Joshua J. Hajicek, Maryam Naghibolhosseini, Simon Henin, and Glenis R. Long ... inserted, the fit in the canal was evaluated for leaks by playing white noise.
PS-176: Similari'es between S'mulus Frequency Otoacous'c Emissions, Distor'on Product Otoacous'c Emission Components, and Audiograms Joshua J. Hajicek, Maryam Naghibolhosseini, Simon Henin, and Glenis R. Long Speech-Language- and Hearing Sciences, The Graduate Center, The City University of New York
This research was supported in part by the Office of Naval Research work unit number WU50904 through the Naval Submarine Medical Research Laboratory. The views expressed in this article are those of the authors and do not necessarily reflect the official policy or position of the US Department of the Navy, US Department of Defense, nor the US government. Parts of this research were also supported by the Provost Digital Innovation Grants and the New Media Lab at the CUNY Graduate Center. We also would like to thank Suzanne Thompson and Shawn Goodman.
2
4
8
60
OAE Frequency (kHz) NN22130604L
30
-10
20
-10
30
-20
40
-30
50
1
2
4
8
60
OAE Frequency (kHz) NN23130605R
30
0
20
-10
30
-20
40
-30
50
-40
60
1
2
4
8
NN24130807
10
0
20
-10
30
-20
40
-30
50
-40
1
2
4
8
NN25130612R
OAE (dB SPL)
-20
40 50
1
2
60
10
0
20
8
60
35
= 0.71
r65
35
= 0.67
0
10
10
0
20
-10
30
-20
40
-30
50
1
2
4
8
NN16130508L
60
-10 r60
20
35
= 0.78
r65
35
= 0.74
0
10
10
0
20
-10
30
-20
40
-30
50
-40
1
2
4
8
60
OAE Frequency kHz NN15130502R
30
35
= 0.24
r65
35
= 0.49
0 10
0
20
-10
30
-20
40
OAE Frequency (kHz)
-30
50
Figure 1. Levels of SFOAEs and DPOAE reflec'on components, audiograms and generator components when L2=60 dB SPL for five par'cipants. Note the generator component is ploaed as a func'on of f2 frequency while the reflec'on component is a func'on of the DPOAE frequency. SFOAEs were measured using s'muli levels from 25-40 dB SPL (in 5 dB increments) while DPOAEs were measured using scissors paradigm (Kummer et al., 1998) with L1= 63, L2 = 60 and L1=65, L2=65 dB SPL. Audiograms were measured using narrowband noise. Noise floors for each measurement are indicated by thin doaed lines. The numbers on the leb axes pertains to OAE amplitude in dB SPL while the numbers on the right axes indicate narrowband thresholds in dB SPL (converted from dB HL using ANSI S3.6 2004). Note that symbols for the thresholds are indicated by a ‘X’ or ‘O’ depending on if the ear was leb or right, respec'vely.
-40
-10
30
-20
40
-30 -40
50
1
2
4
8
OAE: dB SPL
10
60
1
2
4
8
No L2=65 Data N
R 25 N
N
24
13
13
06
08
12
07
R 05 06 13
0.6 0.4 0.2 0
Ear
Figure 4. Correla'ons between each combina'on of DPOAE reflec'on component evoked by either L2 = 60 or 65, and SFOAE evoked by probe level of 35 dB SPL. The legend follows the same format as in figure 3. Note that some L2=65 data is missing for par'cipants NN5 and NN25, therefore, correla'ons between this level and SFOAE are not presented. The light gray line at y = 0.6 represents a correla'on value that we used to separate high and low correla'ons and for which p-values approach 0.05.
