V-PSG observation at the manifestation of panic attacks indicates that panic attacks are likely to occur during the transitional period from stage 2 to stage 3 of non-REM (NREM) sleep but not while awake or during REM sleep (Staner 2003). On the other hand, parasomnias and nightmare disorders occur during stage 4 of NREM sleep and REM sleep, respectively (Staner 2003).
People with PTSD show a high incidence of sleep disorders. A study in the United States investigated 277 adult patients with PTSD and reported that about 93% of the patients had also developed sleep disorders: 56.7% had both insomnias and nightmares, 24.9% had only insomnias and 11.3% had only nightmares (Milanak et al. 2019). Patients without sleep disorders accounted for just 6.9% (Milanak et al. 2019). Note that PTSD may also cause dream enactment and parasomnias (Breen et al. 2018).
PNES are characterized by sudden and time-limited disturbances of motor, sensory, autonomic, cognitive, and/or emotional functions that are often misdiagnosed as epileptic seizures. Although PNES mostly occur during the day, they can occur at any time during the night. PNES always occur during awakening but never during sleep. In contrast to epileptic seizures, PNES are not associated with epileptiform discharges seen in EEG, but are instead derived from psychologic underpinnings (Chen et al. 2017;Gates et al. 1985). Intractable epilepsy is complicated by PNES at high rates. Among patients referred to outpatient epilepsy centers, 5 to 25% are considered to have PNES, while 25 to 40% of patients evaluated in inpatient epilepsy monitoring units for intractable seizures are diagnosed with PNES (Chiba 2019;Szaflarski et al. 2000). Since patients with intractable epilepsy develop high-frequent PNES, it is important to make an accurate differential diagnosis differentiating true seizures from PNES. Dissociative (conversion) disorders also manifest various ABDS that may resemble epileptic seizures and PNES. Dissociative (conversion) disorders are considered to underlie most PNES (Lawton et al. 2008). Table 1 gives a list of potential underlying diseases of ABDS. In actual medical practice, patients suffering from ABDS may have multiple underlying diseases (e.g., parasomnias and PTSD, sleep-related epilepsy and PNES).

Significance of V-PSG and stereo-EEG
The introduction of V-PSG into clinical practice in the 1980's has enabled highly accurate differential analysis for ABDS. Furthermore, it has played an important role in detecting a large body of useful clinical information for treating patients, which includes abnormal sleep architecture, sleep disorders (such as insomnias and sleep apneas), epileptic seizures and interictal/ictal epileptiform discharges which would often be overlooked in the daytime. For instance, the incidence rate of subjective sleep disorders was high among epileptic patients, at approximately 40% (van Golde et al. 2011), and the disorders have been confirmed by several V-PSG diagnoses such as degradation in the quantity and quality of nocturnal sleep. Studies of generalized epilepsy, for instance, have reported reduction in the percentage of NREM sleep (Krishnan et al. 2014), reduction of REM sleep (Mekky et al. 2017;Roshan et al. 2017) and increased time awake after sleep onset (Krishnan et al. 2014;Mekky et al. 2017) in comparison to control populations. In focal epilepsy, there are reports of decreased REM sleep (Parrino et al. 2012a), decreased (Miller et al. 2016) or increased N3 (Parrino et al. 2012a) and increased time awake after sleep onset (Parrino et al. 2012a V-PSG is a highly regarded test in making a differential diagnosis between parasomnias and epilepsy. V-PSG has to simultaneously document video-recorded behaviors, as well as record electrooculograms, electromyograms, and various other biological phenomena. In order to give a differential diagnosis of ABDS, the following information is crucial: under which brain conditions (awakening, NREM sleep, or REM sleep) and under what other conditions (facial expression, body movement, circulatory and respiratory dynamics, etc.) did ABDS occur? We usually continuously record V-PSG using full-montage EEG for 1 to 3 days (9-72 h). Other electrodes such as sphenoidal electrodes may be added to the full-montage electrodes depending on diagnostic purposes.
