Introduction
Noise exposure has been reported to affect the peripheral vestibular system negatively,Reference Golz, Westerman, Westerman, Goldenberg, Netzer and Wiedmyer1, Reference Kumar, Vivarthini and Bhat2 both temporarily and permanently. The mechanism of vestibular impairment has been postulated to be similar to noise-induced hearing loss. Lin and Ferucci demonstrated a 1.4-fold increase in the risk of falls for every 10-dB decrease in hearing sensitivity,Reference Lin and Ferrucci3 suggesting a relationship between the cochlear and vestibular end-organs. Morphological studies have ratified that the otolith-mediated pathway is more susceptible to acoustic-induced damage than the semicircular-canal pathways.Reference Stewart, Holt, Altschuler, Cacace, Hall and Murnane4
The susceptibility of vestibular damage following noise exposure is related to the anatomical proximity of the vestibular end-organs to the acoustic-energy delivery system,Reference Gioacchini, Alicandri-Ciufelli, Kaleci, Magliulo and Re5 the similarity of cellular ultrastructure between the cochlea and vestibular receptors, and the common blood supply of cochlear and vestibular organs.Reference Gale and Forge6 As the hair cells of the vestibular organs share similar morphology to the organ of Corti, they undergo a similar mechanism of transduction that elicits neural activity through the shared vestibulocochlear nerve, eight cranial nerve (CN VIII).Reference Stewart, Holt, Altschuler, Cacace, Hall and Murnane4 The sensitivity of vestibular end organs to sound has been demonstrated through numerous neurophysiological studies.Reference Curthoys, Grant, Pastras, Fröhlich and Brown7
Sohmer et al. postulated that intense impulse sound generates cochlear fluid pressures (condensations and/or rarefactions) through the stapes footplate vibration, which propagates through the inner-ear fluid, thus damaging the vestibular hair cells.Reference Sohmer, Elidan, Plotnik, Freeman, Sockalingam and Berkowitz8
The saccule and inferior vestibular nerve appear to be the most affected peripheral vestibular organs because of their proximity to the cochlea.Reference Golz, Westerman, Westerman, Goldenberg, Netzer and Wiedmyer1, Reference Hsu, Wang, Lue, Day and Young9
On the contrary, noise-induced vestibular and hearing loss has a distinct pathomechanism. Vestibular hair cells and cochlear hair cells are tuned differently.Reference Dessai, Bhat and Kumar10 Cochlear hair cells respond to frequencies between 20 and 20 000 Hz, while vestibular hair cells respond to inputs from 1 to 10 Hz.Reference Hain11
Noise-induced vestibular damage has been reported following both occupation-related exposureReference Tseng and Young12–Reference Viola, Scarpa, Pisani, Petrolo, Aragona and Spadera15 and recreational-induced exposure.Reference Dessai, Bhat and Kumar10, Reference Singh and Sasidharan16, Reference Liao and Young17 This study investigated vestibular function among healthy adolescents and young adult personal listening device users using cervical vestibular-evoked myogenic potential, video head impulse test (VHIT) and subjective visual vertical test (SVV) . The vestibular function was quantified using the cervical vestibular-evoked myogenic potential to assess saccule function, the subjective visual vertical test to assess utricle function and VHIT to assess semicircular canal function.
Methods
The study was conducted in two phases, with participants divided into two groups. The adolescent group included participants aged between 13 and 17 years, and the young adult group involved participants aged between 18 and 25 years.
Participants
Healthy participants aged 13–25 years with a history of earphone usage of more than 1 hour per day for more than 1 year were recruited. The study period was between January 2022 and January 2023. All included participants had normal hearing within the conventional frequencies. Exclusion criteria include participants with active otological complaints, known vestibular conditions and use of ototoxic medications. None of the participants had prior exposure to any other recreational or occupational noise exposure. They answered a survey sent via Google form (Supplementary 1) that aimed to understand their earphone listening behaviour. The self-reported listening volume was an estimated percentage of the volume each participant listens to daily. The participants were instructed to avoid using muscle relaxants for at least 48 hours before cervical vestibular-evoked myogenic potential testing.Reference Dessai, Bhat and Kumar10
Each participant underwent an otoscopic examination followed by tympanometry (Amplivox, Otowave 102, Oxford, UK) to determine the status of their external and middle-ear functions, followed by hearing assessment using pure tone audiometry.
Personal listening device-preferred listening level measurement (40-hour equivalent continuous exposure level)
The subjects’ preferred listening levels were determined based on the methods used in a prior study.Reference Sulaiman, Seluakumaran and Husain18 Each participant listened to a song (‘Just Like Heaven’ by The Cure, 128 kbps bit rate) from a Samsung Note 10 smartphone connected to a pair of earphones (Samsung AKG 3.5) in a quiet room.
