Vagus Nerve Stimulation for Sleep Disorders: Emerging Evidence and Future Directions

Abstract

Sleep disorders are increasingly prevalent clinical conditions that significantly impair physical and mental health. Given the observed limitations of pharmacological therapies, there has been increasing research interest in effective non-pharmacological interventions. Vagus nerve stimulation (VNS), including invasive VNS and non-invasive methods such as transcutaneous auricular VNS (taVNS), has emerged as a promising neuromodulatory strategy. This review summarizes current evidence on VNS for insomnia, central hypersomnia, restless legs syndrome, and sleep-related breathing disorders, highlighting both clinical findings and proposed mechanisms. Relevant literature was identified through searches of PubMed and Web of Science up to May 2026 using terms related to VNS (eg, “vagus nerve stimulation” and “taVNS”) and sleep disorders (eg, “insomnia”, “narcolepsy”, “hypersomnia”, and “sleep-related breathing/movement disorders”). The proposed mechanisms include multilevel regulation of sleep–wake neural pathways; immunomodulation via anti-inflammatory pathways; inhibition of cortical spreading depression; and enhancement of neuroplasticity, cerebral blood flow, and blood–brain barrier integrity. Finally, we discuss key challenges regarding inter-individual variability and stimulation parameter optimization, as well as outline future directions centered on precision neuromodulation. These include the development of intelligent closed‑loop stimulation systems (ie, devices that automatically adjust stimulation parameters based on real time physiological feedback) and the testing of state dependent neuromodulation hypotheses. Taken together, this review not only summarizes current evidence but also proposes a theoretical framework and translational perspective to inform future research and clinical applications.

Plain Language Summary

Why was this review done?

Many people have trouble sleeping, but current treatments (like sleeping pills) are not always effective or can cause side effects. Therefore, researchers are looking for new, drug‑free ways to improve sleep.

What did the researchers do?

We reviewed the scientific literature on vagus nerve stimulation (VNS) – a technique that uses mild electrical pulses to stimulate a nerve in the ear (called taVNS) or in the neck. We collected evidence from clinical studies and laboratory experiments to see whether VNS can help with sleep problems such as insomnia, narcolepsy and restless legs syndrome.

What did they find?

Several studies showed that taVNS can improve sleep quality in people with insomnia, and that it can also reduce daytime sleepiness in narcolepsy. However, the evidence is still limited as many studies are small or have design flaws, so the results must be interpreted with caution. For restless legs syndrome, evidence is limited and conflicting; larger trials are needed.

What do these results mean?

VNS, especially the ear‑stimulation method, may become a useful non‑drug option for certain sleep disorders. More research is needed to confirm its effects and to understand exactly how it works. This review provides a roadmap for future studies and suggests that personalized, “closed loop” devices could make VNS even more effective.

Introduction

Sleep disorders, including insomnia, hypersomnia, sleep-related breathing disorders, and sleep-related movement disorders, represent a major and growing public health challenge.1 Insomnia is characterized by difficulty initiating sleep, maintaining sleep, or early morning awakening. Central hypersomnia predominantly includes narcolepsy and idiopathic hypersomnia; narcolepsy is clinically characterized by uncontrollable sleepiness, cataplexy, hypnagogic hallucinations, and nocturnal sleep disturbance. Sleep-related movement disorders mainly include restless legs syndrome (RLS) and periodic limb movement disorder (PLMD); RLS is characterized by an intense, irresistible urge to move the limbs, which occurs or worsens at rest or during the night, accompanied by deep discomfort in the limbs. Sleep-related breathing disorders (SRBD), also referred to as sleep-disordered breathing (SDB), are most commonly and clinically importantly represented by obstructive sleep apnea (OSA). These conditions are highly prevalent; for example, insomnia alone affects approximately 10% of adults worldwide.2 Sleep disorders are strongly associated with reduced quality of life, increased risks of cardiovascular and metabolic diseases, and a substantial socioeconomic burden. Current treatment strategies primarily include pharmacotherapy and cognitive behavioral therapy for insomnia (CBT-I). However, these approaches have several limitations, including drug resistance, dependence, adverse effects, and limited accessibility to CBT-I. Therefore, novel, effective, and well-tolerated non-pharmacological interventions are needed.

