Recent animal studies provide compelling evidence that even moderate levels of noise exposure can substantially reduce the number of synaptic ribbons between the inner hair cells and auditory nerve fibers (cochlear synaptopathy) without affecting audiometric thresholds (Kujawa & Liberman 2009; Lin et al. 2011; Maison et al. 2013; Furman et al. 2013; Hickox et al. 2017; Valero et al. 2017). This new phenomenon has been referred to as “hidden hearing loss.” The demonstration of reduced and/or shallower growth of auditory brainstem response (ABR) wave I amplitude, and the strong correlation between this amplitude reduction and the number of surviving spiral ganglion neurons suggests that wave I amplitude at supra-threshold levels could be a sensitive noninvasive clinical metric to evaluate and detect cochlear synaptopathy (Liberman & Kuwaja 2017). In addition, the unaltered ABR wave V amplitude in the presence of reduced wave I amplitude suggests central compensation for the reduced peripheral input (Shaheen et al. 2015). While animal studies across several species consistently show ABR wave I change consistent with cochlear synaptopathy, empirical evidence linking noise exposure history and ABR wave I amplitude changes in humans with clinically normal hearing have not been consistently uniform (Stamper & Johnson 2015; Liberman et al. 2016; Prendergast et al. 2017; Grose et al. 2017; Grinn et al. 2017; Fulbright et al. 2017). Although wave I deficits were not observed directly, several studies have reported changes in different metrics that appear to support noise exposure related cochlear synaptopathy. For example, Stamper and Johnson (2015) showed a correlation between Wave I amplitude and noise exposure; Liberman et al. (2016) observed enhanced summating potential/action potential (SP/AP) ratio in their high-risk group (college-age musicians) that was correlated with speech deficits in noise, reverberation, and time compression; Grose et al. (2017) showed reduced wave I/V amplitude in the high-risk group; Valderrama et al. (2018) showed a correlation between wave I amplitude and noise exposure, and a correlation between wave I–V latency and word scores; Ridley et al. (2018) showed a correlation between SP/AP or wave 1 and the threshold-in-noise residual; and recently Mepani et al. (2020) found evidence suggesting neural deficits (elevated SP or SP/AP ratio when wave I was measured AP from the SP shoulder to N1).
Several possible explanations have been advanced in an effort to account for this lack of consensus including, the inherent difficulty in accurately defining the experimental and control groups to be sufficiently distinct based on noise exposure history; differences in susceptibility to noise damage among humans—tough versus tender ears (Subramaniam et al. 1991; Henderson et al. 1996; Attanasio et al. 1999); variability of ABR wave I amplitude in humans ABR due to factors such as age, sex, audiometric thresholds, and head size (Gorga et al. 1985; Mitchell et al. 1989; Burkard & Don 2007); and the possibility that transducer type and stimulus paradigms utilized to date may not have been optimally sensitive to reveal consequences of noise-induced cochlear synaptopathy. In a continued effort to identify more sensitive stimulus manipulations that will reliably identify individuals with cochlear synaptopathy, we examine here the effects of stimulus level (experiment 1) and background noise (experiment 2) on the amplitude and latency of ABR responses obtained from individuals at low-, and high-risk for cochlear synaptopathy based on exposure to loud music.
It is well established that the absolute latency of the ABR components systematically decreases and amplitude increases as stimulus intensity is increased (Picton et al. 1974; Starr & Achor 1975). This intensity-dependent decrease in latency is thought to reflect the increasing basal bias of cochlear regions contributing to the response, and the decreasing threshold of the cochlear excitatory process due to the progressively faster-rising generator potential within the cochlea and a faster development of excitatory postsynaptic potential. The increase in amplitude with intensity reflects both an increase in the number of neural elements and an increase in neural synchrony. However, to date, there are no published reports of a systematic evaluation of the effects of stimulus intensity in individuals with high-risk for cochlear synaptopathy. Because low-spontaneous rate (LSR) and medium-spontaneous rate (MSR) fibers, that presumably encode moderate-to-loud sounds, are implicated in synaptopathy (Furman et al. 2013), we hypothesize that amplitude reduction for ABR wave I in the high-risk group may be limited to moderate and higher level sounds (50 dB nHL and greater) with no differences between the groups at lower levels where responses largely reflect activation of the functionally intact high spontaneous rate (HSR) fibers. It is relevant here to point out that loss of LSR fibers alone is insufficient to appreciably reduce the amplitude of auditory nerve compound action potential (CAP) (Bourien et al. 2014). Whereas the reduced auditory nerve output in the high-risk group may alter the absolute latency (and therefore the inter-peak latencies [IPL] due to increase in synaptic integration time) and amplitude of the later ABR components (waves III and V), it is possible that central compensatory gain mechanisms, well-reported in previous human (Bramhall et al. 2017) and animal studies (Shaheen et al. 2015), may minimize these effects.
It is likely that cochlear synaptopathy degrades neural representation of certain acoustic features important for speech perception, particularly in noise (Lopez-Poveda 2014). There is some indirect evidence supporting the association between changes in different brainstem electrophysiologic metrics and deficits in speech perception in noise (Bharadwaj et al. 2014; Plack et al. 2014; Liberman et al. 2016; Lobarinas et al. 2016; Liberman & Kujawa 2017; Valderrama et al. 2018; Mepani et al. 2020; Bramhall et al. 2019). Here we examine if there is a greater degradation of the ABR components in the presence of noise for the high-risk group. To this end, we examine in Experiment 2, if latency (absolute and IPL) and amplitude of the ABR components show an increased susceptibility to noise in the high-risk group. Increasing levels of background noise have been shown to increase the absolute latency, I–V IPL, and decrease the amplitude of the click-evoked ABR components (Burkard & Hecox 1983,1987). While the amplitude of all ABR components decreases with increasing masker level, wave V latency shows greater latency prolongation compared to the essentially invariant wave I latency, resulting in a longer I–V IPL (Burkard & Hecox 1987; Burkard & Sims 2002). More relevant here are the results of a recent study that evaluated the effectiveness of ABR in noise measures (in both humans and a mice model) to detect cochlear synaptopathy in humans (Mehraei et al. 2016). Their results showed that the human ABR wave V latency change with an increase in ipsilateral broadband masking noise mimicked the growth in the amplitude of wave I between 60 and 100 dB p.e. dB SPL suggesting that differences in latency shift of wave V in noise arise, at least in part, from changes in auditory nerve response. Results from their animal model of noise-induced cochlear synaptopathy showed that ABR wave IV (wave V in humans) latency shift in noise was smaller in animals with synaptopathy. Based on these results these authors reasoned that reduced wave-V latency shift with increasing levels of ipsilateral masking noise could reflect the activity of LSR and MSR fibers given their delayed onset response (Rhode & Smith 1985; Bourien et al. 2014) and a greater resistance to masking (Costalupes 1985; Young & Barta 1986). While these inferences are based on wave V changes, there are no published reports that have specifically evaluated the effects of background noise on human ABR components (waves I and V) in individuals with long-term exposure to loud music. Given that both LSR and MSR fibers exhibit greater suppression, particularly with increasingly higher intensities compared to HSR fibers (Fahey & Allen 1985; Delgutte 1990; Cai & Geisler 1996a,1996b); we hypothesize that the high-risk group will show a smaller wave I amplitude reduction in background noise due to selective loss of LSR and MSR fibers. Based on evidence from the mouse model of noise-induced synaptopathy (Mehraei et al. 2016), we predict a smaller wave V latency shift with increasing noise levels among high-risk group. The reduction of the suppressive masking component may also degrade neural representation of formant peaks in noise (Sachs & Young 1979), thereby producing speech perception deficits in noise.