The participants were instructed to set the volume to their usual listening level without considering their preference for the song and they were blinded from the volume indicator. The actual sound levels in dB Sound Pressure Level (SPL) of the test song were measured using the Knowles Electronics Manikin for Acoustics Research ear-and-cheek simulator (43AG, GRAS Sound and Vibration, Holte, Denmark) and converted into equivalent A-weighted free-field levels. The overall free-field A-weighted sound levels (L) were computed in dBA decibels (A-weighted). The 40-hour equivalent continuous exposure levels were calculated using the formula L + 10 log10(T/40), where L is the free-field corrected listening level (dBA) and T is the mean daily listening duration (in hours) calculated from the total listening duration of the subject in a week.
Vestibular assessments
All the participants were briefed regarding each procedure before starting. Each participant underwent three vestibular assessments: cervical vestibular-evoked myogenic potential to assess the function of the saccule, the subjective visual vertical test to assess utricle function and the Video Head Impulse Test to assess the individual function of semicircular canals and the presence of saccade.
Cervical vestibular-evoked myogenic potential
The cervical vestibular-evoked myogenic potential assessment was conducted following the standardised protocol and normative values established by Isaradisaikul et al.Reference Isaradisaikul, Strong, Moushey, Gabbard, Ackley and Jenkins19 Ipsilateral cervical vestibular-evoked myogenic potential responses were elicited using 500-Hz tone-burst stimuli presented at 98 dB normalised hearing level (nHL) with a repetition rate of 5 per second via insert earphones. A minimum of 200 sweeps were averaged per trial to ensure waveform reliability. The electromyography activity of the sternocleidomastoid muscle was monitored using visual feedback to maintain consistent muscle contraction during head rotation. Biphasic waveforms were identified as P1 (p13) and N1 (n23) based on their latencies, and reproducibility across consecutive runs confirmed response presence. Absence of reproducible P1-N1 waveforms indicated absent cervical vestibular-evoked myogenic potential responses.
Video Head Impulse Test
The Synapsys® VHIT Ulmer device (Synapsys, Marseilles, France) was used in this study to assess all three semicircular canals individually. The methodology used in this study was adapted from previous published studies.Reference Yilmaz, Ila, Soylemez and Ozdek20, Reference Liu, Leng, Zhou, Liu, Wang and Xia21
Vestibulo-Ocular Reflex (VOR) gain for each semicircular canal, the presence of saccade and the asymmetry between canals were determined. The Video Head Impulse Test software calculated the VOR gain according to the head velocity and eye velocity curves.Reference Yilmaz, Ila, Soylemez and Ozdek20 The VOR gain was considered abnormal if it was less than 0.8 for the lateral semicircular canal and less than 0.7 for the anterior and posterior semicircular canals.Reference Liu, Leng, Zhou, Liu, Wang and Xia21
Subjective visual vertical test
The subjective visual vertical test required participants to rotate a projected line until it was aligned with their perception of the true vertical. The test was administered in a dark room using the application Visual Vertical (Clear Health Media, Wonga Park, Australia) on an iPad mini (Apple, Cupertino, CA, USA), which was secured to the bottom of a plastic bucket with two strips of Velcro. The application presented a red line on a black background. The participant rotated the bucket until the red line was aligned with their perception of the true vertical. The subjective visual vertical test was considered abnormal if the score was more than 2.13.Reference Brodsky, Cusick, Kenna and Zhou22
Ethics approval
The study was approved by the University Malaya Medical Centre Ethics Committee (MREC ID NO: 20211121-10791) and was performed in accordance with the Declaration of Helsinki. All individuals completed and signed an informed consent form before inclusion in the study.
Statistical analysis
Statistical analyses were conducted using IBM SPSS version 26.0. The distribution of continuous data was checked using the normality test for skewness, kurtosis and histogram. Continuous data were presented as mean ± standard deviation if normally distributed, otherwise the median (25th percentile, 75th percentile) was used. Categorical variables were presented as frequency and percentage.
The relationship between the level of personal listening device usage and vestibular functions was explored using the Spearman correlation test, while the difference in the level of personal listening device usage between patients with normal and abnormal vestibular functions was tested using the Mann–Whitney U test. One-way analysis of variance was used to explore the difference in the level of personal listening device usage, duration of usage and age of first using a personal listening device between patients who experienced vestibular symptoms. All the tests were two-sided and statistical significance was denoted as p less than 0.05. The Pearson correlation test was performed. If the Pearson correlation coefficient, r, was close to 1, then the correlation between variables was considered significant and strongly correlated.
