Quantitative assessment of human cochlear function by evoked otoacoustic emissions

Deep Learning Models for Predicting Hearing Thresholds Based on Swept-Tone Stimulus-Frequency Otoacoustic Emissions

OBJECTIVES

This study aims to develop deep learning (DL) models for the quantitative prediction of hearing thresholds based on stimulus-frequency otoacoustic emissions (SFOAEs) evoked by swept tones.

DESIGN

A total of 174 ears with normal hearing and 388 ears with sensorineural hearing loss were studied. SFOAEs in the 0.3 to 4.3 kHz frequency range were recorded using linearly swept tones at a rate of 2 Hz/msec, with stimulus level changing from 40 to 60 dB SPL in 10 dB steps. Four DL models were used to predict hearing thresholds at octave frequencies from 0.5 to 4 kHz. The models-a conventional convolutional neural network (CNN), a hybrid CNN-k-nearest neighbor (KNN), a hybrid CNN-support vector machine (SVM), and a hybrid CNN-random forest (RF)-were individually built for each frequency. The input to the DL models was the measured raw SFOAE amplitude spectra and their corresponding signal to noise ratio spectra. All DL models shared a CNN-based feature self-extractor. They differed in that the conventional CNN utilized a fully connected layer to make the final regression decision, whereas the hybrid CNN-KNN, CNN-SVM, and CNN-RF models were designed by replacing the last fully connected layer of CNN model with a traditional machine learning (ML) regressor, that is, KNN, SVM, and RF, respectively. The model performance was evaluated using mean absolute error and SE averaged over 20 repetitions of 5 × 5 fold nested cross-validation. The performance of the proposed DL models was compared with two types of traditional ML models.

RESULTS

The proposed SFOAE-based DL models resulted in an optimal mean absolute error of 5.98, 5.22, 5.51, and 6.06 dB at 0.5, 1, 2, and 4 kHz, respectively, superior to that obtained by the traditional ML models. The produced SEs were 8.55, 7.27, 7.58, and 7.95 dB at 0.5, 1, 2, and 4 kHz, respectively. All the DL models outperformed any of the traditional ML models.

CONCLUSIONS

The proposed swept-tone SFOAE-based DL models were capable of quantitatively predicting hearing thresholds with satisfactory performance. With DL techniques, the underlying relationship between SFOAEs and hearing thresholds at disparate frequencies was explored and captured, potentially improving the diagnostic value of SFOAEs.

Comparing spontaneous and stimulus frequency otoacoustic emissions in mice with tectorial membrane defects

The global standing-wave model for generation of spontaneous otoacoustic emissions (SOAEs) suggests that they are amplitude-stabilized standing waves and that the spacing between SOAEs corresponds to the interval over which the phase changes by one cycle as determined from the phase-gradient delays of stimulus frequency otoacoustic emissions (SFOAEs). Because data characterizing the relationship between spontaneous and evoked emissions in nonhuman mammals are limited, we examined SOAEs and SFOAEs in tectorial membrane (TM) mutants and their controls. Computations indicate that the spacing between adjacent SOAEs is predicted by the SFOAE phase-gradient delays for TM mutants lacking Ceacam16, where SOAE frequencies are greater than ~20 kHz and the mutants retain near-normal hearing when young. Mice with a missense mutation in Tecta (Tecta^Y1870C/+), as well as mice lacking Otoancorin (Otoa^-/-), were also examined. Although these mutants exhibit hearing loss, they generate SOAEs with average frequencies of 11 kHz in Tecta^Y1870C/+ and 6 kHz in Otoa^-/-. In these animals, the spacing between adjacent SOAEs is larger than predicted by the SFOAE phase delays. It is also demonstrated that mice do not exhibit the strong frequency-dependence in signal coding that characterizes species with good low-frequency hearing. In fact, a transition occurs near the apical end of the mouse cochlea rather than at the mid-point along the cochlear partition. Hence, disagreements with the standing-wave model are not easily explained by a transition in tuning ratios between apical and basal regions of the cochlea, especially for SOAEs generated in Tecta^Y1870C/+mice.

Estimating Hearing Thresholds From Stimulus-Frequency Otoacoustic Emissions

It is of clinical interest to estimate pure-tone thresholds from potentially available objective measures, such as stimulus-frequency otoacoustic emissions (SFOAEs). SFOAEs can determine hearing status (normal hearing vs. hearing loss), but few studies have explored their further potential in predicting audiometric thresholds. The current study investigates the ability of SFOAEs to predict hearing thresholds at octave frequencies from 0.5 to 8 kHz. SFOAE input/output functions and pure-tone thresholds were measured from 230 ears with normal hearing and 737 ears with sensorineural hearing loss. Two methods were used to predict hearing thresholds. Method 1 is a linear regression model; Method 2 proposed in this study is a back propagation (BP) network predictor built on the bases of a BP neural network and principal component analysis. In addition, a BP network classifier was built to identify hearing status. Both Methods 1 and 2 were able to predict hearing thresholds from 0.5 to 8 kHz, but Method 2 achieved better performance than Method 1. The BP network classifiers achieved excellent performance in determining the presence or absence of hearing loss at all test frequencies. The results show that SFOAEs are not only able to identify hearing status with great accuracy at all test frequencies but, more importantly, can predict hearing thresholds at octave frequencies from 0.5 to 8 kHz, with best performance at 0.5 to 4 kHz. The BP network predictor is a potential tool for quantitatively predicting hearing thresholds, at least at 0.5 to 4 kHz.

Maximising the ability of stimulus-frequency otoacoustic emissions to predict hearing status and thresholds using machine-learning models

OBJECTIVE

This study aimed to maximise the ability of stimulus-frequency otoacoustic emissions (SFOAEs) to predict hearing status and thresholds based on machine-learning models.

DESIGN

SFOAE data and audiometric thresholds were collected at octave frequencies from 0.5 to 8 kHz. Support vector machine, k-nearest neighbour, back propagation neural network, decision tree, and random forest algorithms were used to build classification models for status identification and to develop regression models for threshold prediction.

STUDY SAMPLE

About 230 ears with normal hearing and 737 ears with sensorineural hearing loss.

RESULTS

All classification models yielded areas under the receiver operating characteristic curve of 0.926-0.994 at 0.5-8 kHz, superior to the previous SFOAE study. The regression models produced lower standard errors (8.1-12.2 dB, mean absolute errors: 5.53-8.97 dB) as compared to those for distortion-product and transient-evoked otoacoustic emissions previously reported (8.6-19.2 dB).

CONCLUSIONS

SFOAEs using machine-learning approaches offer promising tools for the prediction of hearing capabilities, at least at 0.5-4 kHz. Future research may focus on further improvements in accuracy and reductions in test time to improve clinical utility.

Profiles of Stimulus-Frequency Otoacoustic Emissions from 0.5 to 20 kHz in Humans

The characteristics of human otoacoustic emissions (OAEs) have not been thoroughly examined above the standard audiometric frequency range (>8 kHz). This is despite the fact that deterioration of cochlear function often starts at the basal, high-frequency end of the cochlea before progressing apically. Here, stimulus-frequency OAEs (SFOAEs) were obtained from 0.5 to 20 kHz in 23 young, audiometrically normal female adults and three individuals with abnormal audiograms, using a low-to-moderate probe level of 36 dB forward pressure level (FPL). In audiometrically normal ears, SFOAEs were measurable at frequencies approaching the start of the steeply sloping high-frequency portion of the audiogram (∼12-15 kHz), though their amplitudes often declined substantially above ∼7 kHz, rarely exceeding 0 dB SPL above 8 kHz. This amplitude decline was typically abrupt and occurred at a frequency that was variable across subjects and not strongly related to the audiogram. In contrast, certain ears with elevated mid-frequency thresholds but regions of normal high-frequency sensitivity could possess surprisingly large SFOAEs (>10 dB SPL) above 7 kHz. When also measured, distortion-product OAEs (DPOAEs) usually remained stronger at higher stimulus frequencies and mirrored the audiogram more closely than SFOAEs. However, the high-frequency extent of SFOAE and DPOAE responses was similar when compared as a function of the response frequency, suggesting that middle ear transmission may be a common limiting factor at high frequencies. Nevertheless, cochlear factors are more likely responsible for complexities observed in high-frequency SFOAE spectra, such as abrupt amplitude changes and narrowly defined response peaks above 10 kHz, as well as the large responses in abnormal ears. These factors may include altered cochlear reflectivity due to subtle damage or the reduced spatial extent of the SFOAE generation region at the cochlear base. The use of higher probe levels is necessary to further evaluate the characteristics and potential utility of high-frequency SFOAE measurements.

