Abstract

Transcutaneous auricular vagus nerve stimulation (taVNS) is being investigated as a non-invasive neuromodulatory approach for cognitive impairment (CI). This review evaluates the existing preclinical and clinical evidence regarding its potential efficacy and mechanisms of action in conditions such as mild cognitive impairment, post-stroke cognitive impairment, and other forms of CI. Preclinical models suggest that taVNS may influence multiple pathways, including neurotransmitter regulation, neuroinflammation, apoptosis, and synaptic plasticity. Clinically, some small-scale studies report modest improvements in cognitive metrics, but the evidence remains preliminary due to methodological limitations such as small sample sizes, heterogeneous parameters, and short intervention durations. Significant challenges, including the lack of standardized protocols, inadequate sham controls, and an underdeveloped mechanistic understanding, currently hinder the interpretation and translation of findings. Future research necessitates large-scale, rigorously controlled trials and deeper mechanistic studies to determine whether taVNS has a definitive role in the clinical management of cognitive impairment.

1 Introduction

Cognitive impairment (CI) is a common clinical condition characterized by significant decline in one or more cognitive domains, including memory, language, attention, executive function, and orientation (1). This impairment is a core clinical feature of various diseases. In neurodegenerative disorders such as Alzheimer’s disease (AD), CI often serves as a primary symptom and tends to progressively worsen (2). Following cerebrovascular events, it is also a common and severely debilitating complication known as post-stroke cognitive impairment (PSCI) (3). Furthermore, CI is prevalent among patients with psychiatric disorders and is closely associated with various chronic systemic diseases (4). Even during physiological aging, many middle-aged and elderly individuals experience subjective cognitive decline or mild objective cognitive deterioration, which may serve as precursors to more severe CI.

From an epidemiological perspective, the disease burden caused by CI is extremely heavy and continues to intensify with the global aging population (5). Its prevalence increases significantly with age: among those aged 65 and older, the prevalence of dementia is approximately 5–10%, rising to 20–30% or higher in individuals aged 80 and above (6). CI and related dementias not only severely impact patients’ ability to live independently and their quality of life but also impose heavy psychological, physical, and financial burdens on caregivers (7). Taking AD as an example, the global patient population has reached tens of millions, with incidence rates continuing to rise (8). PSCI also occurs in 20–50% of stroke patients (9). Given the high prevalence and widespread nature of CI across diseases, existing pharmacological treatments remain limited in efficacy and associated with significant adverse effects (10). Consequently, developing novel, safe, and effective intervention strategies has become an urgent priority in neuroscience and clinical medicine.

The pathogenesis of CI exhibits high complexity, involving the interplay of multi-system, multi-level pathophysiological processes (11). Its core mechanisms can be summarized into the following four interrelated aspects: (1) Progressive damage to neural structures and functions, primarily manifested as pathological accumulation of abnormal proteins, neuronal degeneration and death, loss of synaptic connections, and impaired cerebral vascular integrity (12, 13); (2) Cerebrovascular dysfunction, where cerebral atherosclerosis, localized infarcts, or microbleeds cause insufficient cerebral perfusion and hypoxic–ischemic injury, subsequently damaging critical brain regions such as the hippocampus and prefrontal cortex that are closely associated with learning and memory (9); (3) Neurochemical imbalance, most notably characterized by significant acetylcholine deficiency within the cholinergic system, alongside dysfunction in glutamatergic excitotoxicity and neurotransmitter systems involving dopamine and serotonin, collectively impairing efficient neural signaling and integration (14, 15); (4) The cumulative effects of systemic risk factors, including age-related declines in cellular repair capacity, carriage of genetic susceptibility genes, persistent neuroinflammatory responses, and oxidative stress-induced damage to biomolecules. These factors mutually exacerbate each other, collectively forming the deep biological basis for progressive cognitive decline (16–18).

Currently, drug intervention remains the primary treatment approach for CI. Among pharmacological therapies, symptom-improving medications such as donepezil and causative-targeted drugs like lecanemab—which has garnered significant attention in recent years—constitute the current treatment regimen (19). However, the former typically only alleviate symptoms without halting disease progression and are often accompanied by certain side effects (20). The latter face core limitations including high cost, potential safety risks such as cerebral edema, and a narrow therapeutic window—being applicable only to early-stage patients (21). Compounding this, clinical diagnosis is often delayed, meaning most patients are diagnosed after missing the optimal intervention window (22). Overall, existing medications are largely confined to symptom management, with significant interindividual variability in efficacy (23). Therefore, exploring safe, effective, and accessible non-pharmacological interventions holds substantial clinical significance and practical value for improving cognitive and emotional functioning.

