Download Effects of a Suppressor Tone on Distortion Product Otoacoustic

Effects of a Suppressor Tone on Distortion Product
Otoacoustic Emissions Fine Structure: Why a
Universal Suppressor Level Is Not a Practical
Solution to Obtaining Single-Generator DP-Grams
Sumitrajit Dhar and Lauren A. Shaffer
Objectives: The use of a suppressor tone has been
proposed as the method of choice in obtaining singlegenerator distortion product (DP) grams, the speculation being that such DP grams will be more predictive of hearing thresholds. Current distortion product
otoacoustic emissions (DPOAE) theory points to the
ear canal DPOAE signal being a complex interaction
between multiple components. The effectiveness of a
suppressor tone is predicted to be dependent entirely
on the relative levels of these components. We examine the validity of using a suppressor tone through a
detailed examination of the effects of a suppressor on
DPOAE fine structure in individual ears.
2000; Kim, Paparello, Jung, Smurzynski, & Sun, 1996;
Kimberley, Hernadi, Lee, & Brown, 1994; Kummer,
Janssen, & Arnold, 1998; Nelson & Kimberley, 1992).
Clinical DPOAE tests are especially useful for newborns and other difficult-to-test populations. However,
the clinical utility of DPOAEs remains limited even
after two decades of research (Gorga, Neely, Bergman,
Beauchaine, Kaminski & Liu, 1994; Heitmann, Waldmann, & Plinkert, 1996; Shera, 2004) primarily due to
significant variability in the correlation between
DPOAE level or threshold and the traditional pure
tone audiogram (see Kummer et al., 1998, for review).
One factor hypothesized to be responsible for this
variability is the quasiperiodic pattern of DPOAE level
fluctuation known as fine structure (Dhar, Talmadge,
Long, & Tubis, 2002; Gaskill & Brown, 1996; He &
Schmiedt, 1993), which can result in peak-to-valley
amplitudes as great as 20 dB in normal-hearing individuals (He & Schmiedt, 1993; Shaffer, Withnell,
Dhar, Lilly, Goodman & Harmon, 2003). This variation, in turn, could result in increased overlap in
DPOAE-level distributions for normal-hearing and
hearing-impaired groups as well as reduced correlation between hearing thresholds and DPOAE level. As
we have gained more insight into the generative mechanisms of DPOAEs in the last few years, the origin of
fine structure has become apparent.
Design: DPOAE fine structure, recorded in 10 normal-hearing individuals with a suppressor tone at
45, 55, and 65 dB SPL, was compared with recordings without a suppressor. Behavioral hearing
thresholds were also measured in the same subjects, using 2-dB steps.
Results: The effect of the suppressor tone on DPOAE
fine structure varied between ears and was dependent on frequency within ears. Correlation between
hearing thresholds and DPOAE level measured
without a suppressor was similar to previous reports. The effects of the suppressor are explained in
the theoretical framework of a model involving
multiple DPOAE components.
Conclusions: Our results suggest that a suppressor
tone can have highly variable effects on fine structure across individuals or even across frequency
within one ear, thereby making the use of a suppressor less viable as a clinical tool for obtaining
single-generator DP grams.
Origin of DPOAE Fine Structure
According to currently accepted DPOAE models,
distortion product (DP) energy is generated as the
result of cochlear nonlinearities at the overlap region of the activity patterns of the two stimulus
tones (f1 and f2, f2⬎f1) (Mauermann, Uppenkamp,
Hengel, & Kollmeier, 1999; Talmadge, Tubis, Long,
& Piskorski, 1998). From the overlap region, the DP
energy travels in two directions on the basilar membrane, basally toward the oval window and apically
(in the case of 2f1-f2) toward the characteristic frequency region of the DP (CFdp). The apically traveling energy from the generator or overlap region is
reflected from the CFdp region as the result of
random inhomogeneities on the basilar membrane
(Talmadge et al., 1998; Zweig & Shera, 1995). The
(Ear & Hearing 2004;25;573–585)
Objectivity, noninvasiveness, and frequency specificity make distortion product otoacoustic emissions
(DPOAEs) a desirable clinical tool for evaluating cochlear status in general and outer hair cell function in
particular (e.g., Dorn, Piskorski, Gorga, Neely, &
Keefe, 1999; Gorga, Nelson, Davis, Dorn, & Neely,
Department of Speech and Hearing Sciences (S.D.), Indiana University, Bloomington, Indiana; and the Department of Speech Pathology and Audiology (L.A.S.), Ball State University, Muncie, Indiana.
0196/0202/04/2506-0573/0 • Ear & Hearing • Copyright © 2004 by Lippincott Williams & Wilkins • Printed in the U.S.A.
573
574
reflected energy travels basally out to the ear canal
along with the initial, basally traveling energy from
the overlap region. Thus, the DPOAE recorded in
the ear canal is a vector sum of two components, one
due to cochlear nonlinearities arising from the overlap region between the stimulus excitations patterns
(the nonlinear component), and the other from the
CFdp due to reflection of energy from random inhomogeneities (the reflection component).
Inherent in the generation mechanisms of the two
components is a fundamental difference in their
phase behavior as a function of frequency for a fixed
f2/f1 sweep. Although the phase of the nonlinear
component is relatively invariant, that of the reflection component changes rapidly as a function of DP
frequency. The fine structure observed in the overall
DPOAE level is due to constructive and destructive
interference between the two components as their
phases vary at different rates with frequency. The
reader is referred to Talmadge, Long, Tubis, and
Dhar (1999) or Shera and Guinan (1999) for a
detailed treatment of the phase behavior of the
different generation mechanisms and their effect on
DPOAE fine structure. Further details about the
above-mentioned model of DPOAEs have been published in analytical form by Talmadge et al. (1998)
and in more descriptive form by Dhar et al. (2002).
Hypothesized Effects of Fine Structure and
Multiple Sources on Threshold Prediction
The apparent discrepancy between DPOAE level
and hearing thresholds may be partially understood
in light of the above model. The natural variation of
DPOAE level has been proposed to be a major factor
in hindering consistent correlation between the two
measures (Shaffer et al., 2003); the fact that two
cochlear regions are represented in the composite
DPOAE signal has been hypothesized to complicate
clinical interpretation (Shera, 2004). Shera (2004)
has recently commented on the “uncontrolled mixing” of two signals from disparate locations on the
basilar membrane due to different mechanisms as a
major limitation to the clinical power of DPOAEs.
Thus, it is not surprising that attempts to correlate
hearing thresholds with DPOAE level have seen
varied success with r values typically between 0.2
and 0.5 either in the positive or negative direction
(e.g., Dorn, Piskorski, Keefe, Neely, & Gorga, 1998;
Gaskill & Brown, 1993; Kimberley et al., 1994).*
*The r2 (multiple regression coefficient) or r (correlation coefficient) values are reported in the literature as either positive or
negative numbers, depending on the particular indices being
compared. Regression or correlation measures are expected to be
negative when DPOAE levels are compared with hearing thresholds. On the other hand, these values are expected to be positive
EAR & HEARING / DECEMBER 2004
The problem of correlating hearing thresholds
and OAE measures is further compounded by differences in individual versus group data and by variability in the direction of correlation. These issues
are examined in greater detail in the Discussion
section of this article.
