Beyond reduced audibility, there is convincing evidence that the auditory system adapts according to the principles of homeostatic plasticity in response to a hearing loss. Such compensatory changes include modulation of central auditory gain mechanisms. Earplugging is a common experimental method that has been used to introduce a temporary, reversible hearing loss that induces changes consistent with central gain modulation. In the present study, young, normal-hearing adult participants wore a unilateral earplug for two weeks, during which we measured changes in the acoustic reflex threshold (ART), loudness perception, and cortically-evoked (40 Hz) auditory steady-state response (ASSR) to assess potential modulation in central gain with reduced peripheral input. The ART decreased on average by 8 to 10 dB during the treatment period, with modest increases in loudness perception after one week but not after two weeks of earplug use. Significant changes in both the magnitude and hemispheric laterality of source-localized cortical ASSR measures revealed asymmetrical changes in stimulus-driven cortical activity over time. The ART results following unilateral earplugging are consistent with the literature and suggest that homeostatic plasticity is evident in the brainstem. The novel findings from the cortical ASSR in the present study indicates that reduced peripheral input induces adaptive homeostatic plasticity reflected as both an increase in central gain in the auditory brainstem and reduced cortical activity ipsilateral to the deprived ear. Both the ART and the novel use of the 40-Hz ASSR provide sensitive measures of central gain modulation in the brainstem and cortex of young, normal hearing listeners, and thus may be useful in future studies with other clinical populations.
The auditory system, like other sensory systems, has an intrinsic ability to adapt to changes in peripheral input. Adaptation to changes in acoustic input is reflective of central auditory system homeostatic plasticity and has been studied in the context of such processes as loudness perception and intensity discrimination (McKinney et al., 1999; Munro, 2008; Munro and Trotter, 2006; Olsen et al., 1999; Philibert et al., 2002; Robinson and Gatehouse, 1995). While central auditory plasticity ideally functions to maintain stable auditory perception, it also may lead to undesirable consequences. For instance, auditory disorders such as tinnitus (phantom sound perceptions, in the absence of an external source of sound) and hyperacusis (abnormal loudness growth and suprathreshold sound sensitivity) have been associated with maladaptive processes in the central auditory pathway, often following a hearing loss (Gu et al., 2010; Noreña, 2011; Noreña and Farley, 2013; Schaette and Kempter, 2009, 2006; Schaette and McAlpine, 2011). An understanding of the adaptive and maladaptive mechanisms associated with plasticity will help improve auditory models that incorporate plasticity effects, shape perspectives on targeted therapies for management of disorders like tinnitus and hyperacusis, and inform clinical guidelines regarding the timeline and expectations for rehabilitation.
A common approach to simulate the sound deprivation effects of peripheral hearing loss in humans has been the use of short term, continuous attenuation through earplugging. Earplugging introduces a temporary, reversible hearing loss and has been useful for investigating subsequent changes in auditory perception and processing in the central auditory pathway (Brotherton et al., 2017, 2016; Formby et al., 2003, 2007; Maslin et al., 2013; Munro and Blount, 2009; Munro et al., 2014). Formby et al. (2003) demonstrated the effects of acoustic attenuation via bilateral earplugging based on changes in categorical loudness judgements as assessed by the Contour Test of Loudness Perception (Contour test; Cox et al., 1997). The Contour test involves presentation of pulsed warble tones ascending in level to a listener who judges the loudness of the tones by reporting one of seven numerical values corresponding to a continuum of categories from “Very soft” through “Uncomfortably loud.” After two weeks of continuous acoustic attenuation from wearing bilateral earplugs, subjects rated the same sounds as being louder than before earplug use. Formby et al. (2003) reported that study participants matched their pre-treatment loudness category ratings at sound levels 5 to 9 dB lower for the higher loudness categories (i.e., Categories 4 to 7), indicating that continuous partial deprivation of acoustic input led to a progressive increase in loudness perception over time, ostensibly attributed to central auditory plasticity or gain. Formby et al. (2007) subsequently replicated their findings in a second study in which participants wore earplugs bilaterally for a 4-week treatment period. Munro et al. (2014), using unilateral earplugging for 1 week, demonstrated a similar reduction in loudness perception in the plugged ear of 5 dB for low (500 Hz) and 8 dB for high (2000 Hz) frequency warble tones.