Depth electrodes or subdural electrodes may be implanted into the brains of patients with intractable epilepsy, who are potential candidates to receive neurosurgery, as an examination before operation. EEG in V-PSG performed for such patients are called stereo-EEG (S-EEG) [that is based on the stereotactic placement of a number of intracerebral multilead electrodes to obtain long-term EEG recording in a 3-D arrangement] (Gibbs et al. 2016). S-EEG can be an invasive test in brain tissues and hence it should be allowed only as an examination before epileptic surgery. Although S-EEG has been used in very restricted settings, it offers precious opportunities to observe how a patient's deep brain structure is involved in epileptic seizures and various ABDS (Gibbs et al. 2016).

Sleep-related epilepsy
Janz (1974) examined 2825 epilepsy patients with generalized tonic-chronic seizures and investigated the association between the seizures and the sleep/wake cycle through clinical observation. Symptoms were classified into three groups: sleep epilepsy with seizures occurring during sleep (44%), awakening epilepsy with seizures occurring shortly after awakening (33%), and diffuse epilepsy with seizures occurring with no correlation to the sleep/waking cycle (23%).
The term sleep-related epilepsy covers both sleep epilepsy (e.g., nocturnal frontal lobe epilepsy) and awakening epilepsy (e.g., juvenile myoclonic epilepsy). Although it was not in ICSD-1 (1990) (American Sleep Disorders Association (Thorpy MJ, Chairman) 1990), it was mentioned in both ICSD-2 (2005) (American Academy of Sleep Medicine 2005) and ICSD-3 (2014) (American Academy of Sleep Medicine 2014). According to Janz (1974), sleep-related epilepsy, which includes both sleep epilepsy (44%) and awakening epilepsy (33%), accounts for a total of 77% of epilepsy incidents. It can be said that epilepsy is a brain disease closely related to sleep.
As we have already seen, Janz (1974) is a pioneer in the clinical study of the association between sleep and epilepsy, providing a conceptual basis for sleep-related epilepsy. Table 2 summarizes sleep-related epilepsy. Seizures caused by sleep-related hypermotor epilepsy (nocturnal frontal lobe epilepsy), a representative of focal epilepsy, occur mostly during NREM sleep and rarely during REM sleep. This will be discussed in more detail later. The expected reason is that thalamocortical hypersynchrony during NREM sleep promotes epileptic seizure generation (Herman et al. 2001).
From nocturnal paroxysmal dystonia to nocturnal frontal lobe epilepsy Lugaresi and Cirignotta (1981) described five patients with frequent episodes occurring in clusters during sleep, characterized by bizarre movements and/or dystonic-tonic posturing of the limbs. These patients showed no interictal and ictal epileptiform discharges in scalp EEG but had a good response to carbamazepine. Based on these findings, the authors came to consider the condition an unusual motor disorder of sleep with an unclear pathophysiology, and labeled the disorder "hypnogenic paroxysmal dystonia", modified later to nocturnal paroxysmal dystonia (NPD) (Lugaresi et al. 1986).
Several years later, Wada (Wada and Purves 1984;Wada 1988) reported that pre-surgical S-EEG evaluation in intractable epileptic patients revealed the epileptiform discharges of frontal origin during the NPD seizures. Similar findings were reported by other investigators (Waterman et al. 1987;Williamson et al. 1985). Furthermore, Tinuper et al. (1990) demonstrated clear-cut epileptiform discharges in the ictal and interictal EEG recordings of three patients previously diagnosed as NPD. These findings suggest that NPD is of epileptic origin and the term was changed to the term nocturnal frontal lobe epilepsy (NFLE).
The V-PSG studies on NPD have greatly contributed to the refinement of the criteria in the ICSD. In particular, it should be noted that NPD, which was originally in the category of parasomnias in ICSD-1 (1990) (American Sleep Disorders Association (Thorpy MJ, Chairman) 1990), was reclassified into the category of NFLE in both ICSD-2 (2005) (American Academy of Sleep Medicine 2005) and ICSD-3 (2014) (American Academy of Sleep Medicine 2014), which implies that without V-PSG, it is extremely difficult to make a differential diagnosis of epilepsy from parasomnias.
