Sleep in Epilepsy Special Interest Group | Brain Flow Dynamics and Sleep Alterations Effect on Epilepsy: From Babies to Late Adults and Breathing to Cognition: The Impact of Sleep on Seizures and Cognitive Function in Epilepsy Through the Lens of Water Exchange Dynamics

Abstract

Decades of research have demonstrated the complex intersection between sleep and epilepsy, which significantly impacts seizure control and epilepsy-related comorbidities. However, further investigation is needed to understand the underlying mechanisms linking sleep and epilepsy. Brain exchanges are dynamic processes that include blood-to-brain transport across the blood–brain barrier (BBB), where metabolism is regulated to sustain neural activity, and brain-to-blood and cerebrospinal fluid clearance via the glymphatic system. Together, the BBB and glymphatic system maintain brain milieu homeostasis, and dysfunction at any level can contribute to neurological disorders, including epilepsy. This review highlights the dynamics of water flow within the brain and explores the interrelationship between exchange and clearance pathways across the lifespan, as presented at the 2025 American Epilepsy Society Sleep and Epilepsy Special Interest Group Session. We provide an overview of (1) sleep-related breathing disturbances, their effects on seizures and cognitive function throughout the lifespan, and their implications for sudden unexpected death in epilepsy; and (2) the role of water dynamics mediated by the BBB and the glymphatic system in modulating sleep and epilepsy. These insights deepen our understanding of epilepsy severity and the role of sleep in modifying disease progression, offering promising directions for future research to advance therapeutic strategies.

Preamble

The complex bi-directional interactions between sleep disorders and epilepsy have been known for many decades. The initial observations described sleep disruption by seizures and higher seizure frequency at times of sleep loss. Further studies have extended these observations by examining nocturnal seizures in relation to breathing and risk of sudden unexpected death in epilepsy (SUDEP). More recently, the theme of sleep as “natural neuroprotector” has emerged. The discovery of the glymphatic system revealed a major path by which sleep disorders may impair cognition and thus further worsen the lives of individuals with epilepsy. This impact may be especially severe in the developing brain. Therefore, in this special interest group session, we addressed the role of common sleep disorders in children and adults, starting with 2 clinical presentations and veering into the role of fluid dynamics, neuroinflammation, and the blood-brain barrier.

Introduction

Sleep disorders affect 60% to 80% people with epilepsy (PWE)—twice as common as the general population—with insomnia being most common (50% PWE), obstructive sleep apnea (OSA) at up to 35%, excessive daytime sleepiness (up to 34%), and restless leg syndrome (up to 28%).^1,2 Poor sleep is associated with higher risk for seizures, poor cognition, and epilepsy comorbidities including depression, metabolic disorders including hypertension, obesity, dyslipidemia, cardiovascular disorders, and mortality from all causes.^3,4

Healthy sleep is crucial for the restorative processes that improve recovery from toxins and insults, and is also important for a healthy memory consolidation. These mechanisms are needed for bidirectional blood and brain exchanges and to perform brain-waste clearance. ⁵ While normal sleep is important for such brain recovery, PWE are more frequently affected by different sleep disorders such as OSA, insomnia, and circadian rhythm disorders (Figure 1).

Figure 1.

Bidirectional relationship between sleep and epilepsy across the lifespan: breathing disorders and water exchange mechanisms.

This graphical abstract illustrates the complex bidirectional relationships between sleep and epilepsy throughout the human lifespan. It highlights how breathing disorders such as obstructive sleep apnea exacerbate both conditions, and how disruptions in blood–brain barrier integrity and glymphatic clearance contribute to their pathophysiology.

Accumulating evidence highlights the relationship between sleep, worsening epilepsy, and harder seizure control. Exchanges between the brain and the blood indirectly fluctuate due to exogenous (light, diet, etc) and endogenous (local molecules at the blood–brain barrier, BBB) stimuli. ⁶

This review represents the highlights of the 2025 American Epilepsy Society Sleep and Epilepsy Special Interest Group Session. We begin with an overview of the clinical importance of sleep in children with epilepsy, then the importance of sleep apnea in adults with epilepsy, importance of impaired arousals in sudden unexpected death in epilepsy (SUDEP), impact of poor sleep on the cognition of older adults with epilepsy, and finally possible sleep-dependent underlying brain and blood exchanges (BBB and the glymphatic system) which could reveal new therapeutic directions.

