| Article Access Statistics|
| Viewed||3572 |
| Printed||85 |
| Emailed||2 |
| PDF Downloaded||25 |
| Comments ||[Add] |
|Year : 2015
: 17 | Issue : 78 | Page
|DPOAE level mapping for detecting noise-induced cochlear damage from short-duration music exposures
Jay C Buckey1, Abigail M Fellows1, Odile H Clavier2, Lindsay V Allen2, Chris A Brooks2, Jesse A Norris2, Jiang Gui3, Deanna K Meinke4
1 Geisel School of Medicine at Dartmouth, Department of Medicine, Hanover, New Hampshire, USA
2 Creare, LLC, Hanover, New Hampshire, USA
3 Geisel School of Medicine, Department of Biomedical Data Sciences, Hanover, New Hampshire, USA
4 University of Northern Colorado, Department of Audiology and Speech-Language Sciences, Greeley, Colorado, USA
Click here for correspondence address
|Date of Web Publication||10-Sep-2015|
Distortion product otoacoustic emission (DPOAE) level mapping provides a comprehensive picture of cochlear responses over a range of DP frequencies and f2 /f1 ratios. We hypothesized that individuals exposed to high-level sound would show changes detectable by DPOAE mapping, but not apparent on a standard DP-gram. Thirteen normal hearing subjects were studied before and after attending music concerts. Pure-tone audiometry (500-8,000 Hz), DP-grams (0.3-10 kHz) at 1.22 ratio, and DPOAE level maps were collected prior to, as soon as possible after, and the day after the concerts. All maps covered the range of 2,000-6,000 Hz in DP frequency and from 1.3 to -1.3 in ratio using equi-level primary tone stimuli. Changes in the pure-tone audiogram were significant (P ≤ 0.01) immediately after the concert at 1,000 Hz, 4,000 Hz, and 6,000 Hz. The DP-gram showed significant differences only at f2 = 4,066 (P = 0.01) and f2 = 4,348 (P = 0.04). The postconcert changes were readily apparent both visually and statistically (P ≤ 0.01) on the mean DP level maps, and remained statistically significantly different from baseline the day after noise exposure although no significant changes from baseline were seen on the DP-gram or audiogram the day after exposure. Although both the DP-gram and audiogram showed recovery by the next day, the average DPOAE level maps remained significantly different from baseline. The mapping data showed changes in the cochlea that were not detected from the DP-gram obtained at a single ratio. DPOAE level mapping provides comprehensive information on subtle cochlear responses, which may offer advantages for studying and tracking noise-induced hearing loss (NIHL).
Keywords: Audiometry, distortion product otoacoustic emission (DPOAE) mapping, music-induced hearing loss, noise exposure, otoacoustic emissions
|How to cite this article:|
Buckey JC, Fellows AM, Clavier OH, Allen LV, Brooks CA, Norris JA, Gui J, Meinke DK. DPOAE level mapping for detecting noise-induced cochlear damage from short-duration music exposures. Noise Health 2015;17:263-72
|How to cite this URL:|
Buckey JC, Fellows AM, Clavier OH, Allen LV, Brooks CA, Norris JA, Gui J, Meinke DK. DPOAE level mapping for detecting noise-induced cochlear damage from short-duration music exposures. Noise Health [serial online] 2015 [cited 2020 Oct 31];17:263-72. Available from: https://www.noiseandhealth.org/text.asp?2015/17/78/263/165037
| Introduction|| |
With distortion product otoacoustic emission (DPOAE) level mapping, both the 2f1 -f2 and the 2f2 -f1 emissions are collected over a range of f2 frequencies and ratios. ,, To create a map, we fix the ratio, then increase the f2 frequency such that the DPOAE frequency will vary over the desired range (e.g., 0.6-6.0 kHz) in user-selected steps (e.g., approximately 44 Hz in this study). At each measurement, both the 2f1 -f2 and 2f2 -f1 emissions are recorded. Then the ratio is increased (e.g., using 0.025 steps from 1.025 to 1.5) and another sweep of f2 is performed. The emissions are plotted as individual f2 /f1 ratio pairs, with the 2f1 -f2 emissions on the top half of the graph and the 2f2 -f1 emissions on the bottom half. The net result is an overall picture of cochlear responses, called a map, that is characteristic of the individual and which can be tracked over time. In a previous cross-sectional study from our research group, we compared DP level maps among the following three groups of subjects: normal hearing subjects, subjects with normal audiograms but a history of noise exposure, and individuals with mild noise-induced hearing loss (NIHL). Compared to the normal hearing group, the NIHL group showed an overall reduction in both DPOAE levels and frequency regions producing detectable DPOAEs.  These data suggest DPOAE level mapping may offer a more sensitive alternative to pure-tone audiometry, or the standard DP-gram, for detecting subclinical changes in cochlear function. , Whether the DPOAE mapping approach improves the ability to detect noise-induced hearing damage in humans, however, has not yet been shown.