The clinical boundaries of NFLE have been mostly refined by the Bologna school (Italy) (Provini et al. 1999(Provini et al. , 2000Tinuper et al. 1990Tinuper et al. , 2005. The clinical symptoms of NFLE, which are similar to those of NPD, are as follows: (1) paroxysmal arousals, (2) hypermotor seizures, (3) asymmetric bilateral tonic seizures, and (4) prolonged epileptic nocturnal wanderings Montagna 1992;Plazzi et al. 1995;Tinuper et al. 2005). The clinical spectrum comprises distinct paroxysmal sleep-related seizures of variable duration (3-120 s, or more) and complexity  ranging from paroxysmal arousals or very brief motor attacks to hypermotor seizures sometimes followed by prolonged complex ambulatory behavior. Provini et al. (1999) examined the clinical and V-PSG findings of 100 consecutive patients with NFLE. NFLE seizures predominate in males (70%). Age at onset of the nocturnal seizures varies, but centers during infancy and adolescence. A familial recurrence of the epileptic attacks is found in 25% of the cases, while 39% of the patients present a family history of nocturnal paroxysmal episodes that fit the diagnostic criteria for parasomnias. Incidence of neuroradiological findings are low. In many patients, ictal (44%) and interictal (51%) EEG findings are uninformative. Marked autonomic activation is a common finding during the seizures. NFLE does not show a tendency to spontaneous remission. Carbamazepine completely abolishes the seizures in approximately 20% of the cases and gives remarkable relief (reduction of the seizures by at least 50%) in another 48%.
Note also that a family was reported to have a high incidence of both (NREM or REM-related) parasomnias and nocturnal frontal lobe epilepsy (Tinuper et al. 2010). Therefore, it is suggested that NFLE and parasomnias share some common pathophysiologic substrate, which requires further investigation.

Sleep-related hyperactive epilepsy: a new concept of NFLE
At a Consensus Conference held in Bologna, Italy in 2014, NFLE was renamed sleep-related hypermotor epilepsy (SHE), based on three critical issues justifying the renaming (Tinuper et al. 2016;Tinuper and Bisulli 2017). First, the term nocturnal was considered misleading because it implies a chronobiological pattern of seizure occurrence, whereas evidence indicates that occurrence in sleep is the most important characteristic rather than the time of day. Second, the term frontal lobe is not always appropriate because the characteristic seizures may also arise from extrafrontal areas. In SHE, approximately 70% of cases have a frontal lobe origin, whereas the remaining 30% of the cases have an extrafrontal origin, coming from the insula, the temporal lobe, as well as the parietal lobe. Third, the term NFLE did not specify the typical clinical semiology involved,  (Tinuper et al. 2016;Tinuper and Bisulli 2017).
SHE diagnosis is primarily based on clinical history. The absence of clear interictal and ictal epileptiform discharges correlates does not necessarily indicate a negative SHE diagnosis. Three different levels of diagnostic certainty have been identified: (1) Witnessed (possible), based on observation by a witness of the core clinical features, but without other sources of evidence. (2) Video-documented (clinical), which involves a high quality video recording of at least one (but preferably two) stereotyped episodes. (3) Video-EEG-documented (confirmed), which requires the V-PSG recording of at least one but preferably two stereotyped events with documented ictal discharge or interictal epileptiform abnormalities (Tinuper et al. 2016;Tinuper and Bisulli 2017).

Sleep and its instability in epilepsy
Fine EEG changes called Cyclic Alternating Pattern (CAP) in intractable epilepsy, which indicate unstable NREM sleep, have been observed at a high rate among patients with intractable epilepsy (Parrino et al. 2012b). It has also been suggested that CAP (the phase A of the CAP) may trigger epileptic seizures (CAP-related seizures) (Halász et al. 2013;Parrino et al. 2012b). These sleep disorders are caused by epileptic seizures, whereas epileptic seizures can also be caused by the disorders. This mutually advancing relation holds between epileptic seizures and sleep disorders.