Vagus nerve stimulation (VNS) has emerged as a promising neuromodulatory strategy to address this therapeutic gap. VNS modulates central neural activity and autonomic function by delivering electrical stimulation to the vagus nerve (VN). VNS can be broadly categorized into invasive VNS (iVNS) and non-invasive VNS (nVNS). iVNS requires surgical implantation of a pulse generator to stimulate the cervical VN and has been approved for conditions such as refractory epilepsy and treatment-resistant depression.3,4 The concept of VNS for seizure control dates back to the 19th century, when massage of the carotid region was observed to reduce seizure activity.5 The first human VNS implantation for epilepsy was performed in 1988,6 and FDA approval followed in 1997 for refractory partial-onset seizures.3 VNS was later approved for treatment-resistant depression in 2005.4 Subsequently, FDA approval was granted for cluster headache in 2017 and for migraine in 2018.7,8 In contrast, nVNS achieves neuromodulation through transcutaneous stimulation approaches, including cervical and auricular methods. Notably, there has been increasing attention to transcutaneous auricular vagus nerve stimulation (taVNS), which targets the auricular branch of the vagus nerve (ABVN) in the external ear, because of its convenience, safety, and tolerability. Building on these developments, subsequent studies have investigated the effects of VNS on sleep regulation through the modulation of sleep–wake circuits, key brainstem nuclei, and neurotransmitter systems. Nevertheless, the therapeutic efficacy of VNS appears to be condition-specific, with negative results reported in some conditions.9,10 This underscores that VNS is not a universally effective therapy and remains an area of active investigation. This review provides a comprehensive overview of the current evidence regarding the use of VNS for sleep disorders.

Moreover, we discuss the multifaceted mechanisms underlying VNS, including neural circuit regulation, anti-inflammatory effects, and enhancement of neuroplasticity. Additionally, we synthesize the available clinical evidence across different sleep disorders, examine current limitations such as inter-individual variability and methodological heterogeneity, and highlight future directions toward precision neuromodulation strategies in sleep medicine.

Clinical Applications of VNS in Sleep Disorders

Clinical Application of VNS in Insomnia

Insomnia, the most prevalent sleep disorder, has increasingly been investigated as a potential therapeutic target for VNS, particularly taVNS.11–17 An RCT by Zhang et al (2024) demonstrated that 8-week taVNS treatment significantly reduced Pittsburgh Sleep Quality Index (PSQI) scores by 8.2 points; compared with 3.9 points in the sham group.12 Moreover, the response rate (≥50% PSQI improvement) was 2.5‑fold higher in the taVNS group (69.4%) than in the sham group (27.8%) and this effect was sustained through 20 weeks of follow‑up.12 Similarly a double-blind RCT by Yeom et al (2025), reported significantly greater improvements with active taVNS than with the sham group across multiple outcomes, including reductions in PSQI and Insomnia Severity Index (ISI) scores, prolonged total sleep time, and improved quality of life.13

Additionally, a growing body of research has investigated the effects of taVNS on sleep architecture.14 Zhang et al (2023) conducted an RCT in patients with high-altitude insomnia using polysomnography (PSG) in combination with subjective sleep assessments.14 PSG data indicated that taVNS significantly shortened sleep latency, improved sleep efficiency, and increased deep sleep duration, whereas total sleep time and wake after sleep onset did not show a significant change.14 These findings suggest that taVNS may improve certain aspects of sleep structure, particularly deep sleep duration, although confirmation in larger studies is needed.

The hyperarousal model is one of the main theories of insomnia pathogenesis, supported by neuroimaging evidence.18 Consistently, functional connectivity (FC) alterations in prefrontal and limbic circuits are thought to reflect this hyperaroused state and correlate with insomnia severity.19 Neuroimaging studies have provided mechanistic insights into the effects of taVNS. A functional magnetic resonance imaging (fMRI) study by He et al (2022) reported that a single session of taVNS was associated with increased amplitude of low‑frequency fluctuations (ALFF) in the left dorsolateral prefrontal cortex (dlPFC) and decreased FC between this region and the bilateral medial prefrontal cortex (mPFC) in patients with chronic insomnia disorder.17 These preliminary findings have not yet been independently replicated and should be interpreted cautiously.A prospective study by Qi et al (2025) reported that taVNS significantly reduced PSQI, ISI, and Self-Rating Anxiety Scale scores.11 They suggested that taVNS alleviates insomnia symptoms by modulating FC between the basal forebrain (BF) and several brain regions, including the visual cortex, sensorimotor cortex, and medial prefrontal cortex (mPFC).11 Furthermore, the study found that baseline FC within the BF–visual circuit, internally validated in an independent cohort, may hold promise as a candidate biomarker for predicting treatment response. However, its clinical applicability remains uncertain due to insufficient sensitivity and specificity, and further validation in external, larger and multicenter studies is needed.

In addition to variable treatment responses, a study by Wu et al (2021) evaluated 40 patients with primary insomnia after 4 weeks of taVNS and found that approximately 26% of patients showed a PSQI improvement rate of less than 25%, classified as non-responders.9 Notably, the non-responder group exhibited higher baseline activity in sensorimotor network regions and lower heart rate variability during taVNS.9 These findings underscore that taVNS is not universally effective across all insomnia patients and highlight the need for predictive biomarkers to guide patient selection.