Both animal (Sergeyenko et al. 2013) and human (Wu et al. 2019) studies have shown that age-related cochlear synaptopathy is seen across frequencies. Most of the previous human published reports employed ER-3A earphones to measure electrophysiological responses. These earphones have a flat frequency response out to 4000 Hz and then rolls off at 36 dB/octave above thus limiting our ability to sample a broader cochlear region contributing to wave I. Also, analysis of derived-band CAPs from the auditory nerve and ABRs (Don & Eggermont 1978) in humans have demonstrated that auditory nerve and early brainstem responses (CAP, ABR wave I and III) largely reflects neural activity from more basal cochlear regions, particularly at higher stimulus levels (Don & Eggermont 1978; Eggermont 1979). Considering these limiting factors, we chose to employ ER-2 insert earphones, which have a relatively flat frequency response out to about 12–14 kHz for 100 μs duration clicks before dropping off at 16 kHz (Elberling et al. 2012). This will enable us to record auditory nerve response reflecting contributions from much higher frequency cochlear regions which may increase the measure’s sensitivity to detect changes in auditory nerve activity consequent to long-term exposure to loud music. Indeed, Elberling et al. (2012) showed that the click spectrum of the ER-2 is flat compared to the ER-3A earphone and gives relatively much higher output in the frequency range above 4000–5000 Hz (30–35 dB higher). Consistent with this they showed that the ABRs elicited using ER-2s were better resolved and robust suggesting increased contribution from more higher frequency cochlear regions.
MATERIALS AND METHODS
A single-blinded design was employed where participants were tested in a random fashion and did not know their group assignment, nor the variables manipulated for data acquisition. For experiment 1, 28 participants were recruited for both the high-risk (14 male and 14 females, mean age: 22.68 years, SD: 2.82) and the low-risk (14 male and 14 females, mean age: 21.13 years, SD: 2.14) group. Fifty-two out of the original 56 participants also participated in experiment 2. Data from two participants were noisy and were excluded. The remaining 50 young adults were considered for experiment 2. With this, for experiment 2, the high-risk group consisted of 25 participants (12 male and 13 females, mean age: 21.28 years, SD: 2.83) and low-risk group consist of 25 participants (12 male and 13 females, mean age: 22.88 years, SD: 2.89). All participants had normal audiometric thresholds (better than 20 dB nHL in both ears) at octave frequencies from 250 to 8000 Hz in both ears and reported no previous history of neurological or psychiatric illnesses. The high-risk group was formed by students who participated in marching band for at least 5 years. The total number of years in marching band for high-risk group ranged from 5 to 15 (mean: 9.54 years, SD: 2.54); mean hours of practice per day: 1.44 hours, SD: 0.75; ratings of hearing protection device use: 15/28 never used, 12/28 rarely used, and only 1/28 almost always used hearing protection device. We chose to use students from marching band because their exposure to loud sounds far exceed the daily permissible sound exposure recommended by the National Institute for Occupational Safety and Health (NIOSH, 1998) and therefore posing a significant risk for noise-induced hearing loss. While we were not able to determine the levels of noise exposure for each musical instrument, our informal measure of the sound levels during a practice session using a smartphone application revealed sound levels peaking between 120 and 125 dB SPL. The noise level exposure in these individuals is similar to or higher than those reported in animal noise exposure studies that produced cochlear synaptopathy (Walter 2011; Skoe & Tufts 2018). In contrast, the individuals comprising the low-risk group had little or no history of recreational or occupational exposure to loud sounds. We are therefore confident that the selected two groups are sufficiently distinct in terms of their noise exposure history. Table 1 summarizes the instruments played by individuals in the high-risk group. Percussion, saxophone, and clarinet were the common instruments played by the group. As most musicians played more than one instrument the total “n” in the instrument played list is higher than the total number of participants. All participants, recruited from the Purdue University student body, were paid and gave informed consent in compliance with a protocol approved by the Institutional Review Board of Purdue University.
TABLE 1. -
Instruments played by participants in the high-risk group
||Total number of participants
|Saxophone (any flavor)
|Clarinet (any flavor)
To ensure functional integrity of peripheral auditory mechanisms and to rule out peripheral factors contributing to ABR changes we obtained thresholds using warbled tones, tympanograms, and distortion product otoacoustic emissions (DPOAE) similar to previous animal and human research.
Air conduction thresholds were measured using warble signals between 250 and 8000 Hz using a clinical audiometer (Grason-Stadler, GSI 61) equipped with an insert earphone (Etymotic, ER-2). A circumaural earphone with extended high frequency (Sennheiser, HDA-200) was used to measure thresholds above 8000 Hz using warbled tones. Normal hearing was defined as warbled tone behavioral thresholds ≤20 dB HL for the octave and inter-octave frequencies between 250 and 8000 Hz for both ears and this was set as an inclusion criterion for the study. A decision was made at the beginning of the research project to test right ear for all monoaural tests (Tympanometry, Otoacoustic emissions, and ABR [discussed later]).
Tympanograms were obtained over a pressure range of +400 to −400 daPa with a positive to negative sweep. The pump speed of 200 daPa/s and probe tone frequency of 226 Hz at 85 dB SPL was employed. All participants had normal tympanograms (Jerger 1970).
Distortion Product Otoacoustic Emissions
DPgrams were obtained with two simultaneously swept primary tones f1 and f2 (f2/f1 = 1.22), presented at 65 dB SPL and 55 dB SPL, respectively. Primary frequencies were swept from f2 = 500–8000 Hz, in 4 logarithmically spaced steps per octave.
For both experiments, ABRs were elicited by clicks (100 µs) using a stimulus generation and EEG data acquisition system (Intelligent Hearing Systems, SmartEP software). In the first experiment ABRs were recorded for click levels ranging from 30 to 90 dB nHL (90 dB nHL = 124 dBpSPL as measured in a 2-cc coupler) presented in 10 dB steps. In experiment 2, we obtained click-evoked ABRs in quiet (70 dB nHL), and in the presence of broadband noise presented at 50, 60, and 70 dB SPL. The click level of 70 dB nHL was chosen to enable identification of all ABR components. The choice of masker levels reflects a range of 50–70 dB SPL that is both tolerable and sufficient to evaluate the effects of masking on latency and amplitude of the ABR components (Burkard & Hecox 1983). In addition, our pilot experiment revealed that no discernible wave I was observed above a noise level of 70 dB SPL. The aim of this second experiment was to determine if the putative loss of synaptic ribbons associated with cochlear synaptopathy increases the susceptibility of the ABR related neural activity to degradation in the presence of background noise.
Participants reclined comfortably in an electro-acoustically shielded booth and were instructed to relax and refrain from extraneous body movement to minimize myogenic artifacts. Participants were told to ignore the sounds they heard and were encouraged to sleep throughout the duration of the recording session. ABRs were recorded from each participant in response to monaural stimulation of the left ear using magnetically shielded insert earphones (Etymotic Research, ER-2). ER-2 transducer with its relatively flat frequency response (Fig. 1) out to about 12–14 kHz (Elberling et al. 2012) enables us to capture neural activity contributing to wave I over an extended higher frequency range, compared to the use of an ER-3A (Don & Eggermont 1978; Eggermont 1979). Stimulus repetition rate was set at 23.3 clicks per second, and rarefaction onset polarity was utilized. Stimulus presentation order was randomized both within and across participants in each experiment.
ABRs were recorded differentially between a non-inverting electrode placed on the high forehead near the hairline, an inverting electrode placed on the left mastoid, and the ground electrode placed on the mid-forehead. All inter-electrode impedances were below 1 kΩ. EEG signals were amplified by150,000 and band-pass filtered between 50 and 5000 Hz (6 dB/Octave RC characteristics). Each ABR response represents an average of 2000 stimulus sweeps digitized over a 12.4 ms analysis epoch using a 40-kHz sampling rate. Stimulus-related epochs contaminated by muscle or movement artifacts were automatically rejected (artifact reject set at ±40 μV) from averaging.
ABR response identification, latency, and amplitude measurements were independently carried out by the first and the second author. There was excellent agreement (98%) for both latency and amplitude measures between the two authors. Two responses obtained for each experimental condition were overlaid to improve accuracy of response detection. Both the absolute and IPL, and the peak-to-trough amplitude of the three dominant ABR components (waves I, III, and V) were measured. Absolute latency of each component reflects the time interval from stimulus onset to the positive peak of each component. IPLs reflects the time-interval between I–III, III–V, and I–V. Amplitude for waves I, III, and, V was defined as the electrical potential difference between the positive peak and the following negative trough. In addition, wave V to wave I amplitude ratio was also derived from each subject across experimental conditions to determine if a history of loud music exposure altered this relationship. For illustration purposes, grand average ABRs were obtained from each group for both experiments.
