Noise Health Home 

[Download PDF]
Year : 2013  |  Volume : 15  |  Issue : 66  |  Page : 315--325

Distortion product otoacoustic emission level maps from normal and noise-damaged cochleae

Deanna K Meinke1, Odile H Clavier2, Jesse Norris2, Robert Kline-Schoder2, Lindsay Allen2, Jay C Buckey3,  
1 University of Northern Colorado, Audiology and Speech-Language Sciences, Greeley, Colorado 80639
2 Creare, Inc., Hanover, New Hampshire 03755
3 Geisel School of Medicine at Dartmouth, Lebanon, New Hampshire 03756

Correspondence Address:
Deanna K Meinke
UNC, Audiology and Speech Language Sciences, Campus Box 140, Gunter Hall 1500, Greeley, Colorado 80639


Distortion product otoacoustic emission (DPOAE) level mapping may be useful for detecting noise-induced hearing loss (NIHL) early. Employing DPOAE mapping effectively requires knowledge of the optimal mapping parameters to use for detecting noise-induced changes. The goal of this project was to show the map regions that differ most between normal and noise-damaged cochlea to determine the optimal mapping parameters for detecting NIHL. DPOAE level maps were generated for the 2f 1 -f 2 and the 2f 2 -f 1 DPOAEs for 17 normal hearing male subjects and 19 male subjects with NIHL. DPOAEs were measured in DPOAE frequency steps of approximately 44 Hz from 0.5 kHz to 6 kHz using constant f 2 /f 1 ratios incremented in 0.025 steps from 1.025 to 1.5 using both unequal-level (L1,L2 = 65,55 dB sound pressure level (SPL)) and equi-level (L1,L2 = 75,75 dB SPL) stimulus paradigms. Maximal responses for the 2f 2 -f 1 emission at L1,L2 = 65,55 dB SPL were found at lower ratios compared to previous studies. The map regions where NIHL eliminated or reduced DPOAE magnitude were identified. DPOAE level mapping using higher-level, equi-level primaries produced significantly more detectable emissions particularly for the 2f 2 -f 1 emission. The data from this study can be used to optimize DPOAE level mapping parameters for tracking noise-exposed subjects longitudinally.

How to cite this article:
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;15:315-325

How to cite this URL:
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 [serial online] 2013 [cited 2023 Jan 29 ];15:315-325
Available from:

Full Text


The ability to detect noise-induced hearing loss (NIHL) early and implement preventive strategies would reduce the social and economic burdens attributed to this disorder. Distortion product otoacoustic emission (DPOAE) level mapping is a potential new approach to detect NIHL and monitor individuals for early signs of cochlear damage from hazardous noise exposure over time. The Knight and Kemp, [1],[2] DPOAE level and phase mapping approach has been used in rabbits to display DPOAE data with normal cochlear status and after noise exposure. [3] These researchers suggested that DPOAE level/phase mapping may provide a sensitive indicator of cochlear damage due to evidence of a strong DPOAE place-fixed component in rabbits with normal cochlear function and in rabbits with NIHL where DPOAEs exist. To date, there are limited human data using the Knight and Kemp, [1],[2] DPOAE level mapping approach in normal and noise-damaged cochleae, perhaps due to the time-intensive nature of the testing paradigm and the customized testing protocols needed to collect comprehensive data sets.

Time-intensive DPOAE level mapping has been utilized to investigate differences in DPOAEs for 20 male subjects with normal hearing and 10 subjects with NIHL. [4] The experimental paradigm used an f 2 /f 1 fixed-ratio approach incremented in 0.025 steps between 1.025 and 1.5 in response to three equi-level primary tone sweeps (L1,L2 = 80,80; 75,75 and 65,65 dB SPL). The extended DPOAE frequency range encompassed 0.5-6.0 kHz in DPOAE frequency steps of 44 Hz. Both wave- and place-fixed emissions were evident for the 2f 1 -f 2 emission and place-fixed emissions were evident for the 2f 2 -f 1 emission in both normal and NIHL subjects with measurable emissions. The outcomes did not support the hypothesis that an increase in place-fixed DPOAEs may occur due to potential changes in cochlear structural irregularities caused by hazardous noise exposure as evidenced in the prior rabbit study. However, in this initial study, DPOAE levels differed for the NIHL group as compared to the normal hearing group in regions that extended beyond the conventional 2f 1 -f 2 optimal f 2 /f 1 ratio of 1.2. These initial data suggested that DPOAE level mapping might provide a more comprehensive evaluation of cochlear function that could be useful in the assessment of NIHL. Investigating this potential application was severely limited by the duration of the initial mapping protocol (45 min/DPOAE level map), which required a high degree of cooperation and tolerance from the subjects. More efficient and flexible instrumentation, test protocols, and analysis techniques were needed.