Results
A total of 131 participants were recruited in this study, with a mean age of 20 ± 2.55 years. The majority of the participants were of Chinese ethnicity (49.6 per cent), with a slight female predominance (59.5 per cent). Underlying co-morbidities were noted in 8.4 per cent of the participants.
Besides online classes (100 per cent), the majority of the participants used personal listening devices for listening to music (91.6 per cent), followed by (23.7 per cent) playing games and (18.3 per cent) watching videos. The preferred music types included pop (78.6 per cent), ballad (29.8 per cent), classical (19.8 per cent) and rock and/or hip-hop and/or metallic (7.6 per cent). The mean age when participants started using a personal listening device was 14.02 ± 2.91 years (12.13 ± 2.21 years for adolescents and 14.57 ± 2.87 years for young adults). The mean duration of personal listening device usage per day was 5.53 ± 2.76 hours, with a slightly higher duration seen in the adolescent group (6.03 ± 0.99 hours) compared with the young adult group (5.39 ± 2.69 hours) (p = 0.261). The mean total years of personal listening device usage was 6.65. In the adolescent group, the age when a personal listening device was first used was earlier (12.13 ± 2.21 years) compared with the young adult group (14.57 ± 2.87 years) (p < 0.001).
As for the type of headphones, in-the-ear headphones were most favoured by both adolescents and young adults (53.4 per cent), followed by on-the-ear headphones (25.2 per cent), ear pods (19 per cent), over-the-ear headphones (5.3 per cent) and bone conduction headphones (1.5 per cent). The preferred volume of personal listening device among adolescents and young adults was 46.76 ± 17.38 % (p = 0.389). The preferred volume level was higher among adolescents (49.17 ± 15.37%) compared with young adults (46.04 ± 17.93) (p = 0.389).
The reported prior vestibular symptoms were higher in the adolescent group than in the young adults (43.3 vs 39.6 per cent) (p = 0.715). The estimated sound exposure level Laeq40h was higher in the adolescent group (58.16 ± 6.51) compared with the young adult group (55.89 ± 6.51) (p = 0.095).
Cervical vestibular-evoked myogenic potential
For the cervical vestibular-evoked myogenic potential, the mean latency of P13 of the right ear was 17.27 ± 2.34 ms, whilst the mean latency of P13 of the left ear was 17.37 ± 2.11 ms. The N23 ratio of the right ear was 23.45 ± 2.70 and of the left ear was 23.48 ± 2.48. The mean amplitude between both groups was slightly higher on the right (44.73 ± 24.26 ms) compared with the left (41.13 ± 23.68 ms). No statistical significance was noted between the adolescent and young adult cervical vestibular-evoked myogenic potential latency, amplitude and asymmetry ratio.
Absent cervical vestibular-evoked myogenic potential was noted in 15 participants, predominantly in the young adult group (73 per cent). Right and left amplitude were lower in adolescents than in young adults, whilst p13 and n23 latency were comparable Table 4. There was a positive correlation between the 40-hour equivalent continuous exposure level (r = 0.406, p = 0.029) and right p13 in the adolescent group (Tables 1 and 4. Figures 1 and 2) show a correlation plot between the 40-hour equivalent continuous exposure level and right and left P13. No correlation was noted among the young adults (Tables 2 and 5).
Association between the 40-hour equivalent continuous exposure level (Laeq40) and the right p13.
Characteristics of the participants
SD = standard deviation.
Association between the 40-hour equivalent continuous exposure level (Laeq40) and the left p13.
Characteristics of personal listening device usage
SD = standard deviation.
Subjective visual vertical test
The mean subjective visual vertical test score was higher in both groups (3.17 ± 1.56°) than the normative level (2.3°). There was no statistically significant correlation between the 40-hour equivalent continuous exposure level and the mean subjective visual vertical test score (p > 0.05).
VHIT
The VHIT revealed a saccade in 22.1 per cent of the participants (26.7 per cent in the adolescent group and 20.8 per cent in the young adult group) (p = 0.496). Each canal’s mean VOR gain value was within the normal level for both the adolescent and young adult groups.
The relationship between the sound exposure level of personal listening device usage and the vestibular function was explored using the Spearman correlation test (Table 3). No statistical significance was noted when cervical vestibular-evoked myogenic potential, and subjective visual vertical test score were compared with individual sound exposure levels.
Vestibular functions
a Independent sample t-test.
b Pearson chi-square test.
c Fisher exact test.
d Mann–Whitney U test. Laeq40 = 40-hour equivalent continuous exposure level; SD = standard deviation.