2 From concept to intervention: the realization of transcutaneous auricular vagus nerve stimulation

The vagus nerve, as the tenth cranial nerve, is the longest and most widely distributed mixed nerve in the human body, characterized by complex anatomy and diverse functions (24). It originates from multiple nuclei within the medulla oblongata, including the ambiguus nucleus, dorsal vagal nucleus, solitary tract nucleus, and spinal trigeminal nucleus. Its fiber composition encompasses four distinct types: special visceral motor, general visceral motor, general visceral sensation, and general somatic sensation. After exiting the cranium through the foramen magnum, it descends within the carotid sheath, with its branches extensively innervating organs in multiple regions including the head, neck, chest, and abdomen (25, 26). The main branches include: (1) the auricular branch, which innervates the skin of the external auditory canal (Figure 1); (2) the pharyngeal branch and superior laryngeal nerve/recurrent laryngeal nerve, which innervate the muscles of the pharynx and larynx; (3) the visceral branches, which extend into the thoracic and abdominal cavities to regulate cardiac, pulmonary, esophageal, and gastrointestinal functions. Functionally, approximately 80% of vagus nerve fibers are afferent, responsible for transmitting diverse sensory information (27, 28). Among these, common somatic afferent fibers project sensory information from the external auditory canal to the trigeminal spinal nucleus, while common visceral afferent fibers collect information from thoracic and abdominal organs and the aortic body, projecting to the caudate nucleus of the solitary tract. Additionally, special visceral afferent fibers originating from the epiglottis are responsible for taste transmission, with their information ultimately projecting to the nucleus tract solitary (NTS) (29). Crucially, the NTS serves as the primary afferent relay station for the vagus nerve. Beyond receiving these signals, it further projects to brainstem nuclei such as the locus coeruleus and parabrachial nucleus. Through extensive connections with higher centers including the thalamus, amygdala, hippocampus, and prefrontal cortex, it participates in regulating diverse physiological and pathophysiological processes, including visceral reflexes, inflammatory responses, emotional cognition, cardiovascular homeostasis, and seizure control (30, 31).

The figure illustrates the distribution of the vagus nerve in the ear. The left diagram shows the names of various parts of the ear; the yellow areas in the right diagram indicate the distribution of the vagus nerve.

Since receiving United States Drug and Food Administration approval in 2005 for treatment-resistant depression, vagus nerve stimulation (VNS) has gradually expanded its indications into the field of CI (32). Accumulating evidence suggests that VNS significantly improves vascular cognitive impairment (33), traumatic brain injury-related cognitive deficits (34), and even AD (35). Despite its promising prospects, the surgical risks, potential infections, and high costs associated with traditional invasive implant-based VNS constitute major barriers to its clinical adoption, driving an urgent need for non-invasive VNS technologies (36).

Based on the unique anatomical feature of the auricular vagus nerve as the sole superficial vagus nerve branch in the human body (37), transcutaneous auricular vagus nerve stimulation (taVNS) serves as an innovative non-invasive neuromodulation technique. By precisely stimulating the auricular vagus nerve branch, it delivers therapeutic effects while successfully overcoming the invasive limitations of traditional VNS. The development of taVNS integrates the wisdom of Eastern and Western medical philosophies. Its conceptual foundation traces back to traditional Chinese medicine’s auricular therapy, practiced for millennia (38, 39). Traditional Chinese medical theory views the auricle as a holographic projection of an “inverted fetus,” connected to all internal organs via meridians. Stimulating specific auricular points is believed to regulate corresponding organ functions (40). This understanding provided the initial theoretical and practical basis for auricular therapy. By the 20th century, modern neuroanatomy achieved a pivotal breakthrough: research confirmed dense innervation of the concha and antrum regions of the human auricle by the auricular branch of the vagus nerve (41). This discovery transformed the traditional “meridian” connection into a defined “nerve” pathway, providing scientific justification for activating the vagus nerve through auricular stimulation. Entering the 21st century, taVNS transitioned from theory to practice. Around 2000, researchers formally introduced the concept of “transcutaneous auricular vagus nerve stimulation,” systematically applying transcutaneous electrical nerve stimulation devices to deliver precise electrical stimulation to the concha region as an alternative to implantable cervical stimulation (42). Subsequent studies using functional magnetic resonance imaging (fMRI) and other techniques confirmed that taVNS effectively activates brainstem nuclei such as the solitary tract nucleus and higher-order brain regions, regulating the autonomic nervous system and inflammatory responses (43–45). As its mechanisms became progressively elucidated, clinical applications expanded from initial indications like epilepsy and depression to encompass migraine, CI, and systemic lupus erythematosus (18, 45–47).