Multivariate Measures
Initial attempts to improve correlation between
DPOAE level and hearing thresholds concentrated on
the use of other indices such as signal-to-noise ratio
(SNR) in isolation or in conjunction with DPOAE level.
However, the use of multiple parameters from the DP
gram did not result in a significant improvement in
the efficacy of DPOAEs as a clinical tool for predicting
hearing thresholds. Researchers have since expanded
the search for the optimal clinical measure to include
DPOAE growth curves (Nelson & Kimberley, 1992)
and multivariate analyses (Dorn et al., 1999). Simultaneous evaluation of multiple DPOAEs such as 2f1-f2
and 2f2-f1 to identify normal versus impaired hearing
status has also been attempted (Gorga, Neely, Dorn &
Hoover, 2000). Use of these statistical tools and multiple indexes typically results in improved predictability of auditory status. For example, Gorga et al. (2000)
reported ROC area curves ⬎0.95% and specificities
⬎92% when distinguishing normal-hearing ears from
those with hearing loss.
Input-Output Functions
DPOAE thresholds predicted from input-output
functions have also been correlated with behavioral
thresholds with limited success. To determine the
DPOAE threshold, the sound pressure levels of the
stimulus tones are varied while keeping their frequencies fixed to obtain an input-output (I/O) curve at an
audiometric test frequency. The stimulus sound pressure level at which the DPOAE level reaches a criterion amount or becomes greater than the noise floor by
a criterion amount is considered the DPOAE threshold. This threshold is then compared with the audiometric threshold at the same frequency. However,
DPOAE thresholds estimated from I/O functions in
this manner have been reported to be highly variable
in both normal-hearing and hearing-impaired populations (Nelson & Kimberley, 1992).
A variation on the method of DPOAE threshold
determination described above is to plot DPOAE
pressure as a function of stimulus pressure (in
when DPOAE thresholds (as measured from an input-output
function) are compared with hearing thresholds. Comparisons
between hearing status and DPOAE results have also been
reported in terms of the area under the receiver operating
characteristics curve.
EAR & HEARING, VOL. 25 NO. 6
Pascals) in place of decibel sound pressure level. The
data are then fit with a linear regression function
and the DPOAE threshold is extrapolated as the
stimulus level at which the DPOAE level (in Pascals) is zero (Boege & Janssen, 2002; Gorga et al.,
2003). Correlation coefficients between DPOAE
thresholds obtained in this manner and hearing
thresholds have been recently reported to be 0.65 for
all frequencies combined (Boege & Janssen, 2002)
and between 0.49 and 0.85 for individual audiometric frequencies (Gorga et al., 2003). This relatively
high degree of correlation can be attributed, partly
at least, to the strict inclusion criteria of regression
functions in the correlational analysis, such as a
SNR⬎6 dB, a standard error ⬍10 dB, and a linear
regression slope ⬎0.2 ␮Pa/dB. The application of
these “screening” criteria resulted in the exclusion of
30% of the I/O functions in one report (Boege &
Janssen, 2002) and 60% of them when a greater
number of hearing-impaired subjects with a wider
range of hearing losses were included in another
(Gorga et al., 2003). The majority of I/O functions
not meeting the SNR criterion were associated with
behavioral thresholds ⬎30 dB HL. Conversely, the
majority of I/O functions not meeting the slope
criterion were associated with normal behavioral
thresholds. It should be noted that exclusion of cases
based on an SNR criterion should not detract from
the applicability of this approach, as these would be
consistent with hearing loss. However, exclusion
caused by noncompliance with the slope criterion
could be attributable to the influence of two components independent of hearing status. While extrapolation of the DPOAE threshold from a linearized
I/O curve appears to result in improved correlation
with hearing thresholds, the false-positive rate, the
condition in which normal ears fail to meet the
criteria, remains high at 19% (Gorga et al., 2003).
Single-Generator DPOAEs
A variety of attempts to predict hearing thresholds
from different DPOAE measures have met with limited success. From the theory of DPOAE generation,
one might predict that correlation of hearing threshold
and DPOAE level would improve if: (1) DPOAE-level
fine structure could be reduced or eliminated, and (2)
the number of basilar membrane regions contributing
to the recorded DPOAE could be limited to one. Suppressor tones, close in frequency to 2f1-f2, have been
used to reduce the level variation of fine structure by
reducing or eliminating the energy reflected at the
CFdp region (Heitmann, Waldmann, Schnitzler, Plinkert, & Zenner, 1998; Talmadge et al., 1999). The
clinical promise of such a procedure to eliminate level
variation was introduced by Waldmann, Heitmann,
575
and Plinkert (1997) in a German publication in which
the authors conceived of a “single-generator [sg]
DPOAE.” The appeal of a single-generator DPOAE is
that the distortion product measured in the ear canal
arises from a single cochlear region, the overlap region
of the stimulus tone excitation patterns, which can be
approximated under typical stimulus conditions as the
f2 characteristic frequency region. Without the reflection component, the ear canal DPOAE signal does not
exhibit the typically observed fine structure. Theoretically, this single-generator distortion product should
correlate better to the behavioral threshold at the f2
frequency. Indeed, GN Otometrics markets a DPOAE
measurement device featuring the “sgDPOAE” test.
The use of a suppressor tone is one of two currently
accepted methods for separation and isolation of the
DPOAEs components, the other being the use of inverse Fast Fourier transform analysis or spectral
smoothing (e.g., Kalluri & Shera, 2001; KonradMartin, Nelly, Keefe, Dorn & Gorga, 2001). The inverse Fast Fourier transform analysis results in more
reliable isolation of DPOAE components but requires
data recorded with very small frequency steps, making
it less suitable for clinical use (Kalluri & Shera, 2001;
Shaffer et al., 2003). Although the use of a suppressor
tone may appear to be an ideal solution to obtaining
DPOAE recordings from one cochlear region and
mechanism, closer examination of the DPOAE generative mechanism reveals complications that need to be
considered.
The effectiveness of a suppressor in altering fine
structure has been reported previously (Heitmann et
al., 1998; Plinkert, Heitmann, & Waldmann, 1997;
Talmadge et al., 1999), along with preliminary evidence of great variety in its effects between individuals
(Shaffer et al., 2003) and across stimulus conditions
(Kalluri & Shera, 2001). The effect of a suppressor tone
of a given intensity is determined purely by the relative magnitudes of the generator and reflection components. The effects of “undersuppression” and “oversuppression” have been presented in analytical form
demonstrating either residual fine structure in the
isolated generator component or modification of the
generator component itself (Kalluri & Shera, 2001). In
case of a relatively dominant reflection component, a
suppressor has also been shown to increase fine-structure depth by equalizing the two components (Talmadge et al., 1999).
