Martin L. Whitehead

Evidence for two discrete sources of 2f1−f2 distortion‐product otoacoustic emission in rabbit: I. Differential dependence on stimulus parameters

The results of studies of the physiological vulnerability of distortion-product otoacoustic emissions (DPOAEs) suggest that the DPOAE at 2f1-f2 in vertebrate ears is generated by more than one source. The principal aims of the present study were to provide independent evidence for the existence of more than one DPOAE source, and to determine the contributions of each to the ear-canal 2f1-f2 signal. To accomplish these aims, specific stimulus parameters were separately and systematically varied to provide detailed parametric information regarding 2f1-f2 DPOAE amplitude and phase in normal ears of awake rabbits. The findings indicate that two discrete sources, demonstrating differential dependence on stimulus parameters, dominate the generation of the 2f1-f2 DPOAE. One source of distortion is dominant above 60-70 dB SPL at moderate primary-frequency separations, and at all stimulus levels when the primary tones are closely spaced. The other source is dominant below 60-70 dB SPL at moderate primary-frequency separations, and may be dominant at all stimulus levels when the primary tones are widely separated in frequency. The results suggest that by varying stimulus parameters, it may be possible to independently study the two generator mechanisms.

Dependence of distortion‐product otoacoustic emissions on primary levels in normal and impaired ears. I. Effects of decreasing L2 below L1

The 2f1-f2 distortion-product otoacoustic emission (DPOAE) is evoked by two primary tones of frequencies f1 < f2, and levels L1 and L2. Previous reports indicate that decreasing L2 below L1 = L2 can; (1) increase DPOAE amplitude in normal ears, and (2) increase the degree to which DPOAE amplitudes are reduced by cochlear trauma. Although both of these factors could be advantageous for clinical applications of DPOAEs, neither has been explored in detail. In the present study, 2f1-f2 DPOAE-amplitude frequency functions were collected from normal and impaired ears of rabbits and humans, with L1 = L2, and with L2 < L1, at each of three values of L1. In rabbits, controlled tonal or noise overexposures were used to produce permanent reductions of DPOAE amplitudes. Comparison of pre- and postexposure DPOAE-amplitude frequency functions demonstrated that the frequency-specific reductions of DPOAEs were enhanced by decreasing L2 below L1. In humans, DPOAE-amplitude frequency functions obtained with the various L1 and L2 combinations were collected from 16 normal ears to provide preliminary normative data for each stimulus-level condition. The L1-L2 that produced the maximum DPOAE amplitude in normal ears was systematically dependent on L1. Thus at most frequencies, decreasing L2 below L1 = L2 substantially reduced mean DPOAE amplitude when L1 > or = 75 dB SPL, but increased mean DPOAE amplitudes at L1 = 65 dB SPL. However, the increase of mean DPOAE amplitude obtained by decreasing L2 below L1 = 65 dB SPL was small, being less than 3.5 dB at most frequencies. More importantly, at L1 = 65 dB SPL, L2 could be decreased considerably below L1 = L2 without reducing mean DPOAE amplitude relative to that at L1 = L2. Inspection of DPOAE-amplitude frequency functions obtained from subjects with mild or moderate sensorineural hearing losses indicated that, in frequency regions of hearing impairment, decreasing L2 below L1 can enhance the degree of reduction of DPOAEs below the corresponding normative amplitudes, without reducing the normative amplitude. It is concluded that decreasing L2 below L1 = L2 has the potential to enhance the performance of DPOAEs in clinical applications.

Evidence for two discrete sources of 2f1−f2 distortion‐product otoacoustic emission in rabbit. II: Differential physiological vulnerability