Finally, a systematic review and meta-analysis by de Oliveira et al (2025) demonstrated greater improvements in PSQI scores and insomnia severity with taVNS than with sham stimulation (mean difference = −3.60; 95% confidence interval: −4.98 to −2.22; p < 0.01).20 Nevertheless, these findings should be interpreted cautiously given the substantial risk of unblinding in most taVNS trials. Active auricular stimulation frequently yields perceivable somatic sensations that may not be adequately replicated by sham stimulation, which increases the risk of performance and detection bias as well as the potential exaggeration of treatment effects. Consistent with these methodological concerns, the Grading of Recommendations Assessment, Development, and Evaluation assessment in the same meta-analysis rated the overall quality of evidence as low to very low because of the risk of bias and imprecision.20

Clinical Application of VNS in Central Hypersomnia

Early observations in patients with epilepsy suggested that VNS might influence alertness.21–23 Subsequent studies have directly evaluated VNS in narcolepsy. A proof-of-concept study by Winter et al (2024) investigated iVNS in patients with narcolepsy who had already received iVNS for epilepsy or depression.24 The study reported a significant reduction in Epworth Sleepiness Scale (ESS)25 scores (15.9 to 9.6 at 6 months, p < 0.001) after 6 months of iVNS treatment.24 However, the small sample size (n=18) and the lack of independent validation warrant cautious interpretation of these findings. Pan et al (2025) conducted the first RCT of taVNS in narcolepsy type 1. Active taVNS significantly reduced ESS scores from 18 to 15 versus 18 to 16 in the sham group (p < 0.05), although both remained pathological (>10).26 Objective measures, however, showed significant improvements in Maintenance of Wakefulness Test (MWT) sleep latency (between-group difference 3.09 min, p = 0.0041) and reduced sleep-onset REM periods (SOREMPs).26

These findings provide preliminary evidence that taVNS may enhance objective wakefulness in narcolepsy, though the clinical relevance of the subjective ESS improvement remains uncertain and requires further validation; cautious interpretation is warranted.

In summary, direct clinical evidence for VNS in central hypersomnia derives primarily from recent studies in narcolepsy. Collectively, these findings suggest that VNS may represent a promising neuromodulatory approach for narcolepsy, but they remain preliminary. Adequately powered, multicenter confirmatory trials are needed to establish its clinical efficacy.

Clinical Application of VNS in Sleep-Related Breathing Disorders (SRBD)

The impact of VNS on SRBD is predominantly adverse. Substantial evidence indicates that iVNS can induce or exacerbate OSA in patients with epilepsy, which is primarily attributed to VNS-induced laryngeal dysfunction. Zambrelli et al (2016) provided key mechanistic evidence demonstrating that iVNS produces persistent, stimulation-locked vocal cord adduction during sleep, which mechanically compromises the upper airway and offers a direct explanation for the high prevalence of OSA in this population.27

This causal relationship has been consistently supported by clinical studies. An early pilot study by Malow et al (2000) reported that iVNS stimulation periods were associated with significant increases in the apnea–hypopnea index (AHI) and oxygen desaturation events.28 These findings are consistent with a recent systematic review by Seth et al (2024), which confirmed that iVNS significantly increases the AHI and prevalence of OSA in patients with drug-resistant epilepsy.29 Clinically, the manifestation can be substantial, as illustrated by Gigandet and Radtke (2024), who reported that iVNS may act as a primary contributor to clinically significant OSA in susceptible individuals.30

Several management strategies have been proposed to mitigate these adverse respiratory effects. The most straightforward approach involves the adjustment of stimulation parameters. Ebben et al (2008) reported a case in which the reduction of stimulation frequency and pulse width during iVNS therapy led to a decrease in AHI and facilitated successful continuous positive airway pressure titration.31 A more advanced strategy involves the use of automated or cyclic stimulation modes. A case report on bi-level VNS therapy demonstrated that a dedicated low-intensity nocturnal stimulation mode could effectively maintain seizure control while minimizing respiratory compromise.32 This approach was further supported by Kim et al (2022), who reported that although iVNS may increase the risk of SRBD, appropriate parameter optimization may improve daytime sleepiness and overall sleep quality, potentially secondary to improved seizure control.33

In conclusion, the management of iVNS-induced SRBD requires a proactive strategy comprising pre-implantation screening, individualized nocturnal stimulation parameters, and regular respiratory monitoring. Future studies are warranted to define the safety profile of non-invasive stimulation and standardize protocol adjustments to optimize the risk–benefit balance in affected patients.