In this report, we ask the general question, can a
universal suppressor tone be used to record sgDPOAEs in a group of normal-hearing individuals? In
attempting to answer this question, we examine the
various effects of a suppressor tone in individual
ears in light of the current theoretical understanding of DPOAE generative mechanisms. As a secondary point, we also examine the effects of using a
576
EAR & HEARING / DECEMBER 2004
suppressor on the correlation between DPOAE levels and behaviorally measured hearing thresholds.
METHODS
Subjects
Auditory thresholds and DPOAE levels were obtained from 20 normal-hearing ears of 10 participants between 18 and 27 yr of age recruited from the
population at the Indiana University. These individuals reported a negative history for otologic disease,
noise exposure, and ototoxicity. Additionally, each
participant exhibited good general health. All participants had normal hearing (pure-tone thresholds
⬍20 dB SPL, re: ANSI, 1989, 250 to 8000 Hz).
Immittance results were consistent with normal
middle ear function. All recordings were obtained in
sound treated audiometric test booths. Guidelines of
the Indiana University Human Subjects’ Committee
were strictly followed.
Pure-Tone Threshold Estimation
Auditory thresholds were obtained using a modified
Hughson-Westlake procedure at eight test frequencies
(250, 500, 1000, 2000, 3000, 4000, 6000, and 8000 Hz)
using ER-3 insert headphones connected to a GSI
(Grason-Stadler Instruments) 61 audiometer using 2
dB steps. Otoscopy and tympanometry (single probe
tone frequency of 226 Hz) were performed on all
subjects to rule out middle ear dysfunction.
DPOAE Recordings
DPOAE recordings were made using a custom-built
experimental system while the participants were
seated comfortably in a recliner inside a double-walled
audiometric test booth. The experimental hardware
was controlled by software developed in the laboratory
running on an Apple Macintosh computer. The MOTU
828 input-output device was used to generate all
signals during the experiment, which were fed to the
participants’ ear canals through a set of ER2 tube
phones after passing through TDT HB6 headphone
buffers. The outputs of the tube phones were coupled
to the ear canal through the probe assembly of an
Etymotic Research ER10 three-port microphone. The
signals recorded in the ear canal by the ER10 microphone were passed through the ER10 preamplifier
and a battery operated Stanford Research SR560 lownoise amplifier/filter before being digitized at a sampling rate of 44100 Hz by the MOTU 828. The SR560
performs narrow-band filtering on the data around the
frequencies of interest. The digitized signal was stored
on disk for off-line analysis.
High-resolution DPOAE recordings were made
from each ear for f2 frequencies between 1000 and
8000 Hz. Stimulus frequencies were chosen to yield
DPOAE data points separated by 0.025 mm on the
basilar membrane, using the Greenwood map (Greenwood, 1990), resulting in frequency spacing of approximately 4 and 18 Hz at DPOAE frequencies of 1000
and 6000 Hz, respectively. Stimulus levels of 65 and 55
dB SPL were used for the lower-frequency (f1) and
higher-frequency (f2) stimulus tones, respectively. The
frequency ratio between the stimulus tones was set at
1.22. Following the DPOAE recording without a suppressor tone, one was added 25 Hz below the DPOAE
at 2f1-f2 and its level (Ls) was varied (45, 55, and 65 dB
SPL) in successive runs. Note that the microphone
probe assembly used in these experiments had three
ports for sound delivery. This allowed us to use three
separate tube phones for the stimulus and suppressor
tones, eliminating any electrical mixing of the signals.
Each data point was recorded for a period of 4 sec.
Estimates of DPOAE level, phase, and noise floor were
made using a custom least-squares-fit (LSF) algorithm
(Long & Talmadge, 1997). In the LSF technique, the
known frequency of the DPOAE is used to estimate the
level and phase of the DPOAE along with the level of
the noise floor at that particular frequency. In contrast, typical FFT-based applications of DPOAE analysis provide an estimate of the noise floor by evaluating frequency bins adjacent to that of the DPOAE.
The instrumentation used for hearing threshold
estimation as well as DPOAE recordings were calibrated in a Zwislocki coupler before the commencement of the experiments. The system distortion
levels for the DPOAE recording setup was lower
than ⫺35 dB SPL in the frequency range of interest.
Calibration in a coupler can lead to variations in
stimulus levels at the tympanic membrane in individual ears. The possible effects of these variations
are included in the Discussion section.
Data Preparation for Analyses
DPOAE levels measured at f2 frequencies of 1000,
2000, 3000, 4000, 6000, and 8000 Hz were used in
these comparisons. These data were “cleaned,” using
a three-step algorithm in which all data points with
a noise floor higher than a preset value were rejected first. The rejection point was determined on a
case-by-case basis but was in the range of ⫺10 dB
SPL for the noisiest of our subjects. The intent was
to select as low of a cutoff as possible for a given ear,
thereby ensuring the preservation of DPOAE level
minima. Median values for DPOAE level and noise
floor were than calculated by using a three-point,
overlapping window. These median values were
then compared, and data points with SNR of less
than 3 dB were eliminated. Finally, the data were
passed through an interpolation algorithm to en-
EAR & HEARING, VOL. 25 NO. 6
577
Fig. 1. Comparison of raw and processed
data from the left ear of subject AC. A
suppressor tone was not used for this
recording. Data were processed by eliminating data points based on signal-tonoise ratio and absolute noise floor. Interpolation was performed to yield a
resolution of 2 Hz. Note “errant” points
in the raw data (open circles) are caused
by “spikes” in the noise floor (dashed
gray line). All such points are eliminated
in the processed data, whereas the overall fine structure is maintained (solid gray
line).
hance the original frequency resolution to a resolution of 2 Hz. This allowed exact matching between
the DPOAE and hearing threshold frequencies.
DPOAE level estimates after the initial LSF analysis are compared with data after the final stage of
interpolation in Figure 1 to demonstrate the validity of
the data processing protocol. The circles represent
initial estimates of DPOAE levels after the LSF analysis, with the noise floor being represented by the dashed
line. The continuous gray line represents DPOAE levels
after the completion of all processing. Note the artificially elevated data points caused by artifactual
noise levels have been eliminated in the processing without altering the general fine structure.
Fine-Structure Depth and Statistical Analysis
To quantify fine-structure depth and the effect
of suppression on fine-structure depth, peak-tovalley measurements of fine structure were made
in three fine-structure periods centered around
each hearing-threshold frequency. Fine-structure
depth (FSdepth) in each period was defined as
avg
FSdepth ⫽ 20 ⴱ log10共Pmax /Pmin
兲,
(1)
where Pmax is the DPOAE amplitude at a maximum
avg
and Pmin
is the average DPOAE amplitude of the
preceding and following minima. The nominal finestructure depth at each frequency was the computed
mean of fine-structure depth across the three periods.
The initial computation of fine-structure depth was
made for the no-suppressor condition. Similar estimates were also obtained for Ls ⫽ 65. Fine-structure
period boundaries from the no-suppressor condition
were used in cases of indistinguishable fine structure
for Ls ⫽ 65. Using a simplified model of three possible
effects of the suppressor on fine structure, we tabulated the number of occurrences in which fine-structure depth increased, decreased, or remained un-
changed when the “no-suppressor” condition was
compared with Ls ⫽ 65. The critical difference in
fine-structure depth between the two conditions at the
95% confidence level was computed at each test frequency by using a general linear regression model.