Sleep Breathing Disturbances in the Developing Brain

Sleep and epilepsy share a multidirectional relationship in children, influencing seizure threshold, cognitive development, and behavioral outcomes. This interplay varies by epilepsy type and patient characteristics. Children with epilepsy, especially those with drug-resistant epilepsy, commonly experience nocturnal seizures and comorbid sleep disorders including fragmented sleep, OSA, insomnia, and parasomnias.^7,8 Nocturnal seizures and/or co-existing sleep conditions lead to sleep deprivation, which can further lower the seizure threshold, impair cognitive function, and negatively affect emotional regulation. The Synaptic Homeostasis Hypothesis suggests sleep scales down synaptic strength to optimize learning and memory during wakefulness. ⁹ Sleep deprivation heightens synaptic excitability, explaining increased seizure risk after prolonged wakefulness. ¹⁰ In pediatric epilepsy, disrupted sleep prevents synaptic homeostasis, worsening seizure control and developmental outcomes. ¹¹

Non-REM (non-rapid-eye movement) sleep potentiates epileptiform discharges via corticothalamic synchrony and triggers seizures in many pediatric epilepsies.^9,10 Sleep-associated epilepsies are activated by N2 and N3 sleep, including self-limited epilepsies of childhood with centro-temporal spikes (SeLECTs) or with autonomic seizures (SeLEAS), and sleep-related hypermotor epilepsy. Some epilepsies, such as juvenile myoclonic epilepsy and generalized tonic-clonic seizures on awakening, occur around the time of waking. Sleep-accentuated epilepsies, developmental/epileptic encephalopathy with spike-wave activation in sleep (DEE-SWAS), Lennox-Gastaut syndrome, and West Syndrome are most activated by N3 sleep. ² The link between spike-wave burden during sleep and development is mixed. In children with DEE-SWAS, a higher spike-wave index (SWI) often predicts poorer outcomes, yet clinical improvement does not always correlate with SWI. ¹² In children with SeLECTs, higher spike-wave density in sleep likely impairs memory consolidation by disrupting slow wave–spindle coupling. ¹³ Children with SeLECTs, however, frequently have learning, memory, and executive function difficulties that persist despite treatment or remission. ¹⁴

Thus, sleep is critical for early brain development. Disruptions, especially early in life, impair synaptic plasticity, cortical maturation, and cognition. Clinicians should remain alert to the high risk of sleep disorders in children with epilepsy and their impact on seizures and learning. Further diagnostic evaluation, including 24-h video EEG, polysomnography, and actigraphy, may be required to characterize the child's specific sleep-related manifestations of epilepsy and comorbid sleep conditions.

Sleep Breathing Disturbances in the Adult

OSA is common in adults with epilepsy, especially among these with drug resistant epilepsy. ¹⁵ Classically, patients present with hypersomnolence, snoring, pauses in breathing observed by a bed-partner, or with complaints of fragmented sleep. However, in a patient with epilepsy, a presenting symptom may be a worsening seizure frequency. OSA should be considered whenever more frequent seizures are reported by someone who was previously better controlled or with a new incidence of epilepsy after the age of 50. ¹⁶ Less common OSA symptoms include headaches upon awakening, characteristically improving over the course of the day, hypertension, nocturnal sleep disruption, or cognitive symptoms. Screening instruments, such as the “STOP-BANG” questionnaire, may help clinicians more quickly identify patients at risk. ¹⁷

Risk factors for OSA include anatomical characteristics (large neck/small chin), as well as higher weight. The cardinal process includes multiple individual event of upper airway collapse, which lead to hypoxemia, and a sympathetic response, often visible as EEG arousal, which terminates the event and restores breathing. Thus, the effect of the events is via distinct and separate mechanisms: hypoxemia/hypercarbia, with potential effects on the vascular endothelium, and CNS response, which disrupts sleep and may lead to neuropsychological consequences.