New technology developed for the present study allows for comprehensive maps to be completed within a clinically reasonable time (maps can be completed in times ranging from 5 min to 25 min depending on the number of points desired), but the optimal parameters to use to collect mapping data for the early detection of NIHL are not known. The objective of this study was to determine the DPOAE frequencies and ratios on a DPOAE level map most affected by noise exposure, and to use this information to develop DPOAE level mapping protocols designed for the early detection and monitoring of cochlear damage from hazardous noise exposure. To accomplish this, mapping data from 17 normal hearing male subjects were compared to data from 19 men with NIHL. We hypothesized that some map regions would show minimal differences and could be omitted from future mapping protocols and that higher-level primaries (e.g., L1,L2 = 75,75) would produce more detectable emissions in the NIHL group. We also hypothesized that the technology developed for this study, combined with the data from the normal and noise-exposed study groups, would yield mapping protocols to assess NIHL that could be completed within a clinically acceptable amount of time.



DPOAE level maps were recorded from male subjects aged 18 to 50 years from the University of Northern Colorado geographical community, which includes veteran/military service members. Inclusion criteria for all subjects included a negative otologic history, no history of neurological disorders, normal otoscopy and normal middle ear status (based upon normative values for static admittance, equivalent ear canal volume and tympanometric peak pressure between −150 daPa to +50 daPa). Additional subject inclusion criteria were dependent upon the experimental grouping; either "normal hearing non-noise exposed" (NHNN) or "NIHL noise exposed." (NIHL) Pure-tone air-conduction thresholds were used to classify the subjects into each experimental group: (1) NHNN where thresholds were ≤20 dBHL for 250-8000 Hz bilaterally and (2) NIHL defined as thresholds ≤20 dBHL at 250-500 Hz and ≤60 dBHL for 1000-6000 Hz. Only one ear was used for the study and right versus left ear was randomized at the time of enrollment for the NHNN group and in cases of bilateral NIHL.

Noise exposure status was determined by an extensive occupational and non-occupational noise exposure case history survey and interview. Subjects classified as NHNN reported either no past or present noise exposures or only rare exposures in the past. Subjects classified as NIHL reported routine and consistent participation in noise hazardous activities in the past and in some cases, ongoing noise exposure at the time of the study enrollment. The use of hearing protection during noise exposed activities varied across subjects and activities, and was typically used either inconsistently or not at all by the majority of NIHL subjects. Trained audiology graduate research assistants and the first author obtained consent, and conducted the case history interviews and experimental testing. The study was approved by the University of Northern Colorado Institutional Review Board.

Audiometry and acoustic-immittance testing

Pure-tone audiograms were obtained manually at 250, 500, 1000, 2000, 3000, 4000, 6000 and 8000 Hz using a Hughson-Westlake procedure performed with an acoustically calibrated Grason Stadler GSI 16 audiometer and TDH-50P supra-aural earphones. [5] Testing was completed in a Tracoustics single-walled sound booth meeting ANSI S3.1-1999 (R2008), [6] permissible ambient noise levels.

Normal tympanometry patterns (Type A) and pressure equalization (−150 to +50 daPa) were verified for the test ear prior to conducting DPOAE measurements using routine acoustic-immittance measurements with a Maico EroScan Pro (i.e., the 226-Hz tympanogram).

 Distortion Product Otoacoustic Emission Level Mapping

DPOAE level maps were collected using the hardware and software developed specifically for this research project in collaboration with Creare Inc., (Hanover, NH). The prototype hearing assessment system specifications are provided in Appendix A.[SUPPORTING:1]

Experimental DPOAE level maps were collected with the human subjects awake and seated in a recliner within the single-walled sound booth after receiving instructions to remain as quiet and still as possible. Subjects were permitted to read or operate electronic games/devices with the audio disabled during the testing. All data were collected during a single 1.5 h experimental session. Rest periods were offered between the two mapping level protocols; however, no subjects requested removal of the probe or a change of position necessitating probe removal and reinsertion. DPOAE probes were sealed in the ear canal using the standard-sized rubberized tube-phone ear tips.