The computed critical values ranged between 4.10 and
4.12. A conservative limit of ⫾2 dB was used as the
benchmark to ascertain change in fine-structure
depth. Instances in which the difference between the
unsuppressed and suppressed conditions did not exceed ⫾2 dB were marked as “unchanged.” Cases in
which fine-structure depth in the unsuppressed condition was greater than 2 dB or smaller than ⫺2 dB for
the suppressed condition were marked as “decrease”
and “increase,” respectively.
Hearing thresholds were compared with DPOAE
levels obtained with various suppressor levels, using a
Pearson product-moment correlation test. In applying
a test of correlation to these data, we make two
assumptions. First, it is assumed that the data recorded from the two ears of a given subject are independent of each other. Second, it is also assumed that
one measured variable (i.e., DPOAE level) is normally
distributed about each value of the other variable
(hearing thresholds in this case). The validity of the
results of our correlation analysis would be compromised to the extent that these assumptions are violated in our data set. The SAS suite of statistical
software was used to perform these analyses.
RESULTS
General Results
The goal of this study was to evaluate the effectiveness of a universal suppressor tone in recording
“single-generator” DP grams. In keeping with this
goal, we present detailed results of the effect of
suppression on DPOAE fine structure in individual
ears as well as averaged results for the group.
578
EAR & HEARING / DECEMBER 2004
Results of a correlational analysis between behavioral hearing thresholds and DPOAE levels are also
presented to satisfy the secondary goal—that of
examining the effect of a suppressor on predictability of hearing thresholds. We begin by presenting
general group data for hearing thresholds as well as
DPOAE levels at the test frequencies.
Mean behavioral hearing thresholds measured in
2 dB steps along with mean DPOAE levels for
different suppressor levels are presented in Figure
2. The traces representing DPOAE level have been
offset along the abscissa to enhance visibility. Noteworthy is the large variance in DPOAE level relative
to the observed variance for hearing thresholds and
the relative insensitivity of this variance to the
presence or level of the suppressor tone.
Effects of Suppressors in Individual Ears
The average depth of level fine structure for all
subjects at each test frequency are presented in
Figure 3. Fine-structure depth for the unsuppressed
condition is represented by the squares; triangles
represent fine-structure depth for Ls ⫽ 65. Average
fine-structure depth is reduced as a result of the
suppressor at each test frequency. The magnitude of
variance is also reduced as a result of the suppressor. The effect of the suppressor was found to be
more varied when results from individual ears were
examined. These variations are reported next.
Figures 4 through 6 show the direct effect of
suppressor level on the level fine structure of individual subjects. Data for subject AC are displayed in
the three panels of Figure 4. The top panel shows
Fig. 2. Mean values of behavioral hearing thresholds and
distortion product otoacoustic emissions (DPOAE) levels for
different suppressor levels. Error bars represent ⴞ1 SD.
Fig. 3. Average fine-structure depth computed over three
periods around each test frequency. Computed fine-structure
depth for the unsuppressed condition (squares) along with Ls
ⴝ 65 (triangles) are presented. Error bars represent ⴞ1 SD.
DPOAE level as a function of frequency without a
suppressor (gray) and with a suppressor at 45 dB
SPL (black). DPOAE level functions for suppressors
of 45 (gray) and 55 dB SPL (black) are displayed in
Fig. 4. Effect of increasing suppressor level on distortion
product otoacoustic emissions (DPOAE) level for subject AC.
DPOAE level functions for two suppressor conditions are
displayed in each panel. Level functions without a suppressor
(gray) and with a suppressor at 45 dB SPL (black) are
displayed in the top panel. Those for suppressor levels of 45
(gray) and 55 dB SPL (black) are displayed in the middle
panel. Bottom panel shows DPOAE level functions for suppressor levels of 55 (gray) and 65 dB SPL (black). Arrow in
middle panel indicates an instance of deepening of fine
structure with the addition of a suppressor.
EAR & HEARING, VOL. 25 NO. 6
579
TABLE 1. Number of subjects in whom fine-structure depth
increased, decreased, or remained unchanged at each frequency
Hz
Increase
Decrease
Unchanged
1000
2000
3000
4000
6000
8000
2
1
0
3
2
3
10
13
11
9
9
10
8
6
9
8
9
7
A conservative critical value of ⫾2 dB at 95% confidence level in a linear regression model
was used to determine direction of change of fine-structure depth. See text for more
details.
Fig. 5. Effect of increasing suppressor level on distortion product
otoacoustic emission level for subject AP. Format is similar to
Figure 4. Subject AP exhibits no fine structure even without a
suppressor. Suppressor tone appears to minimally alter distortion product otoacoustic emission level in this subject.
the middle panel with those for 55 and 65 dB SPL in
the bottom panel. Each panel allows comparison of
DPOAE level between two suppressor conditions.
In general, an overall reduction in fine structure is
observed with increasing suppressor level. Although some “structure” can be observed even
with a suppressor of 65 dB SPL, this is unlike the
quasiperiodic fine structure observed when no
suppressor is present. Also noteworthy is the
increase in the depth of fine structure at frequencies just below 2000 Hz (indicated by an arrow in
the middle panel) for suppressor levels of 55 and
65 dB SPL when compared with the unsuppressed
(and Ls ⫽ 45) condition.
Results from subject AP are displayed in Figure 5.
The format of the figure is similar to that of Figure
4. In this case, fine structure is minimal even when
no suppressor is present, especially above 2000 Hz.
The suppressor does reduce fine structure at frequencies below 2000 Hz but has little impact on
DPOAE level at other frequencies.
DPOAE level for different suppressor levels for
subject KG are displayed in Figure 6. This subject
exhibits very periodic and pronounced fine structure
across the entire frequency range. Introduction of a
suppressor and increasing its level does reduce the
depth of fine structure at most frequencies, although
it is not eliminated even at the highest suppressor
level. The suppressor is most effective in reducing
fine structure at frequencies below 1500 Hz.
The results of categorization of the effects of the
suppressor on DPOAE fine structure are presented in
Table 1. The most common outcome was a reduction in
fine-structure depth. However, fine-structure depth
remained unchanged in as many cases at several test
frequencies. Fine-structure depth increased in as
many as 15% of cases at 4000 and 8000 Hz.
Hearing Thresholds and DPOAE Levels
Fig. 6. Effect of increasing suppressor level on distortion
product otoacoustic emission level for subject KG. Format is
similar to Figure 4. Suppressor appears to have minimal effect
even at the highest level on distortion product otoacoustic
emission fine structure in this subject.
The results of a correlational analysis between
behavioral hearing thresholds and DPOAE levels
are presented in Table 2 and Figure 7. Table 2 shows
the overall correlation for different suppressor levels
when the data are collapsed across frequency. Overall correlation coefficients (r) ranged from 0.12 to
0.03.