To compare the performance of pure-tone, air-conduction threshold audiometry, a DP-gram at a single ratio, and DPOAE level mapping for detecting hazardous sound damage to the cochlea, we tested 13 subjects before and after music concerts using all 3 techniques. We hypothesized that DPOAE level maps would show changes due to noise exposure, and would provide information on nature of the cochlear damage not provided by the other methods.
| Methods|| |
This study included 13 people who were attending music concerts recruited through word of mouth and community posters. All subjects (6 males and 7 females) were asked to abstain from alcohol or other substances during the concert and prior to each testing period. Subjects' ages ranged 22-50 years with an average age of 34 years. Participants were included if they had the ability to perform pure-tone audiometry using the Békésy technique, had detectable DPOAE's in both ears as measured by a standard DP-gram, and had a PTA <15 dB HL in both ears for the frequency region 500-6,000 Hz. All subjects had type A tympanograms and a clear view of the tympanic membrane on otoscopic exam. No cerumen removal was required. All subjects were asked to avoid noise exposure during the period of the study and all reported that they complied with this request.
All tests and procedures were approved by the Committee for the Protection of Human Subjects at Dartmouth. The first screening visit included a computer-based health history and hearing questionnaire, otoscopic exam, cerumen removal, custom probe molding (described below), tympanometry, pure-tone audiometry, and DPOAEs. The initial plan for the study was to have five visits, one within a week of the concert, one immediately before the concert, another within 1 h of the end of the concert, one within 24 h of the concert, and one 7 to 10 days later. But, this involved testing at two locations. The middle three sessions were all performed at the same site under identical conditions, but the first and fifth sessions were performed at different locations from the middle three sessions. Electrical noise influenced the DPOAE mapping tests at the locations used for tests 1 and 5, so for the purpose of this study, only data from the middle three sessions were used.
Each experimental testing session included in order:
- An in-ear noise measurement,
- A frequency sweep (for probe positioning repeatability),
- DPOAE level mapping,
- DPOAEs testing (DP-gram), and
- Pure-tone audiometry. For tests 3-5, the right ear was tested first, audiometry for both ears was completed at the end with the right ear always tested first.
Subjects had all DPOAE testing done with the same noise attenuation technique either in a sound booth (single wall, IAC) or with passive noise-attenuating earmuffs over the ears (David Clark Model 19 A, David Clark Company, Worcester, MA). Audiometry was performed using a Sennheiser HDA-200 headset (Sennheiser, Wedemark, Germany). In-ear noise measurements were used to record sound levels in the ear canal to verify the noise level was consistent across testing sessions. 
Creare testing system
All hearing tests, except for tympanometry, were performed using the Creare Hearing Assessment (CHA) system developed in collaboration with Creare Inc. (Hanover, NH). The system consists of a handheld unit housing a digital signal processor, an audio codec, and power management circuitry with custom firmware to produce calibrated tones and execute the required signal processing and logic for each test. The handheld unit communicates with a PC-based user interface through universal serial bus (USB) communication. The CHA was connected to an ER 10-B+ (Etymotic Research Inc.) probe with ER-2 speakers for the DPOAE tests. Each system was calibrated using a Bruüel and Kjær Type 4157 Ear Simulator/Artificial Ear. A Sennheiser HDA-200 headset was used for the Békésy tests, and calibrated using a Bruel and Kjaer Type 4153 flat plate coupler according to ANSI S3.6.  Pure-tone threshold, DP-gram, and DP level map data were stored in an SQL-Lite database and exported into MATLAB (version 8.2.701 (R2013b)) for analysis.
One known source of variability in standard DPOAE measurements is probe position in the ear canal, which can vary from visit to visit, and lead to artifactual emission shifts. This variability is worse at frequencies of 3 kHz and above, where standing waves in the ear canal can affect the sound pressure levels of the stimulus and of the measured emission. ,, One approach to minimizing this problem is to use custom-molded earpieces, which assist with assuring consistent placement of the DPOAE probe in the same location on repeated testing. The use of custom-molded earpieces was assessed previously in our lab by measuring six subjects repeatedly at five different testing sessions both with and without custom probes.  The custom-molded earpieces had lower variability in the frequency sweeps as compared to the standard probe tips, indicating the probe was placed more consistently with the custom-molded earpieces across sessions. For the custom earpiece, the variation in the frequency sweep across independent insertions was approximately 1 dB from 1 kHz to 12 kHz.
From this experience, we fitted each subject with custom-molded earpieces for this study. The earpieces were made at our test facility using a Radians do-it-yourself custom-molded hearing protection earplug kit (Model CEP001-T, Radians Inc., Memphis, TN). The kit consisted of two materials that were hand-mixed together and pressed into the ear. The impression material extended part way into the ear canal and was molded to fill the concha and fit under the anti-tragus and anti-helix. The material was allowed to set in the ear for 15 min and then finished curing at room temperature for 24 h. The end of the impression material that extended into the ear canal was sliced off to provide a flat surface and a hole was then drilled straight through the silicone earpiece allowing the Etymotic probe tip to be placed in approximately the same location each time. The tip of the Etymotic probe was always positioned to terminate evenly with the canal end of the custom earpiece. A cavity approximately 1-mm deep by 2-mm radius from the tip was carved out around the tip of the probe to allow sound access to the microphone. The channel in the earplug provided a snug fit for the probe tip. Each earpiece was labeled for each subject and reused for each test session.