Studies showed a direct link between OSA and neurodegeneration risk. A recent meta-analysis reported a significantly higher blood T-tau, P-tau, and Aβ40 in patients with OSA than in controls. ¹⁸ In another study of cognitively normal adults, higher OSA severity was associated with a higher incidence of adverse driving behaviors, independently of any Alzheimer's pathology. ¹⁹

In epilepsy patients, more severe OSA-related hypoxemia during sleep is associated with poorer cognitive performances on tests that assess attention/executive functions and verbal memory. Specifically, more severe respiratory events during rapid-eye-movement sleep were associated with worse performances on attention tests. There was also a significant positive correlations between the oxygen saturation levels during rapid-eye movement (REM) sleep and attention/executive tests. Furthermore, lower oxygen levels correlated with a worse verbal memory performances were associated with lower oxygen levels. ²⁰

Treatment of OSA includes lifestyle modifications, oral appliance therapy, medications, surgery, and positive airway pressure which may be challenging in epilepsy patients.^15,21 Addressing OSA can improve overall health and enhance seizure control. In a study of 132 subjects, divided in 3 groups: without OSA, treated OSA (PAP used >4 h, >5night per week), and untreated OSA. There were significantly more subjects with successful outcomes (with ≥50% seizure reduction or seizure-free at both baseline and follow-up) in the group with PAP-treated OSA than in the groups with no OSA and untreated OSA. After adjusting for age, gender, body mass index, AHI, and epilepsy duration, the odds of successful outcomes in subjects in the group with PAP-treated OSA were 9.9 and 3.91 times those of the groups with untreated OSA and no OSA, respectively. ²²

Thus, identifying and treating OSA in epilepsy patients is an important step towards improving sleep continuity and cognitive function and reducing seizure frequency.

Impaired Arousal After Seizure: Mechanisms and Relevance to SUDEP

The CNS responsiveness to major breathing compromise is a life-preserving mechanism, and a critical process that may impact risk of sudden death. The MORTEMUS study described a characteristic sequence of events in SUDEP cases occurring in epilepsy monitoring units. ²³ Fourteen of sixteen SUDEPs occurred at night, and 7 of 10 occurred during sleep. Crucially, detailed review of ictal and postictal physiology revealed a consistent cascade: severe postictal coma or failure to arouse, transient postictal cardiorespiratory abnormalities, progression to terminal apnea, and ultimately asystole. ²³

Arousal from sleep depends on a complex neural network driven largely by the ascending reticular activating system. ²⁴ Reticular neurons in the brainstem activate cortical arousal through both thalamic and extra-thalamic pathways and rely on diverse neurotransmitters, including serotonin (5-HT) from the midbrain and rostral pontine raphe nuclei, cholinergic input from the pedunculopontine and laterodorsal tegmental nuclei of the caudal midbrain and rostral pons, and noradrenergic transmission from the locus coeruleus. ²⁴ Increasing evidence supports a critical role for dorsal raphe nucleus (DRN) 5-HT neurons in CO₂-induced arousal: DRN 5-HT neurons increase their firing in response to elevated PaCO₂ and activate the external lateral parabrachial nucleus (PBel) via 5-HT₂A receptors, though they can also promote arousal through PBel-independent mechanisms. ²⁵ Serotonergic neurons in the medullary raphe exhibit similar CO₂ sensitivity and markedly increase activity during acidosis; these neurons augment ventilation by driving the respiratory network in the ventral medulla and are considered key mediators of central respiratory chemoreception (CCR). ²⁶

Experimental data support a mechanistic link between seizures, serotonergic dysfunction, impaired arousal, and SUDEP. Hippocampal seizures inhibit 5-HT neurons in both the medullary raphe and DRN. ²⁷ Seizures also suppress CCR in animals and humans via inhibition of medullary raphe 5-HT neurons. ²⁸ Conversely, DRN stimulation reduces postictal generalized EEG suppression (PGES)—a proposed SUDEP biomarker—accelerates postictal arousal and decreases seizure-induced mortality.

A recent clinical study from Sainju et al monitored peri-ictal CO₂ and supports impaired CO₂-mediated arousal after generalized convulsive seizures. ²⁹ In this study, the duration of postictal CO₂ elevation was independently associated with delayed recovery of consciousness after adjusting for PGES duration and serotonergic medication use (poster presented at AES 2025 annual meeting). Preictal CO₂ level and convulsion duration were also significantly associated with the duration of postictal immobility, another putative marker of impaired arousal (poster presented at AES 2025 annual meeting).

