The vestibular system, often referred to as the balance system of the inner ear, plays a vital role in motor control, spatial orientation, and overall equilibrium. Though primarily responsible for unconscious motor functions, its influence extends beyond balance to support higher-level brain activities such as conscious spatial awareness, navigation, voluntary movement, visual stabilization, and autonomic regulation. This system comprises the otolith organs (the utricle and saccule), which respond to linear acceleration, and the three semicircular canals, which detect angular motion in different planes. Together, they provide the brain with a comprehensive representation of head motion and position in three-dimensional space [1].

Historically, foundational insights into vestibular responses to auditory stimulation were provided by early studies [2]. Later research demonstrated that intense sound stimuli could evoke short-latency muscle responses, subsequently identified as vestibular-evoked myogenic potentials (VEMPs) [3]. These responses were traced to the vestibular apparatus, even in individuals with auditory impairments, emphasizing their vestibular rather than auditory origin. VEMPs have since become a staple in the clinical assessment of vestibular function. They are commonly classified into cervical VEMP (cVEMP) and ocular VEMP (oVEMP), representing the saccular and utricular reflex pathways, respectively. The cVEMP is recorded from the sternocleidomastoid muscle and is considered a direct reflection of saccular and inferior vestibular nerve activity. It presents a biphasic waveform, typically P13–N23, and is highly sensitive to both aging and pathology [4]. In contrast, oVEMP assesses the utriculo-ocular reflex and is measured from the extraocular muscles, notably the inferior oblique, with a characteristic n10–p15 waveform [5]. Though its amplitude is generally lower than cVEMP, oVEMP is a reliable indicator of utricular function.

Recently, a novel variant known as the masseter VEMP (mVEMP) has gained attention for its potential diagnostic value. Unlike cVEMP and oVEMP, which rely on neck and ocular muscle activity, mVEMP is recorded from the masseter muscle, the principal muscle involved in mastication. The masseter muscle produces short-latency inhibitory electromyographic responses in reaction to vestibular stimuli, characterized by a p11/n15 waveform. This response is believed to be of vestibular origin and shares key similarities with cVEMP, including the requirement for tonic muscle contraction and higher stimulus thresholds. An additional component, p16/n21, is attributed to cochlear activation, indicating that mVEMP reflects both vestibular and auditory inputs [6,7].

The vestibulo-masseteric reflex, from which mVEMP arises, is mediated by a polysynaptic pathway originating from the saccule. Neural signals travel via the inferior vestibular nerve to the medial vestibular nucleus and then cross to the trigeminal motor nucleus, ultimately activating the masseter muscle via the mandibular branch of cranial nerve V. This bilateral projection system allows for symmetrical responses across both sides of the face, offering a unique opportunity to evaluate vestibular integrity in patients for whom cVEMP may be impractical [8].

Although air conduction (AC) is traditionally used to evoke VEMP responses, it requires high sound levels, ranging between 130 and 145 dB SPL, which can be unsafe or ineffective in individuals with conductive hearing loss. Bone conduction (BC) has emerged as a safer and more practical alternative, especially for those with middle ear pathologies such as otosclerosis or chronic otitis media. BC directly stimulates the cochlea and vestibular structures without needing to pass through the external or middle ear, making it particularly suitable for vestibular assessments in a broader population [9].

Despite these advantages, BC-stimulated VEMPs, particularly mVEMPs, remain underutilized due to technical limitations such as the low output and high distortion of existing BC transducers. However, recent studies have demonstrated that mVEMP responses elicited by AC stimulation are reliable and reproducible, with acceptable test–retest consistency and diagnostic value. Given the masseter muscle’s ease of access and the potential applicability in patients with neck muscle impairment, mVEMP represents a promising diagnostic tool [10].

To enhance clinical implementation, it is essential to establish normative values and validate the reliability of BC-stimulated mVEMP using standardized stimuli such as 500 Hz tone bursts. This study aims to evaluate the test–retest reliability of BC mVEMP responses in healthy adults and determine the optimal stimulation threshold for effective vestibular assessment. Establishing these parameters will support the broader application of mVEMP in clinical audiology and vestibular diagnostics. mVEMPs are a promising tool for assessing vestibulo-trigeminal pathway integrity [8]. Recorded from the masseter muscle using zygomatic or mandibular montages, mVEMP reflects both vestibular (p11/n15) and cochlear (p16/n21) responses [6,7]. Tone burst stimuli, particularly at 500 Hz, produce robust responses with optimal amplitude and latency [11,12]. Increased electromyographic (EMG) activity enhances amplitude while latency remains stable [13,14].