Tuning of SFOAEs Evoked by Low-Frequency Tones Is Not Compatible with Localized Emission Generation

Stimulus-frequency otoacoustic emissions (SFOAEs) appear to be well suited for assessing frequency selectivity because, at least on theoretical grounds, they originate over a restricted region of the cochlea near the characteristic place of the evoking tone. In support of this view, we previously found good agreement between SFOAE suppression tuning curves (SF-STCs) and a control measure of frequency selectivity (compound action potential suppression tuning curves (CAP-STC)) for frequencies above 3 kHz in chinchillas. For lower frequencies, however, SF-STCs and were over five times broader than the CAP-STCs and demonstrated more high-pass rather than narrow band-pass filter characteristics. Here, we test the hypothesis that the broad tuning of low-frequency SF-STCs is because emissions originate over a broad region of the cochlea extending basal to the characteristic place of the evoking tone. We removed contributions of the hypothesized basally located SFOAE sources by either pre-suppressing them with a high-frequency interference tone (IT; 4.2, 6.2, or 9.2 kHz at 75 dB sound pressure level (SPL)) or by inducing acoustic trauma at high frequencies (exposures to 8, 5, and lastly 3-kHz tones at 110-115 dB SPL). The 1-kHz SF-STCs and CAP-STCs were measured for baseline, IT present and following the acoustic trauma conditions in anesthetized chinchillas. The IT and acoustic trauma affected SF-STCs in an almost indistinguishable way. The SF-STCs changed progressively from a broad high-pass to narrow band-pass shape as the frequency of the IT was lowered and for subsequent exposures to lower-frequency tones. Both results were in agreement with the "basal sources" hypothesis. In contrast, CAP-STCs were not changed by either manipulation, indicating that neither the IT nor acoustic trauma affected the 1-kHz characteristic place. Thus, unlike CAPs, SFOAEs cannot be considered as a place-specific measure of cochlear function at low frequencies, at least in chinchillas.

Influence of stimulus parameters on amplitude-modulated stimulus frequency otoacoustic emissions

The present study evaluated the influence of suppressor frequency (fs) and level (Ls) on stimulus-frequency otoacoustic emissions (SFOAEs) recorded using the amplitude-modulated (AM) suppressor technique described by Neely et al. [J. Acoust. Soc. Am. 118, 2124-2127 (2005a)]. Data were collected in normal-hearing subjects, with data collection occurring in two phases. In phase 1, SFOAEs were recorded with probe frequency (fp) = 1, 2, and 4 kHz and probe levels (Lp) ranging from 0 to 60 dB sound pressure level (SPL). At each fp, Ls ranged from Ls = Lp to Ls = Lp + 30 dB. Additionally, nine relationships between fs and fp were evaluated, ranging from fs/fp = 0.80 to fs/fp = 1.16. Results indicated that for low suppressor levels, suppressors higher in frequency than fp (fs > fp) resulted in higher AM-SFOAE levels than suppressors lower in frequency than fp (fs < fp). At higher suppressor levels, suppressors both higher and lower in frequency than fp produced similar AM-SFOAE levels, and, in many cases, low-frequency suppressors produced the largest response. Recommendations for stimulus parameters that maximize AM-SFOAE level were derived from these data. In phase 2, AM-SFOAEs were recorded using these parameters for fp = 0.7-8 kHz and Lp = 20-60 dB SPL. Robust AM-SFOAE responses were recorded in this group of subjects using the parameters developed in phase 1.

The relationship between distortion product otoacoustic emissions and extended high-frequency audiometry in tinnitus patients. Part 1: normally hearing patients with unilateral tinnitus

BACKGROUND

The aim of this study was to evaluate distortion product otoacoustic emissions (DPOAEs) and extended high-frequency (EHF) thresholds in a control group and in patients with normal hearing sensitivity in the conventional frequency range and reporting unilateral tinnitus.

MATERIAL/METHODS

Seventy patients were enrolled in the study: 47 patients with tinnitus in the left ear (Group 1) and 23 patients with tinnitus in the right ear (Group 2). The control group included 60 otologically normal subjects with no history of pathological tinnitus. Pure-tone thresholds were measured at all standard frequencies from 0.25 to 8 kHz, and at 10, 12.5, 14, and 16 kHz. The DPOAEs were measured in the frequency range from approximately 0.5 to 9 kHz using the primary tones presented at 65/55 dB SPL.

RESULTS

The left ears of patients in Group 1 had higher median hearing thresholds than those in the control subjects at all 4 EHFs, and lower mean DPOAE levels than those in the controls for almost all primary frequencies, but significantly lower only in the 2-kHz region. Median hearing thresholds in the right ears of patients in Group 2 were higher than those in the right ears of the control subjects in the EHF range at 12.5, 14, and 16 kHz. The mean DPOAE levels in the right ears were lower in patients from Group 2 than those in the controls for the majority of primary frequencies, but only reached statistical significance in the 8-kHz region.

CONCLUSIONS

Hearing thresholds in tinnitus ears with normal hearing sensitivity in the conventional range were higher in the EHF region than those in non-tinnitus control subjects, implying that cochlear damage in the basal region may result in the perception of tinnitus. In general, DPOAE levels in tinnitus ears were lower than those in ears of non-tinnitus subjects, suggesting that subclinical cochlear impairment in limited areas, which can be revealed by DPOAEs but not by conventional audiometry, may exist in tinnitus ears. For patients with tinnitus, DPOAE measures combined with behavioral EHF hearing thresholds may provide additional clinical information about the status of the peripheral hearing.

Characterizing distortion-product otoacoustic emission components across four species

Distortion-product otoacoustic emissions (DPOAEs) were measured as level/phase (L/P) maps in humans, rabbits, chinchillas, and rats with and without an interference tone (IT) placed either near the 2f(1)-f(2) DPOAE frequency place (f(dp)) or at one-third of an octave above the f(2) primary tone (1/3-oct IT). Vector differences between with and without IT conditions were computed to derive a residual composed of the DPOAE components removed by the IT. In humans, a DPOAE component could be extracted with the expected steep phase gradient indicative of reflection emissions by ITs near f(dp). In the laboratory species, ITs near f(dp) failed to produce any conclusive evidence for reflection components. For all species, 1/3-oct ITs extracted large DPOAE components presumably generated at or basal to the IT-frequency place that exhibited both distortion- and reflection-like phase properties. Together, these findings suggested that basal distortion components could assume reflection-like phase behavior when the assumptions of cochlear-scaling symmetry, the basis for shallow phase gradients for constant f(2)/f(1) ratio sweeps, are violated. The present results contradict the common belief that DPOAE components associated with steep or shallow phase slopes are unique signatures for reflection emissions arising from f(dp) or distortion emissions generated near f(2), respectively.

Transient evoked otoacoustic emissions in superior canal dehiscence syndrome

Superior semicircular canal dehiscence syndrome (SCDS) is a clinical disorder that is characterized by vertigo and oscillopsia induced by loud sounds. Transient evoked otoacoustic emissions (TEOAEs) allow to noninvasively check the integrity of the cochlea. The present study aimed at identifying cochlear stress as the result of micro alterations of the cochlear functionality due to anatomic anomaly. 11 SCDS and 10 normal individuals as control group were submitted to history taking, otological examination, basic audiologic evaluation and TEOAEs analysis using the standard wideband protocol and moving time window analysis. Although TEOAEs test results showed no statistically significant difference using the standard protocol, off-line analysis of the waveforms' "effective duration" was statistically significantly shortened (p < 0.0001) when compared to normal ears. In conclusion, dehiscence of bone overlying the superior canal has been shown to have effects on inner ear function in terms of a third mobile window theory, thus altering pressure across cochlear partition with decrease in inner ear impedance.

Evidence for basal distortion-product otoacoustic emission components

Distortion-product otoacoustic emissions (DPOAEs) were measured with traditional DP-grams and level/phase (L/P) maps in rabbits with either normal cochlear function or unique sound-induced cochlear losses that were characterized as either low-frequency or notched configurations. To demonstrate that emission generators distributed basal to the f(2) primary-tone contribute, in general, to DPOAE levels and phases, a high-frequency interference tone (IT) was presented at 1/3 of an octave (oct) above the f(2) primary-tone, and DPOAEs were re-measured as "augmented" DP-grams (ADP-grams) and L/P maps. The vector difference between the control and augmented functions was then computed to derive residual DP-grams (RDP-grams) and L/P maps. The resulting RDP-grams and L/P maps, which described the DPOAEs removed by the IT, supported the notion that basal DPOAE components routinely contribute to the generation of standard measures of DPOAEs. Separate experiments demonstrated that these components could not be attributed to the effects of the 1/3-oct IT on f(2), or DPOAEs generated by the addition of a third interfering tone. These basal components can "fill in" the lesion estimated by the commonly employed DP-gram. Thus, ADP-grams more accurately reveal the pattern of cochlear damage and may eventually lead to an improved DP-gram procedure.