Baseline tympanometry at 226 Hz was performed on both ears using a Madsen Otoflex 100. The Otoflex 100 reported tympanogram type from the location (pressure and static admittance) of the peak of the tympanogram. The pressure limits for type A were −100 to +50 daPa and the static admittance limits for type A were 0.3-2.8 mho. All subjects demonstrated type A tympanograms in order to be enrolled in the study, and had type A tympanograms when they returned for repeat testing.
In-ear noise measurement
The in-ear probe used for the in-ear noise and DPOAE measurements was an Etymotic Research ER-10B+ (Etymotic Research, Inc., Elk Grove Village, IL). This probe contained two speakers and a microphone. For the in-ear noise measurement, the ambient noise spectrum from the microphone was analyzed, and displayed as one-third octave bands on the operator's screen. These data were also stored in the SQL-Lite database. One-third octave bands were calculated following ANSI S1.11-2004 (R2009) "Specification for Octave-Band and Fractional-Octave-Band Analog and Digital Filters." 
Frequency sweep/position check
To assist with consistent probe placement between sessions, a frequency sweep (chirp) was presented in the ear canal (500-7000 Hz). This was displayed and compared to the baseline sweep set on the first baseline visit. If there was a poor seal or the sweep did not align with the baseline, the hearing tester repositioned the earpiece/probe to match the original probe placement chirp response as closely as possible.
At the screening visit, the subject was fitted with an Etymotic probe tip, and an abbreviated screening map was performed at L1, L2 = 65, 65 dB sound pressure level (SPL). One objective of the screening visit was to ensure that the subject had a sufficient number of detectable emissions on the map. This was a subjective evaluation for each subject. If the subject had a limited response to the L1, L2 = 65, 65 dB SPL stimulus levels, another screening map was performed at either L1, L2 = 70, 70 dB SPL or L1, L2 = 75, 75 dB SPL. If the higher levels produced more detectable DPOAEs based on visual inspection, the higher levels were used for that subject throughout the study. Seven subjects were tested at L1, L2 = 65, 65 dB SPL, one subject was tested at L1, L2 = 70, 70 dB SPL, and five were tested at L1, L2 = 75, 75 dB SPL. Equal L1 and L2 levels were used for two reasons. It is an efficient way to collect both the 2f1 -f2 and 2f2 -f1 emissions at the same time during the mapping procedure, and it is a compromise between the optimal parameters needed to generate the 2f1 -f2 and 2f2 -f1 emissions separately. , For all baseline and postconcert measurements, the custom-molded earpiece described above was used for the probe measurements.
Selecting the frequency and ratio range for the protocol involved a balance between making a comprehensive map and test time. Initially testing began with a DP range of 0.6-6.0 kHz and a ratio range of 1.025-1.3. Testing with these initial subjects (subjects 1-3) showed that noise below 2 kHz often reduced the quality of the measurements in that range. Also, the maps showed that extending the frequency and ratio range would likely capture more emissions. So, the protocol was modified twice. For protocol 2 (subjects 4-9), DP frequencies ranged 0.6-6.0 kHz and ratios ranged 1.025-1.4. For protocol 3 (subjects 10-13), DP frequencies ranged 2.0-9.0 kHz and ratios ranged 1.025-1.35. Nevertheless, all mapping protocols included DP frequencies from 2 kHz to 6 kHz, and ratios ranged 1.025-1.3. In other words, all the map protocols included a core set of emissions that would allow comparison across subjects. DPOAE frequency steps of 44 Hz were used, and ratios were increased in 0.025 steps. To create the map, f2 was increased across the selected frequency testing range in 44 Hz steps and at each testing frequency both the 2f1 -f2 and 2f2 -f1 emissions were collected and stored as described graphically previously. 
The first and second protocols had 8 test averages per point. The third protocol used a noise-rejection algorithm (minimum test average 6, maximum test average 10, minimum DP noise floor threshold 10). The number of points in each protocol was as follows: protocol 1 = 2,050, protocol 2 = 2,788, and protocol 3 = 2,560. DPOAE levels were directly plotted in dB SPL and color-coded to facilitate visualization in a DPOAE level map. The map protocols took approximately 10 min per ear to collect.
DPOAEs were collected using a primary f2 frequency sweep from 2 kHz to 10 kHz in 1/10-octave increments (24 points) with a f2 /f1 ratio of 1.22 and stimulus frequency levels of L1 = 65, L2 = 55 dB SPL. The tested frequencies were 2,027 Hz, 2,168 Hz, 2,332 Hz, 2,496 Hz, 2,695 Hz, 2,871 Hz, 3,094 Hz, 3,293 Hz, 3,539 Hz, 3,785 Hz, 4,066 Hz, 4,348 Hz, 4,664 Hz, 4,992 Hz, 5,355 Hz, 5,754 Hz, 6,152 Hz, 6,598 Hz, 7,066 Hz, 7,582 Hz, 8,133 Hz, 8,719 Hz, 9,328 Hz, and 10,020 Hz.