Experiment 1: Effects of Stimulus Level
To evaluate the effects of stimulus level on the ABR components for the two groups, separate two-way (experimental group × stimulus level) mixed-model analysis of variances (ANOVAs) (SAS; SAS Institute, Inc., Cary, NC, USA) were performed on absolute latency, (waves I, III, and V), peak-to-trough amplitude (waves I, III, and V), IPLs (I–III, III–V, and I–V), and amplitude ratio (V/I). Experimental group (low-risk and high-risk) served as the between-subjects factor; subjects nested within the group as the random factor. Stimulus level served as the within-subjects factor. Normality and homogeneity of variance (HOV) assumptions were confirmed prior to statistical inference. All post hoc multiple comparisons were corrected with a Bonferroni significance level set at α = 0.05. The ABR variable (wave I amplitude) which emerged as significant in the group analysis was further explored using bivariate correlations. This correlation analysis examined whether there was a relationship between wave I amplitude (80 dB nHL: representative level), average pure tone detection threshold from 2 to 6 kHz (PTA2–6kHz), and average DPOAE amplitude measured between 2 and 6 kHz (OAE2–6kHz). A discriminant analysis was also conducted to determine the linear combination that best discriminates the two groups employed in this study (low-risk and high-risk). Linear combination considered for comparison were peak amplitude and the combination of wave I amplitude (80 dB nHL), PTA2-6 kHz, and OAE2-6 kHz.
Experiment 2: Effects of Background Noise
Separate two-way (experimental group × masker level) mixed model ANOVAs (SAS; SAS Institute, Inc.) were performed to evaluate the effects of increasing masker intensity on absolute latency (wave I, III, and V), IPL (I–V), relative change (re: the quiet condition) in absolute latency of waves I and V in the presence of masker, peak-to-trough amplitude, and the change in amplitude of waves I, and V in the presence of masking relative to the amplitude in the quiet condition. Analysis of wave III was excluded because discernible wave III component was not reliably identifiable in noise conditions for several participants. Experimental group (low-risk and high-risk) served as the between-subjects factor; subjects nested within the group as the random factor. Masker level (4 levels: quiet, and 50, 60, 70 dB SPL noise) and three levels (3 levels; 50, 60, 70 dB SPL noise condition) for relative change served as the within-subjects factor. In addition, the slope of wave I amplitude change in noise for the two groups was compared using a two-sample t-test. Normality and HOV assumptions were confirmed before statistical inference. All post-hoc multiple comparisons were corrected with a Bonferroni significance level set at α = 0.05 and partial eta squared (ηp2) values were reported to indicate effect sizes.
Pure-tone Audiometry, Tympanometry, and Distortion Product Otoacoustic Emission
Pure-tone audiometric thresholds for both standard and extended high frequencies were within normal limits (Fig. 2, panel A) with no group differences (p > 0.05).
DPOAE amplitudes (Fig. 2, panel B) were not statistically different between the two groups (p > 0.05).
Experiment 1: Effects of Stimulus Level on the ABR Components
Grand average ABR waveforms from each group are plotted as a function of stimulus level in Figure 3. While both groups show the expected decrease in amplitude and latency prolongation as stimulus intensity is decreased, ABR wave I amplitude (indicated by arrows) is smaller for the high-risk group compared to the low-risk group for sound levels greater than 50 dB nHL. In contrast, waves III and V do not appear to show group differences.
Mean absolute latency of ABR response components plotted as a function of click intensity are shown in Figure 4 (panel A). Waves I and wave III were consistently observed in all subjects only for levels 60 dB nHL and higher. For wave I, ANOVA yielded only a main stimulus effect (F3, 162 = 170.78, p < 0.0001, ηp2 = 0.76) where, as expected, latency for 60 dB nHL was longer than the latency for 70 dB nHL, latency for 70 dB nHL longer than latency for 80 dB nHL, and no difference in latency between 80 and 90 dB nHL. Similarly, for wave III, ANOVA yielded only a main stimulus effect (F2, 108 = 44.73, p < 0.0001, ηp2 = 0.453). The stimulus main effect showed that 70 dB nHL elicited a longer response absolute latency than 80 and 90 dB nHL, but latencies at 80 and 90 dB nHL were not different. ANOVA results for wave V also yielded only a stimulus main effect (F6, 324 = 815.51, p < 0.0001, ηp2 = 0.93) where wave V latency decreased with increase in level from 30 to 80 dB nHL and showed no difference between 80 and 90 dB nHL. Collectively, these results suggest that for both groups, the absolute latency of ABR waves I, III, and V decreased in a similar fashion (no group difference) with increase in stimulus level.
ABR IPLs I–III, III–V, and I–V reflect both synaptic integration time and neural conduction times between neural generators along the auditory brainstem. Since clearly discernible ABR wave III could not be obtained in all participants at 60 dB nHL, IPL analysis was restricted to 70, 80, and 90 dB nHL. Mean IPL (I–III; III–V; I–V) plotted as a function of stimulus level for both groups are shown in Figure 4 (panel B). For I–III IPL, ANOVA yielded only a significant group main effect (F1, 54 = 5.53, p = 0.02, ηp2 = 0.245) with no significant stimulus (F2, 108 = 1.43, p = 0.2442, ηp2 = 0.025), The high-risk group exhibited longer I-III IPL relative to the low-risk group. In contrast, III-V IPL, ANOVA yielded only a significant main stimulus effect (F2, 108 = 8.93, p < 0.001, ηp2 = 0.142) where the III-V interval exhibited a “V” shaped function with 80 dB nHL showing the shortest III–V IPL compared to 70 and 90 dB nHL. For I–V IPL, ANOVA yielded both a main effect of stimulus (F3, 162 = 8.81, p < 0.0001, ηp2 = 0.14), and group (F1, 54 = 4.83, p = 0.03, ηp2 = 0.08). The high-risk group exhibited longer I–V IPL relative to the low-risk group. Like the III–V IPL, I–V IPL was shorter for 80 dB compared to the 90 and 70 dB nHL conditions.
ABR Peak to Peak Amplitude and V/I Amplitude Ratio
ABR Peak to Peak Amplitude
Mean peak to peak amplitude of ABR response components plotted as a function of stimulus level for both groups are shown in Figure 5. Analysis was restricted to ABRs above 50 dB nHL, since waves I and III were reliably detected in all participants only at levels 60 dB nHL and higher. For wave I, ANOVA yielded both a stimulus (F3, 162 = 98.89, p < 0.0001, ηp2 = 0.647) and a group effect (F1, 54 = 24.05, p < 0.0001, ηp2 = 0.308). The low-risk group exhibited larger wave I amplitude relative to the high-risk group. The post-hoc multiple comparisons of stimulus main effect revealed a systematic increase in wave I amplitude with stimulus level. For wave III, ANOVA yielded only a stimulus main effect (F2, 108 = 31.95, p < 0.0001, ηp2 = 0.37). The stimulus main effect indicated that 80 and 90 dB nHL elicited larger wave III amplitude than 70 dB nHL. Similarly, ANOVA results for wave V yielded only a main stimulus effect (F6, 324 = 48.24, p < 0.0001, ηp2 = 0.472), with no significant main effect of group (F1, 54 = 0.91, P = 0.3433, ηp2 = 0.016). The stimulus main effect showed an increase in amplitude from 30 to 50 dB nHL, remained unchanged from 50 to 70 dB nHL, and then increased in amplitude from 70 to 80 dB nHL before flattening again from 80 to 90 dB nHL. These results suggest reduced ABR wave I amplitude in the high-risk group compared with the low-risk group with no group differences in response amplitude for waves III and V.
V/I Amplitude Ratio
To satisfy the ANOVA model assumption of HOV for the V/I amplitude ratio data, log10 of the original data (which met the constant variance assumption of ANOVA) is subjected to statistical analysis. ANOVA yielded a main effect of group (F1, 54 = 10.77, p = 0.002, ηp2 = 0.166) and stimulus (F3, 162 = 18.74, p = <0.0001, ηp2 = 0.258). The high-risk group exhibited larger mean V/I amplitude ratio relative to low-risk group (Fig. 6) with the ratio greater at the lowest level compared to the higher stimulus levels. The amplitude ratio decreased with increasing intensity for both groups suggesting a greater growth in wave I amplitude compared to wave V amplitude with increasing level.