Distortion product otoacoustic emission of symphony orchestra musicians before and after rehearsal

The 2f1-f2 distortion product otoacoustic emission (DPOAE) and hearing levels are obtained for 12 normal-hearing symphony orchestra musicians both before and after their rehearsal. The DPOAE fine structures are determined and analyzed according to the character and prevalence of ripples. Hearing levels, DPOAE levels, and DPOAE fine structures before and after rehearsal are similar, indicating that no or marginal temporary change of the state of hearing were caused by the exposure. The data were further compared to similar data for occupationally nonexposed subjects, one group which was age and gender matched, and other two groups of younger individuals (one group with better hearing levels than the other). The data for the age and gender matched group compared well with the musicians data (and the data for the group of better-hearing younger individuals). In general, the analyses of hearing thresholds and DPOAE data thus lead to the same conclusions concerning the state of hearing.

Distortion product otoacoustic emission fine structure analysis of 50 normal-hearing humans

When distortion product otoacoustic emissions (DPOAEs) are measured with a high-frequency resolution, the DPOAE shows quasi-periodic variations across frequency, called DPOAE fine structure. In this study the DPOAE fine structure is determined for 50 normal-hearing humans using fixed primary levels of L1/L2 = 65/45 dB. An algorithm is developed, which characterizes the fine structure ripples in terms of three parameters: ripple spacing, ripple height, and ripple prevalence. The characteristic patterns of fine structure can be found in the DPOAE of all subjects, though the DPOAE fine structure characteristics are individual and vary from subject to subject. On average the ripple spacing decreases with increasing frequency from 1/8 oct at 1 kHz to 3/32 oct at 5 kHz. The ripple prevalence is two to three ripples per 1/3 oct, and ripple heights of up to 32 dB could be detected. The 50 normal-hearing subjects were divided into two groups, the subjects of group A having slightly better hearing levels than subjects of group B. The subjects of group A have significantly higher DPOAE levels. The overall prevalence of fine structure ripples do not differ between the two groups, but are higher and narrower for subjects of group B than for group A.

Audiometric predictions using stimulus-frequency otoacoustic emissions and middle ear measurements

OBJECTIVE

The goals of the study are to determine how well stimulus-frequency otoacoustic emissions (SFOAEs) identify hearing loss, classify hearing loss as mild or moderate-severe, and correlate with pure-tone thresholds in a population of adults with normal middle ear function. Other goals are to determine if middle ear function as assessed by wideband acoustic transfer function (ATF) measurements in the ear canal account for the variability in normal thresholds, and if the inclusion of ATFs improves the ability of SFOAEs to identify hearing loss and predict pure-tone thresholds.

DESIGN

The total suppressed SFOAE signal and its corresponding noise were recorded in 85 ears (22 normal ears and 63 ears with sensorineural hearing loss) at octave frequencies from 0.5 to 8 kHz, using a nonlinear residual method. SFOAEs were recorded a second time in three impaired ears to assess repeatability. Ambient-pressure ATFs were obtained in all but one of these 85 ears and were also obtained from an additional 31 normal-hearing subjects in whom SFOAE data were not obtained. Pure-tone air and bone conduction thresholds and 226-Hz tympanograms were obtained on all subjects. Normal tympanometry and the absence of air-bone gaps were used to screen subjects for normal middle ear function. Clinical decision theory was used to assess the performance of SFOAE and ATF predictors in classifying ears as normal or impaired, and linear regression analysis was used to test the ability of SFOAE and ATF variables to predict the air conduction audiogram.

RESULTS

The ability of SFOAEs to classify ears as normal or hearing impaired was significant at all test frequencies. The ability of SFOAEs to classify impaired ears as either mild or moderate-severe was significant at test frequencies from 0.5 to 4 kHz. SFOAEs were present in cases of severe hearing loss. SFOAEs were also significantly correlated with air conduction thresholds from 0.5 to 8 kHz. The best performance occurred with the use of the SFOAE signal-to-noise ratio as the predictor, and the overall best performance was at 2 kHz. The SFOAE signal-to-noise measures were repeatable to within 3.5 dB in impaired ears. The ATF measures explained up to 25% of the variance in the normal audiogram; however, ATF measures did not improve SFOAEs predictors of hearing loss except at 4 kHz.

CONCLUSIONS

In common with other OAE types, SFOAEs are capable of identifying the presence of hearing loss. In particular, SFOAEs performed better than distortion-product and click-evoked OAEs in predicting auditory status at 0.5 kHz; SFOAE performance was similar to that of other OAE types at higher frequencies except for a slight performance reduction at 4 kHz. Because SFOAEs were detected in ears with mild to severe cases of hearing loss, they may also provide an estimate of the classification of hearing loss. Although SFOAEs were significantly correlated with hearing threshold, they do not appear to have clinical utility in predicting a specific behavioral threshold. Information on middle ear status as assessed by ATF measures offered minimal improvement in SFOAE predictions of auditory status in a population of normal and impaired ears with normal middle ear function. However, ATF variables did explain a significant fraction of the variability in the audiograms of normal ears, suggesting that audiometric thresholds in normal ears are partially constrained by middle ear function as assessed by ATF tests.

Sources and Mechanisms of DPOAE Generation: Implications for the Prediction of Auditory Sensitivity

Otoacoustic emissions (OAEs) have become a commonly used clinical tool for assessing cochlear health status, in particular, the integrity of the cochlear amplifier or motor component of cochlear function. Predicting hearing thresholds from OAEs, however, remains a research challenge. Models and experimental data suggest that there are two mechanisms involved in the generation of OAEs. For distortion product, transient, and high-level stimulus frequency emissions, the interaction of multiple sources of emissions in the cochlea leads to amplitude variation in the composite ear canal signal. Multiple sources of emissions complicate simple correlations between audiometric test frequencies and otoacoustic emission frequencies. Current research offers new methods for estimating the individual components of OAE generation. Input-output functions and DP-grams of the nonlinear component of the 2f2-f2 DPOAE may ultimately show better correlations with hearing thresholds. This paper reviews models of OAE generation and methods for estimating the contribution of source components to the composite emission that is recorded in the ear canal. The clinical implications of multiple source components are discussed.

Patients affected with Fabry disease have an increased incidence of progressive hearing loss and sudden deafness: an investigation of twenty-two hemizygous male patients

BACKGROUND

Fabry disease (FD, OMIM 301500) is an X-linked inborn error of glycosphingolipid metabolism due to the deficient activity of alpha-galactosidase A, a lysosomal enzyme. While the progressive systemic deposition of uncleaved glycosphingolipids throughout the body is known to have protean clinical manifestations, few data are available regarding the cochlear involvement.

METHODS

We non-invasively investigated cochlear functions in 22 consecutive hemizygous males (age 19-64 years, mean 39) affected with classic FD. Conventional audiometry, tympanometry, ABR audiometry, otoacoustic emissions were performed in all patients, together with medical history record and physical examination as part of an exhaustive baseline evaluation prior to enzyme replacement therapy.

RESULTS

A total of 12 patients (54.5%) with classic FD were found to have abnormal audition. Five patients had progressive hearing loss and seven patients (32%) experienced sudden deafness. In addition, a hearing loss on high-tone frequencies was found in 7 out of the 10 remaining patients without clinical impairment, despite their young age at time of examination. The incidence of hearing loss appeared significantly increased in FD patients with kidney failure (P < 0.01) or cerebrovascular lesions (P < 0.01), whereas there was no correlation with left ventricular hypertrophy. In addition, tinnitus aurium was also found in six patients (27%).

CONCLUSION

This is the first evidence of a high incidence of both progressive hearing loss and sudden deafness in a cohort of male patients affected with classic Fabry disease. The exact pathophysiologic mechanism(s) of the cochlear involvement deserves further studies.