Clinical Application of VNS in Sleep-Related Movement Disorders (SRMD)

Research on VNS for restless legs syndrome (RLS) is limited, but interest persists because of the need for alternative therapies for treatment-resistant cases. However, clinical evidence for VNS in RLS remains limited, of low quality, and conflicting. Merkl et al (2007) described an incidental finding in a patient with comorbid major depression and RLS, in whom iVNS for depression led to complete resolution of RLS symptoms, reported by a marked reduction in the periodic limb movement index (PLMI).34 More recently, Hartley et al (2023) conducted a prospective open-label, non-randomised, single-center pilot study of taVNS in 26 patients with severe pharmacoresistant RLS, reported significant reductions in International RLS Rating Scale (IRLS) scores and PLMI, with 50% of participants classified as responders.35 A 6-month follow-up of 15 patients suggested sustained benefits.36 Subsequently, a pilot RCT by Veiz (2023) in 39 RLS patients did not find significant improvements in RLS symptom severity following taVNS.37 This negative finding may be partially attributable to methodological differences from prior positive studies: Veiz used a 30 Hz stimulation frequency at the tragus, whereas Hartley et al applied a lower frequency (2 Hz) at the cymba conchae.

In summary, while early case reports and open-label studies have raised the hypothesis that VNS may influence RLS symptoms, the current evidence remains limited and conflicting, and is insufficient to confirm its efficacy. The discrepancies across studies may be largely attributable to methodological heterogeneity, particularly differences in stimulation parameters and anatomical targets. Future research should systematically compare different stimulation parameters and stimulation site—to identify optimal protocols and determine whether VNS has a genuine therapeutic effect in RLS.

A summary of key clinical studies is provided in Supplementary Table 1.

Proposed Mechanisms of VNS: Current Evidence and Hypotheses

Before describing specific mechanisms, it is important to acknowledge the nature and limitations of the evidence presented in this section. Unless otherwise specified, most mechanistic data derive from preclinical studies or human studies on non-sleep disorders (eg, epilepsy, migraine, and major depressive disorder). Although these findings generate potential hypotheses regarding VNS modulation of sleep–wake circuitry, their direct applicability to sleep disorders in humans remains uncertain. Accordingly, the mechanisms described below should be interpreted as a conceptual framework for future research rather than as established explanations for the clinical efficacy of VNS in sleep disorders. Where appropriate, the strength of evidence is highlighted and existing gaps are identified.

Anatomical and Functional Basis of VNS

The VN, the 10^th cranial nerve, is a mixed nerve comprising approximately 80% afferent fibers and serves as a major communication pathway between peripheral organs and the central nervous system.38 The only cutaneous branch, the ABVN, is distributed within the auricular concha and represents a unique superficial target for stimulation.39 Afferent vagal signals, whether elicited by non-invasive taVNS or iVNS, project to the nucleus tractus solitarius (NTS) in the brainstem, which constitutes the primary central relay for the effects of the VN.38,40 As a major integration hub for visceral sensory information and autonomic regulation, the NTS subsequently disseminates afferent signals to a distributed network of key brain regions, including the parabrachial nucleus (PBN), locus coeruleus (LC), dorsal raphe nucleus (DRN), hypothalamus, thalamus, hippocampus, amygdala, BF, and cerebral cortex.11,40–43 These interconnected structures collectively regulate sleep, arousal, emotion, and inflammatory processes.40,44 The VN–NTS–central network pathway provides the broad neuromodulatory and physiological effects of VNS, may contribute to the modulation of the sleep–wake cycle, higher-order cognitive functions, and inflammatory signaling.5 This anatomical framework may partly explain the wide therapeutic spectrum of VNS and provides a structural basis for its potential state-dependent effects on sleep and wake regulation, as further elaborated in the subsequent sections.

The stimulation target and device configuration for taVNS are illustrated in Figure 1.

Figure 1.

Transcutaneous auricular vagus nerve stimulation (taVNS) device and its application. (A) A patient wearing the device on the left ear during self administered treatment. The electrode is placed on the cymba conchae, the region innervated by the auricular branch of the vagus nerve. (B) Photograph of the taVNS device.

Theoretical Basis: VNS and Sleep–Wake Neural Pathways

By activating the NTS, taVNS may exert effects on downstream neural circuits that are relevant to sleep–wake regulation. The following sections summarize the major mechanistic hypotheses that have been proposed to underlie these effects.