Central auditory plasticity associated with the adaptation to acoustic attenuation is consistent with homeostatic plasticity; a process thought to regulate the functional stability of neural activity in response to a persistent alteration in auditory input. Homeostatic mechanisms have been demonstrated in vitro as a selective strengthening and weakening of synapses that serve to stabilize neuronal activity while maintaining neural network selectivity (Turrigiano, 1999, 2012). Homeostatic plasticity in the intact brain is not likely due to a simple synaptic mechanism but rather a complex interaction among multiple forms of excitatory, inhibitory, and intrinsic plasticity that operate at multiple neuronal levels and on different temporal and spatial scales to ensure compensatory balance across neural circuits (Turrigiano, 2012). The functional consequence of homeostatic plasticity is a change (increase or decrease) in the potentiation of neural activity in the central pathway, referred to as “central gain,” and often described as “an internal volume control” (e.g., Brotherton et al., 2015). Physiological evidence of central gain modulation in the auditory system has been shown in animal models following cochlear damage as increased spontaneous activity in the cochlear nucleus and inferior colliculus (e.g., Kaltenbach et al., 2002; Manzoor et al., 2013), and auditory cortex (e.g., Noreña and Eggermont, 2003), and as enhanced stimulus-driven activity in the central pathway and cortex (e.g., Chambers et al., 2016; Hickox and Liberman, 2014; Salvi et al., 2000). While an internal volume control seems appealing in the case of hearing loss, excessive central gain of spontaneous activity, as well as potential contributions from stochastic resonance, may partially explain the perception of tinnitus (Hébert et al., 2013; Noreña, 2011; Noreña and Farley, 2013; Schaette and Kempter, 2008, 2006; Schaette and McAlpine, 2011; Schilling et al., 2021) and reduced sound tolerance in the case of hyperacusis (Gu et al., 2010; Knudson et al., 2014; Noreña and Farley, 2013). Though there is ample evidence of such gain processes in animal models, there is less understanding about the nature and extent of such a self-regulating gain mechanism in the human central auditory system. While some studies indicate that multiple gain mechanisms may be at work in the human auditory pathway (Maslin et al., 2013a, 2013b; Munro et al., 2014; Parry et al., 2019), it is not clear to what extent each mechanism contributes to the purported gain control following acoustic deprivation. The present study was motivated by these yet unanswered questions as well as the need for identifying reliable measures of central auditory gain modulation for future clinical application.
Although the adaptive plasticity of behavioral loudness judgments in response to earplugging is likely dependent on higher level (i.e., cortical) gain processes within the auditory system, subcortical evidence of gain-related plasticity has been demonstrated in the brainstem following earplugging. In rats, Clarkson et al. (2016) showed that unilateral earplugging for 10 days lead to poorer ABR thresholds but increased relative gain in the cochlear nucleus and lateral lemniscus, as demonstrated by increased ABR amplitude ratios for waves II/I and IV/I. Notably, following earplug removal and 10-days of recovery, the gain increase persisted in the lateral lemniscus but returned to baseline in the cochlear nucleus. In humans, the acoustic reflex threshold (ART) has been used to document changes in brainstem neural function following auditory attenuation via unilateral earplugs (Brotherton et al., 2017, 2016; Maslin et al., 2013; Munro and Blount, 2009; Munro et al., 2014) as well as following auditory enhancement with monaural amplification (Munro et al., 2007; Munro and Merrett, 2013). The acoustic reflex engages a brainstem pathway that includes ipsilateral and contralateral brainstem nuclei (i.e., CN, SOC) and terminates with bilateral contraction of the stapedius muscle in the middle ear (Borg, 1973; Pascal et al., 1998). Activation of this reflex pathway is measured with an ear canal probe system that detects a brief change in the admittance of the middle ear system and can be elicited with a range of stimuli (e.g., pure tones, broadband noise) and sound levels. The ART corresponds to the lowest sound pressure level that elicits a reliable reflex. The earplugging studies have shown that prolonged acoustic attenuation leads to a decrease in the minimum sound pressure level required to elicit the ART relative to baseline measurements (Brotherton et al., 2017, 2016; Maslin et al., 2013; Munro and Blount, 2009), consistent with an increase in central gain. Conversely, enhanced auditory input via short-term and long-term experience with monaural amplification results in an increase in the ART (Munro et al., 2007; Munro and Merrett, 2013), consistent with reduced central gain. While these changes in ART suggest that central gain changes are detectable at an early processing stage in the auditory pathway, it has not been clearly established in humans whether these effects are maintained or further modulated at higher (i.e., cortical) levels of auditory processing in the same individuals (e.g., Maslin et al., 2013b).