Thresholds were measured with the Sennheiser HDA-200 headsets at frequencies of 500 Hz, 1,000 Hz, 2,000 Hz, 4,000 Hz, 6,000 Hz, and 8,000 Hz with a Békésy-like tracking procedure that used fixed-frequency pulsed tones with a duration of 250 ms, rise and fall time of 20 ms, and an interstimulus interval of 500 ms. Subjects were instructed to press the response button on the CHA when the pure-tone was audible and release the response button when the pure-tone was no longer detected. Initially, when the subject pressed the response button, the pure-tone stimuli decreased in 4-dB steps until the first reversal, then 2-dB steps were used subsequently. The first two reversals were rejected as a training trial. Subsequent reversals, where the peak response was lower than the previous valley response, or where a valley response was higher than the previous peak response, were also rejected, and the algorithm restarted counting the number of valid reversals. A total of six valid reversals were counted and averaged. To calculate the threshold, the last six reversal points were averaged.
Subjects wore a noise dosimeter (Noisepro DLX, 3M, St. Paul, MN) during the concert with the microphone attached on their shoulder below their ear and out of range of bumping during head turns. Noise dosimeters were started before subjects left for the concerts and stopped upon return to the testing site. The dosimeter sampling protocol was set to provide measures using American Conference of Governmental Industrial Hygienists (ACGIH) parameters (A-weighting, slow meter response, peak level limit = 140 dB SPL, integrating threshold = 80 dBA, slow max limit = 115 dBA, dynamic range = 70 dB, exchange rate = 3 dB, and criterion level = 85 dBA). This sampling protocol is also consistent with the National Institute for Occupational Safety and Health (NIOSH) recommendations for workplace noise exposure sampling.  The noise exposures are described in terms of run-time, average sound level (Lavg), dose, maximum slow response level, and time-weighted average (TWA) normalized to the standard 8-h exposure period. There was no advance effort to standardize the concert noise exposures, and dosimetry was performed for the purposes of documenting actual exposures for each individual. All subjects were advised that concert sound levels may be potentially hazardous and were given the opportunity to use hearing protection and withdraw from the study in advance.
The pure-tone audiograms and DP-grams were analyzed across time periods using a paired t-test at each frequency. In the DPOAE level maps, points where the DPOAE was less than 3 dB SPL above the noise floor [3 dB signal-to-noise ratio (SNR)] were omitted, as were points corresponding to emissions less than -10 dB SPL. The cutoff of -10 dB SPL was used because this reference point was above the distortion level for the system at all measurement frequencies. Histograms of DPOAE magnitudes (prior to interpolation) were calculated using the following three methods:
- To assess the distribution of DPOAE magnitudes across the entire map, the DPOAEs from the map were sorted into bins corresponding to 0-6 dB, 6-12 dB, 12-18 dB, 18-24 dB, and >24 dB above the -10 dB threshold.
- The DPOAE level maps were also divided into 1/3-octave bands by DP frequency. The emissions within the octave band were averaged to create a histogram of average emission amplitudes by octave band for each time period.
- The ratio scale of the map was divided into equally spaced intervals corresponding to a change in ratio of 0.025. A histogram of average emission magnitude for each ratio band was calculated.
For all the histograms, the values were compared across time periods using paired t-tests.
Twenty-six maps were generated from sampling both the left and the right ears of the 13 participants. All 26 maps were also averaged together to create an overall average map for each time period. To accomplish this, all the maps needed to have the same number of valid response points at identical frequency and ratio pairs. The three mapping testing protocols used (described above), while very similar, did not sample at precisely the same frequency/ratio pairs. So while the map protocols were consistent within each subject, they differed between subjects. To overcome this, each map was fit with a surface using a linear interpolation (using the MATLAB "fit" function), and then the surface was resampled to create a 2500-point map array. These interpolated maps were averaged together to create a single map for each of the baseline, post noise, and next morning time periods. The maps were compared using a two-sample Kolmogorov-Smirnov test. The point-by-point difference between each average map was calculated. The average difference between maps between time points was computed, and points were marked where the difference was greater than two deviations away from the average difference.
As mentioned above, maps were collected at either L1, L2 = 65, 65 dB SPL, L1, L2 = 70, 70 dB SPL, or L1, L2 = 75, 75 dB SPL based on the results from the screening map. The DP-gram, however, was always done at L1, L2 = 65, 55 dB SPL. This could complicate the comparison of the mapping and DP-gram results, since it is possible that the maps might show greater differences than the DP-gram because of the higher stimulus levels. To assess this, a separate analysis on the threshold, DP-gram, and DPOAE mapping data was done using just those seven subjects (14 ears) that had maps at L1, L2 = 65, 65 dB SPL.
| Results|| |
Subject characteristics and sound exposure
[Table 1] shows the characteristics of the tested group. The sound exposures varied between individuals, which was expected since the exposure characteristics of concerts that individuals attended varied and could not be controlled in advance. The live music concerts were held in both indoor and outdoor venues and lasted on average 3 h and 9 min. Overall, the Lavg for the group ranged 89.7-102.2 dBA, with a mean Lavg of 95.6 dBA. TWA exposures ranged 81.3-99.2 dBA with a mean value of 91.0 dBA TWA. The US Occupational Safety and Health Administration (OSHA) 29 CFR 1910.95 limits the slow max noise exposure at 115 dBA.  Four subjects (30.7%) exceeded the 115 dBA slow max limit set by OSHA. The mean slow max was 113.8 dBA for all subjects. Perhaps the more salient measure of noise dose is the variation in noise exposures between subjects when referencing 100% dose as the allowed daily exposure by the NIOSH. For these subjects; noise doses ranged 85-2,658% with a mean dose of 749.4% or the equivalent of a 7-day allowable exposure in the workplace setting.