Paroxysmal arousals (PAs) and minor motor events (MMEs) are mild motor events seen in NFLE. PAs are characterized by sudden and brief arousals (5-10 s) often accompanied by stereotyped movements, dystonic posture, vocalization, frightened facial expression, and/or fear. MMEs are even shorter (2-4 s) movements, often stereotyped, involving the axial musculature or the head and limb (Gibbs et al. 2016). Both PAs and MMEs may go unnoticed by the bed partner or family members (Gibbs et al. 2016). Interictal/ictal epileptiform discharges of PAs and MMEs are often not detectable in scalp EEG recordings (Gibbs et al. 2016).
V-PSG with S-EEG studies revealed that PAs are epileptic seizures because all PAs are associated with epileptiform discharges. However, the same cannot be said of MMEs (Gibbs et al. 2016).
According to a V-PSG with S-EEG study on the relationship between MMEs, epileptiform discharges, and arousal fluctuations during sleep (Terzaghi et al. 2008), the MMEs as well as epileptiform discharges shared a close relationship with arousal fluctuations as depicted by analysis of the CAP, preferably occurring in phase A of the CAP. Therefore, epileptiform discharges-related MMEs seem not to be epileptic seizures but to be nonepileptic motor events which are related to sleep instability (CAP). MMEs may be facilitated, in a nonspecific way, by the presence of epileptiform discharges .
In summary, PAs seem to be epileptic seizures of NFLE (SHE), while MMEs are not epileptic. Several studies suggest that sleep-related epileptiform discharges internally increase sleep instability (CAP rate) that in turn enhance the occurrence of MMEs or other sleeprelated motor events (e.g., parasomnias, periodic limb movement disorder). Conversely, the increased sleep instability (increased CAP rate) would also facilitate the generation of the sleep-related epileptiform discharges.
Delirium, REM sleep behavior disorder, and epilepsy V-PSG may be difficult to carry out in delirious patients because they often exhibit psychomotor excitement. Japanese investigators reported a few V-PSG studies of delirium. This delirium took place during the alcoholic withdrawal period in alcoholics, and during the delirious state immediately after administration of an anticholinergic agent (biperiden) in normal volunteers (Hishikawa 1991;Kojima et al. 2000;Tachibana et al. 1975). These studies revealed peculiar polygraphic readings called stage 1-REM with tonic EMG (stage 1-REM) could be found during delirium. Stage 1-REM is characterized by the features of both stage 1 and stage REM, with concomitant occurrences of low voltage, fast-and-slow mixed frequency EEG, markedly elevated tonic EMG, and markedly increased rapid eye movements (Hishikawa 1991;Tachibana et al. 1975).
It is reported that REM sleep without atonia that is similar to stage 1-REM is observed in REM sleep behavior disorder (RBD) (Ferri et al. 2014;Schenck et al. 1986;Sunwoo et al. 2019) and acute RBD (Provini and Tachibana 2018) that has been associated with various medications or substances, in particular antidepressants, and the abrupt withdrawal from barbiturates, tricyclic antidepressants, monoamine oxidase inhibitors, and alcohol. Therefore, the same pathophysiology of stage 1-REM can be observed not only in delirium but also in RBD and acute RBD.
We confirmed experimentally that the behavioral and video-polygraphic changes induced by biperiden administration in rats are consistent with those of delirium in humans (Tamura et al. 2006). These findings suggest that a biperiden-treated rat is a good animal model for anticholinergic delirium (Tamura et al. 2006). These findings also indicate that anticholinergic mechanisms play an important role in the manifestation of hyperactive delirium, involved with changes in arousal level and REM sleep mechanisms (Tamura et al. 2006).
