Exploring the Feasibility of Mitigating Alcohol‐Induced Instability Using Galvanic Vestibular Stimulation

ABSTRACT

Alcohol consumption significantly impairs postural stability by disrupting vestibular function, increasing the risk of falls. This study investigated the feasibility of using galvanic vestibular stimulation (GVS) to mitigate alcohol‐induced postural instability in healthy adults, employing acute alcohol intake as a transient model of vestibular dysfunction. Sixty participants were assigned to groups with or without alcohol consumption (estimated blood alcohol concentration [BAC] 0.06%–0.08%) and randomly received one of three GVS types: random noise (rnGVS), direct current plus random noise (rn‐dcGVS), or a sham stimulation (fGVS). Static balance was assessed using a stabilometric platform with eyes closed on an unstable surface. As expected, alcohol consumption significantly worsened stabilometric parameters, resembling vestibular pathological hypofunction. fGVS produced no significant changes, ruling out a placebo effect. Notably, rnGVS significantly improved postural stability in subjects who consumed alcohol, reducing sway velocity, maximal anteroposterior (Max AP), and mediolateral displacements (Max ML) during and immediately after stimulation. In contrast, rn‐dcGVS worsened stability in subjects in the no alcohol consuming group with no significant change in the alcohol consuming group. These results demonstrate that rnGVS can effectively attenuate alcohol‐induced postural instability in a transient experimental model of vestibular dysfunction and support the use of acute alcohol intake as a controlled model for investigating vestibular modulation.

Acute alcohol consumption impaired postural stability in healthy subjects, supporting its use as a transient model of vestibular dysfunction. rnGVS improved balance performance, whereas rn‐dcGVS worsened postural stability.

Abbreviations

BAC

blood alcohol concentration

BrAC

breath alcohol concentration

CIOMS

Council for International Organizations of Medical Sciences

COP

center of (foot) pressure

DC

direct current

fGVS

false galvanic vestibular stimulation (Sham stimulation)

GABA

gamma‐aminobutyric acid

GVS

galvanic vestibular stimulation

IQR

interquartile range

Max AP

maximal anteroposterior displacement

Max ML

maximal mediolateral displacement

rn‐dcGVS

random noise + direct current galvanic vestibular stimulation

rnGVS

random noise galvanic vestibular stimulation

p_Hol

p statistics after Holm–Bonferroni correction

RMS

root mean square

SCC

semicircular canals

VOR

vestibulo‐ocular reflex

WHO

World Health Organization

1. Introduction

The vestibular system is involved in spatial perception, postural control, and oculomotor stability. It functions as an inertial reference system that detects linear and angular head accelerations and, through multisensory integration, contributes to postural coordination and locomotion (Cullen 2019). Recent studies have also linked the vestibular system to cognitive processes such as spatial memory and movement perception (Smith et al. 2024).

In vertebrates, the vestibular organ comprises five specialized sensory structures: three semicircular canals (SCC, horizontal, anterior, and posterior), which detect head angular accelerations, and two otolithic organs (utricle and saccule), which detect head linear accelerations (Cullen 2019). Vestibular signals project to the vestibular nuclei, giving rise to the vestibulo‐ocular reflex (VOR) and the vestibulospinal reflex, which are crucial for postural stabilization. At this level, vestibular activity integrates with visual, auditory, and somatosensory inputs and is relayed to higher structures, including the cerebellum, hippocampus, and posterior parietal cortex, supporting cognitive processes related to self‐position and movement (Hitier et al. 2014; Cullen 2016).

Galvanic vestibular stimulation (GVS) is a safe technique that uses modulated electrical currents to evoke neuronal activity in vestibular end‐organs without head motion. GVS is thought to activate both otolithic and SCC afferents, producing oculomotor and postural responses (Dlugaiczyk et al. 2019). Different GVS waveforms have been employed for distinct experimental and clinical purposes, including conventional direct current stimulation in vestibulopathies (Oh et al. 2025). Random noise GVS has been associated with improvements in postural and motor performance in Parkinson's disease (Samoudi et al. 2015) and vestibulopathies (Schniepp et al. 2018; Wuehr et al. 2017, 2024). Noisy direct current GVS has demonstrated lasting reductions in hemispatial neglect in stroke patients (Wilkinson et al. 2014). Additional applications of direct current GVS have been proposed to reduce cybersickness (Groth et al. 2022), as a microgravity countermeasure (Soto and Vega 2024) and for pilot training (Sadovnichii et al. 2021).