BC mVEMP is effective in cases of conductive hearing loss, where AC may fail [15]. Normative values of mVEMP were established in recent studies, one of which noted age-related decline in mVEMP amplitude [16,17]. Clinically, mVEMP shows high diagnostic value in Parkinson’s disease [18], multiple sclerosis [19,20], auditory neuropathy [21], migraine [22,23], and Ménière’s disease [24]. It also aids in evaluating motion sickness susceptibility [25], vestibular neuritis [26], and Wallenberg syndrome [27]. mVEMP demonstrates fair-to-good test–retest reliability [12], and its bilateral nature differentiates it from the ipsilateral cVEMP [28,29]. Its ease of elicitation, especially in pediatric and geriatric cases, supports mVEMP’s inclusion alongside cVEMP and oVEMP in comprehensive vestibular diagnostics [10,16,28].

This normative and reliability study was approved by the Institutional Ethics Committee of the SRM Institute of Science and Technology (Approval No. SRMIEC-ST0125-2008). A total of 45 healthy participants (11 males and 34 females) aged between 18 and 35 years were recruited based on strict inclusion and exclusion criteria. Inclusion criteria involved individuals with normal hearing thresholds (pure-tone average ≤25 dB at 500 Hz, 1 kHz, and 2 kHz), normal oromandibular anatomy, no history of otological or vestibular disorders, and not taking any medication. Participants with conditions such as temporomandibular joint dysfunction, facial paralysis, stroke, recent dental procedures, or any neurological impairment were excluded from the study.

All assessments were performed in a sound-treated room conforming to ANSI S3.1-1977 standards. Audiological evaluations included otoscopy, pure-tone audiometry using the Madsen Astera2 (Otometrics), and middle ear analysis with the Tympstar Pro (Grason-Stadler, Inc.) to rule out conductive components.

Behavioral vestibular screening—including gaze nystagmus, the tandem gait test, and the Fukuda stepping test—was conducted to identify the presence of vestibular pathology. Participants who demonstrated abnormal findings on two or more of these screening measures were excluded from the study. Abnormal findings were operationally defined as follows: for the gaze test, the presence of spontaneous or gaze-evoked nystagmus in any gaze position; for the tandem gait test, inability to complete 10 consecutive heel-to-toe steps without loss of balance, deviation from the line, or the need for external support; and for the Fukuda stepping test, a rotation exceeding 45° or forward displacement greater than 1 m during stepping in place.

For mVEMP recording, participants sat upright and were instructed to clench their teeth to contract the masseter muscle. During mVEMP recording, real-time visual EMG feedback was provided to facilitate consistent activation of the masseter muscle, and participants were instructed to maintain muscle contraction within a target EMG range of approximately 30–70 μV. Muscle activity was continuously monitored using a built-in visual display of EMG deflection on the computer screen. The root mean square (RMS) value of the pre-stimulus EMG signal was used to verify stable muscle activation and to determine EMG magnitude prior to stimulus onset.

No post hoc EMG normalization of mVEMP amplitudes was performed during offline analysis. Accordingly, peak-to-peak P11–N15 amplitudes were analyzed and reported in absolute microvolt (μV) values without normalization to prestimulus RMS EMG. To ensure consistency across recording sessions, participants were re-instructed and re-trained at each visit to achieve and maintain the same target bite force range, rather than relying on performance from the previous session.

BC mVEMPs were elicited using a 500-Hz tone-burst stimulus delivered via a B-71 bone vibrator (RadioEar) connected to the Smart EP four-channel evoked potential recording system (v5.54.10, 2022; Intelligent Hearing Systems [IHS]). Stimuli consisted of 2-0-2 cycle tone bursts presented at a repetition rate of 5.1 stimuli/s with alternating polarity. The stimulus intensity was set at 125 dB SPL, as defined by the IHS system’s internal calibration for BC transducers. This level represents the system-defined electrical drive applied to the bone vibrator in accordance with manufacturer calibration protocols, rather than an absolute airborne sound pressure level or a force-referenced unit (e.g., dB re 1 μN).