Static input-output non-linearity as the source of non-linear effects in maximum length sequence click-evoked OAEs

The application of the maximum length sequence (MLS) technique to the recording of click-evoked otoacoustic emissions (CEOAEs) allows for a reduction in test time by one to two orders of magnitude. This is because the technique permits the use of extremely high click rates, as inter-click intervals are not constrained to be greater than the duration of the response. However, increasing the click rate also causes a progressive reduction in amplitude, or 'suppression', of the CEOAE. The origin of this suppression is unclear, with diverse suggestions in the literature as to its nature and mechanism. This paper presents a simple model of the well-known compressive non-linearity of the CEOAE level function, based on a static amplitude non-linearity within each of a number of narrowband frequency channels. The response of the model to MLS stimulation demonstrates suppression broadly of the form and magnitude previously reported in experimental studies. Furthermore, the model exhibits the generation of additional non-linear components that have been speculated on in connection with CEOAE recordings using the MLS technique. It is concluded that the MLS suppression phenomenon is derived largely, if not entirely, from the static non-linearity of the CEOAE level function. The approach to modelling the phenomenon as described here also bears promise for understanding various aspects of non-linearity in MLS-based CEOAE recordings.

Suppression of tone burst evoked otoacoustic emissions in relation to frequency separation

Tone burst evoked otoacoustic emissions (TBEOAEs) were measured for two tone bursts presented separately and as a two-tone burst complex to examine the linearity of TBEOAE generators for different frequency separations of the stimuli. The stimuli were: (a) tone bursts of 5-ms duration and center frequencies of 1, 1.5, 2 and 3 kHz; (b) complex stimuli with the 1-kHz tone burst combined digitally with each of the other specified tone bursts. Signals were delivered at 70 dB SPL using a non-linear processing method and at 60 dB SPL using a linear method to 21 ears of normally hearing adults. Spectra of TBEOAEs obtained with single-tone bursts were superimposed (composite) and compared to those of the two-tone burst complex. A close correspondence between the composite and complex spectra was present in all ears. However, the components on the higher-frequency slope of the 1-kHz spectral peak were reduced in the complex spectra obtained with a frequency separation of 0.5 kHz when compared to the corresponding composite spectra. The reduction was greater at a stimulus level of 70 dB SPL than with 60 dB SPL. The effect was smaller for a frequency separation of 1 kHz, and almost absent for the tone burst separation of 2 kHz. Thus, suppression leads to weak non-linear frequency superposition for higher-level, closely spaced stimuli.

Identification of neonatal hearing impairment: transient evoked otoacoustic emissions during the perinatal period

OBJECTIVES

1) To describe transient evoked otoacoustic emission (TEOAE) levels, noise levels and signal to noise ratios (SNRs) for a range of frequency bands in three groups of neonates who were tested as a part of the Identification of Neonatal Hearing Impairment multi-center consortium project. 2) To describe the relations between these TEOAE measurements and age, test environment, baby state, and test time.

DESIGN

TEOAEs were measured in 4478 graduates of neonatal intensive care units (NICUs), 353 well babies with at least one risk indicator, and 2348 well babies without risk factors. TEOAE and noise levels were measured for frequency bands centered at 1.0, 1.5, 2.0, 3.0, and 4.0 kHz for a click stimulus level of 80 dB SPL. For those ears not meeting "passing" stopping criteria at 80 dB pSPL, a level of 86 dB pSPL was included. Measurement-based stopping rules were used such that a test did not terminate unless the response revealed a criterion SNR in four out of five frequency bands or no response occurred after a preset number of averages. Baby state, test environment, and other test factors were captured at the time of test.

RESULTS

TEOAE levels, noise levels and SNRs were similar for NICU graduates, well babies with risk factors and well babies without risk factors. There were no consistent differences in response quality as a function of test environment, i.e., private room, unit, open crib, nonworking isolette, or working isolette. Noise level varied little across risk group, test environment, or infant state other than crying, suggesting that the primary source of noise in TEOAE measurements is infant noise. The most significant effect on response quality was center frequency. Responses were difficult to measure in the half-octave band centered at 1.0 kHz, compared with higher frequencies. Reliable responses were measured routinely at frequencies of 1.5 kHz and higher.

CONCLUSIONS

TEOAEs are easily measured in both NICU graduates and well babies with and without risk factors for hearing loss in a wide variety of test environments. Given the difficulties encountered in making reliable measurements for a frequency band centered at 1.0 kHz, its inclusion in a screening program may not be justified.

Nonlinear temporal interactions in click-evoked otoacoustic emissions. II. Experimental data

Click-evoked otoacoustic emissions (CEOAEs) are reduced in amplitude by the presentation of 'suppressor' clicks that either closely lead or follow the stimulus ('test') clicks. A model described in a companion paper (Kapadia and Lutman, Hear. Res. 146 (2000) 89-100) shows that such nonlinear temporal interactions, as previously reported, may be explained in terms of the compressive non-linearity of the CEOAE input-output (I-O) function. This paper presents the results of a detailed parametric investigation into such nonlinear interactions, studied in 12 normal adult ears over a wide range of test and suppressor click levels and inter-click intervals. The results differ from those generated by the model in a number of respects. Principally, maximum suppression is generally obtained for suppressors presented in advance of test clicks, rather than co-incident with the test clicks. The amount of advance depends systematically on the two click levels. The measured suppression can also exceed the theoretical maximum allowed by the model. It is concluded that the nonlinear temporal interactions measured do not simply reflect CEOAE I-O function non-linearity. They may, instead, arise from disturbance of the generator elements from their resting state prior to generation of the CEOAE. These results may also have general implications relating to cochlear responses to transient stimuli and indicate the potential of CEOAEs in probing aspects of cochlear mechanics.

Otoacoustic emissions--an approach for monitoring aminoglycoside induced ototoxicity in children

OBJECTIVES

The early detection of hearing impairment caused by ototoxic drugs, such as aminoglycosides, has been the aim of research world-wide. Histopathological studies have shown that the outer hair cells are the most susceptible cochlear components to injury from ototoxic drugs like aminoglycosides. Otoacoustic emissions reflect the functional status of the outer hair cells and constitute the only non-invasive means of objective cochlear investigation. The aim of this study was to evaluate the potential of otoacoustic emissions in early identification of aminoglycoside-induced cochlear dysfunction. In addition, a comparison with pure-tone audiometry or auditory brainstem responses was performed in order to determine if this test might provide a more reliable method of monitoring early ototoxic insults to the cochlea.

METHODS

Twenty four children receiving gentamicin (4 mg/kg once daily) for 6-29 days were included in the study. Eleven children received gentamicin for up to 7 days (group A), while 13 underwent longer-term therapy lasting 8-29 days (group B). Hearing was serially monitored using transient evoked otoacoustic emissions and pure-tone audiometry (0.25-12 kHz) or auditory brainstem responses for younger or uncooperative children. Transient evoked otoacoustic emissions data were analysed in terms of emission amplitude and response reproducibility as a function of frequency.

RESULTS

All patients yielded a normal baseline audiometric assessment upon hospital admission. For group A patients no significant changes in hearing levels were observed either by pure-tone audiometry (P = 0.2), auditory brainstem responses (P = 0.3) or transient evoked otoacoustic emissions (mean response: P = 0.06, reproducibility by frequency: P > 0.05). For group B patients no significant changes in hearing levels measured by pure-tone audiometry (P = 0.1) or auditory brainstem responses (P = 0.4) were observed. Transient evoked otoacoustic emissions however revealed a statistically significant decrease in the mean response level (P = 0.017) and in the reproducibility over the whole frequency spectrum (1 kHz: P = 0.0057, 2 kHz: P = 0.0247, 3 kHz: P = 0.0134, 4 kHz: P = 0.0049, 5 kHz: P = 0.0019).

CONCLUSIONS

The findings suggest that transient evoked otoacoustic emissions are an extremely sensitive measure of the early effects of aminoglycoside-induced injury to the peripheral auditory system. Therefore, their use is recommended for regular monitoring of cochlear function, in the presence of potentially toxic factors, aiming at prevention of permanent damage.

Frequency-specific information from click evoked otoacoustic emissions in noise-induced hearing loss

Click evoked otoacoustic emissions (CEOAEs), pure-tone audiograms (PTAs), and Bekesy sweep frequency audiograms were recorded from 15 ears of 11 subjects with noise-induced hearing loss. For all ears, hearing threshold levels > or = 30 dB HL were found at the high frequencies. The aims of the study were to examine whether the decomposition of CEOAEs into narrow band components could identify hearing loss in a frequency-specific manner and to what extent audiometric thresholds could be predicted. CEOAEs were parcelled into 0.5-kHz-wide components by means of the wavelet transform. Reproducibility of CEOAE components was compared with audiometric threshold at corresponding frequencies. A general trend of low reproducibility for increasing audiometric thresholds was found. A reproducibility value of 60 per cent was found to best separate normal and elevated thresholds. The presence of a CEOAE component at a given frequency was always associated with audiometric thresholds < or = 20-25 dB HL. On the other hand, the absence of a component was equally associated either with normal or abnormal hearing levels. Large inter-subject variability was observed. A weak linear relationship was found between reproducibility and audiometric thresholds at corresponding frequencies, indicating that analysis of narrow band CEOAE components is valuable for separating normal from hearing-impaired ears but cannot replace the audiogram.