580
EAR & HEARING / DECEMBER 2004
TABLE 2. Pearson correlation coefficients between behavioral
hearing thresholds and distortion product otoacoustic emission
levels collapsed across frequency
Suppressor level
(dB SPL)
Correlation
r
No
45
55
65
0.1246
0.065
0.049
0.029
Figure 7 shows correlation coefficients as a function of suppressor level for each of the audiometric
frequencies. Correlation coefficients (r) ranged from
0.31 to ⫺0.63 when data at each frequency were
evaluated independently. The majority of frequencies exhibited negative correlation of DPOAE level
with behavioral threshold. An exception to this occurred at 4000 Hz, where coefficients were positive
at all suppressor levels.
DISCUSSION
The primary goal of this study was to examine the
feasibility of recording “single generator” DPOAEs
by using a suppressor tone. Specifically, we were
interested in determining if a universal suppressor
could be effective in a group of normal-hearing
individuals, the availability of a universal suppressor being crucial to the clinical viability of such a
protocol. An in-depth examination of the effects of a
suppressor tone on fine structure in individual ears
was conducted to satisfy this goal. We also examined
the effect of a suppressor on the correlation between
behaviorally measured hearing thresholds and
DPOAE level. Great variability was observed, both
in the effects of the suppressor on DPOAE fine
structure in individual ears and in correlation
coefficients.
Fig. 7. Pearson correlation coefficients between behavioral hearing thresholds and
distortion product otoacoustic emission
levels separated by frequency. Statistically
significant data points (p < 0.01) are filled
in.
Individual Variation in Unsuppressed Fine
Structure
Fine-structure depth in the unsuppressed condition varies both within subjects (across frequency)
and between subjects. Within subjects, there is a
tendency for fine structure to be more robust at low
frequencies, with the depth and periodicity of fine
structure tending to diminish or completely “wash
out” at frequencies above approximately 3000 Hz.
Figure 5 shows complete absence of fine structure in
the region above 3000 Hz and Figure 6 shows
reduction in fine-structure depth in that same region. Partial reduction or complete absence of fine
structure in this frequency region occurred in approximately 60% of the 20 ears evaluated. Although
this is a pattern we have frequently observed in the
lab, it does not occur in all subjects. Figure 8
contrasts two subjects, one with a pattern of finestructure depth that decreases at higher frequencies
and one with fine-structure depth that increases
with frequency.
The source of this variability in DPOAE fine
structure with frequency has not been addressed in
the literature. Previous studies on DPOAE finestructure characteristics have reported results from
very few subjects (e.g., Dhar et al., 2002; He &
Schmiedt, 1993; Talmadge et al., 1999) carefully
chosen with pronounced fine-structure characteristics. Although general observations such as finestructure depth as great as 20 dB (He & Schmiedt,
1993) can be found in the literature, a systematic
study of fine-structure characteristics in normalhearing individuals has not been done. The results
of this study show great variability in fine-structure
depth among a group of normal-hearing individuals.
This variability in of itself is curious and warrants
further examination. This variability also lends credence to the hypothesis that the presence of fine
EAR & HEARING, VOL. 25 NO. 6
Fig. 8. Contrasting fine-structure characteristics around 4000
Hz in subject AC and AP. Subject AP shows significant fine
structure when no suppressor is present (gray) at frequencies
below 3000 Hz. This fine structure is eliminated with a 65 dB
suppressor (black) in the top panel. In contrast, subject AC
show significant fine structure only around 4000 Hz without
a suppressor. Fine structure is duly eliminated by a suppressor
(bottom panel).
structure and its variability between individuals has
been a major limiting factor in the full realization of
the clinical potential of OAEs in general and
DPOAEs in particular (Shaffer et al., 2003; Shera,
2004).
Effects of a Suppressor Tone on Fine
Structure
The goal in using a suppressor tone proximal to
2f1-f2 is to eliminate the reflection component from
the ear canal DPOAE signal. Under these ideal
circumstances, the recorded DPOAE would arise
from a single source both mechanistically as well as
spatially, with the source being spatially close to the
basilar membrane location evaluated during puretone audiometry. The basic assumption in advocating the use of a suppressor tone is the possibility of
a “universal” suppressor level that would remain
effective across subjects and frequency. An effective
suppressor would be one sufficiently intense to eliminate the reflection component without altering the
nonlinear component. Our results suggest that such
a “universal” suppressor does not exist and that
there is tremendous variation in the effects of a
suppressor on DPOAE fine structure across individual ears as well as across frequency within an
individual ear. It should be noted that we have used
the presence of fine structure as the only metric to
determine the degree of suppression of the reflection
component. Presence of general structure due to
581
other factors such as variability in the nonlinear
mechanoelectrical transduction process is acknowledged. However, structure caused by mechanisms
other than a “mixing” of two components is differentiated from true fine structure by the characteristic
and consistent spacing across frequency. Given currently accepted OAE theory, it appears reasonable
to equate the presence of fine structure with the
involvement of more than one DPOAE component.
Group data show a general reduction in finestructure depth as well as its variability in each test
frequency (Figure 3). However, data from individual
ears illustrate three different effects that can occur
with the introduction of a suppressor tone. First, as
originally shown by Heitmann et al. (1998), a suppressor tone can diminish fine structure depth. This
effect is illustrated in Figure 4, in which the fine
structure depth is further reduced with each increase in suppressor level. This same figure, however, illustrates another effect. In the frequency
region just below 2000 Hz, fine-structure depth
increases from the unsuppressed condition for suppressor levels of 55 and 65 dB SPL. Second, minimal
fine structure is observed in the unsuppressed condition in Figure 5 above 2000 Hz. The suppressor
has no significant effect on DPOAE level in this case.
Finally, Figure 6 illustrates a case of apparent
ineffective suppression. For this subject, the introduction of a suppressor has minimal effect on the
depth of fine structure, even at the highest suppressor level.
The possible effects of a suppressor on DPOAE
fine structure are modeled in graphical form in
Figure 9. The top panel shows conditions under
which fine structure depth is expected to reduce in
response to a suppressor tone. In this case, the
nonlinear component is relatively larger than the
reflection component such that as the suppressor is
introduced and increased in level (light to dark
arrows), the reflection component is increasingly
attenuated (light to dark bars). This results in
significant reduction and finally elimination of fine
structure (light to dark traces). A similar pattern
was observed in our data in Figure 4. However, some
residual “structure” remains even at the highest
suppressor level in this case. Two possibilities exist
for the source of this residual structure. One source
of the residual structure could be incomplete suppression of the reflection component. The other
equally tenable explanation is that the magnitude of
the nonlinear DPOAE component changes as a function of frequency. If residual structure were eliminated
by a stronger suppressor tone, this would suggest a
case of incomplete suppression. However, a stronger
suppressor could affect the nonlinear component, reducing the overall level of the emission recorded in the
582
Fig. 9. Schematic of possible effects of a suppressor tone on
distortion product otoacoustic emission component magnitudes and fine structure. Vertical arrows represent peak
amplitudes of stimulus-induced traveling waves at the specified frequencies. Shaded boxes indicate component-source
strength (as measured in the ear canal). Peak amplitude and
component strength are displayed against spatial axes showing location within the cochlea. Fine-structure patterns are
displayed as a function of frequency. Top panel shows “ideal”
effect of a suppressor increasing in level, reducing, and finally
eliminating the reflection component and fine structure as a
result. Too strong a suppressor alters the nonlinear component, affecting overall distortion product otoacoustic emission level in the middle panel. Inadequate suppressor results
in residual fine structure (bottom panel). There is an intermediate stage of fine-structure enhancement in this case as well.
ear canal. The source of the residual structure in
Figure 4 is most likely a mixture of incomplete suppression, as well as variation in magnitude of the
nonlinear component. This is suggested by the observation of characteristic fine-structure spacing in a
narrow frequency range below 2000 Hz, whereas the
“structure” above 2000 Hz exhibits much wider periods suggestive of variation in the strength of the
nonlinear component itself.