Acute alcohol consumption provides a well‐characterized, transient, and reversible model of vestibular dysfunction in healthy individuals (Ojanperä et al. 2023). Alcohol significantly alters vestibular function, as evidenced by changes in saccade velocity, reduced smooth pursuit velocity, gaze‐evoked nystagmus, impaired VOR suppression, postural instability, and vertigo (Choi et al. 2022; Han et al. 2020). In healthy individuals, alcohol produces alterations in head impulse testing, video‐oculography, and balance similar to those observed in patients with vestibular or cerebellar dysfunction (Ojanperä et al. 2023; Roth et al. 2014). Due to its hydrophilic nature, alcohol penetrates the brain and the inner ear, affecting multiple proteins, including synaptic receptors and ion channels. In the vestibular organs, alcohol transiently alters SCC cupula density relative to endolymph, producing a “light dome” state, which produces a direction‐changing positional nystagmus and rotatory vertigo (Fetter et al. 1999; Lee et al. 2023). Although this was a hypothetical explanation, recent evidence indicates that certainly, changes in serum osmolality following alcohol ingestion and subsequent change in specific gravity in the perilymph and endolymph may be a cause of positional nystagmus (Han et al. 2020). Besides, GABA receptors are one of the primary targets of alcohol (Dharavath et al. 2023). It is thought that alcohol impairs postural control, partly via effects on GABA receptors in the cerebellum and vestibular nuclei (Mitoma et al. 2021; Müller‐Oehring et al. 2022). Increased GABAergic input to vestibular nucleus neurons contributes to the postural instability observed in individuals under alcohol influence (Ujihara 2000; Nakamura et al. 1989). Also, acute alcohol effects on the vestibulo‐ocular reflex have been thought to be produced by alterations in the velocity storage mechanism (Choi et al. 2022).

This study aimed to evaluate the feasibility of using GVS to modulate postural instability in healthy subjects under control conditions and after acute alcohol consumption. The alcohol model was chosen for its transient impact on vestibular function, providing a controlled model condition to assess GVS‐driven postural adjustments.

2. Material and Methods

2.1. Subjects

Participants were recruited through personal contact between students from the university. Participants were not compensated monetarily for their participation. Sixty healthy subjects (30 men and 30 women) aged 18–29 years old were recruited and evaluated at the Sensory Neurophysiology Laboratory, Institute of Physiology, Benemérita Universidad Autónoma de Puebla, México. Participants were assigned a code to ensure data anonymity and provided written informed consent. The study complied with national and international ethical regulations, including Mexican Official Standard NOM‐012‐SSA3‐2012, the Declaration of Helsinki (2013), and the CIOMS/WHO International Ethical Guidelines (2016). The study was approved by the Ethics Committee of the Institute of Physiology (20250925_ESE_02).

Participants completed a health questionnaire to confirm no history of balance‐affecting pathologies or recent unexplained falls. Blood pressure, heart rate, and blood oxygen saturation were measured for all subjects. Exclusion criteria included presence of metallic implants, any sign of cardiac disease, any evidence of dizziness, or history of vestibular disorders and no previous experience consuming alcohol (never in their life) or a history of alcohol use disorder.

2.2. Galvanic Vestibular Stimulation

GVS was applied bilaterally on the mastoid processes (transmastoid configuration) using a ce‐certified battery‐driven current stimulator (neuroConn DC‐Stimulator Plus, Ilmenau, Germany): Rubber electrodes of 5 × 7 cm (neuroConn) moistened with 0.9% saline solution were placed over the mastoid processes on either side of the head and firmly secured in position by the use of a headband. Participants received one of three stimuli for 30 s: random noise GVS (rnGVS), direct current combined with random noise GVS (rn‐dcGVS), or sham GVS (fGVS). The rnGVS stimuli were generated with the stimulator in the “noise” mode that consisted of a bipolar random noise waveform fluctuating between ± 1.5 mA (zero mean) within a bandwidth of 0–250 Hz. A broadband 0‐ to 250‐Hz noise bandwidth was selected to provide broad‐frequency vestibular stimulation rather than targeting a specific resonance frequency (Inukai et al. 2018; Curthoys et al. 2017). Intensity was selected as a stimulation level to ensure consistent vestibular input across subjects under the challenging experimental conditions used in this study, including alcohol exposure, standing on foam, and eye closure. The safety and tolerability of GVS at 1.5 mA has been established in normal subjects (Pliego et al. 2021), a large cohort of stroke patients and healthy individuals (Utz et al. 2011), and to modulate balance and vestibular perception in clinical populations (Tomioka et al. 2022).