Predicting audiometric status from distortion product otoacoustic emissions using multivariate analyses

OBJECTIVES

1) To determine whether multivariate statistical approaches improve the classification of normal and impaired ears based on distortion product otoacoustic emission (DPOAE) measurements, in comparison with the results obtained with more traditional single-variable applications of clinical decision theory. 2) To determine how well the multivariate predictors, derived from analysis on a training group, generalized to a validation group. 3) To provide a way to apply the multivariate approaches clinically.

DESIGN

Areas under the relative operating characteristic (ROC) curve and cumulative distributions derived from DPOAE, DPOAE/Noise, discriminant function (DF) scores and logit function (LF) scores were used to compare univariate and multivariate predictors of audiometric status. DPOAE and Noise amplitudes for 8 f2 frequencies were input to a discriminant analysis and to a logistic regression. These analyses generated a DF and LF, respectively, composed of a linear combination of selected variables. The DF and LF scores were the input variables to the decision theory analyses. For comparison purposes, DPOAE test performance was also evaluated using only one variable (DPOAE or DPOAE/Noise when f2 = audiometric frequency). Analyses were based on data from over 1200 ears of 806 subjects, ranging in age from 1.3 to 96 yr, with thresholds ranging from -5 to >120 dB HL. For statistical purposes, normal hearing was defined as thresholds of 20 dB HL or better. For the multivariate analyses, the database was randomly divided into two groups of equal size. One group served as the "training" set, which was used to generate the DFs and LFs. The other group served as a "validation" set to determine the robustness of the DF and LF solutions.

RESULTS

For all test frequencies, multivariate analyses yielded greater areas under the ROC curve than univariate analyses, and greater specificities at fixed sensitivities. Within the multivariate techniques, discriminant analysis and logistic regression yielded similar results and both yielded robust solutions that generalized well to the validation groups. The improvement in test performance with multivariate analyses was greatest for conditions in which the single predictor variable resulted in the poorest performance.

CONCLUSIONS

A more accurate determination of auditory status at a specific frequency can be obtained by combining multiple predictor variables. Although the DF and LF multivariate approaches resulted in the greatest separation between normal and impaired distributions, overlap still exists, which suggests that there would be value in continued efforts to improve DPOAE test performance.

Do click-evoked otoacoustic emissions have frequency specificity?

Whether click-evoked otoacoustic emissions (CEOAEs) have frequency specificity is an issue still subject to debate. In order to resolve this issue, changes in the frequency components of the CEOAE power spectrum, together with changes in compound action potential (CAP) thresholds before and after pure-tone exposure in guinea pigs, were examined. Changes in CAP thresholds immediately before and 1 h after exposure were compared with changes in the frequency components in the CEOAE power spectrum before and 1 h after exposure. The ILO 88 was used for measurement of CEOAEs. Total echo energy in the CEOAE power spectrum was converted into frequency bands of 1000 Hz. Shifts in filtered echo power (FEP) levels correlated maximally with those in CAP thresholds at 0.5 kHz above the same frequency. Stepwise regression indicated that only one step could be entered in a linear regression model using the variable of CAP threshold shifts at 0.5 kHz above the same frequency for all FEP shifts except FEP4.5. The remaining variables played a negligible role, since variance no longer changed when they were included in the regression equation. From these results, it was concluded that CEOAEs display frequency specificity. Influence on CEOAEs from higher frequencies is negligible.

Transient evoked otoacoustic emissions in patients with normal hearing and in patients with hearing loss

OBJECTIVES

1) To evaluate transient evoked otoacoustic emission (TEOAE) test performance when measurements are made under routine clinical conditions. 2) To evaluate TEOAE test performance as a function of frequency and as a function of the magnitude of hearing loss. 3) To compare test performance using univariate and multivariate approaches to data analyses. 4) To provide a means of interpreting clinical TEOAE measurements.

DESIGN

TEOAEs were measured in 452 ears of 246 patients. All measurements were made after acoustic immittance assessments, which were used to demonstrate that middle-ear function was normal at the time of the TEOAE test. TEOAE amplitudes and signal to noise ratios (SNRs), analyzed into octave bands centered at 1, 2, and 4 kHz, were compared with the pure-tone threshold at the same frequencies. Data were analyzed with clinical decision theory, cumulative distributions, discriminant analyses, and logistic regressions.

RESULTS

Using univariate analysis techniques, TEOAEs accurately identified auditory status at 2 and 4 kHz but were less accurate at 1 kHz. Test performance was best when audiometric thresholds between 20 and 30 dB HL were used as the criteria for normal hearing. TEOAE SNR resulted in better test performance than did TEOAE amplitude alone; this effect decreased as frequency increased. Multivariate analysis methods resulted in better separation between normal and impaired ears than did univariate approaches, which relied on only TEOAE amplitude or SNR when test frequency band and audiometric frequency were the same. This improvement in test performance was greatest at 1 kHz, decreased as frequency increased, and was negligible at 4 kHz.

CONCLUSIONS

TEOAEs can be used to identify hearing loss in children under routine clinical conditions. Univariate tests accurately identified auditory status at mid and high frequencies but performed more poorly at lower frequencies. The decrease in performance as frequency decreases may be a result of increased noise at lower frequencies but also may be due to properties of the measurement paradigm ("QuickScreen," high-pass filter at 0.8 kHz), which would not be ideal for recording energy around 1 kHz. The improvement in test performance when SNR was used and the interaction of this effect with frequency, however, would be consistent with the view that test performance in lower frequencies is at least partially influenced by the level of background noise. Multivariate analysis techniques improved test performance compared with the more traditional univariate approaches to data analysis. An approach is provided that allows one to assign measured TEOAE amplitudes, SNRs, or outputs from multivariate analyses to one of three categories: response properties consistent with normal hearing; results consistent with hearing loss; hearing status undetermined.

The relation between otoacoustic emissions and the broadening of the auditory filter for higher levels

The active behaviour of outer hair cells (OHCs) is often used to explain two phenomena, namely otoacoustic emissions (OAEs) and the level dependence of auditory filters. Correlations between these two phenomena may contribute to the evidence of these hypotheses. In this study auditory filters were calculated from probe thresholds in notched-noise maskers over a range of at least 25 dB. Further. transient evoked otoacoustic emissions (TEOAEs) were measured at several stimulation levels. Ten normal-hearing and nine hearing-impaired subjects were tested. A linear increase of the width of the auditory filter with 2.2 dB/Hz was found up to a specific saturation level. The group of selected hearing-impaired subjects with mild hearing loss showed no wider than normal auditory filters. As expected, the increase of the width of the auditory filter correlated positively with the level of TEOAEs for click intensities of about 80 dB peak SPL. However, for subjects with TEOAEs wider auditory filters at a masker level of 65 dB/Hz were found for subjects with larger TEOAEs. This result cannot be explained by a model by which the cochlea shows an active behaviour for lower stimulation levels, influencing both the TEOAE levels and the filter skirts, and a passive behaviour for higher stimulation levels.

Multi-variant analysis of otoacoustic emissions and estimation of hearing thresholds: transient evoked otoacoustic emissions

Evaluation of cochlear hearing loss by means of transiently evoked otoacoustic emissions is already established in clinical practice. However, accurate prediction of pure-tone thresholds is still questioned and is still regarded as troublesome. Both click- and tone-burst-evoked otoacoustic emissions at several intensity levels were measured and analysed in 157 ears from normally hearing and 432 ears from patients with different degrees of pure sensory hearing loss using the ILO88/92 equipment. Results of otoacoustic emissions (OAE), elicited by clicks and tone-bursts at centre frequencies from 1 to 5 kHz, were analysed using two different statistical methods. Both multivariate discriminant analysis and forward multiple regression analysis were used to determine which OAE variables were most discriminating and best at predicting hearing thresholds. We found that a limited set of variables obtained from both tone-burst and click measurements can accurately predict and categorize hearing loss levels up to a limit of 60 dB HL. We found correct classification scores of pure-tone thresholds between 500 and 4000 Hz up to 100 per cent when using combined click and tone-burst otoacoustic measurements. Prediction of pure-tone thresholds was correct with a maximum estimation error of 10 dB for audiometric octave frequencies between 500 and 4000 Hz. Measurements of multiple tone-bursts OAEs have a significant clinical advantage over the use of clicks alone for clinical applications, and a good classification and prediction of pure-tone thresholds with otoacoustic emissions is possible.