Promotion and Consolidation of Sleep Pathways

In the context of sleep promotion, VNS may act through multiple mechanisms, including the enhancement of cortical inhibitory tone and modulation of sleep-related oscillatory activity. At the neurochemical level, direct electrophysiological evidence in humans indicates that taVNS rapidly enhances cortical inhibition mediated by gamma-aminobutyric acid (GABA) via both GABA_A and GABA_B receptors.45,46 Beyond these established findings, the broader neuroanatomical architecture of sleep regulation provides additional, albeit largely untested, targets for VNS-mediated modulation. Key sleep-regulatory nuclei, including the ventrolateral preoptic nucleus (VLPO)—a cluster of GABAergic neurons functioning as a principal “sleep switch”, and the hypothalamic orexin system, which stabilizes wakefulness, have been proposed as plausible downstream effectors of VNS.47–50 However, no direct evidence currently links VNS to the modulation of these specific nuclei in the context of sleep. It remains unclear whether VNS influences these pathways by modulating VLPO excitability, attenuating orexinergic drive, or altering histaminergic tone. These potential interactions represent important directions for future hypothesis-driven research.A similar gap exists regarding REM sleep regulation. The sublaterodorsal tegmental nucleus (SLD) is recognized as a central hub for REM sleep generation,51 and VNS has been shown to modulate amygdala activity in animal models.52 However, whether these observations translate to changes in REM sleep architecture in sleep disorder populations remains unknown.

Potential Modulation of Arousal Pathways

In addition to its potential sleep-promoting effects, taVNS engages a distributed network of brainstem and forebrain arousal centers, as suggested by neuroimaging studies.53 This wake-promoting influence is thought to be initiated at the brainstem level, where signals relayed from the NTS may activate key arousal nuclei. The parabrachial nucleus (PBN), a major glutamatergic hub, contributes to wakefulness through projections to the BF and hypothalamus,41,53 and integrates autonomic signals that may modulate sleep architecture in response to internal physiological states.54,55 TaVNS may influence noradrenergic arousal systems.56,57 Long-term VNS increases both 5-HT and dopamine (DA) levels in key forebrain regions, including the mPFC and nucleus accumbens (NAc).58 In patients with primary insomnia, an RCT combined with fMRI reported that taVNS was associated with reduced functional connectivity between the BF and several cortical regions, including the visual, sensorimotor, and medial prefrontal cortices.11 Preclinical studies have shown that VNS can modulate activity in brainstem arousal nuclei, including the LC and DRN, and influence monoaminergic neurotransmitter systems in animal models.43,58 However, the direct relevance of these findings to sleep disorders in humans remains to be established. Taken together, while fMRI evidence in insomnia patients suggests that taVNS modulates BF‑cortical connectivity, the broader arousal network model remains hypothetical and requires direct validation in sleep disorder populations. The key brain structures implicated in the sleep–wake regulatory network are summarized in Figure 2.

Figure 2.

Diagram of key brain structures modulated by VNS in sleep disorders.

Abbreviations: Ach, acetylcholine; Amy, amygdala; BF, basal forebrain; DRN, dorsal raphe nucleus; Glu, glutamate; Hipp, hippocampus; HTh, hypothalamus; LC, locus coeruleus; mPFC, medial prefrontal cortex; NAc, nucleus accumbens; NE, norepinephrine; NTS, nucleus tractus solitarius; PBN, parabrachial nucleus; SMA, supplementary motor area; Th, thalamus; 5-HT, serotonin.

Immunomodulation via Anti-Inflammatory Pathways

Sleep disorders and neuroinflammation are closely interconnected in a bidirectional relationship. Chronic sleep deprivation or disrupted sleep architecture can activate microglia, leading to increased production of pro-inflammatory cytokines such as tumor necrosis factor-alpha (TNF-α) and interleukin-1 beta (IL-1β), which subsequently disrupt sleep homeostasis.59 Preclinical models indicate that VNS attenuates inflammatory responses, which may contribute to improvements in sleep quality. The cholinergic anti-inflammatory pathway represents the most well-characterized anti-inflammatory pathway. In a landmark study, Borovikova et al demonstrated that acetylcholine (ACh), the principal vagal neurotransmitter, significantly suppresses the release of pro-inflammatory cytokines, including TNF, IL-1β, IL-6, and IL-18, from macrophages, while sparing the anti-inflammatory cytokine IL-10.60 Furthermore, direct in vivo VNS can attenuate systemic TNF levels and prevent septic shock, establishing the functional basis of the pathway.60 At the molecular level, many of these effects are mediated via the α7 nicotinic acetylcholine receptor (α7nAChR). Within the central nervous system, VNS modulates microglial activation via α7nAChR, which attenuates lipopolysaccharide (LPS)-induced TNF-α release61 and inhibits activation of the NLRP3 inflammasome, thus exerting potent neuroprotective effects.62 In the periphery, α7nAChR activation suppresses the release of key inflammatory mediators, including TNF-α and high-mobility group box 1 (HMGB1), from macrophages and other immune cells.63,64 However, the precise downstream signaling mechanisms remain unclear. The classical VNS–splenic nerve model of anti-inflammatory signaling is increasingly being refined. Emerging evidence points to a more complex circuitry involving sympathetic nerves and T-cell-mediated interactions within the spleen.65,66 Consistent with this revised framework, Komegae et al demonstrated that selective activation of abdominal vagal afferent fibers suppresses systemic inflammation via a splanchnic sympathetic nerve-dependent pathway, defining a distinct anti-inflammatory mechanism.67 At the intracellular signaling level, VNS-induced anti-inflammatory effects have been associated with the activation of the Janus kinase 2–signal transducer and activator of transcription 3 (Jak2–STAT3) pathway.68 Emerging evidence defines a VN-oxytocin (OT) anti-inflammatory axis, in which vagal afferent signals activate oxytocinergic neurons in the paraventricular nucleus; additionally, OT release subsequently suppresses microglial activation and neuroinflammation.68 However, this pathway remains hypothetical and requires direct in vivo validation. Beyond direct immunomodulation, taVNS may also attenuate the hypothalamic–pituitary–adrenal (HPA) axis through the “ABVN–NTS” pathway, leading to reduced corticotropin-releasing hormone and cortisol secretion, and therefore potentially alleviating stress-induced arousal.59,69

