Acute hearing is vital to musicians, who must rely on their ears to learn, fine-tune, express, and synchronize to music. Given the importance of hearing for musicians, it is perhaps not surprising that increased sound acuity has been observed in musicians across the auditory system. This phenomenon has been observed in various forms, from heightened ability to detect subtle nuances in sounds to fewer age-related changes to central auditory processing and heightened neural acuity to sound, in both active musicians and people with past musical training (past musicians) (Alain et al., 2014; Kraus et al., 2017; Yeend et al., 2017). Yet, as a group, musicians are also at a higher risk of hearing health problems (Jansen et al., 2009). Indeed, musicians at all training levels experience higher noise exposure than the general population (McBride et al., 1992; Miller et al., 2007; Tufts and Skoe, 2018), leading to an increased incidence of noise-induced hearing loss (NIHL) (Jansen et al., 2009). Most types of NIHL are caused by irreversible damage to cochlear hair cells and associated structures (cf. stereocilia) (Chen and Fechter, 2003). Outer hair cells (OHCs), one of two types of cochlear sensory receptors, help amplify low-intensity signals, giving them a vital role in expanding the dynamic range of human hearing (Davis, 1983). However, OHCs are highly susceptible to noise-induced damage, which often affects a particular region of the cochlea and results in frequency-specific hearing loss (Henderson et al., 2006). A loss of hearing acuity is a common clinical concern for musicians, as it can threaten their ability to perceive subtle musical elements, which may, in turn, increase mental distress and decrease social and occupational opportunities to perform (Vogel et al., 2014; Heckman et al., 2021).

In clinical settings, audiological measures of NIHL are commonly limited to pure-tone audiometry. This approach finds the lowest sound intensity a person can hear (their threshold) across a set of frequencies. When hearing is measured through audiometry, NIHL is observed as a loss of sensitivity to frequencies in the 3-6 kHz range (McBride and Williams, 2001). Yet standard audiometry typically only tests as high as 8 kHz, well below the upper-frequency range of the human ear. When extended high-frequency (EHF) audiometry protocols are used to test the upper frequency range of hearing, decreased sensitivity to frequencies greater than 10 kHz are also common in noise-exposed populations (Wang et al., 2000).

While audiometry is currently deemed the gold standard in clinical hearing testing (Durch et al., 2006), it is subjective, requiring the listener to indicate whether they can hear the tone being played. This subjectivity complicates its use in young children or persons with cognitive impairments who may not understand the task (Naito, 2004). But even in healthy young adult populations, audiometry can be subject to response bias and false and exaggerated hearing loss (Peck, 2011). Objective measures are therefore an attractive alternative to conventional audiometry. One such method is otoacoustic emissions (OAEs) testing, an objective measure of OHC function that does not require a behavioral response, limiting the human perceptual and behavioral errors associated with audiometry (Lonsbury-Martin et al., 1991). OAEs are low-intensity (typically below audiometric threshold) sounds (i.e., emissions) that can be detected with a sensitive microphone in the ear canal as the result of non-linear cochlear activity (Kemp, 1978). The strength of the emissions, measured in decibels (dB) and/or signal to noise ratio (SNR), is a reliable indicator of OHC health (Lonsbury-Martin et al., 1991). OAEs can be elicited by various stimuli, with transient and tonal stimuli being common. Transient-evoked OAEs (TEOAEs) use a repeating transient sound such as a click or chirp tone burst to initiate broadband OHC activity. For distortion-product OAEs (DPOAEs), a tonal stimulus consisting of two frequencies is used; this stimulus complex sound produces emissions at non-stimulus frequencies that provide more frequency-specific indications of OHC activity. For both TEOAEs and DPOAEs, weak, absent, or low SNR OAEs may indicate NIHL. (There are some rare auditory pathologies, however, where OAEs are enhanced [Cheatham et al., 2014; El-Badry and McFadden, 2009]). With respect to NIHL and presbycusis, it has been proposed that OAEs provide an earlier indicator of hearing loss than audiometric thresholds (Abdala and Dhar, 2012; Hamdan et al., 2008), although supporting evidence is mixed (Jansen et al., 2009). Nevertheless, the earlier that precursors to NIHL can be detected, the more easily measures can be implemented to slow the progression of irreversible damage to the auditory system.

