Swept – Tone Evoked Otoacoustic Emissions: Stimulus Calibration and Equalization

Project Enquiry:

Fields with * are mandatory

ABSTRACT

Otoacoustic Emissions (OAE) are minute acoustic responses originating from the cochlea as a result of an external acoustic stimulus and are recorded using a sensitive microphone placed in the ear canal. OAEs are acquired by synchronous stimulation with an acoustic click or tone burst and recording of the post-stimulus responses.

This method of acquiring OAEs is known as transient evoked otoacoustic emissions (TEAOE) and is commonly used in clinics as a screening method for hearing and cochlear functionality in infants. Recently, a novel method of acquiring OAEs utilizing a swept-tone, or chirp, as a stimulus was developed. This method  used a deconvolution process to compress the swept tone response into an impulse or click-like response.

Because the human ear does not hear all frequencies (pitches) at equal loudness the swept-tone stimulus was equalized in amplitude with respect to frequency. This equalized stimulus will  be perceived by the ear as equally loud in all frequencies.

In this study a new hearing level  equalized stimulus was designed and the OAE responses were analyzed and compared to  conventional click evoked OAEs. The equalized swept-tone stimulus evoked greater magnitude  OAE responses when compared to the conventional methods. It was also able to evoke responses  in subjects that had little TEOAEs which might fail conventional hearing screening.

Anatomical Cross Section of the Ear, Showing the Three stages of Hearing.

Anatomical Cross Section of the Ear, Showing the Three stages of Hearing.

METHODS AND MATERIALS

Instrumentation:

Existing medical equipment for recording OAE responses is limited only to a couple types of  stimuli, click and pure tones, a relatively short recording window, and typically with 16 bit per sample resolution. The stimulus used for the sTEOAE method is 100 ms in duration and changes its frequency with time (Bennett and Ozdamar, 2010b). For this reason a customized device was used, developed by Bennett and Ozdamar, that allowed the use of a preloaded custom stimulus and recording time up to 1s.

Stimulus Calibration:

For the hearing level (HL) equalization process the standard ISO 226:2003 normal equal loudness level contours were used from which a set of SPL to HL values were obtaine. In  addition to the HL compensation values a set of calibration values were obtained to account for the frequency response of the transducers used in the ER-10D OAE probe. The expected OAE responses ranged up to 5000Hz, so a slight roll off in amplitude was added at very high frequencies, 6000 and 8000Hz.

Courtesy of Intelligent Hearing Systems (IHS) .

Courtesy of Intelligent Hearing Systems (IHS) .

Procedures and Protocol:

The recordings were performed in a single session with the subjects sitting in a soundproof  booth shielded from electromagnetic interference. All of the subjects were tested to  determine the presence of OAEs using clinical equipment and standards. This was done using  IHS USB System and SmartTrOAE software with a 100μs click at intensity of 75 dBSPL, acquired  at a rate of 19.3 clicks per second for 1024 sweep (averages). A subset of 3 subjects was  selected for additional recordings with the following intensities: 75, 65, 55, 45, and 35.

Post-processing and Analysis:

The post processing and analysis of the recordings was also done using custom made software  in Matlab®. After receiving the raw data from the DSP system it was formatted as a swept tone  OAE recording and was stored in a file containing the two raw buffers, information about the  type of recording, the inverse stimulus, calibration parameters, subject name, and other technical parameters regarding the recording.

Visualization of the Hleqs Method, Showing the Stimulus Enveloping, Stimulation of the Ear, Recording, Deconvolution, and Windowing to Obtain the Final Swept-tone Oae Responses.

Visualization of the Hleqs Method, Showing the Stimulus Enveloping, Stimulation of the Ear, Recording, Deconvolution, and Windowing to Obtain the Final Swept-tone Oae Responses.

RESULTS

All of the subjects included in the following studies had normal hearing on both ears, below 30dBHL, and did not exhibit any kind of hearing impairments or losses. Two of the subjects  (males) did not meet the clinical criteria for passing OAE responses at 75dBSPL (DNL) with 100μs click stimulus. One of the subjects (female) had abnormally large OAE responses on both ears. The recordings were performed accordingly with IRB protocols and regulations for responsible human research conduct and all subjects signed a written consent that explained the protocol in detail.

HLeqs: Signal, Noise, and Intensity Characteristics :

For the HLeqs analysis the OAE responses of 11 subjects, 22 ears, were recorded.
With the exception of two, all of the subjects exhibited normal OAE responses and satisfied the clinical passing conditions. This study primarily focuses on the behavior of the HLeqs  responses using the HL swept tone with respect to response frequency and stimulus intensity.

 Hleqs Teoae – Mean of Narrow Band Snr as a Function of Response Frequency for Different Stimulus Intensities.

Hleqs Teoae – Mean of Narrow Band Snr as a Function of Response Frequency for Different Stimulus Intensities.

DISCUSSION AND SUMMARY

Signal, Noise, and SNR Analysis of Hearing Level Equalized OAEs:

This study focused on the characteristics and morphology of the OAE responses acquired with the hearing level equalized swept-tone method. The band and total signal, noise, SNR values  were measured with respect to different stimulus intensities, and compared with the click evoked OAEs.

The general observation regarding the OAE response signal levels is that they decrease  steadily with stimulus intensities from 55 down to 15 dBHL and level off between 15 and 5 dBHL. The response behavior at low stimulus intensities may be explained by the hearing  threshold level or the minimum threshold for evoked OAEs (Bonfils et al., 1988). This is also backed by lack of discrepancy in the responses signal levels between 15 and 5 dBHL.

Source: University of Miami
Author: Todor Mihajloski

Download Project

Project Enquiry:

Fields with * are mandatory

Leave a Comment

Your email address will not be published. Required fields are marked *