In a previous report, it was shown that, in normal rabbit ears, the amplitude and phase of 2f1-f2 distortion-product otoacoustic emissions (DPOAEs) elicited by low-level (< 60-70 dB SPL) stimuli display a differential dependence on stimulus parameters to those evoked by high-level (> 60-70 dB SPL) stimuli, indicating differences in the underlying generation mechanisms. In the present study, the physiological vulnerability of DPOAEs in each of the two 2f1-f2 DPOAE-response regions identified on the basis of differential parametric properties, was characterized. Thus emissions evoked using stimulus levels from 45-75 dB SPL were measured over time upon: (1) induction of lethal anoxia, (2) acute injection of ethacrynic acid, and (3) acute injection of ethacrynic acid 2 h after a single administration of gentamicin. The DPOAEs evoked by low-level stimuli (45 dB SPL) were abolished within 3-4 min of induction of anoxia, whereas DPOAEs evoked by high-level stimuli (75 dB SPL) were unchanged in this period. The high-level emissions decreased with a complex time course postmortem, and demonstrated behaviors, including evidence of susceptibility to fatigue, suggesting a dependence upon a cochlear energy supply. Low-level DPOAEs could be temporarily abolished, with complete recovery, by an acute administration of ethacrynic acid that had little effect on high-level DPOAEs. Treatment with the gentamicin and ethacrynic-acid combination, which would be expected to produce widespread hair-cell damage, eliminated low-level DPOAEs, and greatly reduced high-level emissions. In combination with previously published data, these findings strongly suggest that low- and high-level 2f1-f2 DPOAEs arise from discrete sources. The data are consistent with the proposal that the low-level DPOAE source is an active, micromechanical process, but suggest that the proposed origin of high-level DPOAEs exclusively in the passive macromechanics of the cochlear partition may be incorrect. The elimination of both low- and high-level DPOAEs revealed the presence of a third, residual 2f1-f2 DPOAE component, approximately 75-80 dB below the stimulus-tone levels, that may reflect the true passive-distortion response of the cochlea.

Dependence of distortion-product otoacoustic emissions on primary levels in normal and impaired ears. II. Asymmetry in L1,L2 space.

Previous studies indicate that the amplitude of 2f1-f2 distortion-product otoacoustic emissions (DPOAEs), evoked by two tones of frequencies f1 < f2, demonstrates a complex dependence on the levels (L1 and L2) of the primary tones. In the present study, 2f1-f2 DPOAE amplitudes were measured over a wide range of L1 and L2 in normal human ears, allowing a systematic, level-dependent asymmetry of DPOAE amplitude in L1,L2 space to be characterized. The L1,L2 at which DPOAEs were largest was close to L1 = L2 at high stimulus levels, but moved monotonically toward L1 > L2 as stimulus levels decreased. A related observation was that DPOAE amplitude had a greater dependence on L1 and on L2. These asymmetries were quantified in normal human ears, and compared to the corresponding asymmetries apparent in data from animal models. Recent studies have demonstrated that the reduction of DPOAE amplitude by cochlear trauma is greater when L1 > L2 than when L1 = L2, suggesting that the reduction of DPOAEs by trauma demonstrates an asymmetry in L1,L2 space that is qualitatively similar to that of normative DPOAE amplitude. To investigate this issue, 2f1-f2 DPOAE amplitudes were measured over a wide range of L1 and L2 in rabbit ears pre- and postinjection of the ototoxic loop-diuretic ethacrynic acid. The results indicate that the asymmetry in L1,L2 space of the reduction of DPOAEs by trauma is both qualitatively and quantitatively similar to the asymmetry in L1,L2 space of normative DPOAE amplitude. Specifically, the L1 values that maximized normative DPOAE amplitudes for any specified L2 (or, equivalently, the L1 values that allowed L2 to be minimized for any specified normative DPOAE amplitude) also yielded the greatest reduction of DPOAEs by the diuretic. In humans, the L1 values that maximize normative DPOAE amplitudes for any specified L2 are well approximated by a simple equation, with parameters that vary with frequency and f2/f1. It is suggested that the L1,L2 values defined by this equation may be optimum for use in clinical applications.

Clinical testing of distortion-product otoacoustic emissions.

ABSTRACT Otoacoustic emissions have great promise for use in clinical tests of the functional status of outer hair cells, which represent cochlear structures that make a major contribution to the hearing process. A substantial literature is available concerning the evaluation of outer hair cell function by transiently evoked otoacoustic emissions. However, relatively little attention has been focused on the benefits of testing with distortion-product otoacoustic emissions. The purpose of this presentation is to provide knowledge of the principal advantages offered by distortion-product emissions testing.