Correlation between the Laeq40 and the cervical vestibular-evoked myogenic potential in the adolescent group (age ≤18 years)
Laeq40 = 40-hour equivalent continuous exposure level.
Correlation between the Laeq40 and the cervical vestibular-evoked myogenic potential in the young adult group (age ≥18 years)
Laeq40 = 40-hour equivalent continuous exposure level.
Following the cervical vestibular-evoked myogenic potential, the subjective visual vertical test revealed abnormal results in 15 per cent of the participants. The Video Head Impulse Test function showed abnormal results that affected the posterior semicircular canal (12 per cent), whereas the lateral and anterior canals were equally affected (6.5 per cent).
Differences in the level of personal listening device usage between patients with normal and abnormal vestibular functions were explored. No significant difference was observed between the sound exposure level and subjective vestibular symptoms (p > 0.05).
Discussion
Behaviour and personal listening device usage
We found that adolescent personal listening device users exhibited higher usage of personal listening device, with regard to total hours of usage per day, and were reported to have started using a personal listening device at an earlier age. As for the total years of personal listening device use, the mean personal listening device usage was 6.67 years. The mean age at which personal listening device use was started was 14.73 years.
Studies investigating college student behaviour and habits towards personal listening device usage have demonstrated that most students started using a personal listening device in the fourth to sixth grade of elementary school. You et al. found that 77 per cent of college students had used a personal listening device for more than 2 years, with average daily use between 0.5 and 4 hours.Reference You, Kwak and Han23 The duration of personal listening device usage per day in our study was an average of 5.57 hours per day, which is higher than most studies involving college students. Nonetheless, this could be attributed to online classes and lockdown period during the coronavirus disease 2019 pandemic.
The main reason for personal listening device usage in our study was online classes (100 per cent), followed by listening to music (91 per cent). In a study among Korean college students, personal listening devices were mainly used to listen to music (95.3 per cent). Moreover, these students considered personal listening devices to be convenient (80.1 per cent) and commonly used them when they were feeling bored (77.4 per cent), to help them relax (59.9 per cent), to help them concentrate (54.7 per cent) and to isolate themselves from others (46.2 per cent).Reference You, Kwak and Han23
The type of headphones used was not associated with a higher listening level of preferred volume in our study. Hodgetts et al. found that listeners reported a higher listening level with earbud-type earphones, especially in a noisy environment.Reference Hodgetts, Rieger and Szarko24 Earbuds have been demonstrated to produce a greater output level because of their proximity to the tympanic membrane.Reference Shim, Lee, Koo and Kim25 The maximum output levels measured demonstrated that earbuds produced 100.8 dBA, in-the-ear earphones 102.3 dBA and over-the-ear earphones 96.7 dBA.Reference Portnuff, Fligor and Arehart26 Numerous studies have shown that the over-the-ear earphones have a lower preferred listening level.Reference Hodgetts, Rieger and Szarko24, Reference Henry and Foots27–Reference Stewart, Yu, Huang, Maklad, Tang and Allison29
Noise exposure and vestibular function
It has been demonstrated that the same noise exposure which damages the cochlear hair cells can also damage the vestibular hair cells of the saccule, utricle and semicircular canal.Reference Tseng and Young12, Reference Wang, Hsu and Young30 Several studies have demonstrated that the saccule is more prone to noise-induced damage than the utricle and semicircular canal.Reference Wang, Hsu and Young30, Reference Wang and Young31 Embryologically, the saccule arises from the pars inferior, similar to the cochlea, while the utricle and semicircular canal arise from the pars superior.Reference Hara and Kimura32 The presence of membrane limitans, which separates the pars inferior from the pars superior, can protect the utricle and semicircular canal from the damaging effects of noise exposure.Reference Hara and Kimura32 Moreover, the anatomical proximity of the saccule to the stapes footplate has been attributed to the saccule’s susceptibility to early damage from noise.Reference Akdogan, Selcuk, Take, Erdogan and Dere33
Abnormal cervical vestibular-evoked myogenic potential and noise exposure
Cervical vestibular-evoked myogenic potential is a promising tool for detecting noise-induced vestibular damage, as demonstrated in several studies.Reference Tseng and Young12–Reference Cassandro, Chiarella, Catalano, Gallo, Marcelli and Nicastri14, Reference Ostri, Eller, Dahlin and Skylv34 Abnormal cervical vestibular-evoked myogenic potential findings have been reported in human subjects following noise exposure.Reference Tseng and Young12–Reference Liao and Young17 Similarly, animal-related morphological studies advocate that the saccule is particularly prone to noise and vestibular-related damage.Reference Stewart, Holt, Altschuler, Cacace, Hall and Murnane4 Reported cervical vestibular-evoked myogenic potential abnormality following noise exposure includes reduced amplitude, prolonged latenciesReference Stewart, Holt, Altschuler, Cacace, Hall and Murnane4 and elevated threshold.Reference Akin, Murnane, Tampas, Clinard, Byrd and Kelly13, Reference Viola, Scarpa, Pisani, Petrolo, Aragona and Spadera15 Tseng and Young demonstrated that 70 per cent of the included cohort with noise-induced hearing loss had abnormal cervical vestibular-evoked myogenic potential.Reference Tseng and Young12
Dessai et al. reported reduced amplitude among young personal listening device users (aged between 18 and 29 years) with normal hearing.Reference Dessai, Bhat and Kumar10 Moreover, the authors demonstrated a reducing trend of the peak-to-peak amplitude of p1-n1 with an increase in years of personal listening device usage.Reference Dessai, Bhat and Kumar10 In another study by Singh and Sasidharan that analysed the sacculo-colic pathway among personal listening device users, comparable results were noted, but no statistical significance was reported.Reference Singh and Sasidharan16 This could be attributed to the irritative phase, which may not be long enough to cause a significant reduction in peak-to-peak amplitude.