Inhibition of Cortical Spreading Depression (CSD)

The Vicious Cycle Between Sleep Disorders and CSD

Cortical spreading depression (CSD) is characterized by a slowly propagating wave of neuronal and glial depolarization across the cerebral cortex.70 Sleep disorders, including insomnia, significantly lower the threshold for CSD induction, which increases cortical susceptibility to such events.71 Conversely, CSD can disrupt sleep architecture. Collectively, these observations suggest a bidirectional relationship wherein sleep disturbance enhances CSD vulnerability, which in turn may further exacerbate sleep disruption, establishing a self-reinforcing vicious cycle.72,73

Potential Mechanisms of VNS in CSD Inhibition

VNS may interrupt this cycle through direct suppression of CSD and indirect improvement of the neurophysiological milieu underlying sleep regulation. The direct inhibitory effects of VNS on CSD are primarily mediated by two distinct pathways. Classical Central Pathway: VNS may inhibit CSD via the VN–NTS–LC/DRN axis, ultimately modulating cortical NE and 5-HT levels. These neuromodulators act synergistically in the cortex, significantly elevating the CSD induction threshold while reducing its propagation velocity and spatial extent. Morais et al demonstrated that the inhibitory effects of VNS on CSD depend on a purely central pathway.74 Glu–Tropomyosin Receptor Kinase B (TrkB)–Potassium Chloride Cotransporter 2 (KCC2) Pathway: Liu et al identified an additional molecular mechanism underlying VNS-mediated inhibition of CSD, wherein the activation of NTS Glu neurons promotes TrkB phosphorylation. This enhances the function of KCC2, which reinforces cortical GABAergic inhibition and raises the CSD threshold. This Glu–TrkB–KCC2 signaling cascade represents a non-classical pathway that expands the mechanistic understanding of VNS and provides novel therapeutic targets.75 This pathway has also been associated with the attenuation of CSD-induced neuroinflammation.75

VNS in Enhancing Neuroplasticity

Neuroplasticity is a central process in functional recovery following brain injury. Preclinical studies suggest that VNS can modulate plasticity-related mechanisms. A key pathway involves the activation of a signaling network centered on brain-derived neurotrophic factor (BDNF) and its high-affinity receptor, TrkB. VNS has been shown to induce rapid TrkB phosphorylation, which in turn initiates downstream intracellular cascades such as phospholipase Cγ/protein kinase C (PLCγ/PKC) and rat sarcoma (Ras)/mitogen-activated protein kinase, directly regulating neuronal survival, growth, and synaptic remodeling.76 At the functional level, these molecular processes are associated with enhanced long-term potentiation, a fundamental mechanism underlying learning, memory, and synaptic efficacy in animal models. In ischemic stroke models, VNS activates BDNF signaling in an α7nAChR-dependent manner, promoting axonal sprouting and improves neurological recovery.77 In addition, VNS increases the expression of multiple neurotrophic factors, including basic fibroblast growth factor, in key brain regions such as the hippocampus and cortex, contributing to a neurotrophic environment supportive of neuronal survival and repair.78 These preclinical findings suggest a potential role for VNS in enhancing neuroplasticity through multitarget neurotrophic support. However, direct evidence supporting these mechanisms in sleep disorder populations remains limited.