|Table 1: Age, gender, and concert noise exposure for the 13 subjects. The noise exposure varied widely across subjects, but for this study each subject served as their own control|
Click here to view
In-ear frequency sweep and 1/3-octave band noise results
[Figure 1] shows the average of the in-ear frequency sweep for all 26 ears at the three time periods. The standard deviations from all 26 ears over the 3 time periods were averaged, and the standard deviation of that average is plotted in the gray area in [Figure 1]. The standard deviation of the average standard deviation ranged 0.3-1.9 dB, indicating that the frequency sweeps were highly repeatable and DPOAE probe position was consistent across test sessions.
|Figure 1: Frequency sweep data for three time periods/testing sessions. The three lines represent the average for the 26 ears at each time period. The gray shaded area represents the standard deviation of the average standard deviation over all the ears and time periods. The maximum average standard deviation was 1.9 dB|
Click here to view
[Figure 2] shows the results from the in-ear octave band measurements. The average from all 26 ears for each time period is plotted as separate lines with the error bars showing the standard error of the mean. Except for 10 kHz, where a significant difference between baseline and next day values was seen, there were no significant differences in octave band noise levels across time periods. On the graph, the sound levels from the microphone when recording within a sound booth are plotted as white bars. The gray bars show the minimal permissible ambient noise levels (MPANL) for a sound booth when recording with the ears uncovered. 
|Figure 2: In-ear 1/3 octave band measurements at three time periods. Each line represents the average for the 26 ears at that time period. Data plotted as average ± SEM. Except for 10 kHz, where a significant difference between baseline and next day was seen, there were no significant differences between time periods for the 1/3 octave band measurements. The gray bars show the maximum permissible ambient noise levels (ears uncovered) for each frequency, and the white bars show the results when the 1/3 octave band measurements are made using the same equipment in a sound booth|
Click here to view
Pure-tone and DP-gram results
[Figure 3] summarizes the pure-tone audiometry results for three time periods. The average change in thresholds were modest (1.4 dB at 1,000 Hz, 3.6 dB at 4,000 Hz, and 2.7 dB at 6,000 Hz), but significant (P = 0.01 at 1,000 Hz, P = 0.0001 at 4,000 Hz, and 0.0013 at 6,000 Hz). The ability to detect such small changes was likely due to the use of the 2 dB step Békésy procedure. No significant differences were noted between baseline and the next day values. When the data were analyzed using just the subjects tested at L1 = L2 = 65 dB SPL, the differences between baseline and postconcert measurements were only significant at 4,000 Hz (P < 0.01) and 6,000 Hz (P < 0.01). [Figure 4] shows the results from the DP-gram (2,000-10,000 Hz). The baseline and postnoise time periods were significantly different at two frequencies (f2 = 4,066 P = 0.01, f2 = 4,348 P = 0.04). When the data were analyzed using only the subjects tested at L1=L2=65 dB SPL, no significant differences were seen between different time periods.
|Figure 3: Pure-tone audiometry results for the 26 ears. Data plotted as average ± SEM|
*P < 0.05. The average change in thresholds between baseline and next day were modest but significant. No significant differences were noted between baseline and the next day. When the data were analyzed using just the subjects tested at L1 = L2 = 65 dB SPL, the differences between baseline and postconcert values were only signifi cant at 4,000 Hz and 6,000 Hz
Click here to view
|Figure 4: DP-grams for the 26 ears. Data plotted as average ± SEM. *P < 0.05. The data show the same trend as for the pure-tone audiometry, but the difference between baseline and postnoise time periods were only significantly different at two frequencies. No signifi cant differences were noted between the baseline and the next day values. When the data were analyzed using only the subjects tested at L1 = L2 = 65 dB SPL, no signifi cant differences were seen between any time period|
Click here to view
DPOAE mapping emission magnitude results
As mentioned in the above methods, for each individual map, points where the DPOAE magnitude was less than 3 dB above the noise floor (3 dB SNR) were omitted, as were DPOAEs with magnitudes less than -10 dB SPL. Then the number of DPOAEs with magnitudes between 0-6 dB, 6-12 dB, 12-18 dB, 18-24 dB, and >24 dB greater than the -10 dB threshold were counted and placed in one of the five bins. [Figure 5] illustrates the results. The number of emissions was significantly lower in the 18-24 dB bin (P = 0.03) and >24 dB bin (P = 0.05) after noise exposure. This difference was no longer significant the day after the exposure. Although the changes in the number of emissions in other bins were not significant, the trend shows a shift in emission magnitudes with a tendency for a reduction in the number of high-level emissions and an increase in lower-level emissions after the noise exposure.