Auditory, visual and audiovisual perception of segmental speech features by severely hearing-impaired children

Auditory alone, visual alone and audiovisual recognition of consonant-vowel consonant syllables were measured in 32 severely hearing-impaired children with hearing loss (PTA) in a narrow range around 90 dB HL when using their hearing aids. Multidimensional scaling analysis (INDSCAL) and information transmission analysis (ITA), applied to the confusion matrices obtained from the responses in each presentation mode and for each phoneme category, revealed perceptual dimensions and percentages of transmitted feature information (PTI). These were studied in relation to PTA, the auditory alone score and in relation to the efficiency of the audiovisual interaction (enhancement) over the probalistic summation of the auditory alone and visual alone score. INDSCAL analysis shows that auditory alone recognition of vowels is based on the perceptual dimensions F2 and F1 and that of consonants on the dimensions 'frication' and 'voicing'. In the auditory mode the interpretation of the INDSCAL dimensions in the stimulus spaces is in reasonable agreement with the ITA results. PTI decreases gradually with decreasing auditory alone phoneme score. Audiovisual recognition of vowels is based on a combination of the auditory dimension 'open/closed' (F1), and the visual dimensions 'lip rounding' and 'vertical lip opening'. Audiovisual recognition of initial consonants is based on a combination of the visual dimension 'front/back' and the auditory dimension 'continuance'. Recognition of final consonants is based on a combination of the visual dimension 'front/back' and an uninterpretable dimension. The perceptual dimensions are independent of both the level of the auditory alone phoneme score and audiovisual enhancement. Audiovisual enhancement is mainly a property of an individual and independent of both auditory alone and visual alone scores. ITA analysis, based on a phonological classification of the features, supports the results of the INDSCAL analysis in the auditory alone mode. It is not useful in the description of the audiovisual interaction, probably due to the phonological basis of the feature classification.

Auditory screening in high-risk pre-term and full-term neonates using transient evoked otoacoustic emissions and brainstem auditory evoked potentials

The present report concerns a 3 year, 8 month hearing screening in 1531 high-risk neonates by means of two successive transient evoked otoacoustic emission (TEOAE) recordings followed, in cases of suspected hearing loss, by brainstem auditory evoked potential (BAEP) recording and otolaryngology (ORL) consultation. After TEOAE 1 and 2 and BAEP testing, 1361 infants (88.9%) were declared normal, and 170 (11.1%) suspected of hearing loss. Of these 170, 58 showed bilateral and 26 unilateral impairment. Definite hearing loss on ORL consultation was diagnosed in 14 infants (0.9% of the screened population as a whole); 22 are still being followed, while 86 (5.6%) failed to consult for diagnosis. The mean age on diagnosis of definite hearing loss was 9.9 +/- 4.9 (range 4-20) months. Several auditory function risk factors proved more frequent in deaf than in normal children. Our results show that early hearing loss screening in at-risk neonates needs to be pursued.

On the existence of an age/threshold/frequency interaction in distortion product otoacoustic emissions

Interactions among age, threshold, and frequency in relation to distortion product otoacoustic emissions (DPOAE) have yet to be resolved. The effects of these variables were explored by analyzing DPOAEs in ears with thresholds not exceeding 20 dB HL. Multivariate regression analyses were performed in two different ways. For data to be included in the first analysis, audiometric threshold had to be 20 dB HL or better only at the particular frequency under study, but might exceed 20 dB HL at other half-octave frequencies. Significant main effects were found for age, threshold, and frequency. There was also an age-by-frequency interaction, but a significant age-by-threshold interaction was not observed. DPOAE amplitudes decreased as either age, frequency, or threshold increased. In the second analysis, when a more stringent inclusion criterion was applied (normal thresholds at all frequencies), the main effects for age, threshold, and frequency were not significant. The significant age-by-frequency interaction remained, whereby DPOAE amplitudes decreased as age and frequency increased, but the age-by-threshold interaction again was not significant. The magnitude of DPOAE amplitude change across age, threshold, and frequency and for the age-by-frequency interaction was small but similar for both groups of subjects. Age in association with threshold did not account for observed changes in DPOAE amplitudes for either group. Importantly, the lack of a significant age-by-threshold interaction indicates that there may be processes intrinsic to aging alone that act on DPOAE generation.

COAE thresholds: 1. Effects of equal-amplitude versus subtraction methods

Although research has demonstrated that click-evoked otoacoustic emissions (COAEs) elicited by high-level stimuli are useful for identifying hearing loss, the ability of COAEs to predict behavioral thresholds has not been adequately tested. Results of studies comparing COAE thresholds and behavioral thresholds have been equivocal, perhaps due to the need for a more rigorous approach to COAE threshold estimation. The present study was designed to address several methodological concerns in COAE threshold testing, particularly the effects of two methods of stimulus presentation on COAE testing and threshold calculation. In an attempt to make COAE threshold estimation consistent across participants, COAE threshold calculations were based on mean noise floor levels across participants. COAE and noise floor levels were measured in 15 participants using both equal-amplitude clicks and a subtraction method. Broadband COAEs were analyzed into 1/3 octave bands, so that input/output functions could be examined and COAE thresholds could be calculated for each 1/3 octave band. Comparison of the two stimulus methods indicated several differences. Mean noise floor levels for the equal-amplitude method were approximately 6 dB lower than those measured for the subtraction method across frequency. In many cases COAEs evoked using the equal-amplitude method were higher in amplitude than those evoked using the subtraction method. COAE thresholds measured using the equal-amplitude click stimuli were significantly lower than those measured using the subtraction method. The significantly higher thresholds obtained using the subtraction method may be attributed in part to the reduction of COAE amplitude by the subtraction procedure, and not merely to the higher noise level. Slopes of the input/output functions were not significantly different between the two stimulus methods. These results suggest that the equal-amplitude method is preferable for COAE threshold testing because lower noise floor and larger amplitude COAEs may be obtained in the same test time.

Changes in evoked otoacoustic emissions in the guinea pig after pure-tone acoustic overstimulation

To test if click-evoked otoacoustic emissions (CEOAEs) have frequency specificity, continuous changes in CEOAEs (especially frequency components of the CEOAE power spectrum) after pure-tone exposure in guinea pigs were examined. Pure-tone stimuli (0.5 kHz, 120 dB SPL; 2 kHz, 115 dB SPL; 4 kHz, 110 dB SPL) were given in a closed system for 3 min. After exposure, the frequency components in the CEOAE power spectrum decreased maximally at one-half octave or more above the overstimulation frequency. They partially recovered 2 h after exposure. The time course of compound action potential (CAP) thresholds after exposure was similar to that of the frequency components of the CEOAE power spectrum. It was concluded that some local damage caused by outer hair-cell dysfunction in the guinea pig cochlea can be detected by measuring shifts in frequency components in the CEOAE power spectrum.

Are normal hearing thresholds a sufficient condition for click-evoked otoacoustic emissions?

Transiently evolked otoacoustic emissions (TEOAE) have been reported in several studies as absent in a small minority of normal ears. Other studies have reported TEOAEs in all normal ears. Differences between studies may arise directly from criteria for TEOAE identification, criteria for selection of normals, or statistically due to limited sample sizes. In order to understand and model cochlear processes involved in TEOAE generation, it needs to be known whether the presence of normal hearing leads automatically to generation of TEOAEs. The present study set out to establish in a large sample if any ears could be found that lacked TEOAEs despite normal hearing threshold levels (HTL). A total of 397 ears from highly cooperative adult subjects were examined under laboratory conditions. Using cross correlation between replicate nonlinear waveforms as the criterion, TEOAEs were present in 99.2% of the sample (lower CI 98.1%). However, careful visual assessment of the recorded waveforms for the remaining ears did not unequivocally show absence of TEOAE characteristics in any ear with normal HTLs. While TEOAE strength varies widely among ears, no clear evidence was found to show that TEOAEs can be absent when HTLs are normal.