The reflec'on component (L2 = 60 dB SPL) was generally lower in level than SFOAE evoked with a L35 dB SPL probe. Based on correla'ons (figures 3 and 4), the best matches in level and quasiperodicity were between the reflec'on component evoked by L2 = 60 dB SPL and a probe level of 25 dB SPL. For many par'cipants the best correla'ons between the DPOAE reflec'on component and SFOAE appear when both emissions are robust, and at the lowest s'mulus levels evaluated (L2=60) for DPOAE and (L25 or L30 dB SPL) for SFOAE. Correla'ons below 0.6 (close to p=0.05) were obtained from par'cipants NN7, NN9, NN10R, NN11, NN14, NN15, NN21, and one repeated measure of NN24. These low correla'ons stem from shibs in fine structure between the reflec'on component and SFOAE (e.g. NN21, and NN7 15 to 25 dB) than the reflec'on component. Examples of this can be observed in NN24 from 3-7 kHz, NN21 >2 kHz, NN18 > 2 kHz, NN14 > 2.5 kHz. In cases where the emissions are not as robust (lower SNRs), the most well matched fine structure paaerns occur when the reflec'on and generator components are nearly equivalent in amplitude (e.g. NN11 from 2-3 kHz, NN14 < 2.5 kHz, NN16R around 2 kHz, NN16 L from 1-2.5 kHz, NN17 from 1-4.5 kHz, NN21 from 1-2 kHz, NN25, from 2.2 to 6 kHz, NN26 around 1.6 kHz, NN10L from 1.6-2.9 kHz). In most par'cipants the generator component increases with frequency while the reflec'on component decreases. We kept f2/ f1 constant across frequency. However, it may not be op'mal to maintain a similar frequency ra'o with frequency (e.g. Johnson et al., 2009; Poling et al., 2013). As the frequency ra'o changes the op'mal level ra'o may also change. Obtaining data with op'mal frequency and level ra'os may keep DPOAE component levels more constant as a func'on of frequency. We did not systema'cally evaluate systema'c shibs in the frequency of OAE fine structure so the correla'ons obtained may be subop'mal if there was a small shib in the finestructure between the reflec'on component and the SFOAEs. DPOAE reflec'on components also saturated at rela'vely lower levels than did SFOAEs (Long et al., 2009). Un'l we have established the op'mal s'mulus condi'ons we cannot yet address the poten'al role of the DPOAE reflec'on component as a diagnos'c tool. The rela'onship between audiograms the generator components is complicated. Ploqng the generator component as a func'on of f2 (where it is thought to be generated) might be expected provide a beaer predic'on of the audiogram. However, in our data the audiograms of some par'cipants more closely match the generator component level changes when ploaed as a func'on of 2f1-f2 frequency.
Conclusions:
-10 r60
20
0.8
DISCUSSION AND CONCLUSIONS
-10 r60
0
10
4 NN16130813R
30
-10
20
30
OAE Frequency kHz
OAE Frequency (kHz)
30
-10
20
OAE: dB SPL
10
0
20
-40
Audiogram (dB SPL)
OAE (dB SPL)
0
= 0.5
10
30
-10
20
35
OAE Frequency kHz
OAE Frequency (kHz)
30
r65
0
-40
Audiogram (dB SPL)
OAE (dB SPL)
10
= 0.59
-30
0
10
35
10
-10
20
-10 r60
20
OAE: dB SPL
-40
50
30
65 35 60 35
N 308 N 51 12R N 303 N 71 20R N 303 N 91 21R 30 N N 10 327 N 130 R N 10 328 13 R N N 04 10 11 N 130 R N 11 41 1 N 130 L N 14 329 N 130 R N 15 430 N 130 R N 16 502 N 130 R N 16 813 13 R N N 05 16 08 N 130 R N 17 50 13 8L N N 05 18 09 N 130 R N 21 513 N 130 L N 22 529 N 130 L N 23 60 4 N 130 L N 24 605 13 R 07 N N 24 15R N N 130 25 80 1 30 7 N N 26 612 13 R 08 12 R
-30
Correlation Coefficient
40
NN14130430R
Correlation Coefficients between SFOAE L=35 and DPOAE Reflection Components at L2=65 or L2=60
51
-20
OAE Frequency kHz
N N
30
8
Audiogram: dB SPL
-10
4
Audiogram: dB SPL
20
2
Audiogram: dB SPL
0
1
L
Figure 3. Correla'ons between each combina'on of DPOAE reflec'on components evoked by either L2 = 60 or 65 dB SPL, and SFOAE probe levels of 25, 30, 35, or 40 dB SPL. The numbers in the legend correspond to the level of the DPOAE L2 primary that evoked the reflec'on component followed by the level of the probe used to evoke the SFOAE. Note that some L2=65 data is missing for par'cipant NN25 so no correla'ons were could be calculated for these levels.