Animal data parallel these findings. A study found that seizures impair function of 5-HT neurons leading to prolongation of latency to arousal after CO₂ exposure compared to interictal baseline, and HCVR was also significantly attenuated. ¹²

Together, human and animal studies converge on the conclusion that seizures can impair both HCVR and CO₂-induced arousal—processes largely mediated by brainstem 5-HT neurons. Given the critical homeostatic importance of these systems, their seizure-related dysfunction likely contributes substantially to SUDEP risk.

Sleep, Neuroinflammation, and Epilepsy Comorbidities: Impact on Cognition in Older Adults

Seizures and cognitive impairment can be conceptualized as downstream manifestations of neurovascular unit (NVU) dysfunction. ³⁰ The NVU is essential for maintaining neuronal homeostasis by regulating activity-dependent cerebral blood flow, preserving the brain's immune privilege, delivering critical nutrients, and facilitating the clearance of neurotoxic proteins such as amyloid beta and phosphorylated tau. Notably, vascular risk factors including hypertension, smoking, diabetes, and the APOE4 genotype are associated with increased risk for both cognitive decline and late-onset seizures. ³¹ A unifying mechanism underlying this overlap may be NVU disruption. Impaired clearance mechanisms promote the accumulation of Alzheimer's disease pathology, which is itself pro-epileptogenic. In turn, seizures and cortical hyperexcitability exacerbate the release and spread of these pathogenic proteins, creating a vicious cycle of neurodegeneration and hyperexcitability. ³²

Neuroinflammation plays a central role in the adverse outcomes associated with NVU dysfunction. Disruption of the BBB permits the infiltration of circulating immune cells, such as T cells and macrophages, which in turn activate resident glial cells, including microglia and astrocytes. ³³ This activation leads to the release of pro-inflammatory cytokines, further amplifying the inflammatory milieu. The resulting feedback loop exacerbates BBB permeability, perpetuating neuroinflammatory damage. Notably, albumin extravasation following BBB disruption has been implicated in the initiation and propagation of seizure activity. ³³

Sleep, particularly slow wave sleep (SWS), is essential for maintaining NVU integrity. Emerging evidence suggests that the NVU achieves optimal functionality during sleep. ⁶ Consequently, sleep deprivation impairs NVU function, leading to the accumulation of reactive oxygen species and heightened neuroinflammation. ⁶ Similarly, sleep-disordered breathing, which disrupts sleep architecture and reduces SWS duration, can trigger comparable pathophysiological processes. These changes in turn lead to cognitive dysfunction and seizures as well.

Another downstream consequence of NVU dysfunction is cerebral hypoperfusion, which can lead to white matter ischemic injury (REF). These injuries often manifest as T2 hyperintensities on MRI and are recognized contributors to vascular cognitive impairment. Beyond cognitive decline, white matter damage disrupts structural connectivity between brain regions, an essential substrate for coordinated neural activity. ³⁴ This disruption is thought to underlie “disconnection syndromes,” in which impaired communication between cortical and subcortical regions results in network-level alterations, including hyperexcitability and cortical dysfunction. ³⁴ A higher burden of white matter ischemic disease has also been identified as a risk factor for late-onset epilepsy. ³⁵ Furthermore, white matter injury can affect sleep microarchitecture, impairing the generation of spindles and slow oscillations that are critical for memory consolidation.

Iatrogenic contributions to NVU dysfunction must also be considered. The addition of antiseizure medications to manage epilepsy may inadvertently alter vascular risk profiles and reduce the efficacy of antihypertensive therapies. These medications can affect cognition directly, and may also exert indirect effects through NVU disruption. In particular, enzyme-inducing antiseizure medications have been associated with an increased risk of cerebrovascular disease, potentially compounding NVU vulnerability. ³⁶

Global Brain Dynamic Flows: Common Yet Variable Contributors

Water in the Brain is Tightly Regulated by the BBB and the Glymphatic System

The BBB is a selective interface that regulates transfer of substances between blood and brain tissue. It facilitates the exchange of essential nutrients from the blood to brain tissue, removes toxins from the brain, and protects the brain from harmful substances in the bloodstream. ³⁷ Accumulating evidence demonstrates that BBB dysfunction plays a crucial role in sustaining seizures, beginning even before seizure onset during epileptogenesis. ⁶ Moreover, seizures themselves induce further BBB alterations, that contributes to recurrent seizure patterns, treatment resistance, and potentially epilepsy-related comorbidities such as memory decline. ³⁸