The bone vibrator was calibrated using the IHS system’s built-in calibration procedure, and identical stimulus parameters were maintained across all participants and test sessions. The B-71 vibrator was positioned at a consistent mastoid location, and stable coupling was ensured throughout testing. Electrodes were placed following skin preparation with Nuprep gel. The non-inverting electrode was placed over the lower third of the masseter muscle, the inverting electrode over the midpoint of the zygomatic arch, and the ground electrode on the forehead. Electrode impedance was maintained below 5 kΩ. EMG activity was continuously monitored, and masseter muscle activation was maintained within a target range of 30–70 μV using real-time visual feedback.

Recordings were obtained from both ipsilateral and contralateral stimulation. Each recording included a 50-ms pre-stimulus baseline, a recording window sufficient to capture the mVEMP response, and a total of 200 artifact-free sweeps. To evaluate test–retest reliability, all participants underwent a repeat recording session within a 1-month interval.

Additionally, threshold determination was performed in a subset of 15 participants. This subset testing was undertaken to minimize participant fatigue and because threshold estimation was exploratory in nature. Thresholds were measured during a single testing session only and were not repeated in the second session that was used for test–retest reliability analysis.

The present study evaluated masseter VEMPs elicited exclusively using BC stimulation. AC mVEMP recordings were not obtained from the same participants. Therefore, any comparisons between BC and AC mVEMP findings discussed in this manuscript are based on previously published literature and not on within-subject AC–BC measurements.

In the present study, mVEMPs were consistently elicited in both ears using BC stimulation among 45 healthy young adults aged 18 to 35 years. A characteristic biphasic waveform, consisting of a positive peak (P1) followed by a negative peak (N1), was reliably observed across all participants in both test sessions. Parameters, including P1 and N1 latencies, peak-to-peak amplitude, and asymmetry ratio for both ipsilateral and contralateral stimulations, were compared between Session 1 and Session 2. The positive peak (P1) was consistently observed between 11 and 13 ms, while the negative peak (N1) appeared between 14 and 15 ms. These latencies fell within the expected range, with minor inter-subject variations. The P1 and N1 peaks were clearly identifiable in both ears across all participants, demonstrating excellent waveform reproducibility. The P1 and N1 latencies, peak-to-peak amplitude, and asymmetry ratio for Sessions 1 and 2 are presented in Table 1. Fig. 1 depicts the ipsilateral and contralateral mVEMP waveforms recorded from both ears.

The present study aimed to establish normative data, evaluate test–retest reliability, and determine response thresholds for BC-evoked mVEMPs elicited using a 500-Hz tone-burst stimulus at 125 dB SPL. mVEMP parameters were compared across two test sessions conducted within a 1-month interval. Consistent and reproducible responses were obtained across sessions in all participants. Notably, the positive peak (P1) was reliably identified between 11 and 13 ms, and the negative peak (N1) between 14 and 15 ms, forming a clear P1–N1 waveform complex that is indicative of vestibular origin [6]. Furthermore, all participants demonstrated the presence of the P1–N1 complex, reinforcing the vestibular basis of the recorded responses. The P11–N21 complex, also known as the vestibuloacoustic potential, has been described in earlier literature as a robust and commonly observed response in healthy individuals [12]. The parameters analyzed included P1 and N1 latencies, peak-to-peak (P1–N1) amplitude, and asymmetry ratio, obtained from both ipsilateral and contralateral recordings. The mean P1 and N1 latencies recorded in the present study align with previous studies on AC-evoked mVEMP [8,12,31]. However, these values differed from those reported in a recent study [15], possibly due to differences in bone vibrator models used (B-71 vs. B-81). In the present study, BC stimulus levels were reported using system-defined SPL units provided by the IHS evoked potential system rather than force-referenced calibration. While this ensured consistent within-study comparisons, it may limit direct comparability with studies using force-based calibration or newer transducers such as the B-81. Although vibrator placement and coupling were standardized, small variations in coupling force cannot be entirely ruled out and may have influenced amplitude measures. Future studies employing force-level calibration and updated BC transducers would strengthen cross-study comparability and generalizability. Despite this, the small sample size in their study limits the generalizability of their latency values. A reduction in N1 latency was observed, consistent with findings by Wang, et al. [32], who demonstrated that cVEMP responses to BC stimuli exhibit shorter latencies compared to AC stimuli presented at similar intensities. Therefore, the P1–N1 latency complex observed in the present study aligns more closely with the findings reported by Deriu, et al. [6].