Investigations in animal models have shown that reduced peripheral input due to permanent conductive hearing loss can disrupt cortical function (Xu et al., 2007) without altering brainstem function (Yao & Sanes, 2018). Human subjects with permanent unilateral conductive hearing loss, on the other hand, have shown increased cortical activity (Parry et al., 2019) but reduced or altered hemispheric asymmetry (Maslin et al., 2013a). Moreover, human subjects with temporary, reversible conductive hearing loss from earplugging have shown no change in cortical measures with functional magnetic resonance imaging (fMRI) following treatment (Maslin et al., 2013b). Thus, the cortical effects of peripheral deprivation on central gain modulation in the cortex, in auditory systems with otherwise normal cochlear function, is less than clear, and may be due to differences in measures used to quantify central gain changes across studies.
Some evidence of cortically-mediated gain may be gleaned from changes in auditory middle latency responses (MLRs), reflecting contributions from thalamocortical sources (Musiek and Nagle, 2018), that were measured pre- and post-treatment for mildly hyperacusic hearing-impaired persons using bilateral sound generators over several months (Formby et al., 2017c). In this study, participants simultaneously achieved significant incremental changes in their loudness judgements over the treatment period as well as improvements in MLR latency measures. These promising but limited findings suggest the possibility of identifying neural correlates of central-gain mediated loudness growth in the cortex.
Electrophysiological evaluation of loudness growth in humans has also been investigated using the auditory steady-state response (ASSR), also known as the envelope-following response (EFR). The ASSR has been proposed as a physiological measure that parallels loudness growth (Kubota et al., 2019; Ménard et al., 2008; Van Eeckhoutte et al., 2018a, 2018b, 2016). The ASSR is phase-locked to the envelope periodicity of a modulated tone or noise and can be measured with standard clinical instrumentation with as few as three electrodes or with multi-channel electroencephalography (EEG) systems (Picton et al., 2005, 2003). Clinically, the 80-Hz ASSR has been used to assess brainstem processing as a steady-state alternative to the traditional transient-evoked auditory brainstem response (ABR); however, the utility of this response is manifold. For example, the 80-Hz ASSR has been shown to be a viable method for indexing loudness recruitment resulting from cochlear hearing loss at the level of the brainstem (Kubota et al., 2019) and for indexing loudness perception at higher levels of the auditory system. Ménard et al. (2008) showed that in normal-hearing listeners (i.e., without loudness recruitment) there was a higher correlation between 80-Hz ASSR amplitude growth and behaviorally obtained loudness growth functions than between ASSR amplitude growth and stimulus intensity growth functions. Van Eeckhoutte et al. (2016) also showed that 40-Hz ASSR magnitude growth functions and loudness growth functions were almost identical in shape for normal-hearing and hearing-impaired subjects and that ASSR amplitudes were a good predictor of behavioral loudness scaling results, especially for hearing-impaired subjects. Additionally, the utility of the ASSR has been demonstrated in the animal literature as a potential tool for investigating central gain-related changes in the context of aging and hearing loss in rats. Lai et al. (2017) reported enhanced EFR amplitudes in aged animals (16 to 90 Hz modulation rates) and attributed the increase to a compensatory increase in central gain. Given the results from these studies, the ASSR may be a useful, objective tool for investigating central gain-related changes following both acoustic attenuation and enhancement.
Neural generators of the ASSR are known to exist throughout the auditory pathway; however, contributions from dominant sources can vary depending on the stimulus modulation rate, such that higher modulation rates (e.g., 80 Hz) are thought to reflect dominant activity in the brainstem, whereas slower modulation rates (e.g., 40 Hz) reflect dominant contributions from cortical sources within and beyond the auditory cortex (AC; Herdman et al., 2002; Picton et al., 2003; Luke et al., 2017; Farahani et al., 2017, 2019). As such, the cortically-evoked 40-Hz ASSR would be useful for investigating whether changes in peripheral input (e.g., sound attenuation or enhancement) might alter auditory processing at higher levels of the auditory system, and for further investigating changes in loudness perception that may result from changes in central gain (Formby et al., 2017a).