Bivariate Correlations and Discriminant Analysis
Bivariate correlations were obtained between wave I amplitude (80 dB nHL; pooled across both groups) and mean pure tone threshold between 2 and 6kHz (PTA2–6kHz) as well as mean DPOAE amplitude measured between 2 and 6 kHz (DPOAE2–6kHz) which provided insight whether wave I amplitude is influenced by pure tone thresholds and DPOAE amplitude. The comparisons suggested that wave I amplitude was not correlated with either hearing thresholds or DPOAE amplitude (p < 0.05).
Amplitude (Waves I, III, and V)
Discriminant analysis was used to determine the accuracy with which individual subjects could be classified into their respective groups (low-risk versus high-risk) based on a linear combination of their peak amplitude (waves I, III, and V) for 80 dB nHL click level only. Discriminant analysis by resubstituting method (Osuji et al. 2013) suggested about 75% of subjects were correctly classified into their respective groups (low-risk, 75%; high-risk, 75%). This classification accuracy was reduced by only 7.14% (low-risk, 19/28; high-risk, 19/28) in the cross-validated analysis in comparison to the resubstituting analysis. The pooled within-class standardized canonical coefficients for waves I, III, and V were 0.93, 0.59, and −0.34, respectively, indicating that wave I was the most important variable discriminating between low-risk and high-risk group.
Combination of Wave I Amplitude (80 dB nHL), PTA2–6 kHz, and OAE2–6 kHz
A second discriminant analysis using a linear combination of wave I amplitude (80 dB nHL), PTA2–6kHz, and OAE2–6kHz only revealed subject classification accuracies of 57.14% and 60.71% for low-risk and high-risk, respectively. The pooled within-class standardized canonical coefficients for wave I amplitude, PTA2–6kHz, and OAE2–6kHz, respectively, were 0.98, 0.18, and −0.06, indicating again that wave I was the most important variable in discriminating between low-risk and high-risk group. Taken together, these results suggest that wave I amplitude is not appreciably influenced by either hearing thresholds or DPOAE amplitude and that, wave I amplitude emerged as the most important measure for discriminating the two groups.
Experiment 2: Effects of Background Noise
Grand average waveforms from each group are plotted for the quiet, and noise conditions in Figure 7. For both groups, ABR components are clearly identifiable and appear to show a decrease in amplitude and an increase in latency with an increase in noise level. Vertical dotted line for wave I and wave V shows the differential effects on latency in the presence of noise, specifically, larger latency shift for wave V compared to wave I. While wave I amplitude appears larger for low-risk group (as described in experiment 1) in quiet, it appears to show a relatively greater reduction in the presence of background noise for the low-risk group.
ABR Absolute and Inter-peak Latency
ABR Absolute Latency
Mean absolute latency of ABR response components in quiet and in noise conditions are shown in Figure 8 (panel A). For wave I, ANOVA yielded only a main stimulus effect (F3, 144 = 4.58, p = 0.004, ηp2 = 0.08). Wave I latency in the presence of 60 dB SPL noise was longer than the quiet condition but none of the other stimulus comparisons were significant suggesting that increasing the noise level did not appreciably alter wave I latency. ANOVA results for Wave V also showed only a main stimulus effect (F1, 48 = 147.49, p < 0.0001, ηp2 = 0.75) with wave V latency showing a progressive increase in latency with an increase in noise level. In summary, wave I peak latency changed little with increase in noise level whereas wave V latency showed a progressive increase in latency as the noise level was increased. No group differences were observed for the latency change produced by noise.
I-V Inter-peak Latencies
Mean IPL (I-V) across stimulus conditions for both groups is shown in Figure 8 (panel B). ANOVA yielded only a main stimulus effect (F3, 144 = 79.88, p < 0.0001, ηp2 = 0.625). For both groups, a systematic increase in IPL was observed as the noise level was increased. This increase in IPL with an increase in the level of the background noise is primarily due to the relatively greater latency shift for wave V compared to wave I.
Latency Shift Relative to the Quiet Condition
The latency shift relative to the quiet condition was measured to evaluate if latency changes in noise were different for the two groups. The mean latency shift (referenced to quiet condition) is shown in Figure 8 (panel C). No group differences were observed for wave I and V latency shifts, however, wave V showed relatively greater latency shift with increasing noise levels compared to wave I. For wave I ANOVA revealed a marginally significant stimulus main effect (F2, 96 = 3.16, p = 0.047, ηp2 = 0.06). In contrast, for wave V, ANOVA yielded a robust stimulus main effect (F2, 96 = 67.49, p < 0.0001, ηp2 = 0.584).
Effects of Noise on the Amplitude of ABR Components
Mean amplitude of ABR response components (wave I and wave V) across stimulus conditions shown in Figure 9 (panel A) for both groups. For wave I, ANOVA yielded a main stimulus effect (F3, 143 = 42.83, p < 0.0001, ηp2 = 0474), a main group effect (F1, 48 = 16.30, p = 0.0002, ηp2 = 0.252), and a group × stimulus interaction effect (F3, 143 = 6.55, p = 0.0003, ηp2 = 0.12). By stimulus condition, low-risk group exhibited larger amplitude than the high-risk group for the quiet and 50 dB SPL background noise condition only. By group, the low-risk group showed a systematic decrease in wave I amplitude from quiet to 60 dB SPL background noise condition, however, there was no significant difference between 60- and 70-dB SPL background noise condition. For the high-risk group, the quiet condition had larger amplitude compared to the three background noise conditions (50, 60, and 70 dB SPL) with no significant differences in amplitude between the noise conditions. For wave V, ANOVA yielded only a main stimulus effect (F3, 144 = 25.93, p < 0.0001, ηp2 = 0.351). Wave I amplitude change from 50 dB SPL to 70 dB SPL background noise was smaller for high-risk group compared to the low-risk group (p < 0.05). These results suggest that there is greater decrement of wave I amplitude in the presence of background noise for the low-risk group as indicated by the apparently greater slope in amplitude change but, wave V amplitude reduction in noise was essentially similar for both groups.
Normalized Amplitude Change
In order to make a direct comparison of the amplitude changes in noise for the two groups, amplitude change was measured as the ratio of the ABR peak at each noise condition relative to the quiet condition. The mean amplitude ratio across noise conditions are shown in Figure 9 (panel B). While both waves I and V show increasing amplitude reduction with an increase in noise level, wave I appear to be more susceptible to degradative effects of masking noise than wave V. Interestingly, wave I amplitude reduction in background noise was larger for the low-risk group with no apparent group differences in amplitude reduction for wave V amplitude. For wave I, ANOVA yielded a main stimulus effect (F2, 96 = 12.84, p < 0.0001, ηp2 = 0.021) and a main group (F1, 48 = 4.31, p = 0.04, ηp2 = 0.082) effect. Wave I amplitude reduction was indeed larger for low-risk compared to the high-risk group. Bonferroni comparison of stimulus effect revealed a reduction of wave I amplitude with increasing noise. For wave V, ANOVA yielded only a main stimulus effect (F2, 96 = 12.99, p < 0.0001, ηp2 = 0.213). Overall, these results suggest that wave I amplitude reduction in noise is smaller for the high-risk group compared to the low-risk group, with no group differences observed for wave V amplitude change in noise.
The results of experiment I showed that young normal-hearing adults with long-term loud music exposure showed (1) smaller wave I amplitude at moderate and high sound levels consistent with peripheral neural deficits; (2) wave III and wave V amplitudes similar to the low-risk group suggesting operation of central compensatory gain mechanism(s); (3) enhanced V/I amplitude ratio, again consistent with central compensation for reduced input from the periphery; and (4) longer I–III, and I–V IPL suggesting central neural conduction delays resulting from longer synaptic integration time and/or neural conduction time between wave I generator (distal part of auditory nerve) and the neural generators of the more rostral ABR components III (cochlear nucleus), and V (lateral lemniscus/inferior colliculus) (Moore 1987a,1987b). Alternatively, recent evidence appears to suggest that the changes reflected in wave I may be consequences of changes in SP, which presumably consists of a neural component generated well before the distal portion of the auditory nerve (Pappa et al. 2019), the presumed generator site of wave I.