In the present study, the peak-to-peak amplitude remained consistent across sessions with unilateral stimulation. These findings align with those of previous studies [6,33]. One earlier investigation recorded mVEMP responses using both unilateral and bilateral stimulation and reported no significant difference in the peak-to-peak amplitude between the two conditions [6]. Since normative data for BC-evoked mVEMP are not yet established, the normative values obtained in the present study were compared with those reported for mVEMP elicited through both AC and BC, as well as with cVEMP responses obtained using BC [33].

The present findings indicate that sex does not exert a systematic or clinically meaningful influence on BC-evoked mVEMP response characteristics. Across both test sessions, P1–N1 peak-to-peak amplitudes and asymmetry ratios were comparable between males and females, and the majority of latency measures did not differ significantly by sex. Although isolated latency differences were observed in Session 1—specifically, prolonged ipsilateral P1 latency and contralateral N1 latency in males—these effects were not consistently replicated across sessions or recording conditions, and all latency values remained within the overall normative range with substantial overlap between sexes. As such, these differences should be interpreted as minor and exploratory rather than indicative of a robust sex effect.

These results are broadly consistent with previous studies reporting minimal or inconsistent sex-related differences in mVEMP parameters. Vignesh, et al. [12] similarly found no significant sex differences in latency or asymmetry measures, with the exception of higher EMG-normalized amplitudes in males. In contrast, some investigations have reported earlier P1 and N1 latencies in females at certain stimulus intensities (e.g., De Natale, et al. [31]; Ravichandran, et al. [34]), suggesting that sex-related effects, when present, may be stimulus-dependent and protocol-specific. Importantly, such findings have not demonstrated sufficient magnitude or consistency to warrant routine adoption of sex-specific normative ranges.

Regarding asymmetry ratio, the absence of sex differences in the present study aligns with reports by Vignesh, et al. [12], although other authors have noted sex effects at select intensities (e.g., Kılınç, et al. [16]). Proposed explanations for these discrepancies include anatomical and physiological variations, such as cochlear length differences and trigeminal nerve morphology [35]. Nevertheless, the current data do not support a clinically relevant impact of these factors on BC-evoked mVEMP symmetry.

Taken together, the findings suggest that BC-evoked mVEMP parameters can be interpreted using pooled normative data, without the need for sex-specific reference values. The observed latency differences were small, inconsistent, and based on an uneven sex distribution, which further limits their generalizability. Future studies with larger, sex-balanced samples and effect-size reporting are needed to determine whether subtle sex-related differences in mVEMP latencies have reproducible clinical relevance.

The results of the present study showed no statistically significant interaural differences in any of the mVEMP parameters, including the latencies of P1 (ipsilateral vs. contralateral) and N1 (ipsilateral vs. contralateral), as well as the peak-to-peak P1–N1 amplitude (ipsilateral vs. contralateral). These findings are in agreement with those of Nagarajan, et al. [8], who also reported no significant differences in mVEMP amplitudes or latencies between ipsilateral and contralateral recordings when using both click and tone burst stimuli. The present findings corroborate the notion that mVEMP responses are mediated through a bilateral neural pathway, which explains the absence of consistent ear-related differences in healthy individuals. The lack of significant differences between left and right ear responses can be attributed to the bilateral organization of the vestibulo-masseteric reflex arc, wherein stimulation of one ear activates both masseter muscles through bilateral central projections from the vestibular nuclei to the motor nuclei of the trigeminal nerve [14,36].

Threshold estimation indicated that distinct mVEMP peaks were elicited at an optimal intensity level, below which no identifiable responses were observed in any participants. These findings are consistent with previous studies reporting that the BC mVEMP threshold occurs around 115 dB SPL—a level lower than that required for AC stimuli—highlighting the clinical advantage of BC stimulation, especially for individuals with conductive hearing loss [31,37], reduced neck muscle strength [10], and within the geriatric population [10,16,28].