For the rn‐dcGVS, a direct current (DC) offset was superimposed onto the random noise signal (± 1.5 mA), resulting in a DC biased stochastic waveform ranging from ± 2.5 to 0.5 mA, with a −1 mA mean current. The use of a signal composed of a DC component and a random noise component was motivated by the intention to simultaneously explore two mechanisms potentially relevant to vestibular modulation. The DC component generates sustained polarization of vestibular afferents and the addition of random noise introduces stochastic fluctuations that could facilitate the detection and transmission of vestibular signals. In fact, it has been shown that noisy electrical stimulation produces non‐adapting increase in the discharge rate of afferent neurons, as compared to the DC stimulus that produces responses with a significant decay due to adaptation of the spike generation mechanism (Benavidez 2021; Vega et al. 2022). For fGVS, the experimental setup was similar to the active conditions; electrodes were placed and connected to the stimulator identical to the active stimulation with the device output set to zero, in order to detect expectation‐related effects. Participants were informed that vestibular stimulation would be applied during the experiment but were not informed of the specific stimulation condition administered in each trial. Thus, participants remained blinded to whether they received active or sham stimulation.

Stimulus waveforms were selected based on their functional relevance in modulating vestibular input. The rnGVS was included because it has been shown to improve postural stability (Ko et al. 2020; Inukai et al. 2018). Conventional DC GVS is known to induce polarity‐dependent changes in vestibular afferent firing and vestibulospinal activity, contributing to postural deviations and balance adjustments (Fitzpatrick and Day 2004; Marchand et al. 2025). In addition, combined noisy and direct current vestibular stimulation paradigms have previously been explored in clinical applications such as hemispatial neglect (Wilkinson et al. 2014). Therefore, combining broadband noisy stimulation with a direct current offset was conceived as an exploratory paradigm to assess whether simultaneous noisy and tonic vestibular inputs could produce distinct postural effects.

2.3. Postural Assessment

Static balance was assessed during quiet stance using a force platform (Wii Balance Board, Nintendo) with custom software (Salinas et al. 2009; Soto et al. 2014). Subjects stood on the platform without shoes, with eyes closed (using a black mask), feet apart with the heels together, on a 5‐cm‐thick pad made of closed cell foam plastic, placed between the subjects' feet and the force platform to reduce somatosensory feedback from plantar receptors and enhance the vestibular system involvement in stabilization (Halmágyi and Curthoys 2021). Recordings were always performed in the morning, between 09:00 and 14:00, in a quiet and darkened room; subjects were instructed to maintain their eyes closed, and a mask was added to assure no visual feedback, thus further isolating the vestibular contribution to postural control.

For the stabilometric assessment, center of pressure (COP) displacements was recorded using a force platform with a sampling frequency of 80 Hz over a 30‐s period (Scoppa et al. 2013). Mean velocity was calculated as the total path length of the COP divided by the recording duration. The ellipse area was defined as the 95% confidence surface enclosing the COP oscillations in both the mediolateral (ML) and anteroposterior (AP) axes. The root mean square (RMS) value was obtained by calculating the square root of the mean of the squared distances of the COP relative to its mean position, representing displacement variability. Finally, maximal excursion of the COP along the anteroposterior (Max AP) and mediolateral (Max ML) axes was considered. The analysis of these components provides insight into axis‐specific postural control strategies and has been associated with postural performance and fall risk (Wiśniowska‐Szurlej et al. 2022).

These parameters quantify stability, oscillation magnitude, and balance control speed. Stabilometric measurements were obtained at four time point conditions: pre‐GVS (C), during GVS (S), immediately post‐GVS (P1), and 5‐min post‐GVS (P5) (Figure 1).

FIGURE 1.