|Figure 5: DPOAE emission magnitudes. The overall trend was for a decrease in higher magnitude emissions and an increase in lower magnitude ones. There were signifi cant decreases in the number of emissions in the 18-24 dB SPL bin and in the >24 dB SPL bin between the baseline and postnoise conditions (*). When the data were analyzed using only the subjects tested at L1 = L2 = 65 dB SPL, no signifi cant differences were seen between any time period, although the same trends were apparent|
Click here to view
Emissions analyzed by octave band and ratio
The average emission magnitude across 1/3-octave bands were also computed at each time period. The results for all subjects show a significant decrease in average emission magnitude in the DP frequency bands 2.8-3.5 kHz (P = 0.01), 3.5-4.5 kHz (P = 0.006), and 4.5-5.6 kHz (P = 0.03) from baseline to post [Figure 6]. The differences from baseline were still apparent the next day in the 3.5-4.5 kHz band (P = 0.006) and in the 4.5-5.6 kHz band (P = 0.03). When the data are analyzed using only those individuals with maps where L1 = L2 = 65 dB SPL, the findings for the next day are the same, with significant differences in the 2.8-3.5 kHz (P = 0.04), 3.5-4.5 kHz (P = 0.05), and 4.5-5.6 kHz (P = 0.04) bands.
|Figure 6: Average DPOAE magnitude by 1/3 octave band frequency range. The DPOAE frequencies were divided into 1/3 octave bands and the average DPOAE magnitude within each band was computed. The postnoise values were signifi cantly less than the baseline values in the 2.8-3.5 kHz band, the 3.5-4.5 kHz band, and the 4.5-5.6 kHz band, (shown by the * on the graphs). The changes in the 3.5-4.5 kHz band and the 4.5-5.6 kHz band persisted into the next day (shown by x on graph)|
*P < 0.05 baseline to post, x = P < 0.05 baseline to next day
Click here to view
When the data were divided into bins encompassing f2 /f1 ratio increments of 0.025, and the average DPOAE magnitude within each bin was computed, the results show a significant reduction in DPOAE magnitude after noise exposure for 2f1 -f2 emissions at a ratio of 1.175 (P = 0.03) and 1.025 (P = 0.02). These differences persisted the next day. In fact, the differences between baseline and next day values were significant at ratios of 1.125, 1.15, 1.175, and 1.275. For 2f2 -f1 emissions, the difference between baseline and postnoise exposure was significantly different (P < 0.05) at multiple points between 0 and 1.2 [Figure 7]. At the ratio of 1.05, the difference persisted into the next day. The post-noise and next day values were also significantly different at the ratio of 1.025. When the data are analyzed using those subjects with L1= L2 = 65 dB SPL, the findings the next day are still significant, but at fewer ratios (1.05, 1.025, for the 2f2 -f1 emissions; 1.025, 1.075, and 1.125 for the 2f1 -f2 emissions).
|Figure 7: Average DPOAE magnitude by ratio. The 2f1-f2 emission (right side of graph) was signifi cantly different from baseline after high level music exposure at the 1.175 and 1.025 ratios. For the 2f2-f1 emissions (left side of graph) the average DP amplitude was significantly reduced at the 1.025, 1.05, 1.075, 1.1, and 1.325 ratios *P < 0.05 baseline to post, x = P < 0.05 baseline to next day, +=P < 0.05 post to next day|
Click here to view
Overall DPOAE map comparisons
[Figure 8] illustrates a series of maps before and after surface fitting and interpolation. The left panel is the baseline map, the center map is the postnoise map and the right-most map shows the results from the day after the noise exposure. The top panel shows the map plotted after emissions <3 dB SNR, and emissions <-10 dB have been removed. The bottom panel shows the map after the surface fit and interpolation. [Figure 9] shows the average overall map for all 26 ears in the top panel. The bottom panel has red points plotted where the difference from the baseline map was greater than two standard deviations for the average difference between maps. The differences between the interpolated maps were calculated using the two-sample Kolmogorov-Smirnov test on the maps. The baseline map was significantly different from both the postnoise (P = 0.004) and the next day maps (P = 0.004). The postnoise and next day maps were not significantly different (P = 0.10). When the data were analyzed using only the subjects with L1 = L2 = 65 dB SPL, the baseline and postconcert maps were still significantly different (P = 0.003). Interestingly, the baseline and next day maps were not significantly different when this group is analyzed separately, but the postconcert and next day maps were (P = 0.004).
|Figure 8: Creation of the interpolated maps. The top part of the fi gure shows a series of maps from one ear after the emissions <3 dB SNR, and emissions ≤10 dB SPL have been removed. The bottom part of the figure shows the maps after the surface fi t and interpolation|
Click here to view
|Figure 9: Overall average maps. The top half of the figure shows the average maps created by averaging all 26 interpolated maps. The left panel shows the baseline period, the middle is the postconcert map, and the rightmost is the next day. The changes from baseline are readily apparent visually by comparing the predominance of red/yellow color areas to blue/green color areas. The bottom half of the figure shows the overall map, with those points where the difference from the baseline map was greater than two standard deviations from the average difference between maps marked in red|
Click here to view
| Discussion|| |
In this study, 13 subjects received an acute, high-level music exposure. Although the exposures overall were modest, significant changes from baseline to the postnoise tests could be seen in average pure-tone audiogram, DP-gram, and DPOAE level maps. While the DP-gram and audiogram did not show continued differences from baseline in the next day measures, the DPOAE level mapping measures showed continued significant differences between the baseline and next day maps. These data suggest that the map may be useful for following the recovery after noise exposure, and may be a more sensitive measure of cochlear changes after noise exposure.