Recently, Hatta et al. (2017) suggested that in a randomized placebo-controlled study, suvorexant, a potent and selective orexin receptor antagonist, is effective for the prevention of delirium in elderly patients admitted for acute care. Because the main effect of suvorexant was seen in the sleep-wake cycle of the Japanese version of the Delirium Rating Scale-Revised-98 (Hatta et al. 2017) and the drug has not shown cholinergic affinities, the reservation/restoration of sleep-wake cycle may have priority over cholinergic neurotransmission in delirium prevention.
Epilepsy may coexist with delirium and RBD. In patients with epilepsy, it is well known that delirium is frequently observed during the ictal and postictal period. Manni et al. (2007) reported that RBD episodes were found to coexist with epilepsy in 10 out of 80 (12.5%) elderly subjects (aged 60 or over). Iranzo et al. (2006) also reported five patients with potassium channel antibody-associated limbic encephalopathy who showed simple or complex partial seizures of epilepsy with epileptiform discharges in the temporal regions during wakefulness as well as RBD episodes confirmed by V-PSG. It has been suggested that REM sleep inhibits epileptic phenomena due to a desynchronization of cortical cellular discharges and the physiological muscle atonia. Interestingly, a number of experimental studies showed that in feline REM sleep without cortical EEG desynchrony and REM sleep without atonia induced pontine dissociation techniques facilitate seizures of both generalized and limbic seizure models (Shouse 2002). Although the pathophysiological mechanisms of the coexistence of RBD and epilepsy is still unclear, functional abnormalities of the limbic system and the brainstem is assumed to be responsible for the coexistence (Iranzo et al. 2006).

Conclusion
ABDS show a myriad of symptoms ranging from simple and minor motor activities to complex and intense behaviors. Their underlying diseases are also diverse, which include psychiatric disorders not listed in in the ICSD-3 such as delirium, panic disorder, PTSD, PNES and dissociative (conversion) disorder. Since ABDS may severely affect the QOL of patients and their families, correctly giving an early diagnosis of the underlying disease(s) is extremely important. In clinical practice, it is imperative to properly build V-PSG findings during the manifestation of ABDS.
Since the pathophysiology behind ABDS is not fully understood, the present paper focused on parasomnias and epilepsy, the underlying diseases of ABDS, and discussed some important achievements in ABDS research.
The progress of the two research fields of sleep medicine and epileptology has been advancing with the recent development of long-term monitoring using V-PSG with full-montage EEG, various brain imaging technologies, and genetic screening systems. Great attention has been paid to the close relationship between the two fields, accumulating fruitful research findings. S-EEG has deepened our understanding of deep brain structure which scalp EEG could not capture.
Previous studies on V-PSG have made great contributions to the refinement of the criteria in the ICSD-1, − 2, and − 3. In particular, it should be noted that nocturnal paroxysmal dystonia, which was originally in the category of parasomnias in ICSD-1 (1990), was reclassified into the category of NFLE in both ICSD-2 (2005) and ICSD-3 (2014), which implies that without V-PSG, it is extremely difficult to make a differential diagnosis of parasomnias from epilepsy.
The underlying diseases behind ABDS turned out to be more than targets of differential diagnoses and have some pathophysiologic similarity. Recently, fine EEG changes called CAP in intractable epilepsy, which indicate unstable NREM sleep, have been observed at a high rate among patients with intractable epilepsy. It has also been suggested that CAP may trigger epileptic seizures (CAP-related seizures). These sleep disorders are caused by epileptic seizures, whereas epileptic seizures can also be caused by the disorders. This mutual relation holds between epileptic seizures and sleep disorders.
To sum up, it has been suggested that sleep disorders and epilepsy are closely associated with each other. Sleep epileptology, which covers the interface between sleep medicine and epileptology, will enable us to deepen our understanding of the semiology of ABDS, refine the diagnostic criteria, and take further steps towards unraveling the pathophysiology behind ABDS. In order to develop treatments for patients suffering with ABDS and improve their QOL, it is absolutely essential for sleep epileptology to establish its own field by accumulating findings and achievements.