This study found abnormal cervical vestibular-evoked myogenic potential in 61.5 per cent (59 per cent right and 64 per cent left). The right and left amplitudes are lower in adolescents, although this difference is not statistically significant (p > 0.05). We also found that the duration of personal listening device usage per day was higher in adolescents than in young adults (p > 0.05). We postulate that prolonged personal listening device usage could lead to reduced amplitude. We found a positive correlation between the 40-hour equivalent continuous exposure level (r = 0.406, p = 0.029) and the right p13 among the adolescent group. Latency abnormality from noise exposure postulates that neural conduction response along the sacculocolic pathway is affected.
Subjective visual vertical test and noise exposure
The otolith organ, especially the utricle, plays a key role in determining the perception of gravitational vertical. The subjective visual vertical test’s role as a marker for vestibular compensation following deafferentation among acute vestibular injury has been documented.Reference Totilienė, Uloza, Lesauskaitė, Damulevičienė, Kregždytė, Kaski and Ulozienė35, Reference Li, Xu, Li, Gao, Shang and Tian36 Very few studies have examined the extent of subjective visual vertical test involvement among subjects exposed to chronic noise.Reference Das, Kalidoss, Bakshi and Ramesh37 In our study, we found no statistical significance between the subjective visual vertical test scores for the adolescent and young adult groups (p > 0.05), which could indicate that utricle damage from noise exposure may require more prolonged exposure.
VHIT and noise exposure
We found no statistical significance between the VOR gain of all six semicircular canal and personal listening device usage. Yilmaz et al. reported a statistical significance between VOR gain among patients exposed to industry-related noise exposure with hearing loss, with impairment found in the unilateral lateral semicircular canal in 68.2 per cent of patients and in the bilateral lateral canal in 27.2 per cent of patients,Reference Yilmaz, Ila, Soylemez and Ozdek20 with none of the subjects with canal deficits complaining of vertigo due to the compensatory mechanism of the central nervous system.Reference Yilmaz, Ila, Soylemez and Ozdek20
Strengths and relevance of this study
Despite existing publications on the effects of personal listening device usage on vestibular and balance function, the field remains insufficiently underexplored. This study contributes valuable insights by demonstrating that listening behaviours, along with the characteristics of personal listening device usage, can be associated with vestibular impairment in this population.
Limitation
There are several limitations to this study. The college students may have underestimated the preferred volume level, leading to an overall sound exposure level that was higher than intended. The vestibular assessment did not include other relevant investigations, such as the ocular vestibular-evoked myogenic potential to assess the utricle and the caloric test to assess the low-frequency semicircular canal, which could have been more representative.
• Noise exposure from PLD usage results in subjective vestibular impairment both in adolescents and young adults
• PLD usage causes saccular dysfunction evidently by an increase in p13 latency
• Adolescents have higher sound exposure levels compared to young adults
• Noise exposure can result in vestibular impairment before causing hearing impairment
• Vestibular assessment should be performed together with hearing assessment when assessing the effect of noise exposure
Conclusion
Personal listening device exposure results in vestibular damage, notably saccule damage. Cervical vestibular-evoked myogenic potential latency was reported to be affected by listening levels among adolescents.
Acknowledgements
The authors acknowledge funding from a UMSC C.A.R.E Research Grant ((UMG005C-2023).
Competing interests
None declared.