To help in the prevention of NIHL and to identify individuals at increased risk, various tools have been developed to estimate noise exposure. A quick method for estimating a person's noise exposure is to use a questionnaire, such as the Noise Exposure Questionnaire (NEQ) (Johnson et al., 2017). The NEQ is a roughly 5-minute survey on noise exposure that asks participants how often they engaged in different noisy activities such as power tool use, sporting events or nightclub attendance, musical instrument practice, music listening over earphones, and noisy jobs over the past 12 months. Responses to the NEQ are scored into an annual noise exposure estimate (Johnson et al., 2017). The NEQ has been used by other groups to assess noise exposure in a variety of populations, from college students to military personnel (Bernard et al., 2019; Bhatt, 2017; Brungart et al., 2019; Grinn et al., 2017; Washnik et al., 2020, 2016). Several recent studies have used the NEQ to evaluate noise exposure in student musicians (Washnik et al., 2020, 2016). A more time-intensive but objective approach to estimating personal noise exposure is personal noise dosimetry, which involves a small body-worn recording device that logs environment sound levels. In research on musicians, dosimetry measurements are often conducted during rehearsals or performances, which restricts dosimetry to a period of a few hours (Miller et al., 2007) or individual days (Washnik et al., 2016). Extending the dosimetry recording window to a longer period, a week in our case, can yield greater insight into routine exposure, from social or other activities, that may occur outside of scheduled rehearsals (Tufts and Skoe, 2018). Since each of these noise exposure measures has its own limitations, using both an objective and subjective measure can provide greater insight into individual noise exposure profiles and therefore better predict the risk to the health of outer hair cells.

Studies of outer cell function have yielded mixed results when musicians are compared to non-musicians. Some have shown weaker or more variable OAE amplitudes in musicians (Hamdan et al., 2008; Høydal et al., 2017), presumably due to subclinical loss from increased noise exposure. Yet others have found no difference in OAEs amplitudes between musicians and non-musicians (Couth et al., 2020; Henning and Bobholz, 2016; Liberman et al., 2016; Møllerløkken et al., 2013; Reuter and Hammershøi, 2007), with some even finding a trend for musicians to have stronger OAEs on average (although to our knowledge there are no reports of this trend reaching a statistical significance) (Perrot et al., 1999; Wang et al., 2019). In studies that have recorded OAEs in the presence of a suppressor stimulus, musicians have been found to have finer cochlear tuning (Bidelman et al., 2016) and also stronger medial olivocochlear reflexes (MOCRs) (Bidelman et al., 2017; Brashears et al., 2003; Micheyl et al., 1997, 1995; Perrot et al., 1999; Perrot and Collet, 2014). (In the MOCR paradigms, reflexes are measured as the difference between the OAE amplitude with and without a contralateral stimulus, with a greater a difference indicative of stronger reflexes).

The heterogeneity in findings may be due to differing methodology and recruitment approaches. Some studies comparing OAEs between musicians and non-musicians focused on specific types of musicians such as rock musicians (Høydal et al., 2017) or professional musicians (Hamdan et al., 2008), whereas others did not target a specific group of musicians (Couth et al., 2020). Studies also varied in their sample sizes, with some having limited sample sizes (n = ∼12-30: e.g., Henning and Bobholz, 2016; Møllerløkken et al., 2013; Reuter and Hammershøi, 2007) and others having considerably larger sizes (n = ∼100-130, e.g., Couth et al., 2020; Emmerich et al., 2008; Yeend et al., 2017; n > 300, e.g., Jansen et al., 2009). Those with larger sample sizes generally have a wide age span (∼20-60 years) creating the potential for confounding age effects (e.g., Emmerich et al., 2008; Jansen et al., 2009). While some studies measured noise exposure objectively via dosimetry (e.g., Emmerich et al., 2008) or through extensive interviews (e.g., Couth et al., 2020), others assumed higher levels of noise exposure in musicians without directly measuring it (Liberman et al., 2016).

To address the limitations of this prior work, we recruited a large sample of similarly aged college students (ages 18-23), all attending the same university and having audiometric thresholds in the clinically normal range (≤ 25 dB HL for frequencies up to 8 kHz). The study most comparable to the current study in size and age range is from Couth et al., 2020 (18-27 years), although they included a small number of musicians with unilateral hearing loss in the sample. To define “musicians”, we followed the “six-year criterion” established by Zhang and colleagues (2020). In their meta-analysis of 90+ papers published between 2011-2017, Zhang and colleagues found a consensus point in the literature for “musicians” to be defined as having at least 6 years of musical experience. To encourage more standardization across studies and promote replicability, we followed this guideline and defined musicians as anyone with 6 or more years of musical training and non-musicians as anyone with 0-5 years of training.

The most novel element of our study is the inclusion of both active and past musicians in the musician sample. Including past musicians in the musician sample allowed us to examine differences in OAEs associated with musicianship while minimizing the effects of ongoing/recent noise risks to the auditory system from loud music activities. In all participants, we estimated annual noise exposure levels using the NEQ, and in the active musicians and a small subset of the other participants, we used a combination of the NEQ and one week of personal noise dosimetry. In all participants, we also measured TEOAEs to a broadband chirp stimulus. This gave us an estimate of OHC integrity in the 1-5 kHz range, a range overlapping the frequency band where NIHL often presents in audiometry. Results over this frequency range were also investigated in a subset of participants on whom DPOAE data was available. With the DPOAE protocol, emissions in the extended high frequency range—i.e. those not captured by the TEOAE protocol or the audiometry—were also available for exploratory examination. We predicted that active musicians would, as a group, have higher noise exposure estimates based on the NEQ and personal dosimetry than non-musicians. With respect to OAEs, we predicted that if any effects did emerge, OAE emissions would be weaker in the musician group by virtue of having higher levels of lifetime noise exposure.