60
Audiogram: dB SPL
10
OAE: dB SPL
10
-40
Audiogram (dB SPL)
OAE (dB SPL)
0
04
Ear
1
20
N
0
23
10
N N
25
0
06
30
0
13
35
= 0.43 = 0.47 = 0.56 = 0.58
22
25
40
No L2=65 Data
1
30
r65 r65 r65 r65
N
50
10
35
= 0.46 = 0.5 = 0.6 = 0.63
N
-30
20
40
L
40
r60 r60 r 60 r60
0.2
29
-20
-10
05
30
-40
RESULTS
ACKNOWLEDGEMENTS
-10
NN21130529L
30
0.4
13
20
0.6
21
0
0.8
No L2=65 Data
10
60 25 60 30 60 35 60 40 65 25 65 30 65 35 65 40
N
10
Correlation Coefficients between SFOAE L25-L40, and DPOAE Reflection Components at L2 = 65 or 60
N
RL2 65 G L2 60 Noise Thresh
Correlation Coefficient
0
Audiogram: dB SPL
20
1
loca'ons of peaks and nulls across levels and OAE types. Correla'ons were then calculated between all available SFOAE levels and the two levels of DPOAE reflec'on components. Note that the generator component (shown in the red dashed line) is ploaed as a func'on of 2f1-f2 when L2 = 60 dB SPL. Ploqng the generator as a func'on of OAE frequency shows the dependence of the reflec'on component on the generator component level. The correla'on between SFOAEs and DPOAE reflec'on components are given in the upper right of each plot and are summarized in figures 3 and 4.
OAE: dB SPL
SF L40 RL2 60
Par0cipants: 19 par'cipants between the ages of 18-40 were recruited for this study.
The quasiperiodic level varia'on in SFOAEs and the DPOAE reflec'on components were similar in most ears. SFOAEs were frequently larger in amplitude than the DPOAE reflec'on component, but the rela'onship between individual audiograms and DPOAE components was complex. In a subset of par'cipants, threshold sensi'vity appeared to depend on DPOAE reflec'on component level while in other par'cipants, or at other frequencies, it depended on the generator component. In order to explore this rela'onship further we ploaed the generator as a func'on of f2 frequency (figure 1) and as a func'on of 2f1-f2 frequency (figure 2). Figure 1. illustrates the similari'es and differences in the quasiperiodicity between SFOAE and DPOAE reflec'on components. The hypothesis is that these emissions arise from the same underlying cochlear mechanism (e.g. Talmadge et al., 1998; Shera and Guinan, 1999; Talmadge et al., 2000). It is interes'ng to note that the reflec'on component was saturated in level (as was the generator component (not shown)) whereas the SFOAE only saturates in some subject dependent frequency regions.
-10
Audiogram (dB SPL)
SF L30 SF L35
METHODS
Hearing thresholds were es'mated using narrowband noise audiograms measured at standard inter-octave intervals from 0.25-8 kHz. Audiograms were converted from dB HL to dB SPL. An audiogram was not collected from NN18. Par'cipants were required to have at least three hearing thresholds < 20 dB HL between 1000 and 8000 Hz. Otoscopy was used to verify a healthy tympanic membrane and that ear canals were free of excess cerumen. Middle ear status was verified via tympanometry and tympanic peak pressure was required to be within +/- 50 daPa of ambient pressure. Procedure: All data was recorded in an IAC double-walled sound booth. Par'cipants were seated in a comfortable reclining chair in the reclined posi'on. The ear with beaer thresholds was measured or, if the thresholds were similar, the right ear was selected. Aber the OAE probe was inserted, the fit in the canal was evaluated for leaks by playing white noise at 65 dB SPL and performing a fast Fourier Transform. These “fit-checks” were performed periodically to verify that the probe was stable throughout the measurements. DPOAEs were measured with L1= 63, L2 = 60 and L1=65, L2=65 dB SPL using the scissors paradigm (Kummer et al., 1998), and f2/f1 = 1.22. The primaries were presented 48 'mes for each level. Pilot data from 6 addi'onal par'cipants (not included here) revealed that these levels evoked a reflec'on component that was similar in level to SFOAE evoked with a 35 dB SPL probe over the 2f1-f2 frequencies of 1-3 kHz. For the par'cipants presented here, we swept the primaries from 1.5 to 12.55 kHz. Sweep rates were 2s/octave (0.5 octaves/second) for a total dura'on of 6.12s. SFOAEs were measured at 35 dB SPL in all ears. Probe