Experimental timeline. After electrode placement, subjects stood on the balance platform for four consecutive 30‐s recordings: baseline control (C), during GVS (S), immediately post‐GVS (P1), and 5‐min post‐GVS (P5).

2.4. Alcohol Consumption

Subjects in the alcohol consumption group consumed vodka (40% alcohol; 2.2 mL/kg) diluted with an equal amount of tonic water over 10‐min. Blood alcohol concentration (BAC) was estimated using a portable digital breathalyzer based on electrochemical fuel cell technology (RD900 Alcohol Tester, Shenzhen Red Dragon Instrument Co., China). This device detects breath alcohol concentration (BrAC, expressed in mg/L of expired air). The instrument automatically converts BrAC to an estimated BAC value, expressed as grams of alcohol per 100 mL of blood (g/100 mL). For the present study, the BAC display mode was used.

Possible mild clinical effects such as disinhibition and mild impairment of cognitive and motor functions were monitored. All experimental sessions were supervised by a medical doctor, and there was always an investigator with experience in vestibular neurophysiology and galvanic vestibular stimulation. In cases of moderate symptoms, participants were offered rest, sugar‐sweetened beverages, and continuous monitoring (Mirijello et al. 2023; Sarasa‐Renedo et al. 2014).

2.5. Experimental Design and Data Analysis

A total of 60 subjects were randomly assigned to one of two groups: with (n = 30) or without (n = 30) alcohol consumption. Each subject was then randomly assigned to one of three GVS stimulus subgroups (rnGVS, rn‐dcGVS, or fGVS). Within each subgroup, 10 participants consumed alcohol, and 10 did not consume alcohol. A between‐subjects design was used to minimize adaptation and motor learning associated with the repeated task. Subjects were exposed to a single round of GVS and to the experimental set only once.

A formal a priori power analysis was not conducted because this study was designed as an exploratory feasibility study intended to evaluate the effects of different GVS configurations on alcohol‐induced postural instability in healthy subjects. Sample size is comparable to previous exploratory studies investigating GVS effects on postural control and vestibular modulation (Ko et al. 2020; Schniepp et al. 2018). This represents a limitation of the study, as discussed below.

Data normality was assessed using the Shapiro–Wilk test. Comparisons between groups (alcohol consumption versus no alcohol consumption) were performed using the Mann–Whitney U test, given that the groups were independent and the data were not normally distributed. This analysis was limited to testing the a priori hypothesis that alcohol consumption increases postural oscillation compared to control subjects, thus minimizing the risk of a Type I error. Based on the hypothesis that each GVS condition would modify postural oscillation compared to the control condition, within‐subject comparisons across time points (Control vs. S, P1, and P5) were performed using Wilcoxon signed‐rank tests. Statistical comparisons are reported with both the uncorrected p value (p) and the Holm–Bonferroni corrected p value (p_Holm). Within each subgroup, three planned comparisons were performed per variable (Control vs. S, P1, and P5), and Holm–Bonferroni correction was applied across these three pairs to control the familywise error rate. Results with p_Holm ≤ 0.05 are considered statistically significant after correction; results with p ≤ 0.05 but p_Holm > 0.05 are reported as marginal and interpreted with caution, as they carry a higher risk of Type I error. Both values are reported to allow assessment of effect magnitude and to avoid obscuring potentially meaningful trends that may be detectable with larger samples. Effect size was quantified using the rank‐biserial correlation coefficient r, calculated as r = |Z| / √n, where Z is derived from the Wilcoxon test statistic and n is the number of paired observations. Values of r ≥ 0.1, ≥ 0.3, and ≥ 0.5 are conventionally interpreted as small, medium, and large effects, respectively.

Results are presented in Raincloud‐box plots showing all data points, distribution, the median and the first and third quartile. Medians and interquartile ranges (IQR) are reported as measures of central tendency and variability. Data tables are included in the supporting information.

Baseline demographic characteristics (age, weight, and height) were compared between the alcohol and no alcohol groups. Variables that met assumptions of normality were analyzed using independent samples t test, whereas nonnormally distributed variables were analyzed using Mann–Whitney U test.