The researchers subjectively determined an adequate seal in the ear canal by visualization, tug-test for resistance, and subjective feedback from the subject. To assist with proper probe placement, a frequency sweep (chirp) was presented in the ear canal before the DPOAE testing. The results from three chirps (500 Hz to 5000 Hz) at 65 dB SPL were averaged, smoothed, and displayed to the research assistant. In the case of a poor seal, the probe was reseated and the chirps were repeated. Every effort was made to assure an adequate first seal and the final acoustic sweep test was stored as a baseline reference for future DPOAE level-mapping testing. If the probe did need to be reinserted, the baseline acoustic frequency-sweep spectrum could be used as a reference and the probe could be adjusted until the sweeps were consistent with baseline measures. For this cross-sectional study, however, the probes did not need to be reseated and baseline acoustic frequency-sweeps were consistent between mapping primary tone stimulus levels (L1,L2 = 65,65 and 75,75).

DPOAEs were measured in DPOAE frequency steps of approximately 44 Hz for both the 2f 1 -f 2 and 2f 2 -f 1 emissions from 0.5 kHz to 6 kHz using constant f 2 /f 1 ratios incremented in 0.025 steps from 1.025 to 1.5. Over this parameter space, f 2 ranged from 0.258 kHz to 18.023 kHz while f 1 ranged from 0.388 kHz to 12.016 kHz. This produced a map composed of 5160 points. Using a total of 4 averages per point, each map required 25 min to complete. DPOAE level maps were collected for two primary tone levels: 65,55 dB SPL (consistent with conventional DP-Grams), and 75,75 dB SPL. There was no evidence of artifactual distortion at levels above −20 dB SPL when these high-level primary stimuli were measured in an artificial ear (Brüel and Kjaer Model 4157, Nærum, Denmark).

[Figure 1] provides a general description of the plotting to orient the reader to the data presentation. The 2f 1 -f 2 and 2f 2 -f 1 DPOAEs were shown within the same area plot, consistent with the Knight and Kemp, [1] and Martin, [3] data presentations. The 2f 1 -f 2 DPOAE was plotted in the upper half of the panel and was contiguous with the 2f 2 -f 1 emission in the lower half of the same panel frame [Figure 1]. Emission levels were mapped (z-axis) by aligning vertically on the y-axis the DP (distortion product) frequency (~500 Hz to 6000 Hz) and in ascending ratio order on the x-axis for both the 2f 1 -f 2 and 2f 2 -f 1 DPOAEs [Figure 1]a. This 3-dimensional approach (DP Frequency, DP Level and f 2 /f 1 ratio) was then color coded for DPOAE level [Figure 1]b in 0.8 dB SPL color gradients and converted to a 2-dimensional plot [Figure 1]c. DPOAE level data were plotted using warmer colors (reds, yellows) for high-level emissions and cooler colors (light green, light blue) for low-level emissions. DPOAEs that most likely reflect noise floor (NF) values were plotted as a dark blue color. The color range was set from −20 dB to 30 dB SPL. Values in the map either lower or higher than these bounds were displayed using the value for the lower or upper bound limit respectively.{Figure 1}

When the DP level was plotted as in [Figure 1]c, the map tended to show a concentration of higher level DPs at the lower DP frequencies, which were also contaminated by higher NF levels. To illustrate this more clearly, the data were also presented as a signal-to-noise ratio (SNR) map. In this case, the DP minus NF was plotted, rather than just the DP level [Figure 1]d. The color range was set from 0 dB to 30 dB for SNR and values either lower or higher than these bounds were displayed using the value for the lower or upper bound limit respectively.

The trajectory of the more commonly referenced f 2 frequency is provided in [Figure 1]c for 3 kHz and 6 kHz, since this is a region typically damaged by hazardous noise exposure. These plots span a larger frequency space and utilize a larger frequency increment (44 Hz versus 12 Hz) when compared with the Knight and Kemp's, [2] report's figures.

Statistical analysis

Each map was a matrix of 129 frequencies and 40 ratios (5160 DPOAEs). To create average maps, the individual DP levels (for a level map) and the individual DP minus NF values (for an SNR map) were averaged point-by-point across subjects within an experimental group. For the average SNR maps, the percentage of overall points greater than or equal to 3 dB SNR, 6 dB SNR and 9 dB SNR were calculated for each group.