Existing clinical studies indicate that taVNS significantly improves symptoms such as CI, demonstrating efficacy comparable to VNS (48). It offers advantages including non-invasiveness, safety, and ease of operation, showcasing broad clinical potential. However, current evidence primarily stems from small-sample, short-term trials, with a lack of high-quality research systematically validating the long-term efficacy and neural mechanisms of taVNS for treating CI (49, 50). This review aims to synthesize existing clinical and experimental evidence to comprehensively evaluate the efficacy of taVNS in treating cognitive impairment, explore its mechanisms of action, and provide theoretical foundations and practical guidance for advancing the clinical translation of this technology. Specifically, given that current clinical trials primarily focus on functional outcomes and often lack direct evidence of neural changes, we will also review pertinent preclinical animal literature in Section 5. Specifically, while cognitive impairment is also a prevalent feature in psychiatric disorders such as depression, this review primarily focuses on cognitive decline stemming from neurodegenerative, vascular, and systemic organic pathologies. In these contexts, cognitive dysfunction serves as a primary target for intervention rather than a secondary outcome of mood dysregulation.

3 Methods3.1 Search strategy

A systematic literature search was conducted across four major electronic databases—PubMed, Embase, Web of Science, and Cochrane Library—from their inception until November 10, 2025. The aim was to identify clinical studies investigating the use of taVNS in populations with CI. The search strategy combined Medical Subject Headings (MeSH) with free-text keywords, focusing on terms and their variations such as “transcutaneous auricular vagus nerve stimulation,” “taVNS,” “transcutaneous vagus nerve stimulation,” “tVNS,” “auricular transcutaneous vagus nerve stimulation,” “atVNS,” “cognitive impairment,” “mild cognitive impairment,” “post-stroke cognitive impairment,” “vascular cognitive impairment,” “Alzheimer’s disease,” “dementia.” Study types included randomized controlled trials (RCTs), single-arm trials, case reports, and non-randomized controlled trials.

3.2 Research screening process

The screening process was conducted in two phases. First, three reviewers (DP, JS, and WM) independently screened titles and abstracts of all retrieved records. If eligibility could not be determined from the title and abstract alone, the full text was obtained for further assessment. Discrepancies were resolved through discussion or by consulting a fourth reviewer (HS). To improve efficiency and accuracy, retrieved records were imported into EndNote software (Analytics Inc., Philadelphia, United States) for deduplication and initial screening. Eligible records were then transferred to Zotero 5.0 (Digital Scholar LLC, Vienna, United States) for full-text review and final inclusion confirmation. No restrictions were placed on publication date or language.

Inclusion criteria were as follows: (1) Original clinical research articles; (2) Studies involving participants with cognitive impairment associated with neurological, vascular, or systemic organic pathologies; (3) Intervention involving taVNS or transcutaneous vagus nerve stimulation (tVNS) applied to the auricular region; (4) Study designs including RCTs, single-arm trials, case reports, or non-randomized controlled trials; (5) Articles published in English.

Exclusion criteria included: (1) Non-original research (e.g., reviews, meta-analyses, editorials, conference abstracts); (2) Studies using non-electrical auricular interventions; (3) Studies utilizing transcutaneous cervical vagus nerve stimulation (tcVNS) or other non-auricular stimulation sites, to maintain anatomical homogeneity; (4) Lack of relevant cognitive or safety outcome data; (5) Inaccessible data after contacting authors; (6) Studies not focused on cognitive impairment; (7) Studies focusing primarily on CI secondary to psychiatric disorders (e.g., depression, schizophrenia) to maintain pathophysiological homogeneity.

