Stimulus frequency otoacoustic emissions evoked by swept tones

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.

Analysis of cross-talk cancellation of bilateral bone conduction stimulation

When presenting a stereo sound through bilateral stimulation by two bone conduction transducers (BTs), part of the sound at the left side leaks to the right side, and vice versa. The sound transmitted to the contralateral cochlea becomes cross-talk, which can affect space perception. The negative effects of the cross-talk can be mitigated by a cross-talk cancellation system (CCS). Here, a CCS is designed from individual bone conduction (BC) transfer functions using a fast deconvolution algorithm. The BC response functions (BCRFs) from the stimulation positions to the cochleae were obtained by measurements of BC evoked otoacoustic emissions (OAEs) of 10 participants. The BCRFs of the 10 participants showed that the interaural isolation was low. In 5 of the participants, a cross-talk cancellation experiment was carried out based on the individualized BCRFs. Simulations showed that the CCS gave a channel separation (CS) of more than 50 dB in the 1-3 kHz range with appropriately chosen parameter values. Moreover, a localization test showed that the BC localization accuracy improved using the CCS where a 2-4.5 kHz narrowband noise gave better localization performance than a broadband 0.4-10 kHz noise. The results indicate that using a CCS with bilateral BC stimulation can improve interaural separation and thereby improve spatial hearing by bilateral BC.

Usefulness of phase gradients of otoacoustic emissions in auditory health screening: An exploration with swept tones

Otoacoustic emissions (OAEs) are low-level sounds generated by the cochlea and widely used as a noninvasive tool to inspect cochlear impairments. However, only the amplitude information of OAE signals is used in current clinical tests, while the OAE phase containing important information about cochlear functions is commonly discarded, due to the insufficient frequency-resolution of existing OAE tests. In this study, swept tones with time-varying frequencies were used to measure stimulus frequency OAEs (SFOAEs) in human subjects, so that high-resolution phase spectra that are not available in existing OAE tests could be obtained and analyzed. The results showed that the phase of swept-tone SFOAEs demonstrated steep gradients as the frequency increased in human subjects with normal hearing. The steep phase gradients were sensitive to auditory functional abnormality caused by cochlear damage and stimulus artifacts introduced by system distortions. At low stimulus levels, the group delays derived from the phase gradients decreased from around 8.5 to 3 ms as the frequency increased from 1 to 10 kHz for subjects with normal hearing, and the pattern of group-delay versus frequency function showed significant difference for subjects with hearing loss. By using the swept-tone technology, the study suggests that the OAE phase gradients could provide highly sensitive information about the cochlear functions and therefore should be integrated into the conventional methods to improve the reliability of auditory health screening.

Towards effective assessment of normal hearing function from ABR using a time-variant sweep-tone stimulus approach

Objective. The auditory brainstem response (ABR) audiometry is a means of assessing the functional status of the auditory neural pathway in the clinic. The conventional click ABR test lacks good neural synchrony and it mainly evaluates high-frequency hearing while the common tone-burst ABR test only detects hearing loss of a certain frequency at a time. Additionally, the existing chirp stimuli are designed based on average data of cochlear characteristics, ignoring individual differences amongst subjects.Approach. Therefore, this study designed a new stimulus approach based on a sweep-tone concept with a time variant and spectrum characteristics that could be customized based on an individual's cochlear characteristics. To validate the efficiency of the proposed method, we compared its performance with the click and tone-bursts using ABR recordings from 11 normal-hearing adults.Main results. Experimental results showed that the proposed sweep-tone ABR achieved a higher amplitude compared with those elicited by the click and tone-bursts. When the stimulus level or rate was varied, the sweep-tone ABR consistently elicited a larger response than the corresponding click ABR. Moreover, the sweep-tone ABR appeared earlier than the click ABR under the same conditions. Specifically, the mean wave V peak-to-peak amplitude of the sweep-tone ABR was 1.3 times that of the click ABR at 70 dB nHL (normal hearing level) and a rate of 20 s^-1, in which the former saved 40% of test time.Significance. In summary, the proposed sweep-tone approach is found to be more efficient than the traditional click and tone-burst in eliciting ABR.