3. Results

All participants completed the study. No adverse events were observed during the study. No participant reported pain, skin lesions, burns, or requested discontinuation of stimulation. Mild transient tingling beneath the electrodes was reported by about 10% of the participants. All reported sensations were mild and resolved spontaneously. No significant differences were found between the alcohol and no alcohol groups in age, weight, or height (all p > 0.05), indicating that the groups were similar with respect to baseline demographic characteristics (Table 1).

TABLE 1.

Values are presented as median [IQR]. Age and height were compared using independent samples t tests, whereas weight was compared using a Mann–Whitney U test.

Variable	No alcohol	Alcohol	p
Age (years)	24 [4.8] (18–29)	23 [5.5] (18–29)	0.270
Weight (kg)	66.5 [17.9] (49–126)	65.9 [26.8] (47–112)	0.971
Height (cm)	167 [10.5] (150–183)	165 [7] (148–183)	0.658

As expected, alcohol consumption significantly impaired postural balance: Sway velocity, ellipse area, RMS, maximal anteroposterior displacement (Max AP), and maximal mediolateral displacement (Max ML) were significantly higher in the alcohol group compared to the nonalcohol group at baseline before any GVS (Mann–Whitney U), with large size effect (r = 0.54–0.84). The least increase was found in the sway velocity (1.3 times) and greatest increase in the area of ellipse (three times larger) (Table 2 and Figure 2).

TABLE 2.

Effect of alcohol consumption on stabilometric parameters. Values show the median and [IQR]. Statistical comparisons between the alcohol (n = 30) and no‐alcohol (n = 30) groups were performed using the Mann–Whitney U test. All parameters were significantly greater in the alcohol group, with a large effect size (r = 0.54–0.84).

	N	Sway velocity (cm/s)	Area of ellipse (cm2)	RMS (cm)	Max AP (cm)	Max ML (cm)
No alcohol	30	4.3 [1.2]	6.9 [5.6]	1.1 [0.4]	6.2 [3.1]	1.9 [0.8]
Alcohol	30	5.4 [2.6]	20.8 [22.4]	2.0 [0.8]	10.8 [5.1]	3.2 [1.4]

p	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001
Size effect (r)	0.54	0.84	0.8	0.76	0.64

FIGURE 2.

Representative stabilometric recordings. Registers were mean‐centered, showing the sway path (solid line) and the 95% confidence ellipse (dashed line). (A) Recording (30 s) from a subject who did not consume alcohol. (B) Recording from a subject after alcohol consumption. The sway in the subject after alcohol consumption was approximately double that of the subject without alcohol.

3.1. Effect of fGVS

Alcohol consumption significantly impaired the stabilometric parameters. In the alcohol group, parameters were all significantly higher than in the no‐alcohol group (p ≤ 0.01), the sway velocity increased from a median of 4.4 and interquartile range of [0.4] to 5.2 [1.9], ellipse area from a median of 6.5 [4.4] to 21.3 [22.7], RMS from a median of 1.1 [0.4] to 2.1 [0.9], Max AP from a median of 6.2 [3.0] to 10.1 [4.1], and Max ML from a median of 1.8 [0.7] to 3.4 [1.5]. fGVS produced no significant effect in any stabilometric parameter in either group and across all recordings (C, S, P1, and P5) (Figure 3 and Table S1). This confirms the absence of a placebo effect related to electrode placement or experimental setup.

FIGURE 3.

Raincloud‐box plots comparing the effect of fGVS on postural stability in subjects with and without alcohol consumption. In this and following graph, each raincloud plot shows the raw data distribution (rain), box plot (median, IQR, and whiskers), and half‐violin density estimate of postural sway parameters: (A) sway velocity, (B) ellipse area, (C) Max AP, and (D) RMS. Parameters in each condition remains stable within measurements (C, S, P1, and P5); in the alcohol group, all parameters were significantly higher than in the no‐alcohol group (p ≤ 0.01). The lack of significant temporal changes confirms the absence of a placebo effect or spontaneous recovery during the experimental session. In these and following graphs, blue: no alcohol; red: alcohol.

3.2. Effect of rnGVS

In subjects without alcohol consumption, rnGVS (n = 10) produced no significant changes in mean sway velocity, ellipse area, and Max ML across any time point (all p_Holm > 0.400). A marginal decrease of RMS from a baseline median of 1.1–1.0 cm in P5 (−9.4%; p = 0.037, r = 0.66) and at Max AP also showed decreased from a baseline of 5.7 cm to 4.9 cm in P5 (−14.0%; p = 0.027, r = 0.69). However, these differences did not survive Holm–Bonferroni correction (p_Holm = 0.111 and p_Holm = 0.082, respectively).