tones were con'nuously swept in frequency from 1000-8000 Hz. Responses from 128 sweeps were recorded. To reduce interference from the probe tone during analysis, a swept suppressor tone was presented along with the probe on every other sweep. The suppressor was 55 dB SPL and was maintained 50 Hz above the probe frequency. In a subset of five par'cipants, SFOAEs were collected with addi'onal probe levels of 25, 30, and 40 dB SPL. The level of the suppressor was always 20 dB above the probe. Equipment: S'muli were generated, and the ear canal response recorded, using custom Mac sobware (OSX) interfaced with a MOTU 828 Firewire Audio Interface (Cambridge, MA). The s'muli were delivered to the ear canal using Etymo'c ER-2 insert earphones coupled to the OAE probe. Ear canal pressure responses were recorded with either an Etymo'c ER10A or ER10B+ microphone/preamplifier system connected to a Stanford SR560 low-noise preamplifier. The output of the SR560 was connected to the MOTU, which digi'zed the signal at a sampling rate of 44.1 kHz. The digi'zed data was stored on a Mac computer for offline analysis. The ER10A and ER10B+ OAE probe microphone responses were calibrated (before LSF analysis) using the procedure specified by Siegel (2007) and Rasetshwane and Neely (2011). Responses from the same ear but across probes were well matched aber microphone calibra'on. Data Analysis: Analysis was performed using a least squares fit (LSF) analysis (see Talmadge et al., 1997; Long et al., 2008) on the ar'fact free averages. Ar'facts were reduced by down weigh'ng noisy segments before averaging. For the LSF, a model is generated that includes the expected components in the recordings (e.g. probe or primaries, suppressor, and significant OAE components). A least squares fit determines the complex amplitude coefficients for each component in the model by minimizing the mean squared error between the model and the recording. Noise floors were determined by subaveraging recordings into two buffers and subtrac'ng them. The same LSF was then performed on the residual which provides a conserva've es'mate of the noise floor. For SFOAE, the LSF analysis used a modified latency func'on (Naghibolhosseini et al., 2014). The level of the probe was reduced before analysis by subtrac'ng separate averages of sweep responses with and without the suppressor. Subtrac'on produced residual that had significantly reduced probes. A LSF was performed on the residuals to es'mate the levels and phases of each component.
NN21130529L
30
Audiogram (dB SPL)
There are two major classes of otoacous'c emissions (OAEs), reflec'on source (a.k.a. place fixed) and nonlinear source OAEs (wave fixed) (e.g. Talmadge et al., 1998; Shera and Guinan, 1999). Each source arises from different processes and cochlear regions. Reflec'on OAEs, such as s'mulus frequency OAEs (SFOAEs) arise from coherent reflec'on in the region near the OAE best frequency place. Distor'on product OAEs (DPOAEs) consist of at least two components. One depends on cochlear nonlinearity and is generated where the two primaries overlap (the nonlinear source). Distor'on product energy generated in the overlap region travels basally to the ear canal (generator component) and apically to its best frequency place where it is coherently reflected producing a second component with the same frequency (DPOAE reflec'on component). The major difference between the SFOAE and the DPOAE reflec'on component is the source of the energy that travels to the OAE best frequency region. Consequently, it is expected that there would be a close rela'onship between the quasiperodicity of SFOAEs and the DPOAE reflec'on components. The amount of the energy generated by the nonlinear interac'on reaching the OAE best frequency region will depend on the health of the cochlea in the overlap region and the amount of amplifica'on occurring as it travels to the DPOAE best frequency region. The difference in amplitude of the two DPOAE components is thus predicted to depend on cochlear health in different regions and is expected to be related to the hearing status reflected in an audiogram.
