Influence of sleep on seizures and interictal epileptiform discharges in epilepsy
Article information
Abstract
Sleep significantly influences seizure occurrence and interictal epileptiform discharges (IEDs) in patients with epilepsy. Sleep-related epilepsy, where seizures occur exclusively or predominantly during sleep, has been observed in various epilepsy syndromes. Understanding the influence of sleep on seizures and IEDs is crucial in the diagnosis, classification, and management of epilepsy. Although there is a bidirectional relationship between sleep and epilepsy, this review focuses on the influence of sleep on seizures and IEDs in epilepsy. Seizures are more common during non-rapid eye movement (NREM) sleep, particularly during stage N2, and are suppressed during rapid eye movement (REM) sleep. Sleep also activates IEDs, increasing the diagnostic yield of EEG recordings. The rate of IEDs increases during NREM sleep, reaches its maximum during stage N3, and decreases during REM sleep. Sleep affects the electrical field of IEDs, with an increase of spiking fields during NREM sleep and a decrease during REM sleep. In the localization of epileptogenic foci, REM sleep is less sensitive but more specific than NREM sleep. Thalamocortical EEG synchronization during NREM sleep and desynchronization during REM sleep underlie their opposing effects on seizures and IEDs. Accumulating evidence has suggested an antiseizure effect of orexinergic antagonism in animal studies. Interventions that promote REM sleep, including orexinergic antagonists, should be studied in the future as novel treatment strategies for epilepsy.
Introduction
It has long been recognized that the sleep-wake cycle influences seizure occurrence in epilepsy. In some patients with epilepsy, seizures occur exclusively or predominantly (>90%) during sleep, a situation referred to as sleep-related epilepsy. Self-limited epilepsy with centrotemporal spikes (previously known as benign childhood epilepsy with centrotemporal spikes), self-limited epilepsy with autonomic seizures (Panayiotopoulos syndrome), and sleep-related hypermotor epilepsy are well-known epilepsy syndromes that are strongly related to sleep. Prevalence of sleep-related epilepsy varies between 7.5% and 45% depending on the case definition [1]. Langdon-Down and Brain [2] analyzed the incidence of 2,524 seizures in 66 patients with epilepsy and found that 24.2% had seizures exclusively or predominantly at night (nocturnal), 42.5% exclusively or predominantly during the day (diurnal), and 33.3% during either day or night (diffuse). Notably, nocturnal seizures had two time peaks; 2 hours after going to bed and 2 hours before waking. However, a diagnosis of pure sleep epilepsy does not guarantee the absence of seizures while awake, with awake seizure incidence ranging from 7.5% to 19% during two years of follow-up in patients with pure sleep epilepsy [3,4]. Higher baseline seizure frequency and sudden withdrawal of antiseizure medication were associated with a higher incidence of awake seizures in patients with pure sleep epilepsy [3].
Relationship of Seizures to Sleep Stages
Sleep is divided into rapid eye movement (REM) and non-rapid eye movement (NREM) stages. When scoring sleep stages from polysomnography (PSG) recordings, REM sleep is scored as stage R, and NREM sleep is further divided into stages N1, N2, and N3. Each sleep stage has unique electroencephalographic (EEG) and neurophysiological characteristics, which are attributed to neural mechanisms that regulate the sleep-wake cycle [5]. Previous studies have shown that epileptic seizures commonly occur during NREM sleep, particularly during stage N2, and are suppressed during REM sleep. In an overnight video-EEG-PSG study, 95% of seizures occurred in NREM sleep (61% in N2, 20% in N1, and 14% in N3) [6]. The number of seizures per time spent in each sleep stage was also highest in stage N2 at 0.42/hour, followed by N1 at 0.36/hour, N3 at 0.30/hour, and R at 0.06/hour. Another study showed similar results; sleep seizures were most common during stage N2 (68%), followed by N1 (23%) and N3 (9%), and there were no seizures during REM sleep [7]. Moreover, focal to bilateral tonic-clonic seizures (previously known as secondarily generalized tonic-clonic seizures) were most common in stage N2 (45.5%), followed by N3 (29.4%) and N1 (21.5%). Regarding the location of seizure onset, seizures during sleep were more common in frontal lobe epilepsy (FLE; 57.1%) than in temporal lobe epilepsy (TLE; 43.5 %) and occipital lobe epilepsy (13.3 %). A study comparing the distribution of seizures in patients with FLE and TLE revealed consistent findings [8]. Seizures during sleep were more frequently observed in FLE (61.1%) than in TLE (10.9%). Overall, 94.5% of sleep seizures occurred in stage N2, and there was no difference in the distribution of seizures according to sleep stage between FLE and TLE. Taken together, these findings indicate that NREM sleep, particularly stage N2, promotes seizures, whereas REM sleep inhibits seizures. The seizure-promoting effect of NREM sleep is more pronounced in FLE than in other focal epilepsies.