The middle panel of Figure 9 illustrates a reflection component that is weaker in magnitude compared with the nonlinear component, resulting in
less pronounced fine structure in the unsuppressed
case. The introduction of a suppressor effectively
eliminates the reflection component and the fine
structure. Once again, further increase in the suppressor level could affect the nonlinear component
EAR & HEARING / DECEMBER 2004
causing a reduction in the overall magnitude of the
DPOAE recorded in the ear canal. This effect is
demonstrated to some extent in Figure 5 in our data
set, in which minimal fine structure is observed
below 2000 Hz only. The suppressor is very effective
in eliminating the weak fine structure below 2000
Hz but has no significant effect on the observed level
above that frequency. The suppressor could also
have no effect at all on fine structure in a given
ear—possibly due to two scenarios. First, the reflection component could be so large that the suppressor
does not alter it even at its highest level. This
appears to be unlikely. Second, the reflection component could be very small or absent in a given ear,
thereby eliminating any effect of the suppressor on
the signal recorded in the ear canal.
The bottom panel of Figure 9 illustrates a case in
which the unsuppressed reflection component is larger
in magnitude than the nonlinear component. Introduction of a suppressor initially reduces the magnitude of
the reflection component making it equal to that of the
nonlinear component. Equalization of the two DPOAE
components, however, results in “deepening” of fine
structure. Indeed, this effect has been demonstrated in
a previous report (Talmadge et al., 1999) as well as in
the frequency range below 2000 Hz in Figure 4 in our
data (indicated by an arrow in the middle panel).
Further increase in the suppressor level attenuates
the reflection component to a greater extent but is
unable to eliminate it completely. This results in
residual fine structure even at the highest suppressor
level, as demonstrated in Figure 6 in our data set. The
source of the residual fine structure is not in question
here, as its distinct periodicity is maintained for the
unsuppressed as well as all the suppressor conditions.
A variety of effects of a suppressor tone on DPOAE
fine structure can be predicted even from a simplified
two-source model based on the relative magnitudes of
DPOAE components. We have demonstrated the main
possibilities in our data set and the reader is directed
to Talmadge et al. (1999) and Kalluri and Shera (2001)
for a more formal presentation of the model and the
errors that can result in attempting to isolate the
source components. A count of each of these effects at
each test frequency has been presented in Table 1. It is
evident that fine-structure depth decreased in a majority of the ears at most test frequencies. However,
the impracticality of a universal suppressor is demonstrated in fine structure remaining unchanged in a
large number of ears and fine-structure depth actually
increasing in as many as 30% of the ears at two test
frequencies. One could argue that a suppressor of
greater magnitude would reduce the number of cases
in which fine-structure depth remains unchanged.
However, such increased suppressor levels would alter
the level of the nonlinear component in the ears in
EAR & HEARING, VOL. 25 NO. 6
which the reflection component was already substantially suppressed or eliminated altogether.
Although our data present a complex picture of the
effects of a suppressor on the reflection component of
DPOAEs, another avenue to examining this phenomenon could be through suppression studies of stimulus
frequency otoacoustic emissions (SFOAEs). The reflection component of DPOAEs and SFOAEs are thought
to be generated by the same mechanism (Shera &
Guinan, 1999); the similarity between these two emissions has been demonstrated before (Kalluri & Shera,
2001). Unfortunately, suppression characteristics of
SFOAEs have not been studied as extensively as those
of other emission types. However, previous work in
this area presents a complex picture of the effects of a
suppressor on SFOAEs—similar to the one presented
here. Suppressors at frequencies higher than the stimulus frequency appear to be more effective for low-level
suppressors. In contrast, low-frequency suppressors
are more effective for suppressors at higher levels
(Kemp, Brass, & Souter, 1991). Distant frequencies
also appear to play a role in suppression of SFOAEs, as
evidenced through the results of using narrow-band
noise for suppression (Souter, 1995). When using puretone suppressors, Brass and Kemp (1993) found suppressor tones higher than the stimulus tone by 12 dB
to completely eliminate SFOAEs in some cases. Establishing an estimate of the “stimulus level” is, however,
nontrivial in case of the reflection component of
DPOAEs. This essentially would be the apically traveling component from the overlap region between the
traveling wave patterns of the stimulus tones. All
factors determining the magnitude of this component
remain under study at this time.
Hearing Thresholds and DPOAE Level
When the data were collapsed across frequency
yielding an overall correlation for each suppressor
level, behavioral thresholds and DPOAE level were
positively correlated but were not statistically significant. Even so, the correlation coefficient for the
unsuppressed condition fell within the range of
values observed in previous studies (e.g., Gaskill &
Brown, 1993) and generally supports the view that
DPOAE level, as measured clinically, is not a good
predictor of hearing thresholds. If the presence of
fine structure is truly a confounding factor in obtaining good correlation between hearing thresholds and
DPOAE level, then introduction of a suppressor tone
should improve correlation coefficients. This is
clearly not observed in the overall correlation data.
In fact, correlation coefficients decreased (approaching zero) with increasing suppressor level. The direction of change in the correlation with increasing
suppressor level was not completely unanticipated.
583
We expected DPOAE level and hearing thresholds to
be negatively correlated, assuming that high
DPOAE level is an indicator of good hearing sensitivity (and hence lower behavioral thresholds) and
to become increasingly more negative as fine structure is suppressed. From this perspective, the
change in the overall correlations with increasing
suppressor level was in the expected direction.
Correlation between hearing thresholds and
DPOAE level was also examined independently at
audiometric test frequencies. The range of variability
in correlation coefficients observed across frequencies
clearly explains the lack of statistical significance in
overall correlation when frequency was collapsed. Of
the six frequencies analyzed, five showed predominately negative correlation of hearing threshold and
DPOAE level with statistically significant correlations
occurring for 1000 and 8000 Hz. At 4000 Hz, correlation coefficients were positive for all conditions. This is
especially curious, given the traditionally strong correlation reported at this frequency in the literature.
Although we cannot fully explain these results at 4000
Hz, the difference between these data and those reported previously could be the fact that only normalhearing ears were included here. If indeed variability
in fine structure is greatest at 4000 Hz for normalhearing ears, correlation between hearing thresholds
and DPOAE levels could be adversely affected. On the
other hand, if minimal fine structure is observed at
this frequency with even small degrees of hearing loss,
including these ears in a data set would lead to better
correlation. Two frequencies (6000 Hz and 8000 Hz)
exhibited increasingly negative coefficients with increasing suppressor level as initially expected. In general, these findings do not support the idea that
introduction of a suppressor tone will improve correlation of hearing thresholds and DPOAE level.