The results of experiment II demonstrated that: (1) for both groups, increasing noise level produced greater latency shift for wave V compared to wave I (therefore greater I–V IPL), suggesting cumulative effects of synaptic delay along the auditory brainstem; (2) the amplitude decrement with increase in background noise level for wave I amplitude was relatively smaller for the high-risk group with no difference in amplitude change for wave V for both groups suggesting that masking effects are reduced for individuals in the high-risk group; and (3) there was no difference in wave V latency shift in noise between low-risk and high-high risk group, unlike previous studies.
Is the Reduction in Wave I Amplitude an Indicator of Reduced Auditory Nerve Output Associated With Exposure to Loud Music?
Our observation of reduced ABR wave I amplitude in the high-risk group is consistent with both animal studies with noise-induced, and age-related cochlear synaptopathy (Kujawa & Liberman 2009; Lin et al. 2011; Furman et al. 2013; Sergeyenko et al. 2013), and few human studies in individuals with history of noise exposure (Stamper & Johnson 2015; Bramhall et al. 2017). Interestingly, although it was not measured in Liberman et al. (2016) research, Figure 2A of the manuscript seem to support the observation of this study that wave I (N1-P1) is reduced in the high-risk group. While the reduction in wave I amplitude observed here in humans cannot be confirmed with certainty to be a consequence of synaptopathy, harvested human temporal bone studies nevertheless do show reductions in both synaptic ribbons and spiral ganglion cells with age, that parallel age-related reduction in ABR wave I amplitude (Makary et al. 2011; Konrad-Martin et al. 2012; Wu et al. 2019). Thus, the reduced wave I amplitude in human studies may be interpreted to indicate loss of synapses and/or reduced auditory nerve fibers consistent with the data from the animal models of noise-induced cochlear synaptopathy. Consistent with this viewpoint, it is likely that the smaller ABR wave I amplitude observed for the high-risk group at moderate to high levels in this study may indeed reflect the consequences of loss of synapses at the IHCs due to long-term exposure to loud music. Specifically, the reduced amplitude may reflect both a reduction in the number of neural elements active and the degree of synchrony in these active neural elements. However, we cannot rule out the possibility that this wave I amplitude reduction may be associated with changes in OHC function not revealed in our normal DPOAEs, damage to IHCs or the auditory nerve unrelated to the IHC-auditory nerve synapse.
Several human studies have failed to find any relationship between ABR wave I amplitude and noise exposure history (Grinn et al. 2017; Grose et al. 2017; Prendergast et al. 2017). These findings have been interpreted to suggest non-prevalence of noise-induced cochlear synaptopathy in young audiometrically normal-hearing individuals or the low exposure is insufficient to cause cochlear synaptopathy. The current study enrolled students who participated in marching band for at least 5 years in the high-risk group which led to clear separation of noise exposure distribution between high and low-risk groups. Further, these results appear to suggest that synaptopathic deficits manifest only in individuals with long-term higher-level noise exposure (including the high-risk individuals in this study) and not in individuals with exposure to common recreational noise (Grinn et al. 2017). Several other factors could, at least in part, account for the differences in results across human studies. One source of variability is the validity of the group assignment based on questionable and subjective questionnaires/interviews aimed at estimating the degree of lifetime noise exposure. That is, the low-risk and the high-risk group may not be sufficiently independent. Second, most of the previous human research employed ER-3A earphones to measure ABRs. These transducers have flat frequency response only out to about 4000 Hz with a 36 dB/octave roll-off beyond 4000 Hz and thereby reducing or eliminating higher frequency cochlear contributions to click ABR in general, and wave I in particular. This is particularly important for ABR wave I, as the analysis of derived-band auditory CAPs and ABRs in humans (Eggermont 1979; Don & Eggermont 1978) have demonstrated that the auditory nerve generated response component (CAP, and ABR wave I), particularly at high intensities, largely reflects synchronized neural activity from the cochlear high-frequency regions. In this study, the use of ER2 transducers (flat and smooth amplitude response out to 10,000 Hz), permitted us to record ABR wave I that reflects contributions from higher frequency cochlear regions and thereby increase the sensitivity of this measure (Elberling et al. 2012).
Do the Increased I–III and I–V Interpeak Latencies in the High-risk Group Reflect Synaptic and/or Neural Conduction Delays Consequent to Post Loud Music Exposure?
In addition to the reduction in wave I amplitude at high sound levels, the high-risk group showed longer I–III and I–V IPLs. These results primarily suggest that synaptic integration time and/or neural conduction time between the generators of wave I (distal part of auditory nerve) and III (cochlear nucleus) is longer for the high-risk group which manifests as longer I–III and I–V IPLs. This longer conduction time may reflect reduced auditory nerve output which in turn may have introduced synaptic delays for more rostral responses (waves III and V). The prolongation of I–III and I–V IPLs with normal III–V is characteristic of the auditory nerve and lower brainstem space-occupying lesions-typically interpreted as neural conduction delays due to a conduction block (Hirsch & Anderson 1980; Schwaber & Hall 1992). While synaptopathy is not a space-occupying lesion, the reduced number of auditory nerve synapses may contribute to these observed delays via increased synaptic integration time and/or neural conduction time post auditory nerve generation of wave I. It is plausible that the reduced auditory nerve input prolongs synaptic integration time at more rostral levels due to disruption in neural timing. Since neural conduction time, reflected in the IPLs, is dependent on both fiber diameter and extent of myelination, it is tempting to speculate if the increased IPLs represent consequences of cochlear synaptopathy that produce anterograde changes (including reduced fiber diameter and/or demyelination) along the auditory pathways in the brainstem rostral to the auditory nerve. There is some empirical support for this possibility, at least at the auditory nerve level. Tagoe et al. (2014) observed a permanent decrease in thickness of myelin sheath with extended exposure to the high-intensity sound which manifested as decreased wave I amplitude and longer wave I latency. We propose that similar changes at more central levels in the lower brainstem consequent to cochlear synaptopathy could be contributing to the longer I–III and I–V IPLs reported here. Nevertheless, these changes could be the consequences of changes occurring earlier than the distal portion of the auditory (the presumed generator site of wave I) in the neural components of the SP (Pappa et al. 2019).
Does the Lack of Change in Amplitude for Waves III and V in the High-risk Group Reflect the Operation of Central Compensatory Gain Mechanisms?
Despite reduced wave I amplitude for the high-risk group there were no group differences in amplitude for the later waves, III and V. Consequently, the V/I amplitude ratio (V/I) was enhanced in the high-risk group. The lack of reduction in ABR wave III and V amplitudes in the high-risk groups in the presence of a reduced auditory nerve drive is similar to observations in animal studies of synaptopathy. For example, in mice with noise- or age-related cochlear synaptopathy, no decrease in wave V amplitude was observed in the presence of wave I amplitude reduction (Sergeyenko et al. 2013; Hickox & Liberman 2014; Möhrle et al. 2016). This suggests that despite a significant reduction in the peripheral output, the rostral brainstem response is very similar. These results suggest that some form of the gain control mechanism (homeostatic mechanisms) is operating between the auditory nerve and brainstem to compensate for the reduced input. In general, the neuron’s receptive field is determined by the balance of excitatory and inhibitory inputs especially at higher centers in the auditory neuraxis. There is evidence of neurotransmitter-mediated inhibition beginning at the level of ventral and dorsal cochlear nucleus regulating input to the higher levels of the auditory system (Caspary et al. 1983,2005,2008). Because inhibition suppresses the discharge rate of neurons, the selective loss of inhibition (disinhibition) results in an increase in the discharge rate. It is hypothesized that it is this disinhibition which results in sufficient increase in discharge rate (renormalization of neuronal response magnitude) at the level of brainstem to compensate for reduced auditory input following peripheral insult such as noise or age-induced synaptopathy (Salvi et al. 2000; Chambers et al. 2016; Möhrle et al. 2016; Salvi et al. 2017). In addition, individuals with normal pure-tone thresholds who report tinnitus show smaller wave I amplitudes, but similar or larger wave III and V amplitudes compared with their non-tinnitus counterparts (Gu et al. 2012; Schaette & McAlpine 2011). Although the participants in the tinnitus studies were not evaluated for noise exposure history, tinnitus has been proposed as a potential perceptual consequence of synaptopathy, and the pattern of ABR amplitudes observed in the individuals with tinnitus is very similar to what was observed in the present study (Gu et al. 2012; Schaette & McAlpine 2011). The absence of a change in amplitude for the later ABR waves has been interpreted as evidence of either hyperactivity or loss of inhibition compensating for reduced input from the auditory periphery. Thus, the lack of group differences for ABR amplitudes III and V may suggest the operation of central compensatory gain mechanisms (Chambers et al. 2016; Eggermont 2017; Salvi et al. 2017; Sheppard et al. 2017). The V/I amplitude ratio is higher for the lowest sound level (60 dB nHL) compared to the higher sound levels. One plausible explanation for this could be that with increasing sound level wave I amplitude increases relatively more compared to wave V amplitude. It should be noted here that the clinical utility of V/I ratio measure may be limited given the larger variability associated with this measure.