Effect of Sleep on Interictal Epileptiform Discharges
EEG is an essential tool for the diagnosis, classification, and management of epilepsy. Interictal epileptiform discharges (IEDs) on EEG are highly specific to each patient, and sleep is the most powerful and best documented activator of IEDs [9]. It has been established that EEG recordings during sleep increase the rate of diagnosis in patients with suspected epilepsy. In 1947, Gibbs and Gibbs [10] reported that waking EEG detected IEDs in 36% of patients, which increased to 82% in sleep EEG. Independent of the activating effects of sleep deprivation [11], sleep itself has been shown to activate IEDs in approximately one-third of patients with epilepsy, and the rate of diagnosis increased to 90% in patients with sleep-related epilepsies [12]. A stereo-EEG study in patients with drug-resistant focal epilepsy demonstrated that the rate of IEDs increased during NREM sleep, reached its maximum during stage N3, and dropped during REM sleep to a level lower than that during wakefulness [13]. Changes in the IED rate by sleep stage were most prominent in the frontal and limbic areas. A meta-analysis of 10 studies on focal epilepsy found a 2.22-fold increase in the rate of focal IEDs during stage N3 and a 1.11-fold decrease during stage R compared with wakefulness [14].
Sleep affects not only the rate of IEDs but also the electrical fields of IEDs. Sammaritano et al. [15] examined changes in IEDs during sleep and wakefulness in 40 patients with TLE and found an increase of spiking fields (i.e., irrigative zone) during NREM sleep and a restriction of the fields during REM sleep. The appearance of bilateral independent IEDs increased during NREM sleep, whereas only unilateral spikes appeared during REM sleep. In the localization of epileptogenic foci, REM sleep is less sensitive but more specific than NREM sleep. These findings were confirmed in a source localization study using high-density EEG [16]. The spatial extent of IEDs during REM sleep was 24% less than that during NREM sleep. Moreover, the IED sources obtained during REM sleep were more concordant with the ictal onset zone than those obtained during NREM sleep.
Variation in IEDs by sleep-wake cycle was also observed in generalized epilepsy. A previous study of 12 patients with absence seizures showed that the IED rate was highest during stage N3 and lowest during REM sleep [17]. The morphology of IEDs also changed with sleep and wake state; generalized IEDs became more irregular, disorganized, and slower in frequency as NREM sleep deepened. Similarly, in patients with idiopathic generalized epilepsy, the rate of IEDs during NREM sleep was 14-fold higher than during wakefulness [18]. A meta-analysis of 256 patients with idiopathic generalized epilepsy demonstrated that, compared to wakefulness, the generalized IED rate was 2.03-fold higher during stage N3 and 3.25-fold lower during REM sleep [14].