Analysis of Click and Swept-Tone Auditory Brainstem Response Results for Moderate and Severe Sensorineural Hearing Loss

INTRODUCTION

Auditory brainstem response (ABR) is one of the commonly used methods in clinical settings to evaluate the hearing sensitivity and auditory function. The current ABR measurement usually adopts click sound as the stimuli. However, there may be partial ABR amplitude attenuation due to the delay characteristics of the cochlear traveling wave along the basilar membrane. To solve that problem, a swept-tone method was proposed, in which the show-up time of different frequency components was adjusted to compensate the delay characteristics of the cochlear basilar membrane; therefore, different ABR subcomponents of different frequencies were synchronized.

METHODS

The normal hearing group, moderate sensorineural hearing loss group, and severe sensorineural hearing loss group underwent click ABR and swept-tone ABR with different stimulus intensities. The latencies and amplitudes of waves I, III, and V in 2 detections were recorded.

RESULTS

It was found that the latency of each of the recorded I, III, and V waves detected by swept-tone ABR was shorter than that by click ABR in both the control group and experimental groups. In addition, the amplitude of each of the recorded I, III, and V waves, except V waves under 60 dB nHL in the moderate sensorineural hearing loss group, detected by swept-tone ABR was larger than that by click ABR. The results also showed that the swept-tone ABR could measure the visible V waves at lower stimulus levels in the severe sensorineural hearing loss group.

CONCLUSION

Swept-tone improves the ABR waveforms and helps to obtain more accurate threshold to some extent. Therefore, the proposed swept-tone ABR may provide a new solution for better morphology of ABR waveform, which can help to make more accurate diagnosis about the hearing functionality in the clinic.

Variable-rate frequency sweeps and their application to the measurement of otoacoustic emissions

Swept tones allow the efficient measurement of otoacoustic emissions (OAEs) with fine frequency resolution. Although previous studies have explored the influence of different sweep parameters on the measured OAE, none have directly considered their effects on the measurement noise floor. The present study demonstrates that parameters such as sweep type (e.g., linear or logarithmic), sweep rate, and analysis bandwidth affect the measurement noise and can be manipulated to control the noise floor in individual subjects. Although responses to discrete-tone stimuli can be averaged until the uncertainty of the measurement meets a specified criterion at each frequency, linear or logarithmic sweeps offer no such flexibility. However, measurements of the power spectral density of the ambient noise can be used to construct variable-rate sweeps that yield a prescribed (e.g., constant) noise floor across frequency; in effect, they implement a form of frequency-dependent averaging. The use of noise-compensating frequency sweeps is illustrated by the measurement of distortion-product OAEs at low frequencies, where the ear-canal noise is known to vary significantly.

Frequency specificity and left-ear advantage of medial olivocochlear efferent modulation: a study based on stimulus frequency otoacoustic emission

The medial olivocochlear (MOC) bundle is an auditory nucleus that projects efferent nerve fibers to the outer hair cells (OHCs) for synaptic innervation. The aim of the present study was to investigate the possible existence of frequency and ear specificity in MOC efferent modulation, as well as how MOC activation influences cochlear tuning. Stimulus frequency otoacoustic emissions (SFOAEs) were used to study MOC efferent modulation. Therefore, the current experiment was designed to compare the degree of SFOAE suppression in the both ears of 20 individuals at 1, 2, 4, and 8 kHz. We also compared changes in Q10 values of SFOAE suppression tuning curves at 1, 2, and 4 kHz under contralateral acoustic stimulation (CAS) and no-CAS conditions. We observed a significant reduction in SFOAE magnitude in the CAS condition compared with the no-CAS condition at 1 and 2 kHz in the left ear. A significant difference in CAS suppression was also found between the left and right ears at 1 and 2 kHz, with larger CAS suppression in the left ear. CAS further produced a statistically significant increase in the Q10 value at 1 kHz and a significant reduction in Q10 values at 2 and 4 kHz. These findings suggest a left-ear advantage in terms of CAS-induced MOC efferent SFOAE suppression, with larger MOC efferent modulation for lower frequencies, and cochlear tuning was sharpened by means of MOC activation at lower frequencies and broadened at higher frequencies.