Effects of Masking Noise on ABR
Overall the latency prolongation and amplitude reduction observed for the ABR components in noise are essentially consistent with previous results (Burkard & Sims 2002; and Burkard & Hecox 1987 [all ABR waves]) and Burkard & Hecox (1983) (wave V only). Specifically, increasing the level of broadband masking noise increases click-evoked ABR absolute latencies and inter-peak latency, and decreases amplitude of all components (Burkard & Hecox 1983,1987). These authors suggest that increasing masker level stresses synaptic processes and likely increase the synaptic integration time that is cumulative along the several synapses leading to the wave V generator(s). The reduction in amplitude reflects a decrease in both the number of neural elements responding and a disruption of neural synchrony in the presence of noise.
Why is the Amplitude Reduction in Noise for Wave I Smaller for the High-risk Group Compared to the Low-risk Group?
Surprisingly, the high-risk group exhibited a smaller change in wave I amplitude with increase in noise level. While counter-intuitive, similar findings were observed in a prior study comparing ABR responses in background noise between young and older adults (Burkard & Sims 2002). Burkard & Sims (2002) reported that wave I amplitude did not decrease in the older adults until the noise was very high, whereas in younger adults, masking noise produced a decrease in wave I amplitude at substantially lower levels of masking noise. Also, they observed, like in this experiment, wave V amplitude decreased by a similar proportion for both younger and older adults. Similar to selective loss of LSR and MSR fibers with noise-induced cochlear synaptopathy, aging also results in greater loss of LSR and MSR fibers (Schmiedt et al. 1996).
Since our results in experiment 1 showed smaller wave I amplitude for the high-risk group suggesting a loss of LSR and MSR fibers, we expected the high-risk group would be more susceptible to masking noise. Instead, our results showed a smaller reduction in wave I amplitude in noise. It is not clear why the high-risk group would be less susceptible to noise-induced amplitude reduction. To the extent that response amplitude reflects both the number of neural elements responding and the neural synchrony of the responding elements, the relatively smaller change in response amplitude for the high-risk group would suggest a reduced susceptibility to masking. That is, the magnitude of suppressive masking, that kick in at moderate to high levels, may be different in these two groups. Specifically, the reduction in LSR and MSR fibers in the high-risk group may reduce the suppressive masking effects, particularly at moderate levels of our masking noise (Delgutte 1990; Rhode 1978). That is the shallower slope of amplitude reduction for the high-risk group may suggest that the upward spread of masking, largely due to suppression, is more gradual (Delgutte 1990). Similar findings were reported by Sachs and Young (1979) studying nonlinearities on speech encoding in the auditory nerve. They observed LSR fibers to show the most striking two-tone suppression effects for vowels to retain formant peaks at higher levels. One other possibility is that the shallower function in noise for the high-risk group may simply indicate a restricted dynamic range of the much smaller response consequent to reduced number of neural elements rather than a greater resilience to the degradative effects of masking. While it is likely that the relatively smaller wave I amplitude in high-risk group reflects reduced number of neural elements, the clear difference in trajectory of the amplitude behavior between the two groups with increasing background noise level does suggest a differential effect of noise that cannot be simply accounted for by reduced dynamic range of the smaller response. The fact that wave I in the high-risk group is clearly discernible and reliably measurable even at the highest background noise level further reinforces this point of view. To summarize, reduced/shallower wave I reduction in noise for the high-risk group may be related to a reduction in MSR and LSR fibers and or the absence of the central component of suppression (Rhode 1978; Delgutte 1990; Cai & Geisler 1996). Additionally, there is also the possibility of HSR fiber damage/loss following noise over-exposure that produces cochlear synaptopathy. Recordings from auditory nerve fibers following moderate noise exposure in chinchillas, show reduced driven rates in HSR fibers too (Muthaiah et al. 2017), suggesting that the pathophysiology of hidden hearing loss could be more complex.
Given the observation that suppression at higher moderate to high levels helps in retaining clear formant peaks at higher levels (Sachs & Young 1979), it is tempting to speculate if the reduced suppressive masking effects observed here for the high-risk individuals may indeed degrade representation of formant peaks in the presence of noise thereby producing deficits in speech perception in noise. While other electrophysiologic metrics have also advanced an association between noise exposure and speech-in-noise perception (Liberman et al. 2016,Valderrama et al. 2018; Mepani et al. 2019) suggesting functional deficits, the fundamental difficulty is that they do not provide a coherent explanation of how a limited population neural activity synchronized to the onset of a brief stimulus could be used to predict speech perception in noise which presumably recruits a vast network involving cortical and subcortical structures. Plainly put, the neural activity underlying the brainstem electrophysiologic measures used and the neural networks involved in speech perception in quiet and noise are not isomorphic.
No Differences in Latency Measures (Absolute Latency, I–V Latency, Latency Shift in Noise) Between Low and High-risk Group
While we did find differences in amplitude, we did not find group differences in peak latencies (I and V), inter-peak latency (I–V), and latency shift for waves I and V. These results suggest that the cochlear regions contributing to the response (affecting mostly the absolute latencies of the ABR component) and the neural conduction time (reflecting both synaptic integration time and neural conduction) are essentially similar for both groups and are not altered by noise exposure. This is contrary to the findings of Mehraei et al. (2016), where a 32 kHz tone pip elicited ABR showed reduced wave-IV latency shift with increase in masking noise level for cochlear synaptopathic mice model. The analysis of derived-band ABRs (Don & Eggermont 1978; Eggermont 1979) in humans have demonstrated that wave V receives contributions from a broad cochlear region. Thus, the wave V recording in our study would represent primary activity along a broader cochlear region which obscures the latency differences between groups compared to the mice study where ABR wave IV was measured for more place-specific 32 kHz tone pip. Also, it is around this region (32 kHz) the synaptopathic changes are most readily observed in mice models. Future studies employing high-frequency tone pip need to be considered to make any definitive conclusions regarding the usefulness of broadband masking noise to diagnose cochlear synaptopathy in humans. By comparing the effects of broadband masking noise on the click ABR in young and older adults, Burkard & Sims (2002) observed very similar latency shift of wave V in young adults and older adults with normal or near-normal hearing sensitivity. A recent study by Viana et al. (2015) using postmortem analysis of human temporal bone, and Sergeyenko et al. (2013) using age-graded series of mice observed age-related synaptopathy in the IHC area similar to noise-induced synaptopathy. This further highlights the importance of recording place-specific ABR to evaluate latency shift in masking noise.