Mechanisms by Which Sleep Affects Epilepsy
NREM sleep is characterized by synchronization among the brainstem reticular activating system, thalamus, and cortical pyramidal neurons, which becomes more pronounced as sleep progresses from stage N1 to N3. In addition, hypersynchronization of a neuronal population is one of the key mechanisms for the generation of epileptic seizures, which explains the activation of seizures and IEDs during NREM sleep. Cortical EEG recordings during NREM sleep are characterized by slow oscillations (SOs), which are slow waves of less than 1 Hz that originate from cortical neurons [19]. They appear only during NREM sleep, and their rate of occurrence progressively increases as sleep deepens. NREM SOs consist of rhythmic alternations between depolarized “up” and hyperpolarized “down” states and reflect variations in the resting membrane potentials of thalamic and cortical neurons. Cortical SOs are synchronized with other NREM-sleep oscillations including sleep spindles in thalamocortical circuits and sharp-wave–ripples in the hippocampus and contribute to the integrity of NREM sleep [20]. Precise SO synchronization is also implicated in information transfer from the hippocampus to the cerebral cortex and memory consolidation during NREM sleep [21]. Impairment of temporal coupling between cortical SOs and sleep spindles was noted in neurodegenerative diseases [22].
In an intracerebral EEG study of eight patients with drug-resistant focal epilepsy, Frauscher et al. [23] found that the numbers of IEDs and high-frequency oscillations (HFOs) increased proportionally to the amplitude of SOs during NREM sleep. Furthermore, IEDs and HFOs occur most often in SOs at the transition from the up to the down state rather than at the peak of the up state, supporting the idea that hypersynchronization contributes more to epileptic activities during NREM sleep than excitability does. However, it remains unclear why seizures most frequently occur during stage N2, even though IEDs predominate during stage N3. Deepening of NREM sleep is characterized by an increase in both EEG synchronization and cortical SOs, which is attributable to progressive hyperpolarization of thalamocortical neurons [24]. Furthermore, hyperpolarization of thalamocortical neurons contributes to development of physiological SOs into paroxysmal epileptic-like spike-and-wave complexes during NREM sleep. Therefore, a possible explanation for the discrepancy in sleep depth between IEDs and seizures is that the levels of synchronization and hyperpolarization that maximally activate epileptic networks differ between the two phenomena. It should also be emphasized that IED occurrence is not directly related to seizure onset. After adjusting for sleep depth, the rate of IEDs during sleep did not increase or decrease before seizures [25]. The lack of a temporal relationship supports the notion that different pathomechanisms underlie IEDs and seizures.
Antiepileptic Effect of Rapid Eye Movement Sleep
In contrast to NREM sleep, REM sleep suppresses seizures and IEDs. REM sleep is characterized by low-amplitude, mixed-frequency EEG activity; intermittent REMs; and loss of skeletal muscle tone. Unlike NREM sleep, EEG desynchronization during REM sleep represents widespread cortical activation, mainly resulting from depolarization of the thalamus by ascending projections from cholinergic REM-on neurons, such as the laterodorsal tegmental and pedunculopontine tegmental nucleus [26]. On the other hand, muscle atonia during REM sleep is caused by descending projections from glutamatergic REM-on neurons, such as the sublaterodorsal tegmental nucleus in the dorsal pons [27]. In an experiment using a feline generalized epilepsy model, loss of thalamocortical EEG desynchronization induced by systemic atropine abolished the inhibitory effects of REM sleep on seizures [28]. However, although electrolytic pontine lesions created REM sleep atonia, they did not affect seizure susceptibility. Therefore, the antiepileptic properties of REM sleep are attributed to ascending cholinergic projections originating from brainstem REM-on neurons.