3.3 Data extraction

Data extraction and quality assessment were performed independently by two reviewers (JS and DP). For studies with missing or unclear information, corresponding authors were contacted via email to obtain additional details. Extracted data included: (1) First author and publication year; (2) Study population characteristics; (3) Stimulation site; (4) Intervention protocol details; (5) Primary cognitive outcome measures. Extracted data were summarized in tabular form (Table 1) to facilitate comparison and synthesis (Figure 2).

ReferencesCharacteristics (n)Study
designtreatmentTreatment durationMain siteStimulation parametersAssessment Tools (Objective/Subjective)Key Outcomes (Pre vs. Post / Score Change)ConclusionsWang et al. (48)MCI (60)RCT (Sham-controlled)taVNS vs. Sham24 weeksConcha (taVNS) vs. scaphoid (sham)Mixed frequencies: 20 Hz for 10s & 100 Hz for 50s;
Intensity: 0.6–1.0 mAObj: MoCA-B, AVLT-H Subj: PSQIMoCA-B: Significant improvement in taVNS group vs. Sham (p < 0.05), AVLT-H: N5 and N7 scores significantly increased (p < 0.001)1. Significantly improved global cognition and memory in MCI
2. Intervention was safe with one minor adverse event;
3. Represents an effective non-pharmacological option for MCIMurphy et al. (63)MCI (50)RCT (Sham-controlled)tVNS vs. ShamAcute (During MRI scan)Active: Left tragus; Sham: Left ear lobeFrequency: 20 Hz; Pulse width: 50 μs; Intensity: set at 80% of individual discomfort threshold (Mean: 7.3 mA)Obj: Resting-state fMRIfMRI: Altered functional connectivity in semantic and salience networks (temporal/parietal lobes) and from hippocampi to cortical/subcortical clusters vs. sham.1. tVNS modifies activity in brain networks associated with AD.
2. Provides evidence of afferent target engagement in MCI patients.Chen et al. (52)PSCI (1)Single-arm pilottaVNS8 weeksRight earMixed frequencies: 20 Hz for 7 s / 4 Hz for 3 sObj: MoCA, STT, DTIMoCA: Score improved from 7 (baseline) to 11 (post-treatment), STT: Completion time decreased significantly1. Feasible and potentially improves global cognition and executive function in long-term PSCI
2. Induced white matter remodeling
3. Mood and sleep improvements observedWang et al. (54)CKD related cognitive impairment (36)RCT (Sham-controlled,)taVNS vs. Sham2 weeksLeft cymba conchae25 Hz;
Pulse width: 200 μs;
Intensity: max 10 mAObj: MoCA, Subj: VAS-FMoCA: taVNS prevented the intradialytic decline seen in Sham (p < 0.001), VAS-F: Fatigue significantly reduced (p = 0.004)Improved cognitive performance and reduced fatigue
2. Alleviated intradialytic decline in cerebral oxygenation
3. Improvement in cerebral rSO₂ partially mediated cognitive benefits
4. Promising non-pharmacological intervention for MHD patientsDolphin et al. (62)Amnestic MCI (40)RCT (Sham-controlled, crossover)tVNS vs. Sham vs. BaselineAcute (Single 60-min session)Auricular branch8 Hz; Individual amplitude (mean 2.5 mA)Obj: FNAT, Spatial navigation task, HemodynamicsFNAT: Recall accuracy significantly improved vs. baseline/sham (p = 0.01). Spatial navigation: Time significantly reduced vs. baseline/sham.1. tVNS is safe and tolerable in aMCI.
2. Acute tVNS improves spatial and associative memory.Guo et al. (51)MCI (1)Case reporttaVNS24 weeksBilateral conchaSparse/dense wave (20/100 Hz);
Intensity: 3–8 mAObj: MoCA-B, AVLT, fMRIMoCA-B: Score improved from 19 to 24, AVLT: Scores improved (e.g., N5 from 4 to 3, N7 from 10 to 19)1. Improved cognitive function and depressive symptoms
2. Increased neural activity in right temporal pole and left medial orbitofrontal gyrus
3. Benefits persisted for 24 weeks post-treatmentLi et al. (53)PSCI (30)Single-arm pilottaVNS + TOT3 weeksLeft cymba conchae25 Hz, 30s on/30s off
Pulse width: 500 μs; Intensity: 1–10 mAObj: MoCA, FMA-UE fNIRSMoCA: Significant improvement compared to baseline and Sham (p < 0.05), FMA-UE: Upper limb function improved (p < 0.05)1. Superior to sham + TOT in improving upper limb function and cognition
2. Enhanced autonomic homeostasis and corticospinal excitability
3. fNIRS revealed task-specific cortical activation
4. Promotes recovery via multi-pathway neuroregulation

Applications of taVNS in clinical cognitive disorders.