Our findings of the study of reduced wave I amplitude in the high-risk group are consistent with the animal model of cochlear synaptopathy involving selective damage of LSR and MSR fibers and few human studies. These differences could be because of the methodology employed in this study; targeting participants which provided clear separation of noise distribution between groups and the frequency response of earphone which captured high-frequency contributions of wave I and the use of noise masking to separate the groups. But, without postmortem examination by harvesting human temporal bone (the gold standard for identifying synaptopathy) with different noise exposure background, no direct inferences can be derived for the presence/extent of cochlear synaptopathy in high-risk group with high sound over-exposure history. It is proposed that similar cochlear synaptopathic changes would be manifested in individuals with the traditional hearing loss (Liberman & Kujawa 2017). Unfortunately, ABR wave I is difficult to measure in hearing-impaired populations limiting the clinical use of this measure. However, the use of either tiptrodes and particularly the more invasive intrameatal tympanic membrane electrode will provide substantial enhancement in wave I amplitude thereby increasing the feasibility of measuring wave I as well as the SP.
Results of experiment II showed a smaller reduction in wave I amplitude in noise. To the extent that response amplitude reflects both the number of neural elements responding and the neural synchrony of the responding elements, the relatively smaller change in response amplitude for the high-risk group would suggest a reduced susceptibility to masking. One plausible mechanism would be that suppressive effects that kick in at moderate to high levels are different in these two groups, specifically, the reduction in LSR and MSR fibers in the high-risk group may reduce the suppressive masking effects, particularly at moderate levels of the masking noise. A similar reduced effect of noise on wave I amplitude was observed by Burkard & Sims (2002) in aged adults with near normal-hearing using tympanic membrane electrode.
More stimulus manipulations need to be attempted to derive measures which affect later components which could be measured even at threshold levels. In spite of reduced wave I amplitude, the later waves were very similar between groups suggesting central compensation due to inhibitory circuitry changes. Larger scale datasets with different noise exposure background, longitudinal measurements (changes due to recreational over-exposure by studying middle-school to high-school students enrolled in marching band) with an array of behavioral and electrophysiological tests to understand the complex pathogenesis of sound over-exposure damage in normal-hearing individuals.
Thanks to Dr. Jackson Gandour for his assistance with statistical analysis.
Attanasio G., Barbara M., Buongiorno G., et al. (Protective effect of the cochlear efferent system during noise exposure. Ann N Y Acad Sci, 1999). 884, 361–367
Bourien J., Tang Y., Batrel C., et al. (Contribution of auditory nerve fibers to compound action potential of the auditory nerve. J Neurophysiol, 2014). 112, 1025–1039
Bharadwaj H. M., Verhulst S., Shaheen L., et al. (Cochlear neuropathy and the coding of supra-threshold sound. Front Syst Neurosci, 2014). 8, 26
Bramhall N., Beach E. F., Epp B., Le Prell C. G., Lopez-Poveda E. A., Plack C. J., Schaette R., Verhulst S., Canlon B. (The search for noise-induced cochlear synaptopathy in humans: Mission impossible? Hear Res, 2019). 377, 88–103
Bramhall N. F., Konrad-Martin D., McMillan G. P., Griest S. E. (Auditory brainstem response altered in humans with noise exposure despite normal outer hair cell function. Ear Hear, 2017). 38, e1–e12
Burkard R., Hecox K. (The effect of broadband noise on the human brainstem auditory evoked response. II. Frequency specificity. J Acoust Soc Am, 1983). 74, 1214–1223
Burkard R., Hecox K. E. (The effect of broadband noise on the human brain-stem auditory evoked response. III. Anatomic locus. J Acoust Soc Am, 1987). 81, 1050–1063
Burkard R. F., Sims D. (A comparison of the effects of broadband masking noise on the auditory brainstem response in young and older adults. Am J Audiol, 2002). 11, 13–22
Burkard R.F, Don MBurkard R. F., Don M., Eggermont J. JThe auditory brainstem response.Auditory Evoked Potentials2007). Wolters Kluwer/Lippincott Williams & Wilkins.254–274
Cai Y., Geisler C. D. (Suppression in auditory-nerve fibers of cats using low-side suppressors. I. Temporal aspects. Hear Res, 1996a). 96, 94–112
Cai Y., Geisler C. D. (Suppression in auditory-nerve fibers of cats using low-side suppressors. II. Effect of spontaneous rates. Hear Res, 1996b). 96, 113–125
Caspary D. M., Havey D. C., Faingold C. L. (Effects of acetylcholine on cochlear nucleus neurons. Exp Neurol, 1983). 82, 491–498
Caspary D. M., Ling L., Turner J. G., Hughes L. F. (Inhibitory neurotransmission, plasticity and aging in the mammalian central auditory system. J Exp Biol, 2008). 211Pt 111781–1791
Caspary D. M., Schatteman T. A., Hughes L. F. (Age-related changes in the inhibitory response properties of dorsal cochlear nucleus output neurons: role of inhibitory inputs. J Neurosci, 2005). 25, 10952–10959
Chambers A. R., Resnik J., Yuan Y., et al. (Central Gain Restores Auditory Processing following Near-Complete Cochlear Denervation. Neuron, 2016). 89, 867–879
Costalupes J. A. (Representation of tones in noise in the responses of auditory nerve fibers in cats. I. Comparison with detection thresholds. J Neurosci, 1985). 5, 3261–3269
Delgutte B. (Two-tone rate suppression in auditory-nerve fibers: dependence on suppressor frequency and level. Hear Res, 1990). 49, 225–246
Don M., Eggermont J. J. (Analysis of the click-evoked brainstem potentials in man using high-pass noise masking. J Acoust Soc Am, 1978). 63, 1084–1092
Eggermont J. J. (Narrow-band AP latencies in normal and recruiting human ears. J Acoust Soc Am, 1979). 65, 463–470
Eggermont J. J. (Effects of long-term non-traumatic noise exposure on the adult central auditory system. Hearing problems without hearing loss. Hear Res, 2017). 352, 12–22
Elberling C., Kristensen S. G., Don M. (Auditory brainstem responses to chirps delivered by different insert earphones. J Acoust Soc Am, 2012). 131, 2091–2100
Fahey P. F., Allen J. B. (Nonlinear phenomena as observed in the ear canal and at the auditory nerve. J Acoust Soc Am, 1985). 77, 599–612.
Fulbright A. N. C., Le Prell C. G., Griffiths S. K., Lobarinas E. (Effects of recreational noise on threshold and suprathreshold measures of auditory function (Vol. 38, pp. 298–318). 2017). Presented at the Seminars in hearing, Thieme Medical Publishers.
Furman A. C., Kujawa S. G., Liberman M. C. (Noise-induced cochlear neuropathy is selective for fibers with low spontaneous rates. J Neurophysiol, 2013). 110, 577–586
Gorga M. P., Worthington D. W., Reiland J. K., et al. (Some comparisons between auditory brain stem response thresholds, latencies, and the pure-tone audiogram. Ear Hear, 1985). 6, 105–112
Grinn S. K., Wiseman K. B., Baker J. A., Le Prell C. G. (Hidden hearing loss? No effect of common recreational noise exposure on cochlear nerve response amplitude in humans. Front Neurosci, 2017). 11, 465
Grose J. H., Buss E., Hall J. W. III. (Loud music exposure and cochlear synaptopathy in young adults: isolated auditory brainstem response effects but no perceptual consequences. Trends Hear, 2017). 21, 2331216517737417
Gu J. W., Herrmann B. S., Levine R. A., Melcher J. R. (Brainstem auditory evoked potentials suggest a role for the ventral cochlear nucleus in tinnitus. J Assoc Res Otolaryngol, 2012). 13, 819–833
Henderson D., Subramaniam M., Spongr V., Attanasio G. (Salvi R. J., Henderson D. (Eds.), Biological mechanisms of the “toughening” phenomenon. In Auditory Plasticity and Regeneration (pp. 1996). Thieme Medical Publishers.143–154).
Hickox A. E., Larsen E., Heinz M. G., et al. (Translational issues in cochlear synaptopathy. Hear Res, 2017). 349, 164–171
Hickox A. E., Liberman M. C. (Is noise-induced cochlear neuropathy key to the generation of hyperacusis or tinnitus? J Neurophysiol, 2014). 111, 552–564
Hirsch A., Anderson H. (Audiologic test results in 96 patients with tumours affecting the eighth nerve a clinical study with emphasis on the early audiological diagnosis. Acta Oto-Laryngologica, 1980). 89(Suppl 369), 1–26
Jerger J. (Clinical experience with impedance audiometry. Arch Otolaryngol, 1970). 92, 311–324
Konrad-Martin D., Dille M. F., McMillan G., et al. (Age-related changes in the auditory brainstem response. J Am Acad Audiol, 2012). 23, 18–35; quiz 74.