REM sleep can be divided into tonic and phasic REM sleep depending on the absence or presence of REMs, respectively [29]. Cholinergic neurotransmission and EEG desynchronization increase more highly during phasic REM sleep than during tonic REM sleep. A scalp-intracranial EEG study demonstrated that IEDs, ripples, and fast ripples were suppressed more highly during phasic REM sleep than during tonic REM sleep [30]. Tonic REM sleep accounted for 61% of IEDs, while phasic REM sleep accounted for 39%. The discrepancy was more pronounced for fast ripples; 82% occurred during tonic REM sleep whereas only 18% occurred during phasic REM sleep. Similar results were found in patients with electrical status epilepticus during sleep. The spike index in phasic REM sleep was reduced by 87% compared with tonic REM sleep in that specific sleep-related epilepsy syndrome [31]. These results underpin the hypothesis that thalamocortical EEG desynchronization is responsible for the antiepileptic effects of REM sleep [14]. Spatial and temporal summation of aberrant depolarization and recruitment of surrounding neurons are essential for the generation of IEDs and seizure initiation. Therefore, cortical desynchronization during REM sleep impairs the pathomechanisms of epilepsy, suppressing the occurrence of both IEDs and seizures.
Role of the Orexin System in Epilepsy
Orexin, also known as hypocretin, is a hypothalamic neurotransmitter that regulates the sleep-wake cycle [32]. Orexinergic neurons promote arousal and sustain wakefulness by activating wake-promoting monoaminergic neurons. The selective loss of orexinergic neurons causes narcolepsy type 1 (NT1), which is a rare neurological disease characterized by excessive daytime sleepiness and cataplexy [33]. Orexin is also implicated in the regulation of REM sleep; during wakefulness, orexinergic neurons suppress REM sleep by inhibiting REM-on neurons and activating REM-off neurons [26]. Conversely, the activity of orexinergic neurons is absent during REM sleep. Cataplexy, a distinctive symptom of NT1, is a manifestation of REM sleep dysregulation resulting from the disinhibition of REM sleep due to the loss of orexin signaling during wakefulness.
Considering the antiepileptic effect of REM sleep, manipulation of the orexin system could be applied as a new treatment strategy for epilepsy. Accumulating evidence has suggested the antiseizure effect of orexinergic antagonism in animal studies [34]. Orexin receptor antagonists significantly reduced the severity and duration of pentylenetetrazole-induced seizures in rats [35]. Intraperitoneal injection of almorexant, a dual orexin receptor antagonist (DORA), reduced both the incidence of stage 2 to 5 seizures and overall seizure burden by 75% in a Kcna1-null mouse model of TLE [36]. Almorexant also reduced REM sleep latency, and there was a positive correlation between REM sleep latency and seizure severity. Furthermore, in the Kv1.1 knockout mouse model, intraperitoneal administration of DORA not only reduced seizures but also improved cardiorespiratory phenotypes related to sudden unexpected death in epilepsy and mortality [37]. However, the antiseizure effect of orexinergic antagonists has not yet been studied in patients with epilepsy. Only a few clinical studies have measured cerebrospinal fluid (CSF) orexin levels in epilepsy patients. These studies showed a decrease in CSF orexin levels in patients with epilepsy compared to controls, which was more pronounced after repetitive seizures or status epilepticus [38,39]. These results are interpreted as an effect of seizures on the orexin system and may be related to the postictal sleepiness commonly observed after generalized tonic-clonic seizure or status epilepticus. The hypothesis that orexinergic antagonism promotes REM sleep with diffuse cortical synchronization, consequently increasing the seizure threshold, should be verified through future clinical studies.
Summary
The impact of sleep on seizures and IEDs in epilepsy was reviewed in this study. Sleep and sleep deprivation are useful in increasing the accurate diagnosis of epilepsy. Identifying the close relationship between seizures and the sleep-wake cycle can help classify epilepsy syndromes. NREM and REM sleep have opposing effects in epilepsy, with NREM sleep activating seizures and epileptiform discharges on EEG, and REM sleep suppressing these. Thalamocortical EEG synchronization during NREM sleep and desynchronization during REM sleep underlie the promotion and suppression of seizures, respectively. In this regard, interventions that promote REM sleep, including orexinergic antagonists, are potential novel treatment strategies for epilepsy.
Notes
Conflicts of Interest
No potential conflict of interest relevant to this article was reported.