SF L25
OAE (dB SPL)
ABSTRACT
In figure 2, data was smoothed with a 1/12th octave band median filter. Smoothing reduced the impact of small shibs in the
60
OAE Frequency kHz
Figure 2. SFOAEs and DPOAE reflec'on components, generator component (ploaed as a func'on of 2f1-f2 frequency) and audiograms, for five ears from four par'cipants (note R and L ears from NN16). Data was smoothed with 1/12th octave median filter in order to reduce the effects from small shibs in nulls and/or peaks. Pearson correla'ons are shown in the upper right corner of each plot. Correla'on coefficients were computed for each combina'on of SFOAE and DPOAE reflec'on component levels with the format of ra b, where ‘a’ and ‘b’ are DPOAE’s L2 and the SFOAE probe level, respec'vely.
As predicted from their similar origins, the DPOAE reflec'on component and SFOAE have similar quasiperodicity within individual ears, especially when both reflec'on source emissions are robust. The rela'onship of the DPOAE components to the audiogram depends on the rela've levels of each DPOAE component.
REFERENCES
Johnson, T. A., Neely, S. T., Kopun, J. G., Dierking, D. M., Tan, H., and Gorga, M. P. (2010). “Clinical test performance of distor'on-product otoacous'c emissions using new s'mulus condi'ons,” Ear and Hearing, 31, 74–83. doi:10.1097/AUD.0b013e3181b71924 Kalluri, R., and Shera, C. A. (2013). “Measuring s'mulus-frequency otoacous'c emissions using swept tones,” J. Acoust. Soc. Am, 134, 356–368. doi:10.1121/1.4807505 Kummer, P., Janssen, T., and Arnold, W. (1998). “The level and growth behavior of the 2 f1-f2 distor'on product otoacous'c emission and its rela'onship to auditory sensi'vity in normal hearing and cochlear hearing loss,” J. Acoust. Soc. Am, 103, 3431–3444. Long, G. R., Jeung, C, and Talmadge,C. L., (2009). “Concepts and Challenges in the Biophysics of Hearing” Edited by N.P. Cooper & D.T. Kemp. World Scien'fic Press, Singapore. pp 203-208. Long, G. R., Talmadge, C. L., and Lee, J. (2008). “Measuring distor'on product otoacous'c emissions using con'nuously sweeping primaries,” J. Acoust. Soc. Am, 124, 1613–1626. doi:10.1121/1.2949505 Naghibolhosseini, M., Hajicek, J. J., Henin, S., and Long, G. R. (2014). “Discrete and swept-frequency SFOAE with and without suppressor tones,” In Associa'on for Research in Otolaryngology, Bal'more, 1. Presented at the Associa'on for Research in Otolaryngology. Rasetshwane, D. M., and Neely, S. T. (2011). “Calibra'on of otoacous'c emission probe microphones,” J. Acoust. Soc. Am, 130, EL238–6. doi:10.1121/1.3632047 Shera, C. A., and Guinan, J. J., Jr. (1999). “Evoked otoacous'c emissions arise by two fundamentally different mechanisms: a taxonomy for mammalian OAEs,” J. Acoust. Soc. Am, 105, 782–798. Siegel, J. H. (2007). “Calibra'ng Otoacous'c Emissions,” In M. S. Robineae and T. J. Glaake (Eds.), Otoacous'c Emissions, New York, 3rd ed., pp. 403–427. Talmadge, C. L., Long, G. R., Tubis, A., and Dhar, S. (1999). “Experimental confirma'on of the two-source interference model for the fine structure of distor'on product otoacous'c emissions,” J. Acoust. Soc. Am, 105, 275–292. Talmadge, C. L., Tubis, A., Long, G. R., and Piskorski, P. (1998). “Modeling otoacous'c emission and hearing threshold fine structures,” J. Acoust. Soc. Am, 104, 1517–1543. Talmadge, C. L., Tubis, A., Long, G. R., and Tong, C. (2000). “Modeling the combined effects of basilar membrane nonlinearity and roughness on s'mulus frequency otoacous'c emission fine structure,” J. Acoust. Soc. Am, 108, 2911–2932.