Visualization of the onset of distortion‐product otoacoustic emissions, and measurement of their latency

This paper describes a method for visualization of the onset of distortion-product otoacoustic emission (DPOAE) waveforms in the time domain. The DPOAE waveforms are obtained using ensemble averaging of samples of microphone output. A rectangular sample window is used, and the primary tones are turned on within the sample window. The phases of the primary tones (f1 and f2) are varied systematically between samples in such a way that the primary tones, and all DPOAEs (e.g., 2f2-f1, 3f1-2f2, 2f1), except the DPOAE of interest (e.g., 2f1-f2), are cancelled in the ensemble average. Visualization of the DPOAE onset allows measurement of the onset latency (OSL) of the DPOAE. These direct measurements of OSL are compared to phase-gradient latencies (PGLs) in the same ears determined by measuring the phase change of the DPOAE as a function of DPOAE frequency. The direct measures of OSL vary from > 10 to < 1 ms, decrease with increasing frequency and increasing stimulus level, and are shorter in rabbits than humans. The direct measures of OSL are, in general, quantitatively similar to PGL estimates, but there are exceptions. Visualization of DPOAE onset also allows quantification of DPOAE rise times, and reveals phase and amplitude changes of the DPOAE that occur several milliseconds after onset in rabbits and humans. It is proposed that the phase and amplitude changes result from vector summation of multiple components of the DPOAE signal, each with a different latency.

Sensitivity of distortion-product otoacoustic emissions in humans to tonal over-exposure: time course of recovery and effects of lowering L2.

An important concern of industrial hearing-conservation programs is detecting the onset of noise-induced hearing loss. If it can be shown that otoacoustic emissions are sufficiently sensitive to reliably detect auditory fatigue and the permanent hearing loss that eventually develops, they could become an important part of the hearing-conservation test battery. The present study in humans was designed to examine the influence of overall primary-tone level and the effects of lowering the f2 primary on the sensitivity of distortion-product otoacoustic emissions (DPOAEs) to acoustic overstimulation. One ear from each of 14 subjects with normal hearing was exposed to a 105-dB SPL pure tone at 2.8 kHz for 3 min using a protocol consisting of distinct pre-exposure, exposure, and post-exposure periods. As a quantitative index of the functional status of the outer hair cells, 2f1-f2 DPOAEs were monitored systematically over time using four stimulus-test conditions consisting of either one of two levels of equilevel primary tones, or one of two levels of offset primaries, with L2 set 25 dB lower than L1. The overall finding was that the DPOAE protocol incorporating both the lowest level of stimulation and an f2-primary tone that was 25 dB below the level of the f1 stimulus [i.e., L1 (55 dB SPL) - L2 (30 dB SPL) = 25 dB] was most sensitive to the exposure effects. The results establish that DPOAEs elicited with unequal, in contrast to equal-level primaries, have comparable signal-to-noise ratios, but are considerably more sensitive to reductions in emission levels induced by exposure to short-lasting, moderately intense tones. The recovery of DPOAE amplitudes over the first 15 min post-exposure appeared to be roughly linear in log time and, in many cases, could be closely approximated by fitting a logarithmic curve to the post-exposure data. From these functions, the initial amount of loss (y-intercept) and the slope of recovery were identified as potential measures of vulnerability to acoustic exposure in that these variables appeared to be related to the susceptibility of some of the subjects, who also participated in a subsequent experiment on the behavioral effects of the exposure stimulus. Finally, compared to behaviorally measured temporary threshold shift (TTS), the time course of the recovery for DPOAEs was very similar, suggesting that, with the appropriate parameters, DPOAEs can be as sensitive to TTS as routine pure-tone audiometry.