CONCLUSIONS
Results of the correlational analysis between hearing thresholds and DPOAE levels showed great variability across frequency. While the results at 1000 and
3000 Hz showed correlation (between ⫺0.55 and
⫺0.63) comparable to the best of previous reports, the
correlation was dismal at best for other frequencies
(e.g., 4000 Hz). When examining the overall results, it
is evident that better correlation between hearing
thresholds and DPOAE levels has been reported in
several studies even without the use of a suppressor
(e.g., Dorn et al., 1998, 1999; Gaskill & Brown, 1993;
Gorga et al., 2000). Although a few of the above studies
used multiple measures (e.g., multiple orders of
DPOAEs or multiple measures from a single DPOAE
order), the most important difference between these
studies and the current one is perhaps in the sample
584
size. Although large variations in DPOAE level are
observed in individual ears, averaging data across a
large number of ears essentially low-pass–filters the
data. The effect of natural level variation is still
evident when the level ranges of these data sets are
considered. Such larger variance leads to significant
overlap between normal-hearing and hearing-impaired distributions (Gorga, Neely, Ohlrich, Hoover,
Redner & Peters, 1997). This effect is evident in our
data set in Figure 2, where different suppressor levels
appear to make negligible difference in the averaged
DPOAE levels. The correlation values reported here
would predictably improve with the use of a larger
sample size. However, the problem of comparing individual data with such a normative benchmark would
remain.
It is evident from these results that obtaining a
single-generator DP gram by using a suppressor
tone does not universally lead to better correlation
between hearing thresholds and DPOAE level. Although the use of a suppressor tone in recording a
single-generator DP gram is attractive at the outset,
the effects turn out to be rather complex, being both
varied and largely unpredictable. Such variability of
the effects of a suppressor on DPOAE fine structure
limit development of a clinically viable protocol.
It is perhaps also worthwhile to question the limits
of DPOAEs as a predictor of hearing thresholds.
DPOAEs are without a doubt a good indicator of outer
hair cell status in a given cochlea. A behavioral hearing loss however can encompass much more than
outer hair cell damage. It has previously been demonstrated using psycho-physical estimates of auditory
filter shapes, growth of forward masking, and loudness functions (all presumably measures of outer hair
cell integrity) that these measures, although highly
correlated between themselves, are not very highly
correlated to audiometric thresholds (Moore, Vickers,
Plack, & Oxenham, 1999). Additionally, hearing
thresholds and DPOAE levels are typically obtained at
dramatically different input levels to the cochlea.
Given the input-dependent nature of the cochlea’s
response to stimulation, DPOAEs as measured for the
DP gram and hearing thresholds might be evaluating
different aspects of cochlear function. Finally, the
efferent system may actively attenuate the response of
the cochlea during DPOAE measurements but not
during hearing threshold estimation (Guinan, Backus,
Lilaonitkul, & Aharonson, 2003). Both DPOAEs and
hearing thresholds are known to demonstrate fine
structure or microstructure, and these patterns are
not matched in frequency in a given ear. Therefore,
correlation between these two measures can only be
expected at frequencies where their structures overlap
(Mauermann, Relationship between hearing thresholds and DPOAE levels, personal communication,
EAR & HEARING / DECEMBER 2004
2004). Given these critical differences in cochlear response during hearing threshold and DP gram measurement, it might be prudent to re-examine DPOAE
recording strategies as well as the role of DPOAEs in
the scheme of audiological diagnostics.
The results from this study exhibit some peculiarities that need to be noted. First, we observed
hearing thresholds at 8000 Hz that were significantly better than those recorded at adjacent frequencies such as 3000, 4000, and 6000 Hz. We
calibrated our equipment with due diligence before
the initiation of these experiments and rechecked
the calibration once we noticed this trend in hearing
thresholds. No errors in calibration were observed.
As mentioned before, the results of the correlation
analysis at 4000 Hz were contrary to previously
reported results. We are unable to fully explain this
observation. Errors caused by our calibration technique (iso-voltage) would be predicted to be greatest
at frequencies around 3000 and 4000 Hz. It is
unknown whether errors in calibration of signal
levels at the tympanic membrane contributed to
these results. These peculiarities notwithstanding,
we have clearly demonstrated the ineffectiveness of
using a universal suppressor in recording “single
generator” DPOAEs in normal-hearing individuals.
Our results also appear to indicate little improvement in correlation between hearing thresholds and
DPOAE levels on using a suppressor tone.
ACKNOWLEDGMENTS
The authors wish to thank Ms. Bernadette Milone and Elizabeth
Thompson for their assistance in data recording and analysis. This
research was supported by NIH/NIDCD grant R03 DC005692–
01A1 to S.D. and a grant from the Deafness Research Foundation
(S.D. and L.A.S.). Our thanks to Professors Arnold Tubis and Robert
Withnell for useful comments on earlier drafts of the paper. The
ideas presented here have been heavily influenced by numerous
discussions with our mentors Professors Glenis Long, Carrick Talmadge, and Arnold Tubis. Finally, the paper was much improved
due to the input received from Drs. Michael P. Gorga, Manfred
Mauermann, and two other anonymous reviewers.
Address for correspondence: Sumit Dhar, Department of Speech
and Hearing Sciences, Indiana University, Bloomington, IN
47405.
Received February 23, 2004;. accepted July 25, 2004.
REFERENCES
Boege, P., & Janssen, T. (2002). Pure-tone threshold estimation
from extrapolated distortion product otoacoustic emission I/Ofunctions in normal and cochlear hearing loss ears. Journal of
the Acoustical Society of America, 111, 1810 –1818.
Brass, D. N., & Kemp, D. T. (1993). Suppression of stimulus
frequency otoacoustic emissions. Journal of the Acoustical
Society of America, 93, 920 –939.
Dhar, S., Talmadge, C., Long, G., & Tubis, A. (2002). Multiple
internal reflections in the cochlea and their effect on DPOAE
EAR & HEARING, VOL. 25 NO. 6
fine structure. Journal of the Acoustical Society of America,
112, 2882–2897.
Dorn, P. A., Piskorski, P., Gorga, M. P., Neely, S. T., & Keefe,
D. H. (1999). Predicting audiometric status from distortion
product otoacoustic emissions using multivariate analysis. Ear
Hear, 20, 149 –163.
Dorn, P. A., Piskorski, P., Keefe, D. H., Neely, S. T., & Gorga,
M. P. (1998). On the existence of an age/threshold/frequency
interaction in distortion product otoacoustic emissions. Journal of the Acoustical Society of America, 104, 964 –970.
Gaskill, S. A., & Brown, A. M. (1993). Comparing the level of the
acoustic distortion product 2f1-f2 with behavioural threshold
audiograms from normal-hearing and hearing-impaired ears.
British Journal of Audiology, 27, 397– 407.