Influence of medial olivocochlear efferents on the sharpness of cochlear tuning estimates in children

The present study objectively quantified the efferent-induced changes in the sharpness of cochlear tuning estimates and compared these alterations in cochlear tuning between adults and children. Click evoked otoacoustic emissions with and without contralateral broadband noise were recorded from 15 young adults and 14 children aged between 5 and 10 yrs. Time-frequency distributions of click evoked otoacoustic emissions were obtained via the S-transform, and the otoacoustic emission latencies were used to estimate the sharpness of cochlear tuning. Contralateral acoustic stimulation caused a significant reduction in the sharpness of cochlear tuning estimates in the low to mid frequency region, but had no effect in the higher frequencies (3175 and 4000 Hz). The magnitude of efferent-induced changes in cochlear tuning estimates was similar between adults and children. The current evidence suggests that the stimulation of the medial olivocochlear efferent neurons causes similar alterations in cochlear frequency selectivity in adults and children.

Effects of contralateral acoustic stimulation on otoacoustic emissions induced by swept tones

The medial olivocochlear complex (MOC) is an auditory nucleus that projects efferent nerve fibers to control the behaviors of both sides of the cochlea. Otoacoutsic emissions (OAEs) are by-products the activities of the outer hair cells (OHCs) in the cochlea and could be used as a noninvasive way to study the efferent control of the MOC. However, existing results regarding the efferent control are quite controversial and often restricted to a rather limited frequency range. In this study, a new method of measuring stimulus frequency otoacoustic emissions (SFOAEs) with the presence of a contralateral acoustic stimulation (CAS) was proposed to study the efferent control over the cochlea. SFOAEs were measured with swept tones with time varying frequencies so that SFOAE spectra with and without the presence of the CAS could be compared with high frequency-resolution. The results showed that there was consistent decrease in the amplitude of the swept-tone SFOAEs across a wide frequency range from 0.5 to 8 kHz when the CAS was presented, suggesting an outstanding attenuation of OHC activities by the efferent control from the MOC. The SFOAE decrease with the presence of the CAS might provide a new approach to measure the strength of the efferent control and to evaluate the functional status of the central auditory pathway.

An in-situ calibration method and the effects on stimulus frequency otoacoustic emissions

Background

The interference between the incoming sound wave and the acoustic energy reflected by the tympanic membrane (TM) forms a standing wave in human ear canals. The existence of standing waves causes various problems when measuring otoacoustic emissions (OAEs) that are soft sounds closely related with the functional status of the inner ear. The purpose of this study was to propose an in-situ calibration method to overcome the standing-wave problem and to improve the accuracy of OAE measurements.

Methods

In this study, the sound pressure level (SPL) at the TM was indirectly estimated by measuring the SPL at the entrance of the ear canal and the acoustic characteristics of the earphone system, so that sound energy entering the middle ear could be controlled more precisely. Then an in-situ calibration method based on the estimated TM SPL was proposed to control the stimulus level when measuring the stimulus frequency otoacoustic emissions (SFOAEs) evoked by swept tones. The results of swept-tone SFOAEs with the in-situ calibration were compared with two other calibration methods currently used in the clinic.

Results

Our results showed that the estimate of the SPL at the TM was rather successful with the maximal error less than 3.2 dB across all the six subjects. With the high definition OAE spectra achieved by using swept tones, it was found that the calibration methods currently used in the clinic might over-compensate the sound energy delivered to the middle ear around standing-wave frequencies and the SFOAE amplitude could be elevated by more than 7 dB as a consequence. In contrast, the in-situ calibration did not suffer from the standing-wave problem and the results could reflect the functional status of the inner ear more truthfully.

Conclusions

This study suggests that calibration methods currently used in the clinic may produce unreliable results. The in-situ calibration based on the estimated TM SPL could avoid the standing-wave problem and might be incorporated into clinical OAE measurements for more accurate hearing loss screenings.