Locus of generation for the 2 f1−f2 vs 2 f2−f1 distortion-product otoacoustic emissions in normal-hearing humans revealed by suppression tuning, onset latencies, and amplitude correlations

The present study used distortion-product otoacoustic emission (DPOAE) suppression tuning curves (STCs), DPOAE onset latencies (OLs), and DPOAE amplitude correlations to investigate the locus of generation of the 2f1-f2 DPOAE versus the 2f2-f1 DPOAE in humans. The results of the tuning study revealed that, for the 2f1-f2 DPOAE, the tips of the STCs tuned consistently below the geometric-mean (GM) frequency of the primary tones. In contrast, for the 2f2-f1 DPOAE, STCs tuned above the GM of the primaries, with 50% of the tip frequencies at, or above, the 2f2-f1 frequency place. When the average ratio of the 2f2-f1 to the 2f1-f2 tip frequencies was computed, a factor of 1.44 provided an estimate of the frequency shift needed to align the two DPOAE generation sites. Other results showed that OLs for the 2f2-f1 DPOAE were uniformly shorter than those for the 2f1-f2, with differences at the low frequencies amounting to as much as 6-7 ms. Further, for both DPOAEs, curves describing latency decreases as a function of increasing GM frequencies were best fit by power functions. Shifting the GM frequency producing the 2f2-f1 DPOAE by a factor of 1.6 caused the latency distributions for both DPOAEs to overlap thus resulting in a single function that described cochlear delay as a function of GM frequency. Finally, for each GM frequency in the DP-gram, sliding correlations from 108 normal ears were performed on both DPOAEs by holding the primaries producing the 2f1-f2 DPOAE constant, while all 2f2-f1 DPOAE amplitudes were successively correlated with the 2f1-f2 amplitudes. This procedure demonstrated that, for a given GM frequency producing the 2f1-f2, the correlations between the two DPOAEs peaked when the primaries of the 2f2-f1 were at a GM frequency that positioned the 2f2-f1 frequency place near the GM of the primaries that produced the 2f1-f2 DPOAE. As a whole, the above findings strongly suggest that the 2f2-f1 DPOAE in humans is generated basal to the primary-tone place on the basilar membrane.

Spontaneous Otoacoustic Emissions in Different Racial Groups

To determine if there are racial differences in the prevalence of spontaneous otoacoustic emissions (SOAEs), both ears of 20 Negro, 20 Asian and 20 Caucasian subjects were examined for the presence of SOAEs. Within each racial group, equal numbers of normally hearing males and females were tested. Significant differences in the occurrence of SOAEs were found between the three racial groups, with Negroes expressing more SOAEs than Caucasians, and Asians demonstrating an intermediate number of these emissions. In support of previous observations, more emissions were recorded from female than from male ears, and a significant correlation of the number of emissions in the two ears of an individual was also noted.

Time-windowing of click-evoked otoacoustic emissions to increase signal-to-noise ratio

Objective To investigate the effects of decreasing the response-window duration on the signal-to-noise ratio (Sm) of click-evoked otoacoustic emissions (CEOAEs). Design The ILOSS (Otodynamics, Ltd.) was used to measure CEOAEs from 149 normal adult ears, and 75 adult ears with high-frequency sensorineural hearing loss. Data were collected using the default response window of 2.5 to 20.5 msec post-click. Each response was rewindowed, post-hoc, from 2.5 to 7.5 msec, 2.5 to 9 msec, 7.75 to 14.25 msec, and 13 to 19.5 msec post-click. For each window, spectra of the CEOAE and of the background noise were determined. The S/N was estimated by subtracting the noise level from the CEOAE amplitude. Results The 13- to 19.5-msec window contained little CEOAE energy relative to earlier windows. Relative to the 2.5- to 20.5-msec window, the 2.5- to 7.5- and 2.5- to 9-msec windows reduced noise levels more than CEOAE amplitudes, yielding increased S/N, and greater “reproducibility” values. The increased S/N of the 2.5- to 7.5- and 2.5- to 9-msec windows allowed measurement of greater CEOAE-amplitude reductions in the impaired ears relative to the normal ears. With short-duration windows, click-presentation rate could be increased, allowing more responses to be averaged in a given time, thus further decreasing noise levels. Although click rate was not varied in the present study, the decrease of noise levels is predictable. Accounting for this factor, it is expected that a specified S/N would be obtained about five times faster using the 2.5- to 7.5-msec window with a 7.5-msec interstimulus interval, than when using the default window. Conclusions Decreasing the response-window duration substantially increases the measurement efficiency of CEOAEs in adults, and thus may enhance clinical-test performance.