Transcutaneous Auricular Vagus Nerve Stimulation, taVNS; Task-Oriented Training, TOT; Chronic Kidney Disease, CKD; Post-Stroke Cognitive Impairment, PSCI; Mild Cognitive Impairment, MCI; Functional Near-Infrared Spectroscopy, fNIRS; Regional Oxygen Saturation, rSO₂; Maintenance Hemodialysis, MHD; MoCA-B, Montreal Cognitive Assessment-Basic; AVLT-H, Auditory Verbal Learning Test; PSQI, Pittsburgh Sleep Quality Index; STT, Shape Trails Test; DTI, Diffusion Tensor Imaging; RAVLT, Rey Auditory Verbal Learning Test; VAS-F, Visual Analog Scale of Fatigue; fMRI, Functional Magnetic Resonance Imaging; FMA-UE, Fugl-Meyer Assessme; Obj, Objective; Subj, Subjective.

Flow chart of study selection.

4 Applications of taVNS in clinical cognitive impairment

In recent years, taVNS has gained prominence in the treatment of CI. Table 1 summarizes the characteristics and key findings of the included studies. We explicitly categorized outcome measures into objective cognitive tasks and subjective patient-reported outcomes to provide a granular view of efficacy. Quantitative changes in scores and statistical significance are detailed in the table. Clinical studies indicate that taVNS demonstrates therapeutic efficacy across multiple types of CI, including mild cognitive impairment (MCI) (48, 51), post-stroke cognitive impairment (PSCI) (52, 53), and chronic kidney disease (CKD)-related cognitive impairment (54). Although most current research focuses on taVNS’s improvement of mild cognitive impairment, studies on other types of CI remain relatively scarce (Table 1; Figure 3). Although current research predominantly focuses on MCI, emerging preclinical and clinical investigations suggest that taVNS may also hold potential in dementia-related contexts, such as vascular cognitive impairment and dementia (55, 56). However, robust clinical evidence specifically targeting Alzheimer’s disease or other major dementia subtypes remains limited, and further high-quality trials are needed to establish its efficacy and mechanisms in these populations (Table 1; Figure 3).

This figure summarizes the application of taVNS in clinical cognitive impairment.

4.1 Application of taVNS in mild cognitive impairment

MCI refers to a clinical state where patients exhibit impairment in memory or other cognitive functions, yet their ability to perform daily activities remains largely unaffected, falling short of the severity of dementia. It is considered a transitional phase between normal aging and dementia (57). Among individuals aged 65 and older, the annual incidence of MCI ranges from approximately 5 to 10%. Annually, about 10–15% of these patients progress to dementia, a rate significantly higher than the 1–2% observed in the general elderly population (58, 59). Therefore, the MCI stage, particularly its early phase, is recognized as a “golden window” for intervention. By controlling modifiable risk factors and adopting healthy lifestyles, it is possible to delay or even prevent the progression of some MCI cases to dementia (60).