Kujawa S. G., Liberman M. C. (Adding insult to injury: cochlear nerve degeneration after “temporary” noise-induced hearing loss. J Neurosci, 2009). 29, 14077–14085
Liberman M. C., Epstein M. J., Cleveland S. S., Wang H., Maison S. F. (Toward a Differential Diagnosis of Hidden Hearing Loss in Humans. PLoS One, 2016). 11, e0162726
Liberman M. C., Kujawa S. G. (Cochlear synaptopathy in acquired sensorineural hearing loss: Manifestations and mechanisms. Hear Res, 2017). 349, 138–147
Lin H. W., Furman A. C., Kujawa S. G., Liberman M. C. (Primary neural degeneration in the Guinea pig cochlea after reversible noise-induced threshold shift. J Assoc Res Otolaryngol, 2011). 12, 605–616
Lobarinas E., Spankovich C., Le Prell C. G. (Evidence of “hidden hearing loss” following noise exposures that produce robust TTS and ABR wave-I amplitude reductions. Hear Res, 2017). 349, 155–163
Lopez-Poveda E.A. (Why do I hear but not understand? Stochastic undersampling as a model of degraded neural encoding of speech. Front Neurosci, 2014). 8, 348
Maison S. F., Usubuchi H., Liberman M. C. (Efferent feedback minimizes cochlear neuropathy from moderate noise exposure. J Neurosci, 2013). 33, 5542–5552
Makary C. A., Shin J., Kujawa S. G., et al. (Age-related primary cochlear neuronal degeneration in human temporal bones. J Assoc Res Otolaryngol, 2011). 12, 711–717
Mehraei G., Hickox A. E., Bharadwaj H. M., et al. (Auditory brainstem response latency in noise as a marker of cochlear synaptopathy. J Neurosci, 2016). 36, 3755–3764
Mepani A., Kirk S., Hancock K., et al. (Middle ear muscle reflex and word recognitionin “normal-hearing” adults evidence for cochlear synaptopathy? Ear Hear, 202041, 25–38
Mitchell C., Phillips D. S., Trune D. R. (Variables affecting the auditory brainstem response: audiogram, age, gender and head size. Hear Res, 1989). 40, 75–85
Möhrle D., Ni K., Varakina K., et al. (Loss of auditory sensitivity from inner hair cell synaptopathy can be centrally compensated in the young but not old brain. Neurobiol Aging, 2016). 44, 173–184
Moore J. K. (The human auditory brain stem: a comparative view. Hear Res, 1987a). 29, 1–32
Moore J. K. (The human auditory brain stem as a generator of auditory evoked potentials. Hear Res, 1987b). 29, 33–43
Muthaiah V. P. K., Walls M., Heinz M. (Effects of cochlear-synaptopathy inducing moderate noise exposure on auditory-nerve-fiber responses in chinchillas. J Acoust Soc Am, 2017). 141, 3814
Osuji G. A., Onyeagu S. J., Ekezie Z. Z. (Comparison of Jackknife and resubstitution methods in the estimation of error rates in discriminant analysis. Int J Math Stat St, 2013). 129–37
Pappa A. K., Hutson K. A., Scott W. C., et al. (Hair cell and neural contributions to the cochlear summating potential. J Neurophysiol, 2019). 121, 2163–2180
Picton T. W., Hillyard S. A., Krausz H. I., Galambos R. (Human auditory evoked potentials. I. Evaluation of components. Electroencephalogr Clin Neurophysiol, 1974). 36, 179–190
Plack C. J., Barker D., Prendergast G. (Perceptual consequences of “hidden” hearing loss. Trends in Hearing, 2014). 18, 2331216514550621
Prendergast G., Guest H., Munro K. J., et al. (Effects of noise exposure on young adults with normal audiograms I: electrophysiology. Hear Res, 2017). 344, 68–81
Rhode W. S. (Some observations on cochlear mechanics. J Acoust Soc Am, 1978). 64, 158–176
Rhode W. S., Smith P. H. (Characteristics of tone-pip response patterns in relationship to spontaneous rate in cat auditory nerve fibers. Hear Res, 1985). 18, 159–168
Ridley C. L., Kopun J. G., Neely S. T., et al. (Using thresholds in noise to identify hidden hearing loss in humans. Ear Hear, 2018). 39, 829–844
Sachs M. B., Young E. D. (Encoding of steady-state vowels in the auditory nerve: Representation in terms of discharge rate. J Acoust Soc Am, 1979). 66, 470–479
Salvi R. J., Wang J., Ding D. (Auditory plasticity and hyperactivity following cochlear damage. Hear Res, 2000). 147, 261–274
Salvi R., Sun W., Ding D., et al. (Inner hair cell loss disrupts hearing and cochlear function leading to sensory deprivation and enhanced central auditory gain. Front Neurosci, 2016). 10, 621
Schaette R., McAlpine D. (Tinnitus with a normal audiogram: physiological evidence for hidden hearing loss and computational model. J Neurosci, 2011). 31, 13452–13457
Schmiedt R. A., Mills J. H., Boettcher F. A. (Age-related loss of activity of auditory-nerve fibers. J Neurophysiol, 1996). 76, 2799–2803
Schwaber M. K., Hall J. W. (Cochleovestibular nerve compression syndrome. I. Clinical features and audiovestibular findings. Laryngoscope, 1992). 102, 1020–1029
Sergeyenko Y., Lall K., Liberman M. C., Kujawa S. G. (Age-related cochlear synaptopathy: an early-onset contributor to auditory functional decline. J Neurosci, 2013). 33, 13686–13694
Shaheen L. A., Valero M. D., Liberman M. CTowards a diagnosis of cochlear acceleration of cochlear synaptopathy in “recovered” ears. J Neurosci, 2015). 35, 7509–7520
Sheppard A. M., Chen G. D., Manohar S., et al. (Prolonged low-level noise-induced plasticity in the peripheral and central auditory system of rats. Neuroscience, 2017). 359, 159–171
Skoe E., Tufts J. (Evidence of noise-induced subclinical hearing loss using auditory brainstem responses and objective measures of noise exposure in humans. Hear Res, 2018). 361, 80–91
Stamper G. C., Johnson T. A. (Auditory function in normal-hearing, noise-exposed human ears. Ear Hear, 2015). 36, 172–184
Starr A., Achor J. (Auditory brain stem responses in neurological disease. Arch Neurol, 1975). 32, 761–768
Subramaniam M., Campo P., Henderson D. (The effect of exposure level on the development of progressive resistance to noise. Hear Res, 1991). 52, 181–187
Tagoe T., Barker M., Jones A., et al. (Auditory nerve perinodal dysmyelination in noise-induced hearing loss. J Neurosci, 2014). 34, 2684–2688
Valderrama J. T., Beach E. F., Yeend I., et al. (Effects of lifetime noise exposure on the middle-age human auditory brainstem response, tinnitus and speech-in-noise intelligibility. Hear Res, 2018). 365, 36–48
Valero M. D., Burton J. A., Hauser S. N., et al. (Noise-induced cochlear synaptopathy in rhesus monkeys (Macaca mulatta). Hear Res, 2017). 353, 213–223
Viana L. M., O’Malley J. T., Burgess B. J., et al. (Cochlear neuropathy in human presbycusis: confocal analysis of hidden hearing loss in post-mortem tissue. Hear Res, 2015). 327, 78–88
Walter J. S. (Sound exposures of high school marching band
members during band camp: a pilot study. J Band Res, 2011). 46, 18–59
Wu P. Z., Liberman L. D., Bennett K., et al. (Primary neural degeneration in the human cochlea: Evidence for hidden hearing loss in the aging ear. Neuroscience, 2019). 407, 8–20
Young E. D., Barta P. E. (Rate responses of auditory nerve fibers to tones in noise near masked threshold. J Acoust Soc Am, 1986). 79, 426–442