Modulation of Cerebral Blood Flow (CBF) by VNS

CBF modulation has been proposed as a potential mechanism by which VNS may exert therapeutic effects in certain sleep disorders; however, direct evidence in sleep disorder populations remains limited. Several sleep disorders, including idiopathic REM sleep behavior disorder and OSAS, are associated with CBF dysregulation.79–81 VNS promotes cerebral vasodilation and increases global CBF through suppression of sympathetic activity and enhancement of parasympathetic tone, based on findings primarily derived from studies in patients with epilepsy.82–84 However, current evidence regarding the effects of VNS on CBF is largely confined to epilepsy and depression populations, with comparatively limited research on sleep disorders. Notably, the effects of VNS on CBF do not appear to represent a simple tonic increase in cerebral perfusion. Although VNS may not significantly alter resting CBF, it can augment task-evoked CBF responses in a dose-dependent manner when paired with cognitive activity.84 These findings suggest that VNS may enhance neurovascular coupling efficiency in healthy individuals and in patients with epilepsy. Nevertheless, whether similar effects contribute to improvements in sleep architecture or sleep-related neurophysiology remains unclear. Overall, CBF modulation represents a plausible mechanistic pathway through which VNS could potentially improve multiple sleep disorders. Further studies are warranted to characterize disorder-specific neurovascular responses and determine their relevance to clinical sleep outcomes.

Preservation of Blood–Brain Barrier (BBB) Integrity

The blood-brain barrier (BBB), a core component of the neurovascular unit, is crucially involved in maintaining cerebral homeostasis and regulating sleep–wake cycles.85 In animal models of cerebral ischemia and traumatic brain injury, VNS has been shown to preserve BBB integrity. These effects are mediated by the cholinergic anti-inflammatory pathway and involve the reduction of BBB permeability, suppression of matrix metalloproteinase-2/9 expression, and attenuation of reactive astrogliosis.86 In addition, preclinical studies suggest that VNS may enhance glymphatic clearance of metabolic waste by promoting cerebrospinal fluid and interstitial fluid exchange,87 which is essential for brain restoration during sleep. Collectively, these findings suggest that VNS may exert beneficial effects in sleep disorders through the stabilization of BBB function and enhancement of glymphatic activity. However, direct validation in sleep disorder models and clinical populations remains necessary.

Conclusions

Current evidence positions VNS as a promising yet experimental approach for sleep disorders. Transcutaneous VNS (taVNS) has demonstrated sustained improvements in chronic insomnia for up to 20 weeks, although the evidence quality remains low due to methodological limitations, particularly unblinding. Preliminary findings for central hypersomnia and restless legs syndrome are encouraging but limited, requiring independent validation in larger cohorts. Notably, invasive VNS carries a documented risk of worsening obstructive sleep apnea, underscoring modality-specific safety concerns. While preclinical data suggest potential neuromodulatory mechanisms, direct evidence linking these pathways to clinical sleep outcomes is lacking. Collectively, rigorous, large-scale, and sham-controlled trials are urgently needed to establish definitive efficacy and safety profiles.

Challenges and Future Perspectives

Despite the promising therapeutic potential of VNS for sleep disorders, several challenges remain before widespread clinical implementation can be achieved. First, the overall quality of evidence is limited. For taVNS in insomnia, the evidence is rated as low to very low by GRADE assessment, primarily due to methodological limitations and high risk of unblinding.20 For central hypersomnia, only one small RCT in narcolepsy type 1 is currently available, providing only preliminary evidence.26 In restless legs syndrome, the evidence is even more conflicting—one open-label study reported positive results, while a subsequent pilot RCT found no significant improvement.36,37 Second, clinical responses to VNS exhibit substantial interindividual variability, and reliable predictive biomarkers are lacking.11 Third, standardized stimulation protocols have not been established. Fourth, the inherent difficulty of maintaining effective blinding in device-based trials complicates the interpretation of subjective outcomes, as active taVNS often produces noticeable sensations that sham procedures cannot fully replicate.20 Fifth, safety profiles differ across modalities: iVNS carries a well-documented risk of inducing or exacerbating OSA,29 whereas current research on VNS in sleep disorders has largely shifted to non-invasive transcutaneous approaches, particularly taVNS. taVNS appears safe in the short term, although long-term safety data are lacking.88