The DPOAE level maps also provided an overall picture of which regions in the cochlea were affected. The frequency bands and ratios where the most significant changes occurred could be identified. The visual presentation of the mapping data also offers an improvement over the limited scope DP-gram. A comparison of [Figure 4] and [Figure 9], for example, shows the advantages of presenting the DPOAE data as an image, rather than as a plot of mean data points at selected frequencies.
Noise exposure, frequency sweeps, and 1/3 octave band noise
The noise exposure dose varied widely among the subjects. On average, the subjects were exposed to 95.5 dBA during the concerts, which according to the NIOSH recommendations, would be a permissible noise exposure for 42 min. Therefore, the subjects on average exceeded the permissible exposure 7.5 fold (749%). Only one subject had an exposure less than 100% dose (85%). Two subjects had very high-level exposures resulting in doses of 2,511% and 2,658%. One objective was to determine if DPOAE level mapping might be more sensitive than other audiological techniques, and a mild or moderate noise exposure helps to show this. The fact that changes could be seen on the audiogram, DP-gram, and map indicate that there was temporary auditory damage even though most of the exposures were not particularly extreme and were representative of workplace noise exposures. Subjects were counseled regarding their noise exposures and advised to utilize high-fidelity hearing protectors for concert sound exposures in the future.
The in-ear probe position frequency sweeps were consistent over time, indicating that the custom-molded probes helped to insure proper probe placement from visit to visit, and maintained a consistent stimulus level across measurements. As expected, there was more variability in the frequency sweep data at the highest frequencies. Nevertheless, the variability within the frequency sweeps stayed within a narrow range. When the standard deviations for all subjects over all time periods at each frequency were averaged, and after the standard deviation of that average was calculated, the results showed that the maximum variation of the standard deviation was less than 2 dB. Although further testing with the custom-molded approach is needed, these initial data are encouraging and suggest an improvement in stabilizing DPOAE measurements over time for individual subjects.
The in-ear 1/3 octave band noise levels were also consistent over time. This indicates that the external ambient noise was well-controlled from test to test, and the effect of physiological noise from each subject was minimized by each person serving as their own control. A major change in noise within the ear canal between testing periods can reduce the detection of low-level DPOAEs and influence the overall results from DPOAE testing. The fact that no major differences in 1/3 octave band levels across time periods was seen, provides confidence that the results were not due to differences in the noise environment during the testing.
DPOAE level mapping and threshold audiometry results
The Békésy audiometry results were important findings of this study. The audiograms showed small but significant differences between baseline and postconcert measurements. Interestingly, the frequencies where the audiometry results were different (4-6 kHz) roughly matched the DP frequencies where significant changes were seen. The ability to detect the small differences seen on the threshold audiograms was likely due to the 2 dB step size used with the Békésy audiometry. The calibration of the Sennheiser HDA-200 headsets showed that they could reliably produce stimuli within 1 dB of the desired level. The results suggest that Békésy audiometry, using a smaller step size, can detect changes due to moderate noise exposures that may be missed using conventional audiometry with 5 dB step size.
The major difference between the DPOAE map measures and audiometry in the present study was in the next day data. While the audiogram had recovered back to baseline by the day after the noise exposure, significant changes could still be detected within the DPOAE level map. The next day mapping data were significantly different from baseline when the data were analyzed by frequency, by ratio, or as an overall map. The only case where this was not true was for the overall interpolated maps for the smaller subset of subjects (n = 7) who were tested at L1 = L2 = 65 dB SPL. This may be due to a loss of statistical power related to a smaller sample size. When the baseline and postconcert maps were compared using the Kolmogorov-Smirnov statistic, the changes were not significant. These results suggest that further work is needed on the statistical analysis of overall map results. The analyses done by dividing the map into 1/3 octave bands, or into bins spanning ratios of 0.025, provided the same results over time, regardless of whether the data were analyzed using all the subjects or only those subjects who had maps done at L1 = L2 = 65 dB SPL. The overall map analysis, however, weighs all regions of the map equally. This could present a problem, since this analysis does not focus on those regions most likely to show changes or regions where DPOAEs are more robust. Further work is needed to determine the best approach to comparative map analysis across groups.
DPOAE level map findings
In our previous cross-sectional study, we showed that individuals with NIHL had, on average, reduced DPOAE amplitudes over a wide range of frequencies and ratios.  The data from this previous study suggested that there may be benefit in examining more than just the optimal DPOAE level at a single ratio when monitoring individuals for changes over time. In the present study, when the data were analyzed by ratio, significant changes were seen postconcert at 1.175, slightly below the optimal 2f1 -f2 DPOAE ratio of 1.2, and at several other ratios when considering the day after concert sound exposure. For the 2f2 -f 1 emission, differences were also seen within several ratio bands. These results support testing at ratios other than the optimal 1.2 ratio for detecting changes due to sound exposure.
The sound exposure could not be standardized among the subjects, and subjects likely differed in their susceptibility to cochlear damage from sound. Nevertheless, the main objective of the study was to compare results between different measurement techniques in terms of the ability to detect auditory change from hazardous sound exposure. In addition, the time course for recovery from the hazardous sound exposure may differ among subjects. The fact that some subjects may have had more sound exposure than others does not change the main result, which was the detection of changes the day after the sound exposure that were not seen on the audiogram or DP-gram.