Gaskill, S. A., & Brown, A. M. (1996). Suppression of human
acoustic distortion product: dual origin of 2f1-f2. Journal of the
Acoustical Society of America, 100, 3268 –3274.
Gorga, M. P., Neely, S. T., Bergman, B. M., Beauchaine, K. L.,
Kaminski, J. R., & Liu, Z. (1994). Towards understanding the
limits of distortion product otoacoustic emission measurements.
Journal of the Acoustical Society of America, 96, 1494 –1500.
Gorga, M., Neely, S., Dorn, P., & Hoover, B. (2003). Further
efforts to predict pure-tone thresholds from distortion product
otoacoustic emission input/output functions. Journal of the
Acoustical Society of America, 113, 3275–3284.
Gorga, M. P., Neely, S. T., Ohlrich, B., Hoover, B., Redner, J., & Peters,
J. (1997). From the laboratory to clinic: A large scale study of
distortion product otoacoustic emissions in ears with normal hearing
and ears with hearing loss. Ear and Hearing, 18, 440–455.
Gorga, M. P., Nelson, K., Davis, T., Dorn, P. A., & Neely, S. T. (2000).
Distortion product otoacoustic emission test performance when
both 2f1-f2 and 2f2-f1 are used to predict auditory status. Journal
of the Acoustical Society of America, 107, 2128 –2135.
Greenwood, D. D. (1990). A cochlear frequency position function
for several species—29 years later. Journal of the Acoustical
Society of America, 87, 2592–2605.
Guinan, J. J., Backus, B. C., Lilaonitkul, W., & Aharonson, V.
(2003). Medial olivocochlear efferent reflex in humans: Otoacoustic emission (OAE) measurement issues and the advantages of stimulus frequency OAEs. Journal of the Association
for Research in Otolaryngology, 4, 521–540.
He, N.-J., & Schmiedt, R. A. (1993). Fine structure of the 2f1-f2
acoustic distortion product: Changes with primary level. Journal of the Acoustical Society of America, 94, 2659 –2669.
Heitmann, J., Waldmann, B., & Plinkert, P. K. (1996). Limitations in the use of distortion product otoacoustic emissions in
objective audiometry as the result of fine structure. European
Archives of Otolaryngology, 253, 167–171.
Heitmann, J., Waldmann, B., Schnitzler, H. U., Plinkert, P. K., &
Zenner, H. P. (1998). Suppression of distortion product otoacoustic emissions (DPOAE) near 2f1-f2 removes DP-gram fine
structure— evidence for a secondary generator. Journal of the
Acoustical Society of America, 103, 1527–1531.
Kalluri, R., & Shera, C. A. (2001). Distortion-product source
unmixing: A test of the two-mechanism model for DPOAE
generation. Journal of the Acoustical Society of America, 109,
622– 637.
Kemp, D. T., Brass, D. N., & Souter, M. (1991). Observations on
simultaneous SFOAE and DPOAE generation and suppression. 202–209.
Kim, D. O., Paparello, J., Jung, M. D., Smurzynski, J., & Sun, X.
(1996). Distortion product otoacoustic emission test of sensorineural hearing loss: performance reharding sensitivity, speci-
585
ficity and receiver operating characteristics. Acta Otolaryngologica, 116, 3–11.
Kimberley, B. P., Hernadi, I., Lee, A. M., & Brown, D. K. (1994).
Predicting pure tone thresholds in normal and hearing impaired ears with distortion product emission and age. Ear and
Hearing, 98, 197–210.
Konrad-Martin, D., Neely, S. T., Keefe, D. H., Dorn, P. A., &
Gorga, M. P. (2001). Sources of distortion product otoacoustic
emissions revealed by suppression experiments and inverse
fast Fourier transforms in normal ears. Journal of the Acoustical Society of America, 109, 2862–2879.
Kummer, P., Janssen, T., & Arnold, W. (1998). The level and
growth behavior of the 2f1-f2 distortion product otoacoustic
emission and its relationship to auditory sensitivity in normal
hearing and cochlear hearing loss. Journal of the Acoustical
Society of America, 103, 3431–3444.
Long, G. R., & Talmadge, C. L. (1997). Spontaneous otoacoustic
emission frequency is modulated by heartbeat. Journal of the
Acoustical Society of America, 102, 2831–2848.
Mauermann, M., Uppenkamp, S., Hengel, P. W. J. van, & Kollmeier,
B. (1999). Evidence for the distortion product frequency place as a
source of distortion product otoacoustic emission (DPOAE) fine
structure in humans. I. Fine structure and higher-order DPOAE
as a function of the frequency ratio f2/f1. Journal of the Acoustical
Society of America, 106, 3473–3483.
Moore, B. C. J., Vickers, D. A., Plack, C. J., & Oxenham, A. J. (1999).
Inter-relationship between different psychoacoustic measures assumed to be related to the cochlear active mechanism. Journal of
the Acoustical Society of America, 106, 2761–2778.
Nelson, D. A., & Kimberley, B. P. (1992). Distortion-product
emissions and auditory sensitivity in human ears with normal
hearing and cochlear hearing loss. Journal of Speech-Language
Hearing Research, 35, 1142–1159.
Plinkert, P., Heitmann, J., & Waldmann, B. (1997). Single generator distortion products (sgDPOAE). precise measurements
of distortion product otoacoustic emissions by three tone stimulations. HNO, 45, 909 –914.
Shaffer, L. A., Withnell, R. H., Dhar, S., Lilly, D. J., Goodman,
S. S., & Harmon, K. M. (2003). Sources and mechanisms of
DPOAE generation: Implications for the prediction of auditory
sensitivity. Ear and Hearing, 24, 367–379.
Shera, C. A. (2004). Mechanisms of mammalian otoacoustic
emission and their implications for the clinical utility of otoacoustic emissions. Ear and Hearing, 25, 86 –97.
Shera, C. A., & Guinan, J. J. (1999). Evoked otoacoustic emissions
arise by two fundamentally different mechanisms: Ataxonomy
for mammalian oaes. Journal of the Acoustical Society of
America, 105, 782–798.
Souter, M. (1995). Suppression of stimulus frequency otoacoustic
emissions by contralateral noise. Hearing Ressearch, 91, 167–177.
Talmadge, C., Long, G. R., Tubis, A., & Dhar, S. (1999). Experimental confirmation of the two-source interference model for
the fine structure of distortion product otoacoustic emissions.
Journal of the Acoustical Society of America, 105, 275–292.
Talmadge, C., Tubis, A., Long, G. R., & Piskorski, P. (1998).
Modeling otoacoustic emission and hearing threshold fine
structures in humans. Journal of the Acoustical Society of
America, 104, 1517–1543.
Waldmann, B., Heitmann, J., & Plinkert, P. (1997). Distorsionsproducte (sgDPOAE): Entwicklung eines neuen prazisionsme
systems. Audiologische Akustic, 1, 22–31.
Zweig, G., & Shera, C. A. (1995). The origins of periodicity in the
spectrum of evoked otoacoustic emissions. Journal of the
Acoustical Society of America, 98, 2018 –2047.