Frequency-specific identification of hearing loss using transient-evoked otoacoustic emissions to clicks and tones

Transient-evoked otoacoustic emissions (TEOAE) to clicks and to 500- and 2000-Hz brief tones were measured in 72 normal-hearing and hearing-impaired subjects (86 ears). The TEOAE's reproducibility parameter was used for the analyses. The purpose of the investigation was to determine which stimuli best predicted the presence of sensorineural hearing loss in a frequency-specific manner at 500, 1000, 2000, and 4000 Hz. Analyses of the TEOAEs filtered into frequency-specific bands showed that separation of normal and impaired ears at 1000, 2000 and 4000 Hz was best achieved by TEOAEs evoked by clicks. Identification of hearing loss at 500 Hz was best obtained using 500-Hz tone-evoked TEOAEs filtered using a band centered at 500 Hz. Octave- and half-octave-wide bands identified hearing loss equally well. An analysis sweep time of 20 ms provided slightly better results compared to 30 ms, except for 500 Hz, where the 30-ms sweep time slightly improved the identification of hearing loss. Increases in the audiometric criterion did not yield better test performance once hearing loss exceeded 20 dB HL. The findings from this study suggest that the combination of bandpass-filtered TEOAEs to clicks and TEOAEs to 500-Hz tones identifies with reasonable accuracy ears with sensorineural hearing loss at 500, 1000, 2000, and 4000 Hz.

Limitations in the use of distortion product otoacoustic emissions in objective audiometry as the result of fine structure

There is strong evidence for a link between intact cochlear function and otoacoustic emissions (OAE). However, all attempts to find a close correlation between auditory thresholds and amplitudes of distortion product otoacoustic emissions (DPOAE) have failed. As an explanation for these findings, we have studied DPOAE fine structure and its dependence on increasing primary sound levels. Errors due to different calibrations of equipment for measuring DPOAE and auditory thresholds were also investigated. DPOAE were measured in 16 subjects using a frequency range of 500-1000 Hz. Frequencies were changed in 12.5 Hz steps at primary levels of 55, 60, 65 and 70 dB SPL. DPOAE amplitudes were found to vary by up to 20 dB for a frequency step of 50 Hz. Some fine structures showed narrow dips that shifted in frequency and diminished in amplitude with increasing primary levels. These findings demonstrated that sampling DPOAE amplitudes at widely spaced frequencies gave incomplete information about true course. DPOAE growth functions measured close to a dip in the DPOAE fine structure were rendered useless by interference with either the frequency shift or amplitude variations of the dip at different primary levels.

Transient-evoked otoacoustic emissions and contralateral suppression in patients with unilateral tinnitus

Transient-evoked otoacoustic emissions (TEOAE) have been recorded from 20 patients with unilateral tinnitus and symmetrical hearing up to 2000Hz. Responses were digitally low-pass filtered at 2000Hz. The function of the medial olivocochlear system (MOC) was examined by contralateral acoustic stimulation (CAS) with broadband noise at 50 and 70 dB SPL. Emission amplitude was significantly lower in tinnitus ears in the 10-15ms part of the response. No difference in contralateral suppression of amplitude could be found. The latency shifts were small and were not helpful in separating tinnitus from non-tinnitus ears. Analysis on different parts of the response window may be a useful tool for separating tinnitus from non-tinnitus ears. The lack of efferent effect differences could result from influence of stimuli on tinnitus mechanisms.

Moving time window analysis of transiently evoked otoacoustic emissions

We have developed a technique for off-line analysis of transient otoacoustic emissions. The correlation of two curves was calculated in a moving time-window of 1 ms time sections. By this method it is possible to recognize high-correlation parts, time sections with low correlation, and thus the duration of the emission can be determined. The mean absolute duration of transiently evoked otoacoustic emissions in 36 normal ears was 15.1 +/- 0.69 ms (mean +/- SE). The duration was significantly shorter (9.34 +/- 0.43) in 75 ears with different grades of sensorineural hearing loss. The breakdown of this group into sub-groups according to the width of the high frequency band with elevated threshold showed a correlation between the severity of the hearing loss and the absolute duration.

[Deafness in the neonatal period: basis for screening]

Deafness must be recognized in infancy in order to reduce auditory disability to a minimum. To achieve this, it is important to implement screening programmes as soon after birth as possible. In the United States, the Joint Committee on Infant Hearing recommended in 1982 that identification of hearing loss should be screened in the neonatal period. This early detection is now considered critical for optimal rehabilitative outcome. This paper presents the "state of art" neonatal screening principles and procedures. In France, neonatal screening programs for auditory dysfunction are not consistent with these principles. Evoked otoacoustic emissions represent an important advance in screening for hearing loss in normal neonates and babies from neonatal intensive care units. This method records very low intensity sound energy released by the cochlea in response to a brief sound stimulation. These otoacoustic emissions show promise as a rapid, cost-effective means of quickly discharging all babies with normal peripherical auditory systems.

Distortion product otoacoustic emissions and sensorineural hearing loss

As other types of otoacoustic emissions, distortion product otoacoustic emissions (DPOAEs) allow the exploration of the active cochlear mechanisms known to take place in the outer hair cell system. Most authors consider that 2f1-f2 DPOAEs are generated in a cochlear region corresponding to the geometric mean (GM) of the primary frequencies. To verify the relevance of this hypothesis in clinical practice, DPOAEs were recorded at seven different frequencies, ranging from 0.5 to 4 kHz, in 81 hearing-impaired patients and in 24 normally hearing subjects. To test the hypothesis that DPOAEs reflect the hearing threshold at the frequency of the GM rather than at the 2f1-f2 frequency, this study compares the 2f1-f2 frequency and the GM of the primaries to the frequency of hearing loss. DPOAEs can be used to explore a large range of frequencies, especially at high frequencies, but responses at low frequencies are less reliable due to noise contamination. Secondly, DPOAEs can be recorded in ears that have a hearing threshold as high as 65 dB HL at the frequency corresponding to the GM of the primaries. Finally, DPOAE recordings show frequency specificity: i.e., hearing loss at a specific frequency correlates best with DPOAEs whose GM of primary frequencies corresponds to the frequency of the hearing loss. However, this frequency specificity is still unsatisfactory and decreases as the levels of primaries increase above 60 dB SPL. Moreover, DPOAE amplitude is too variable to predict hearing loss at a particular frequency, whereas DPOAE threshold allows a correct prediction of abnormal auditory threshold in more than 80% of the cases at frequencies above 1 kHz.

Chirp evoked otoacoustic emissions

The principles of short frequency sweeps (chirps) and their application to evoke transiently evoked otoacoustic emissions (TEOAE) are developed in comparison to using standard click stimuli. In contrast to click stimuli, chirp signals have the advantage of stimulating a freely selectable frequency range. In addition, chirp signals contain more energy than a click stimulus with the same maximum amplitude. The effects of different stimuli on TEOAE were investigated in normal hearing and hearing-impaired subjects. Using wide-band chirp signals yields a better signal-to-noise ratio compared to click stimulation. In addition, the stimulation of selected regions of the basilar membrane with frequency-limited chirps evokes TEOAE with frequency components that lie within the stimulated frequency range. The characteristic fine structure of this spectrum was found to be independent of the stimulus applied. The utilization of chirp stimuli appears to be useful for evoking TEOAE in, e.g., clinical applications.

Ipsilateral suppression effects on transient evoked otoacoustic emission

The spectral properties of click- and tone-evoked otoacoustic emission (OAE) under ipsilateral simultaneous tonal masking conditions as well as the changes of click-evoked OAE under ipsilateral forward masking by clicks were studied in normal-hearing subjects. It was found that (i) transiently evoked OAE (TEOAE) spectra consist of a number of peaks individual to the subject and distributed along a sufficiently wide cochlear region; (ii) each spectral peak is excited only when the stimulus energy was concentrated within the frequency range covering the frequency of the particular peak; (iii) different TEOAE spectral peaks can be masked independently under simultaneous tonal masking conditions; (iv) the TEOAE tuning curve shape is more closely related to the spectrum of TEOAE but not to that of the stimulus; (v) forward masking of TEOAE generation is most pronounced during the first few milliseconds after masker click onset and become significantly smaller with the longer latency. It is suggested that: (i) TEOAE is generated by a number of local generators individually distributed along the cochlear partition within a sufficiently wide region and characterized by different power; (ii) possible efferent effects in the ipsilateral TEOAE suppression are negligible as compared to the TEOAE reduction of exclusively cochlear origin.

[Otoacoustic emissions in the human]

Otoacoustic emissions are sounds emitted by the cochlea, basically deriving from the active micromechanical properties of the outer hair cells of the organ of Corti. As they can be recorded painlessly and non-intrusively, they provide a good means of studying human cochlear functioning. In this report, the main types of otoacoustic emission are described, with their characteristics and relation to cochlear functioning. The contribution of otoacoustic emission studies to the physiology of the medial olivocochlear system is discussed, this being the only sensitive and non-intrusive way of studying this system, the function of which remains uncertain.