Another limitation is the lack of multiple baseline measurements for the DPOAE map. The initial study design included two baseline maps prior to noise exposure and an additional map 1 week after the noise exposure. These other maps, however, were affected by electrical noise at the other testing locations used for those measurements. Despite this, the use of the custom-molded probes with the frequency sweeps likely led to consistent probe positioning and minimized stimulus variability. The in-ear 1/3 octave band measurements indicated that changes in the noise environment within the ear canal did not influence the results. Also, variability within the DPOAE maps would likely be random, and so would uniformly reduce the possibility of detecting a significant difference between average maps at any time period. Nevertheless, future studies are needed to verify the repeatability of the DPOAE level maps over time. The sequence of tests may have influenced the outcomes. But, since the mapping was performed last in the sequence, the inability of the audiogram or DP-gram to detect changes cannot be attributed to test sequence.
Future work might examine whether the mapping approach here is optimal for detecting NIHL. For example, after identifying the ratios that show the greatest DPOAE changes, it may be possible to create a map of input-output functions at these ratios. In other words, the ratio could be fixed, while frequency and stimulus intensity are varied to create a three dimensional (3-D) input-output DPOAE surface. A data set generated in this way could possibly offer additional options for monitoring NIHL effects.
| Conclusion|| |
DPOAE level maps showed changes the day after loud music exposure that were not apparent on either the audiogram or the standard DP-gram. Also, the DPOAE maps provide additional information on the regions of the cochlea most affected by the hazardous sound exposure when compared to Békésy 2 dB step-size audiometry and 2f1 -f2 DP-grams. Changes were concentrated within DPOAE frequencies from 2.8 to 5.6 kHz, and were seen at multiple f2 /f1 ratios. Not all regions of the map are equally important for analysis, and investigation of the optimal statistical techniques to use are ongoing. The maps provide considerable information on cochlear responses, but the optimal analysis techniques are still under development.
This work is supported by Office of Naval Research Grant No. N00014-09-1-0859. We thank Dr. Brenda Lonsbury-Martin and Dr. Glen Martin for their help with the testing and developing of the DPOAE mapping hardware and protocols. We would like to thank all the individuals who participated in the study.
Financial support and sponsorship
This work was supported by grant N00014-09-1-0859 from the Office of Naval Research.
Conflicts of interest
There are no conflicts of interest.
| References|| |
Knight RD, Kemp DT. Indications of different distortion product otoacoustic emission mechanisms from a detailed f1,f2 area study. J Acoust Soc Am 2000;107:457-73.
Martin GK, de La Garza A, Stagner BB, Lonsbury-Martin BL. Detailed DPOAE level/phase maps in normal and noise-damaged rabbit ears: Insights into generation processes. Assoc Res Otolaryngol 2005;28:52.
Meinke DK, Clavier OH, Norris J, Kline-Schoder R, Allen L, Buckey JC. Distortion product otoacoustic emission level maps from normal and noise-damaged cochleae. Noise Health 2013;15:315-25.
Buckey JC, Fellows AM, Jastrzembski BG, Maro II, Moshi N, Turk M, et al
. Pure-tone audiometric threshold assessment with in-ear monitoring of noise levels. Int J Audiol 2013;52:783-8.
American National Standard. Specification for Audiometers. Melville, NH: Secretariat, Acoustical Society of America; 2004.
Whitehead ML, Stagner BB, Lonsbury-Martin BL, Martin GK. Effects of ear-canal standing waves on measurements of distortion-product otoacoustic emissions. J Acoust Soc Am 1995;98:3200-14.
Scheperle RA, Goodman SS, Neely ST. Further assessment of forward pressure level for in situ
calibration. J Acoust Soc Am 2011;130: 3882-92.
Scheperle RA, Neely ST, Kopun JG, Gorga MP. Influence of in situ
, sound-level calibration on distortion-product otoacoustic emission variability. J Acoust Soc Am 2008;124:288-300.
Norris J, Clavier OH, Meinke DK, Fellows A, Kline-Schoder R, Buckey JC. Use of custom-Molded Earpieces for DPOAE Measurements. New Orleans, LA: National Hearing Conservation Association (NHCA); 2012. p. 46.
American National Standard. Specification for Octave-Band and Fractional-Octave-Band Analog and Digital Filters. Melville, NH: Secretariat, Acoustical Society of America; 2009.
Fitzgerald TS, Prieve BA. Detection of hearing loss using 2f2-f1 and 2f1-f2 distortion-product otoacoustic emissions. J Speech Lang Hear Res 2005;48:1165-86.
Horn JH, Pratt SR, Durrant JD. Parameters to maximize 2f2-f1 distortion product otoacoustic emission levels. J Speech Lang Hear Res 2008;51:1620-9.
NIOSH. Criteria for a Recommended Standard Occupational Noise Exposure. Cincinnati, OH: National Institute for Occupational Safety and Health; 1998.
American National Standard. Maximum Permissible Ambient Noise Levels for Audiometric Test Rooms. New York, New York: Secretariat, Acoustical Society of America; 1999. p. 10.
Jay C Buckey
The Geisel School of Medicine at Dartmouth, One Medical Center Drive, Lebanon, New Hampshire - 03756
Source of Support: None, Conflict of Interest: None
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8], [Figure 9]