A double-blind randomized controlled trial by Wang et al. systematically evaluated the efficacy and safety of taVNS in MCI patients. Sixty MCI patients were randomly assigned to either the taVNS group or a sham stimulation group for 24 weeks of home-based self-administration (61). Results demonstrated significant improvements in overall cognitive function and memory in the taVNS group compared to sham stimulation, with good safety profiles. Only one participant with a history of tympanic membrane perforation reported mild, reversible discomfort. This study suggests taVNS, as an effective, safe, and easily scalable non-pharmacological therapy, may help delay the progression of MCI to dementia. Guo et al. explored the efficacy and neural mechanisms of taVNS in patients with MCI comorbid with depression through a case study (51). A 71-year-old female patient underwent 24 weeks of taVNS intervention combined with functional magnetic resonance imaging (fMRI) assessment. Post-treatment, the patient demonstrated significant improvements in both cognitive function and depressive symptoms. fMRI further revealed enhanced functional activity in brain regions associated with cognition and emotion, including the right temporal pole and left medial orbitofrontal cortex. Symptoms remained remitted during follow-up, indicating that taVNS holds potential therapeutic value for MCI patients with depressive symptoms. This provides preliminary clinical and neuroimaging evidence supporting the application of this therapy in managing cognitive-emotional comorbidity. In addition to long-term interventions, recent studies using the broader terminology of tVNS have highlighted the acute benefits of auricular stimulation. For instance, single-session tVNS has been shown to rapidly improve spatial and associative memory in patients with amnestic MCI (62), while concurrent functional neuroimaging reveals that it immediately modulates functional connectivity within semantic and hippocampal networks, providing direct evidence of afferent target engagement (63).

4.2 Application of taVNS in post-stroke cognitive impairment

PSCI is one of the most common complications following stroke (3). Clinical observations indicate that over one-third of patient’s exhibit significant cognitive impairment in the early post-stroke period, and the severity of post-stroke cognitive impairment is closely associated with patient prognosis (64). Compared to non-stroke individuals and stroke survivors without cognitive impairment, patients with PSCI experience accelerated cognitive decline (65). Therefore, early identification of its severity and implementation of targeted interventions can significantly improve overall patient outcomes.

Wang et al. conducted an 8-week home-based taVNS intervention on a 71-year-old male patient with persistent PSCI 2.5 years after stroke onset (52). The patient was assessed at four time points using cognitive scales and diffusion tensor imaging (DTI). Results showed that after 8 weeks of taVNS intervention, the patient’s Montreal Cognitive Assessment score improved, and the completion time for the Shape Trajectory Test-B was significantly reduced. DTI further revealed improved white matter integrity in the dorsolateral prefrontal cortex (DLPFC), a region closely associated with cognitive function, along with an increase in the number of white matter fiber tracts connecting bilateral DLPFC. This study demonstrates that taVNS not only helps improve cognitive function in long-term PSCI patients but also promotes white matter structural remodeling, offering a promising new strategy for cognitive rehabilitation in home-based stroke patients. Li et al. conducted a randomized, double-blind, sham-controlled trial involving 30 subacute stroke patients (53). Over 3 weeks, participants received taVNS combined with task-oriented training. Results demonstrated that compared to the sham stimulation group, the taVNS group achieved significant improvements in cognitive function and limb motor function among patients with PSCI. Functional near-infrared spectroscopyc (fNIRS) data further demonstrated that taVNS activated the prefrontal cortex, DLPFC, and primary motor cortex—regions associated with cognition and motor function. This study indicates that taVNS promotes post-stroke recovery of cognitive and upper limb function through multi-level mechanisms by activating cognition-motor-related brain regions and facilitating network functional reorganization. It provides novel neurobiological insights and intervention pathways for stroke rehabilitation.

4.3 Application of taVNS in chronic kidney disease-related cognitive impairment

Maintenance hemodialysis (MHD) is the primary renal replacement therapy for end-stage CKD (66). In recent years, the concept of the “kidney-brain axis” has gained increasing attention (67). Research indicates that hemodynamic fluctuations and osmotic pressure changes during hemodialysis (HD) may impose additional stress on the brain, potentially inducing ischemic injury and neurovascular damage, ultimately leading to HD-associated CI (68, 69). Reports indicate that the incidence of CI in MHD patients reaches 50 to 70%, and it has been identified as an independent risk factor for mortality (54, 70). Further studies using magnetic resonance imaging and spectroscopy during dialysis have confirmed the association between acute brain injury associated with maintenance hemodialysis and cognitive decline (71). Therefore, early intervention and treatment are crucial for improving patient prognosis.

Wang et al. conducted a randomized, single-blind, sham-stimulation-controlled trial involving 36 MHD patients to investigate taVNS intervention (54). Results demonstrated that compared with the sham-stimulation group, taVNS significantly improved cognitive function and reduced fatigue during dialysis. fNIRS monitoring revealed that taVNS effectively mitigated dialysis-induced decreases in DLPFC cerebral oxygen saturation, with significant correlations between cerebral oxygenation changes and cognitive performance. This study suggests that taVNS may alleviate cognitive decline in MHD patients by improving cerebral hypoxia during dialysis, offering a potential non-pharmacological intervention strategy for preventing and treating dialysis-related brain injury.