Emerging research indicates that sleep plays a critical role in maintaining BBB function. During sleep, BBB permeability is regulated to allow optimal brain homeostasis. ⁶ Studies showed that sleep deprivation or disruption can impair BBB function by increasing its permeability and potentially allowing harmful substances to cross. ³⁹

On the other hand, the glymphatic system ensures “waste” clearance to avoid toxin accumulation. The glymphatic system is a brain-wide network that promotes eliminating neurotoxic substances through the cerebrospinal fluid along perivascular spaces. This system is most active during non-REM sleep highlighting the importance of sleep in maintaining brain homeostasis. ⁴⁰ A recent study showed reduced glymphatic flow in patients with drug-resistant temporal lobe epilepsy and hippocampal sclerosis, compared to healthy controls. ⁴¹ In epilepsy, seizures may disrupt the glymphatic flow, sustaining this vicious circle where impaired clearance exacerbates seizure recurrence. Pre-clinical research supports that glymphatic system disruption was associated with increasing seizure frequency and cognitive delays. ⁴²

In summary, accumulating evidence supports a complex interplay between sleep, the BBB, and the glymphatic system in which dysfunction of any component may initiate a cascade that sustains seizures and contributes to cognitive decline in epilepsy.

Non-Contrast Water Exchange Imaging of the BBB and the Glymphatic Flow

Various MRI techniques have been employed to assess the integrity of the BBB and the blood–cerebrospinal fluid barrier (BCSFB) in humans. These methods can be broadly categorized based on the type of tracer used: water-based tracers and gadolinium-based contrast agents (GBCAs).

Water-based techniques leverage the distinct MRI properties of water protons across different tissue compartments. When water, or more specifically protons, cross barriers such as the BBB or BCSFB during MRI data acquisition, they experience differing MRI environments in each compartment. This differential allows for the measurement of barrier permeability. Commonly assessed MRI parameters include T1, T2, and diffusion properties. Some sequences are combined with arterial spin labeling (ASL), which magnetically labels inflowing blood and tracks signal changes downstream. ⁴³ Diffusion-prepared ASL exploits differences in water diffusivity between capillaries and tissue. ⁴⁴ T2-based (sometimes combined with T1) ASL utilizes differences in water relaxivity between arterial blood and tissue.^45,46 Additionally, certain sequences directly measure differences in water diffusivity between blood and brain parenchyma without ASL, such as filter-exchange imaging ⁴⁷ and vascular exchange imaging, ⁴⁸ using diffusion gradients and variable mixing times to assess water exchange across tissue barriers.

In contrast, GBCA-based techniques involve intravenous injection of gadolinium contrast, with MRI signal changes monitored before and after administration. These methods are predominantly applied to evaluate BBB integrity rather than BCSFB. The premise is that when the BBB is intact, GBCAs remain confined within blood vessels; however, BBB disruption permits gadolinium to leak into brain parenchyma, altering local T1 and T2* relaxation properties and enabling quantification of barrier leakage. Dynamic contrast-enhanced MRI is commonly used to assess T1 change, ⁴⁹ while T2/T2*-weighted imaging can also be employed to monitor these alterations. ⁵⁰

Conclusion and Future Directions

Addressing and ultimately “curing” epilepsy requires a comprehensive understanding of the complex interplay between sleep breathing disturbances and epilepsy. The heterogeneity of epilepsy has posed significant challenges for clinical research. However, a unifying mechanism across epilepsy types, causes, and locations appears to involve dysfunction of the BBB and the glymphatic system—key components of global brain dynamics essential for maintaining healthy brain function. These systems influence blood chemistry, which is modulated by sleep states and seizure occurrence, and is often dependent on specific sleep stages.

This insight opens new opportunities for future research that transcend the traditional heterogeneity of epilepsy. Targeting restoration of BBB integrity, microvascular health, and enhancement of brain perfusion and clearance of neurotoxic proteins such as beta-amyloid and tau may offer promising therapeutic strategies. Such interventions could act as disease-modifying treatments by restoring healthy brain exchanges, maintaining tissue homeostasis, and promoting effective toxin clearance.

Ultimately, these advances have the potential to reduce epilepsy comorbidities—including epilepsy-associated dementia—and improve treatment responsiveness, including responsiveness to antiseizure medications.