Threshold estimation was performed in a limited subset of 15 participants and during a single session only, as it was exploratory in nature and intended to minimize participant fatigue. Accordingly, the small sample size and absence of repeated threshold measurements may limit the generalizability of the threshold findings. Future studies with larger samples are needed to establish reliable normative threshold values.

Wiener-Vacher, et al. [38] conducted a study to establish normative data for cVEMP using both BC and AC stimuli in typically developing children aged 6 months to 15 years. Their findings indicated that the average BC cVEMP threshold in healthy children was lower. Similarly, Welgampola, et al. [39] reported that cVEMP thresholds were generally lower when elicited by BC stimuli compared to AC stimuli.

A limitation of the present study is that AC and BC mVEMP responses were not obtained from the same participants. Consequently, direct within-subject comparisons between AC and BC stimulation could not be performed, and conclusions regarding BC stimulation as an alternative to AC stimulation are based on comparisons with previously published literature rather than on direct experimental evidence from the current cohort. Therefore, the findings of this study should be interpreted as establishing normative characteristics and test–retest reliability for BC-evoked mVEMPs, rather than demonstrating superiority of BC over AC stimulation.

While BC stimulation offers practical advantages in certain clinical conditions, such as conductive hearing loss or middle-ear pathology, the present results do not support definitive conclusions regarding its superiority over AC stimulation. Instead, BC-evoked mVEMP should be considered a complementary assessment tool within the broader vestibular test battery.

Future studies should include both AC- and BC-evoked mVEMP recordings obtained from the same participants, which would allow direct comparison of response characteristics, diagnostic sensitivity, and clinical applicability across stimulation modes. Such within-subject designs would provide stronger evidence to guide clinical decision-making regarding the relative utility of AC and BC stimulation in vestibular assessment.

Across the two test sessions, measurements of mVEMPs demonstrated consistent response patterns, with no statistically significant variations observed between sessions. Furthermore, ICC analysis indicated fair-to-good reliability for P1 and N1 latencies, P1–N1 amplitude, and the asymmetry ratio across both ipsilateral and contralateral recordings. Despite the consistent test–retest measurements obtained, no statistically significant differences were detected between the two sessions.

Previous investigations into the test–retest reliability of mVEMP have reported comparable findings. Ramesh, et al. [10] observed fair-to-good reliability for P11 and N21 latencies, while parameters such as peak-to-peak amplitude and asymmetry ratio demonstrated lower consistency. Their study concluded that mVEMP represents a reliable assessment tool. Similar repeatability has been documented for cVEMP responses to BC stimuli. Earlier studies reported higher ICC values for P1/N1 amplitude with BC stimulation, along with more consistent P1 and N1 latency measures across sessions [33,40], possibly due to the stable and reproducible placement of the B71 bone vibrator on the mastoid across trials. Supporting evidence was also provided by Welgampola, et al. [39], who reported an ICC for P1/N1 amplitude in response to BC stimulation, which aligns with the findings of the present study.

This study provides an initial investigation into the test–retest reliability of BC-evoked mVEMPs in healthy young adults. The results demonstrated fair reliability across two sessions, with no statistically significant differences in BC-evoked mVEMP parameters between sessions. Although minor and isolated sex-related latency differences were observed, these effects were inconsistent across sessions and did not substantially alter overall BC-evoked mVEMP response characteristics; therefore, the present findings support the use of pooled normative data rather than sex-specific reference values.

The amplitude ratios, P1–N1 amplitudes, and latencies all showed consistent performance, indicating that BC mVEMP has the potential to be a repeatable and useful measure in clinical settings. However, certain limitations of the study, including the use of the B71 RadioEar transducer due to limited equipment availability and the absence of standardized methods for maintaining consistent muscle contraction, may have influenced the outcomes. The possibility of muscle fatigue due to repeated clenching could also have contributed to variability in the results. To enhance the reliability of BC mVEMP recordings, future studies should consider the use of bite blocks or soft materials between the teeth during clenching to ensure uniform muscle activation. Expanding the study population to include a broader age range and individuals with vestibular disorders will help in establishing normative data and diagnostic relevance. Furthermore, comparative studies with other VEMP modalities and further exploration of factors such as anatomical variation, sex, and muscle fatigue will be essential to refine this method.

In conclusion, this study supports the feasibility and fair reliability of BC mVEMP testing in healthy young adults and highlights the need for continued research to optimize protocols and broaden clinical applicability.