In subjects who consumed alcohol (n = 10), rnGVS significantly reduced postural oscillation during and immediately after stimulation (Figure 4 and Table S2).

FIGURE 4.

Raincloud‐box plots comparing the effect of rnGVS on postural stability in subjects with and without alcohol consumption. (A) Sway velocity. (B) Area of ellipse. (C) Max AP. (D) RMS. In each case, recordings were performed in C, S, P1, and P5. In subjects without alcohol consumption, RMS and Max AP showed marginal reductions at P5. In subjects that consumed alcohol, the rnGVS significantly reduced the mean sway, ellipse area, and Max AP. The clearest effect was on the sway velocity and the Max AP parameters. Colors—blue: no alcohol; red: alcohol.

Mean sway velocity decreased from a baseline median of 6.7 to 5.2 cm/s during stimulation (−21.5%; p_Holm = 0.041, r = 0.73) and further to 5.0 cm/s immediately after stimulation P1 (−25.3%; p_Holm = 0.041, r = 0.76). A partial rebound was observed at P5 (5.7 cm/s; p_Holm = 0.232, r = 0.40).

Max AP displacement showed the most robust effects: from a baseline of 15.0 to 11.0 cm during stimulation (−26.7%; p_Holm = 0.029, r = 0.79) and 11.1 cm at P1 (−26.4%; p_Holm = 0.039, r = 0.73), with partial rebound at P5 (13.2 cm; p_Holm = 0.492, r = 0.24). Max ML decreased after ending the stimulation at P1 from control of 3.3 [1.0] to 2.8 [0.6] cm (−17.9%; p = 0.037, r = 0.66) and 2.6 [1.0] cm (26.9%; p_Holm = 0.041, r = 0.76).

Ellipse area showed a directional reduction from a baseline of 30.9–22.3 cm² during stimulation (−27.7%; p = 0.06, r = 0.60) and 22.9 cm² at P1 (−26.0%; p = 0.049; r = 0.63), though these did not survive Holm correction (p_Holm = 0.146). RMS displacement decreased 9.4% at P1 from 2.5 to 2.3 cm (p = 0.037; r = 0.66). Both variables showed large size effects.

Comparison of the effect of rnGVS in those who consumed alcohol with the no alcohol group revealed that at S (p = 0.104) and at P1 (p = 0.140), the mean sway velocity was no longer significantly different from the no alcohol group, indicating a transient partial functional normalization. However, other parameters remained significantly elevated relative to the no alcohol group at all‐time points (all p ≤ 0.05; Table S2), indicating that although rnGVS meaningfully attenuated alcohol‐induced instability, it did not fully restore sober‐level balance within the observation window.

3.3. Effect of rn‐dcGVS

In subjects without alcohol consumption, rn‐dcGVS (n = 10) significantly worsened postural stability during stimulation. Ellipse area increased from a median of 8.1 cm² at baseline to 12.0 cm² (48%; p_Holm = 0.041, r = 0.76). Marginal increases were also observed at P1 (11.9 cm²; 46.9%; p = 0.049, r = 0.63) and at P5 (9.3 cm²; 14.8%; p = 0.037, r = 0.66), but these did not survive Holm–Bonferroni correction (p_Holm = 0.074 for both time points). RMS displacement increased from 1.3 to 1.5 cm (15%; p_Holm = 0.041, r = 0.76). Max AP displacement increased from 6.8 to 8.1 cm that survived Holm–Bonferroni correction (19%; p_Holm = 0.018, r = 0.82), confirming a robust anteroposterior destabilization during active rn‐dcGVS. Mean sway velocity showed a tendency to increase from 4.2 to 4.4 cm/s (5%; p = 0.049, r = 0.63) without surviving Holm–Bonferroni correction (p_Holm = 0.146). These destabilizing effects were most pronounced during the stimulation period, with a partial return toward baseline after stimulation ended (Figure 5 and Table S3). Max ML marginally increased from a baseline of 1.9 to 2.6 cm at S (36.8%; p = 0.037) but did not survive Holm–Bonferroni correction (p_Holm = 0.111).