To translate the potential of VNS into clinical practice, future research should address these challenges through several priority directions. First, large-scale, rigorously designed, multicenter RCTs with extended follow-up are urgently needed. These trials should incorporate more effective sham controls to minimize placebo effects and maintain blinding integrity. Systematic optimization of stimulation parameters is also required to establish standardized protocols for each indication. In particular, dose–response designs comparing multiple active frequencies (eg, 1–2 Hz as a low-dose control, 6–10 Hz as an intermediate dose, and 20–25 Hz as a high dose) would help establish a true frequency–response relationship. Such multi-arm designs not only strengthen blinding by ensuring comparable somatosensory perceptions across groups but also help distinguish non-specific placebo effects from genuine frequency-dependent neuromodulation. Second, the development of validated predictive biomarkers would enable personalized patient selection and treatment stratification. Third, intelligent closed-loop neuromodulation systems that integrate real-time physiological biomarkers to deliver adaptive stimulation represent a particularly promising direction. These systems may be further refined through respiratory-gated approaches such as RAVANS.89–91 Fourth, a deeper mechanistic understanding—combining neuroimaging in humans and cell-specific circuit interrogation in animal models—is essential. One intriguing but unproven hypothesis is that VNS may exert state-dependent effects on sleep and wakefulness, based on preclinical observations that VNS effects vary with baseline brain state.43,92 This hypothesis warrants direct testing in well-controlled trials using objective physiological measures (eg, PSG, MWT). Finally, exploring the combination of taVNS with behavioral (eg, CBT-I) and pharmacological (eg, orexin receptor-targeting drugs) interventions may offer broader therapeutic benefits than single-modality approaches. Future trials should consider systematic OSA screening to minimize confounding. For taVNS studies, we propose a stepwise approach using validated questionnaires, followed by HSAT or PSG when indicated. For iVNS, routine pre-implantation PSG screening remains standard practice.

Funding Statement

The authors disclosed receipt of the following financial support for the research, authorship, and publication of this article: Brain Science and Brain-like Intelligence Technology -National Science and Technology Major Project under Grant 2021ZD0204303.

Abbreviations

ABVN, auricular branch of the vagus nerve; ACh, acetylcholine; AHI, apnea–hypopnea index; BF, basal forebrain; CBT I, cognitive behavioral therapy for insomnia; CBF, cerebral blood flow; CSD, cortical spreading depression; DA, dopamine; DMN, default mode network; DRN, dorsal raphe nucleus; EDS, excessive daytime sleepiness; EEG, electroencephalogram; ESS, Epworth Sleepiness Scale; fMRI, functional magnetic resonance imaging; GABA, gamma aminobutyric acid; Glu, glutamate; HMGB1, high mobility group box 1; HPA, hypothalamic–pituitary–adrenal; IL 1β, interleukin 1 beta; IL 6, interleukin 6; iVNS, invasive vagus nerve stimulation; Jak2–STAT3, Janus kinase 2 – signal transducer and activator of transcription 3; LC, locus coeruleus; LPS, lipopolysaccharide; mPFC, medial prefrontal cortex; MWT, Maintenance of Wakefulness Test; NAc, nucleus accumbens; NE, norepinephrine; NLRP3, NLR family pyrin domain containing 3; NREM, non‑rapid eye movement; nVNS, non invasive vagus nerve stimulation; NTS, nucleus tractus solitarius; OSA, obstructive sleep apnea; OT, oxytocin; PBN, parabrachial nucleus; PLCγ/PKC, phospholipase Cγ/protein kinase C; PLMI, periodic limb movement index; PSG, polysomnography; PSQI, Pittsburgh Sleep Quality Index; PTSD, post‑traumatic stress disorder; PVT, paraventricular thalamus; RAVANS, respiratory‑gated auricular vagal afferent nerve stimulation; RCT, randomized controlled trial; REM, rapid eye movement; RLS, restless legs syndrome; SAS, Self Rating Anxiety Scale; 5 HT, serotonin; SLD, sublaterodorsal tegmental nucleus; SOREMP, sleep onset REM period; SRBD, sleep‑related breathing disorder; SRMD, sleep related movement disorder; taVNS, transcutaneous auricular vagus nerve stimulation; TBI, traumatic brain injury; TMN, tuberomammillary nucleus; TNF‑α, tumor necrosis factor alpha; TrkB, tropomyosin receptor kinase B; TST, total sleep time; VLPO, ventrolateral preoptic nucleus; VN, vagus nerve; VNS, vagus nerve stimulation; WASO, wake after sleep onset; α7nAChR, α7 nicotinic acetylcholine receptor.

Data Sharing Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Ethical Approval and Informed Consent

Not applicable. This is a review article and does not involve original research with human participants or animals.

Author Contributions

Xinyi Wang: Writing – original draft, Data curation. Qingqing Sun: Data curation, Investigation, Writing – review & editing. Lijia Cai: Data curation, Investigation, Writing – review & editing. Yaru Wang: Writing – review & editing, Visualization. Yanan Zhang: Writing – review & editing, Visualization. Zan Wang: Conceptualization, Supervision, Project administration, Funding acquisition, Writing – review & editing.

All authors gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

Prof. Dr. Zan Wang reports support for the study from Brain Science and Brain-like Intelligence Technology -National Science and Technology Major Project.The authors declared no further conflicts of interest with respect to the research, authorship, and/or publication of this article.