To identify specific cortical generators of the ASSR, a variety of imaging and source localization techniques have been used to evaluate cortical and subcortical contributions to the ASSR. Reyes et al. (2004) used positron emission tomography (PET) to demonstrate peak activation in the right and left primary auditory cortices, the left medial geniculate body, as well as the right middle frontal gyrus in response to a unilaterally presented (right ear) amplitude modulated (AM) tone. Both EEG and magnetoencephalography (MEG) studies of the ASSR also have demonstrated contributions from the left and right primary and secondary auditory cortices and some non-primary cortical sources (Farahani et al., 2019, 2017; Gutschalk et al., 1999; Herdman et al., 2002; Luke et al., 2017; Pantev et al., 1996; Roß et al., 2005). Luke et al. (2017) used EEG to investigate source localization of the ASSR for acoustic and electric (i.e., cochlear implant) hearing and showed a complex pattern of hemispheric laterality that depended on both modulation rate and ear of stimulation. Specifically, for those with acoustic hearing, a right hemisphere dominance was observed for 40-Hz modulation rates regardless of whether the stimulus was presented monaurally to either ear or diotically to both ears. For those with electric hearing, the hemisphere contralateral to the implanted ear was preferentially activated. Similarly, investigations of permanent unilateral hearing loss have shown hemispheric differences in the distribution of cortical activity depending on the ear of stimulation (Han et al., 2021). If the ASSR is to serve as an index of cortical change in central gain following changes in peripheral input, it is important to evaluate contributions from both right and left hemispheres, especially if the experimental paradigms induce differences across ears.
The present study was designed to investigate effects of short-term, unilateral auditory deprivation via earplugging on central gain changes across the auditory system using both subjective and objective outcome measures. The test battery reflects a combination of outcome measures (Contour test, ARTs) previously used to index auditory changes over time following different acoustic treatments, as well as an electrophysiological correlate of central gain (ASSR). To our knowledge, the latter has not been investigated following continuous acoustic attenuation in humans. If changes are observed in the cortically-evoked ASSR following earplugging, then both the ART and ASSR will inform current perspectives on central gain modulation in the auditory pathway from brainstem to cortex.
Here, we hypothesize that unilateral earplugging increases central gain, resulting in a reduction in the ART in the plugged ear following unilateral auditory attenuation, consistent with previous literature (Brotherton et al., 2019, 2017, 2016; Munro et al., 2014). We also hypothesize that unilateral earplugging modulates central gain in the cortex as reflected in altered loudness judgements and cortically-evoked ASSR measures. It is unknown whether central gain changes in the ART brainstem measures will be mirrored in the cortical ASSR measures or if more complex patterns will emerge. To examine potential effects across the entire system, it is of interest to investigate changes in both the plugged ear and not-plugged ear to determine possible cross-ear effects, as has been reported by Munro et al. (2009, 2014), as well as potential hemispheric asymmetries (e.g., Han et al., 2021; Maslin et al., 2013a; Xie et al., 2019).
A total of 12 normal-hearing participants were enrolled (11 females, 1 male; median 21.4 years, range 20-35 years). Each participant underwent a standard intake protocol including pure tone air- and bone-conduction audiometry, immittance testing (tympanometry, ARTs), and speech-in-noise testing using the Hearing-In-Noise Test (HINT). Normal hearing sensitivity was defined as pure-tone thresholds <25 dB HL at octave frequencies from 250 Hz to 8000 Hz, bilaterally, with no asymmetry between ears
Average attenuation values from earplug attenuation measurements during the first visit are shown in Fig. 3, with the maximum attenuation (mean = -32.7 dB, ± 1.29 dB, SEM) observed at 3000 Hz. There is a possibility that the low-frequency attenuation (below 1000 Hz) is underestimated due to potential slit leaks caused by probe tube placement between the wall of the ear canal and the earplug. As noted above, participants consistently inserted the earplugs and achieved attenuation values of ± 3
We hypothesized that unilateral earplugging would reduce the ART in the plugged ear following unilateral auditory attenuation. Short-term sound attenuation from unilateral earplug use resulted in an average decrease of 8 to 10 dB in the ART for the plugged ear during the two-week period of earplug use relative to baseline (see Fig. 4). There was a slight but non-significant increase in ARTs in the not-plugged ear, resulting in an average asymmetry between ears of about 8.5 dB at Weeks 1 and 2
Peter Hutchison: Conceptualization, Methodology, Software, Formal analysis, Investigation, Data curation, Writing – original draft, Visualization, Conceptualization, Methodology, Software, Formal analysis, Investigation, Data curation, Writing – original draft, Visualization. Hannah Maeda: Conceptualization, Methodology, Data curation, Writing – review & editing, Conceptualization, Methodology, Data curation, Writing – review & editing. Craig Formby: Conceptualization, Methodology, Writing –
The authors would like to thank the many members of the Auditory & Speech Sciences Laboratory at USF who helped with this study including a special thank you to Dr. Fernanda Magliaro Aburaya for valuable assistance in data collection and analysis. This work was supported in part by the National Institutes of Health [NIH/NIA P01-AG09524].