The percentage of points at or above the 3 cut-off values (≥3 dB, ≥6 dB and ≥9 dB) were compared between the NHNN and NIHL groups using the non-parametric statistics (Kruskal-Wallis test). To compare the average NHNN and NIHL maps, the maps were first median filtered, and then compared point-by-point using a one-way ANOVA with an alpha of 0.05. A compensation for multiple comparisons was not used, since the objective of the point-by-point comparison was to outline areas on the map with the highest likelihood of showing significantly different DPOAE levels between the NHNN and NIHL groups. The goal was to create a map of the points where the two average SNR maps differed significantly, with the understanding that some of these points might represent type I errors. This overall approach provides a way to define the areas of the map most likely to show differences between a NHNN and NIHL group, which is needed to provide the greatest opportunity to detect subtle early changes in mapping for noise-exposed individuals.



A total of 36 male subjects participated in the study (NNHN n = 17; NIHL n = 19). The mean age of the NHNN group was 28.2 years ± 8.8 years. The NIHL group was slightly older with a mean age of 35.7 years ± 9.3 years. Unpaired Student's t-testing reveals a statistically significant age difference between the two groups using an alpha of 0.05 (P = 0.018).

Pure-tone audiometry

All subjects had measureable thresholds between 250 Hz and 8000 Hz. Average audiometric thresholds for test ears in NHNN and NIHL groups are provided in [Figure 2]. The NHNN group was comprised of 53% right and 47% left ears while the NIHL group was comprised of 37% right and 63% left ears. The average thresholds from 250 Hz to 8000 Hz for the NIHL group were significantly poorer than the NHNN group (P < 0.01).{Figure 2}

DPOAE equal-ratio level and signal-to-noise ratio maps

Qualitative observations: General

[Figure 3] provides a single subject example of a level and SNR map for a NHNN subject. [Figure 4] provides an example for a NIHL subject. The maps for the individual subjects were unique, but on average patterns of DPOAE response were evident. [Figure 5] shows the average level and SNR maps for the NHNN and NIHL groups.{Figure 3}{Figure 4}{Figure 5}

A qualitative review of the DPOAE level maps revealed obvious visual differences between the experimental groups. The NHNN group had more high-level DPOAEs, a wider band of ratios in the 2f 1 -f 2 area where DPOAEs were detected, and a larger area where 2f 2 -f 1 DPOAEs were detected compared to the NIHL group.

Qualitative observations: Amplitude features

The DPOAE level data are consistent with a band of maximal response centered at the f 2 /f 1 frequency ratio of 1.2 across DP frequencies for all subjects at both L1,L2 = 65,55 and 75,75. This is consistent with the findings of Harris et al., at L1,L2 = 65,55. [7]

The 2f 2 -f 1 emissions were present over a smaller area of the map as illustrated in [Figure 5]. For the NHNN group, the optimal 2f 2 -f 1 emission for 65,55 level maps occurred at an f 2 /f 1 ratio of 1.025 and for the 75,75 map, a ratio of 1.075.

At both stimulus levels, there was visual evidence of a low-frequency color band (reds) or vertical stripe at the lowest DP frequencies [Figure 5]a, c, e and g, which was not apparent in the SNR map. This color band did not occur when the system was tested using a cavity indicating that low frequency physiologic noise was the likely source of the response observed in the level maps. The SNR maps further suggest that the low-frequency response is dominated by noise.

Quantitative observations: Amplitude maps

For the NHNN subjects at L1,L2 = 65,55 dB SPL, the highest DPOAE level response occurs near the f 2 /f 1 ratio of 1.275 for the lowest DP frequencies (<2000 Hz) but occurs near a ratio of 1.15 for the highest DP frequencies (>5000 Hz). At primary levels of 75,75 dB SPL, the highest response occurs at ratios of about 1.3 at low DP frequencies, and 1.15 at highest DP frequencies. This is similar to the DP frequency/ratio function reported by Harris et al. [7] Higher-level DP amplitudes corresponded with higher SNRs at the same ratio and DP frequency. When the primary tone levels were increased to L1,L2 = 75,75 dB SPL [Figure 5]e and f the averaged response region was enhanced with higher level emissions for the 2f 1 -f 2 region and a broad response for the 2f 2 -f 1 region of the map. The higher DP levels at low frequencies were also correlated to relatively high SNR levels for the 2f 1 -f 2 emission at ~ 1.2 ratios, but not at other ratios or for the 2f 2 -f 1 emission [Figure 5]e and f. This would indicate the presence of DPOAEs around the 1.2 ratio, but not at other ratios (despite the high levels shown at all ratios at low frequencies on the map), at low frequencies (<2000 Hz).