Peripheral analysis of frequency in human ears revealed by tone burst evoked otoacoustic emissions

Otoacoustic emissions were evoked in the same ears with single tone bursts at 1, 2 and 3 kHz and with a complex stimulus consisting of a digital addition of the three tone bursts. Stimuli were presented at 75, 59 and 37 dB SPL to 28 ears of human subjects with normal hearing. The purpose was to determine if comparisons of responses to the complex stimulus with a posthoc addition of responses from single tone bursts could delineate features of cochlear frequency analysis of short-duration signals. For processing of the data, the results from the individual tone bursts were combined offline to form a composite response. This was then compared with the response obtained with the complex stimulus. Results revealed close correspondence between the spectra of the complex and composite responses in all ears despite interindividual differences in response morphology. Correlations between the complex and composite waveforms exceeded 80% for all stimulus levels. Subtractions of the two spectra revealed that the majority of the differences occurred at frequencies on the high-frequency slopes of the 1- and 2-kHz spectral peaks. This was due to a reduction in energy for the responses obtained with the complex stimulus. There was little variation between the two response types in the peak frequencies of their spectra, in the energy at frequencies on the lower frequency sides of the spectral peaks at 1 and 2 kHz, or in the spectral components at 3 kHz. Results reveal characteristics of the analysis of frequency in the preneural stages of cochlear processing.

Interrelations between psychoacoustical tuning curves and spontaneous and evoked otoacoustic emissions

With the hypothesis that cochlear active mechanisms are the origin of otoacoustic emissions (OAEs) and of the high frequency selectivity exhibited by the ear, psychoacoustical tuning curves (PTCs), transiently evoked otoacoustic emissions (TEOAEs), and spontaneous otoacoustic emissions (SOAEs) have been examined in 50 normal hearing subjects. Using a clinical simplified method, PTCs were successively assessed at three frequencies--1, 2 and 4 kHz--in each subject. The results showed the existence of significant differences in the quality of tuning (Q10dB) of the PTCs between, first, subjects having SOAEs and subjects having no SOAEs (Student's t-test; p < 0.05; df = 35) and, second, subjects having large TEOAEs and subjects having small TEOAEs (Student's t-test; p < 0.05; df = 14). Nevertheless, these significant differences did not appear for all the frequencies studied: the frequency selective relationship between PTCs and OAEs mainly involved the 2 kHz zone. Such results are discussed according to the specificities of the clinical method used for PTC measurement as well as to the spectral characteristics of OAEs.

Frequency analysis of transiently evoked otoacoustic emissions in sensorineural hearing disturbance

Transiently evoked otoacoustic emissions (TEOAE) are present in nearly all normal hearing ears and recognized as useful for objective auditory screening. In the spectral analysis of TEOAE with click stimuli, the normal power spectrum of TEOAE has yet to be obtained. In this study, the normal range of the power spectrum was determined for 42 adults with normal hearing and compared with sensorineural hearing disturbance data. The power spectrum of TEOAE in normal hearing adults sloped down at high frequencies and was the same for right and left ears in both males and females. For comparison with sensorineural hearing disturbance, a significant correlation between the audiogram and power spectrum of TEOAE was sought. Frequency specificity was noted in the power spectrum of TEOAE with click stimulation.

Characteristics of click-evoked otoacoustic emissions in ears with normal hearing and with noise-induced hearing loss

The clinical application of click-evoked otoacoustic emissions (EOAE) in the assessment of noise-induced hearing loss (NIHL) was examined in a group of 72 ears with NIHL and 61 ears with normal hearing (NH). The characteristics of the EOAE in ears with NIHL significantly differed from the NH, according to all EOAE parameters tested in the present study. The mean overall EOAE level was lower and the mean EOAE nonlinearity threshold was worse in the NIHL group. In 95% of the NH ears the EOAE spectrum range was wide, while in 91.5% of the NIHL ears the range was narrow. Moreover, in 94% of the ears with NIHL, the frequency at which the hearing loss began (BHL) was at or above the frequency of the last peak in the EOAE spectrum (FLP). Furthermore, combination of EOAE spectral measures correctly discriminate on average 93.5% of ears with NH from NIHL (sensitivity) and 92% of ears with NIHL from NH (specificity). In contrast, the nonlinearity threshold and the overall level of EOAE yielded lower specificity of less than 33%. It was therefore concluded that EOAE spectrum may serve as a useful and objective tool in screening adults with suspected noise-induced high frequency hearing loss.

Temporal patterns of transient-evoked otoacoustic emissions in normal and impaired cochleae

The spatial distribution of outer hair cells that participate in generating transient-EOE frequency-components has been investigated in man. According to several models (e.g. Wilson (1990) Hear. Res. 2, 527-532; Zwicker (1986) J. Acoust. Soc. Am. 80, 154-162; Wilson and Kemp (Eds.), Cochlear Mechanisms, Structures, Functions and Models, Plenum Press, NY), EOEs result from interferences between broadly distributed contributions, responsible for their long frequency-dependent delay. This work presents an analysis of the temporal patterns of click- and tone-burst-EOEs in human ears when contributions to EOEs are reduced by noise-induced lesions with audiometric notches centred around 4 kHz (N = 46). Although the auditory thresholds at the frequencies of the studied EOE-components were always normal, these components exhibited drastic and predictable changes compared to normal control ears (N = 40). (1) Their temporal pattern at the highest EOE frequency fmax just below the audiometric notch appeared to be determined by the cochlear state at high frequencies (6 to 8 kHz). Either it was normal and the EOE exhibited a complicated beat-structure, or it was impaired and the time envelope of the EOE was simple. In contrast, any type of time pattern could be observed in normal ears. (2) The temporal patterns of EOEs one octave below fmax always presented many beats and short delays. The proposed interpretation is that contributions to a transient-EOE component at frequency f can come from distant basal cochlear areas, i.e. more than 1 and sometimes 1.5 octaves from the place tuned to f. Therefore, the possible relationships between transient-EOEs and tuning mechanisms which presumably involve only a small number of OHC need further investigations.

Clinical monitoring using otoacoustic emissions

Damaging influences to the cochlea are a leading cause of sensorineural hearing loss. Examples include acute or chronic noise exposure and cochleotoxic drugs such as aminoglycosides. Typically, once damage has occurred, the cochlea cannot recover. Therefore, prevention is critical. If damaging influences cannot be avoided, then secondary prevention or early detection of cochlear hearing loss becomes important. Ideally, methods for the detection of cochlear damage should be as specific and as sensitive as possible. Otoacoustic emissions satisfy these criteria and offer a means of testing aspects of cochlear function in a non-invasive and objective way. Evoked otoacoustic emissions measured either after transient stimuli or during two-tone stimulation (distortion-product otoacoustic emissions) are the types most commonly used for clinical purposes. They are stable over time within individual ears and their repeatability has been established under conditions of clinical testing using commercial equipment. Thus, they are well suited as an effective means of monitoring subtle changes in cochlear status. The possibility of making non-invasive repeated measures of cochlear function has led to the widespread use of otoacoustic emissions in animal experiments. Influences of development, anoxia, anaesthesia, noise, and drugs have been monitored. Preliminary studies in humans demonstrate that cochlear damage due to ototoxic drugs such as aminoglycosides or cisplatin and due to noise exposure can be detected using otoacoustic emissions. Comparison of such results to those available using pure-tone audiometry indicates a greater sensitivity of otoacoustic emissions for detecting early cochlear damage.(ABSTRACT TRUNCATED AT 250 WORDS)

Frequency specificity of human distortion product otoacoustic emissions

The amplitudes and growth functions of distortion product otoacoustic emissions (DPOE) at 2f1-f2, elicited by two primary tones f1 and f2 with a constant frequency ratio f2/f1 = 1.23 and varying geometric mean values 1, 1.5, 2, 4, 6 and 8 kHz, were measured for 25 normal subjects and 50 patients with sensorineural hearing loss. Partial correlations between DPOE amplitudes and auditory thresholds (0.25 to 8 kHz, half-octave steps) were examined. The amplitude of DPOE evoked by low-intensity primary tones (at or below 62 dB SPL) was strongly correlated only with the auditory threshold at their mean frequency, and DPOE disappeared for local hearing losses larger than about 30 dB. Moreover, DPOE amplitudes did not depend on the basal cochlear state. Confounding effects of middle ear transmission and aging were not significant in this set of experiments. When elicited by higher intensities of primary tones (72 dB SPL), DPOE exhibited a more complex and non-local behavior, and their sensitivity to hearing loss decreased. These results suggest that when low-intensity primaries are used, DPOE patterns provide frequency-specific information on the local cochlear state.