FIGURE 5.

Raincloud‐box plots of the effect of rn‐dcGVS on postural stability in subjects with and without alcohol consumption. Significant differences were found in all conditions between alcohol versus no alcohol. rn‐dcGVS increased sway velocity (A), ellipse area (B), Max AP (C), and RMS (D), in no alcohol subjects. In (B), ellipse area increased during rn‐dcGVS and remained significantly large trough P5. Colors—blue: no alcohol; red: alcohol.

In subjects who consumed alcohol (n = 10), rn‐dcGVS did not produce significant changes in mean sway velocity, ellipse area, or RMS at any time point (all p ≥ 0.131). Numerically, ellipse area increased from 17.8 to 25.7 cm² during stimulation and mean sway velocity from 5.5 to 6.6 cm/s—consistent in direction with the destabilizing effect observed in the no‐alcohol group; however, the larger within‐group variability under alcohol conditions precluded statistical significance.

4. Discussion

This study shows the feasibility of using rnGVS to modulate transient postural instability induced by acute alcohol intoxication in healthy subjects. Acute alcohol exposure served as a reversible and experimentally controlled model of vestibular dysfunction. Our findings further support the validity of this model (Ojanperä et al. 2023), which is useful for probing the efficacy and mechanisms of GVS. Alcohol acts as a vestibular depressant, disrupting sensory integration and motor coordination, thereby increasing instability and likelihood of falls (Hafström et al. 2014; Ikeda et al. 1980; Marczinski and Mearns 2019). Our findings show that rnGVS significantly improved postural stability in subjects exposed to alcohol.

A potential mechanism accounting for the effect of rnGVS as compared with rn‐dcGVS relates to vestibular afferent neuron response to electrical stimulation. Typically, DC pulses produce a marked adaptation of afferent neuron discharge rate, which may lead to a nearly complete adaptation during sustained stimulation. In contrast, noisy electrical stimuli did not induce such adaptation in the response of vestibular afferents (Benavidez 2021). Consequently, rnGVS may preserve vestibular afferent responsiveness throughout stimulation. In addition, stochastic resonance (SR) has been proposed as a mechanism by which externally applied noise enhances weak sensory signals (Mulavara et al. 2011; Wuehr et al. 2017), potentially contributing to the improvements observed during stimulation.

Alcohol is known to alter cerebellar and vestibular network function through multiple neurophysiological mechanisms, including modulation of GABAergic transmission, which contributes to impaired balance control (Nakamura et al. 1989; Ujihara 2000; Mitoma et al. 2021). Within this altered state, rnGVS may facilitate vestibular signal processing by modulating the excitability of vestibular‐cerebellar networks. In fact, the carry‐over effects observed at P1 and P5 are more consistent with short‐term plasticity within central vestibular pathways, which may outlast the stimulation period. This is supported by previous reports of poststimulation improvements following vestibular noise stimulation, with effects persisting from minutes to hours after stimulation (Fujimoto et al. 2016, 2018; Inukai et al. 2020), suggesting that short‐term vestibular plasticity contributes to the full pattern of GVS effects.

Regarding the stimulus amplitude, evidence from clinical and experimental studies suggests that beneficial effects of rnGVS are not strictly restricted to subthreshold amplitudes. Significant postural improvements have been reported using a range of stimulation intensities, including amplitudes comparable to those employed in the present study (Fujimoto et al. 2019, 2025; Nguyen et al. 2023). Moreover, a safety study demonstrated that GVS at 1.5 mA is well tolerated, with only mild, transient itching and tingling and no serious adverse events (Utz et al. 2011). Furthermore, Tomioka et al. (2022) successfully used 1.5‐mA bipolar GVS to modulate subjective visual vertical and sitting balance in stroke patients.