5 The underlying mechanism of taVNS in cognitive impairment

While clinical studies have confirmed the therapeutic potential of taVNS, the precise cellular and molecular mechanisms driving these improvements remain difficult to fully explore in human subjects due to ethical and technical constraints. To address this gap, this section synthesizes evidence from animal models to provide a biological explanation for the clinical benefits observed in Section 4. As illustrated in Figure 4 and Table 2, taVNS exerts multidimensional neuroprotective and restorative effects across various cognitive impairment models. Its effects are not dependent on a single pathway but are achieved through five interconnected core mechanisms thoroughly validated by preclinical research, presented here in the order they appear in the figure: anti-inflammation and immune modulation, pro-angiogenesis, white matter repair and myelination, enhanced autophagy and cellular clearance, and enhanced neural plasticity. These mechanisms precisely target key pathophysiological links in CI. Through synergistic effects, they improve the brain microenvironment, protect neuronal function, and ultimately promote cognitive recovery. The following sections detail the specific actions and molecular basis of each mechanism sequentially, supported by preclinical evidence from diverse cognitive impairment models.

In preclinical models of cognitive impairment, taVNS exerts neuroprotective effects through five core pathways: (1) Anti-inflammation via cholinergic/α7nAChR-mediated microglial modulation, inhibiting cytokines; (2) Pro-angiogenesis through VEGF upregulation, enhancing cerebral perfusion; (3) White matter repair by promoting oligodendrocyte differentiation and MBP expression; (4) Enhanced autophagy via AMPK/SIRT1 activation, reducing oxidative stress and apoptosis; and (5) Enhanced neural plasticity through upregulated neurotrophic factors, improving LTP, dendritic spine density, and neurogenesis.

ReferencesRodent modelsDeviceInitial timetaVNS parameterStimulation siteKey Biomarkers/EffectsResults and conclusionCai et al. (75)Male Sprague–Dawley ratstaVNS5 min before surgeryFrequency: 20 Hz
Intensity: 0.5 mA
Pulse width: 0.5 ms
Stimulation duration: 30-s train repeated 4 times with 5-min intervalsLeft cavum conchaα7nAChR expression; Pro-inflammatory cytokines; NF-κB activation; Apoptosis markers; AD-related pathology1. Improved spatial memory
2. Reduced neuroinflammation
3. Decreased hippocampal apoptosis
4. Attenuated AD-related pathologyChoi et al. (89)Male C57BL/6 mice with tBCCAO-induced VCItaVNSPost-surgery; Behavioral tests on day 3 (short-term) or day 7 (long-term)Frequency: 20 Hz
Intensity: 1 mA
Pulse width: 330 μs
Duration: 20 min/sessionCymba concha and dorsal auricle of left earCSF circulation; Heart rate; Cognitive behaviorBrambilla-Pisoni et al. (84)Adult male CD-1 miceatVNSImmediately or 3 h after NOR trainingFrequency: 20 Hz; Intensity: 1 mA;
Pulse width: 320 μs;
Duration: 30 minLeft auricular conchac-Fos-based functional connectivity network; Novel object recognition memory performance1. Enhanced memory persistence with critical time window
2. Unaltered regional neuronal activation in 30 brain regions
3. Reorganized functional brain networkZhou et al. (76)Aged male Sprague–Dawley ratstaVNS24 h before anesthesia, for 5 daysFrequency: 10 Hz
Intensity: 1 mA
Pulse width: Not specified
Stimulation duration: 30 min/day for 5 consecutive daysLeft concha auricularisApoptosis; Necroptosis; Microglial activation; Cognitive function

Improved learning and memory (MWM)

Activated cholinergic neurons; Reduced apoptosis and necroptosis

Suppressed microglial activation and neuroinflammation

Effects blocked by cholinergic lesion

Nazari et al. (90)Male Wistar rats with morphine dependence and withdrawaltaVNSPost-cessation, for 14 daysFrequency: 5, 20, 100 Hz
Intensity: 2 mA
Pulse Width: 500 μs
Pattern: 30s ON, 5 min OFF
Total Duration: 30 min/dayLeft auricular concha region