Conversely, rn‐dcGVS impaired postural stability in the no‐alcohol group and showed a nonsignificant trend in the alcohol group. We hypothesize that the added direct current component produces a vestibular signal analogous to a sustained tilt stimulus, thereby activating vestibulo‐spinal reflexes. Unlike rnGVS, which has a zero mean and produces no net directional vestibular drive, rn‐dcGVS generates a persistent asymmetry between the cathode and anode sides, resulting in net excitation on one side and inhibition on the other. Such asymmetrical activation is known to induce body sway and postural displacement in the direction of the vestibular bias (Fitzpatrick and Day 2004; Lopez and Cullen 2024). Furthermore, the direct current component may promote adaptation of vestibular afferent firing over time, whereas noisy stimulation has been shown to maintain afferent responsiveness more consistently than sustained DC stimulation (Benavidez 2021). Together, these mechanisms may contribute to the poorer postural performance observed during rn‐dcGVS.

GVS has been shown to enhance spatial memory and to facilitate cognitive recovery in patients with acute unilateral peripheral vestibulopathy, with no adverse effects reported (Oh et al. 2025). GVS has also shown significant modulatory effects in patients with vestibulopathies, improving their posture or body balance (Fujimoto et al. 2018; Wuehr et al. 2024); this effect is particularly greater in those with postural instability (Fujimoto et al. 2019) and lasting longer in older adults (Fujimoto et al. 2016). Similarly, in our subjects, we observed rnGVS‐induced improvement specifically in alcohol‐consuming subjects but not in the subjects without alcohol consumption.

Although this study provides positive evidence, there are limitations that should be considered. The sample size was relatively small, limiting statistical power and increasing Type II error risk, which may have underpowered the detection of potentially meaningful effects. In fact, the data hint at potential sex differences in alcohol effects, warranting future investigation with larger samples. Long‐term effects of GVS were not assessed but are relevant for clinical device development.

4.1. Conclusions

Our results demonstrate that the use of alcohol under controlled circumstances constitutes a model to study transient vestibular dysfunction. We found that rnGVS improves postural stability in subjects under alcohol influence. rnGVS enhances postural stability during and shortly after stimulation, whereas rn‐dcGVS increased instability, suggesting it is not suitable for postural stabilization devices.

These findings support the potential of rnGVS as a noninvasive tool for improving postural control and vestibular function, while further validating alcohol‐induced instability as a reversible experimental model of vestibular impairment.

Future studies should investigate whether the effects of rnGVS can be optimized by different stimulation parameters, evaluate repeated stimulation sessions, and try to clarify the neural mechanisms underlying the observed improvements in postural stability.

Author Contributions

Nancy Barranco: investigation, methodology, data curation, formal analysis, writing – original draft. Gustavo Arellano: data curation, validation, formal analysis, investigation. Karla Pineda: investigation, methodology, validation. Enrique Soto: writing – original draft, writing – review and editing, conceptualization, methodology, investigation, validation, supervision, resources.

Funding

This work was supported by the Secretaría de Ciencia, Humanidades, Tecnología e Innovación (SECIHTI‐Mexico) (CBF‐2025‐I‐462) and BUAP‐VIEP Cuerpos Académicos (BUAP‐CA‐119).

Conflicts of Interest

The authors declare no conflicts of interest. However, E.S. wishes to disclose that he holds patents related to GVS‐based devices for medical and space applications.

Supporting information

Table S1: Effect of fGVS on subjects with (n = 10) and without alcohol (n = 10) consumption. Values are the median and the [IQR]. C: control; S: during GVS; P1: immediately post‐GVS; P5: 5 min post‐GVS.

Table S2: Effect of rnGVS in stabilometric recordings parameters in subjects with (n = 10) and without alcohol consumption (n = 10). Values are median [IQR]. C: control; S: during GVS; P1: immediately post‐GVS; P5: 5 min post‐GVS. ** p_Holm ≤ 0.05, * p ≤ 0.05.

Table S3: Effect of rn‐dcGVS on the stabilometric parameters in subjects with (n = 10) and without (n = 10) alcohol consumption. Values are median [IQR]. C: control; S: during GVS; P1: immediately post‐GVS; P5: 5 min post‐GVS. ** p_Holm ≤ 0.05, * p ≤ 0.05.

Barranco, N. , Arellano G., Pineda K., and Soto E.. 2026. “Exploring the Feasibility of Mitigating Alcohol‐Induced Instability Using Galvanic Vestibular Stimulation.” European Journal of Neuroscience 64, no. 2: e70627. 10.1111/ejn.70627.

Data Availability Statement

The dataset supporting the findings of this study has been deposited in Figshare (available under DOI: 10.6084/m9.figshare.32573088).