For subjects with NIHL, lower level amplitude 2f 1 -f 2 DPOAE responses (L1,L2 at 65,55 dB SPL) were evident at all DP frequencies for the 1.2 ratio region when compared to NHNN subjects. The range of ratios that exhibit a response is narrower as well. For both groups, the region of relative higher DPOAE levels was near 3-4 kHz, which correlates to an f 2 range of ~4.5-6 kHz when referencing an f 2 /f 1 of 1.2. There was little or no response evident for the 2f 2 -f 1 region of the map when using moderate level primary tones. Stimulation with high level primary tones of L1,L2 = 75,75 dB SPL elicited higher level DPOAE responses in impaired ears. For high level primary tones, a broader frequency and ratio region of response for the 2f 1 -f 2 and the 2f 2 -f 1 emissions became evident at the DP frequencies of 3-4 kHz.



SNR averages for each group were analyzed to determine relative areas of the map with significant levels of emissions. [Figure 6] shows the results for each experimental group and for each emission type. Each plot shows the cumulative number of points with an SNR at or above 3 dB, 6 dB and 9 dB as a function of primary tone level and experimental group for each emission type. [Table 1] summarizes the statistical significance of the differences.{Figure 6}{Table 1}

For the NHNN group, approximately half the points in the upper portion of the map have an SNR of at least 3 dB and 25% of all points are above 9 dB SNR indicating a high probability of DPOAE presence for the L1,L2 = 65,55 dB SPL protocol. In contrast, the NIHL group shows fewer points when referencing a SNR of 3 dB (46% versus 62% for NHNN) and less than half as many as for the NHNN group with an SNR of 9 dB. At L1,L2 = 75,75 dB SPL stimulus levels, both group show increases in the number of points with higher SNR values, but with the same relative differences evident between the two groups.

In general, the number of points with SNR of 3 dB or more is 30-50% less on the lower half of the map (2f 2 -f 1 ) than the upper half of the map (2f 1 -f 2 ), for either group. The difference between the two stimulus level protocols is more striking for the 2f 2 -f 1 emission. Indeed, for either group, the number of points with 9 dB SNR or higher quadrupled when increasing stimulus levels from 65, 55 dB to 75, 75 dB SPL (from 6.5% to 30% of the measurements for the NHNN group, and from 3.5% to 15% for the NIHL group). [Table 1] provides the P values when the NHNN and NIHL results are compared using the Kruskal-Wallis test. The percentage of points at each SNR threshold was significantly lower (P < 0.05) for the NIHL group for all comparisons, except for values ≥3 dB at 65,55 for the 2f 2 -f 1 emission. The most significant differences were seen for the percentage of points ≥9 dB at L1,L2 = 75,75 for the 2f 2 -f 1 emission.

The box and whisker plots [Figure 6] illustrate the variability within the NHNN and NIHL groups. Although the medians for the NHNN and NIHL groups were significantly different at each cut-off level (≥3 dB, ≥6 dB, or ≥9 dB), the interquartile range overlapped at all SNR cut off levels.

Statistical comparison of averaged DPOAE level maps

To reduce noise within the averaged DPOAE level maps, the data points were spatially averaged using a 3 × 3 median filter and then a one-way ANOVA was applied to each DPOAE level in the map. The results are plotted in [Figure 7].{Figure 7}

For the 2f 1 -f 2 DPOAE elicited with a moderate level primaries (65,55 dB SPL), the area of differentiation occurred at ratios between 1.125 and 1.275 at DP frequencies ranging between 1 kHz and 5 kHz, with the largest differences occurring between 2.5 kHz and 4.5 kHz. This correlates to an f 2 frequency range of about 3-6.8 kHz at a 1.2 ratio. For the 75,75 dB SPL protocol, a larger region of significance was evident. In addition, one small area of difference between the two groups appeared at the higher frequencies, where DP values are greater than 5 kHz and at ratios between 1.25 and 1.3 (corresponding to f 2 values of 8-9.5 kHz at these ratios).

For the 2f 2 -f 1 DPOAE there was minimal evidence of significant differences at the 65,55 dB SPL primary tone levels; however, at the L1,L2 = 75,75 dB SPL stimulus level there were lobes of responses that differed significantly between the two experimental groups. Areas of particular significance occur at DP frequencies ranging between 4 kHz and 5 kHz at small ratios as well as above 5.5 kHz at ratios ranging between 1.025 and 1.35.

Technical development

As part of this study, the technology was further developed to minimize the time required for each individual map. A new DPOAE mapping system using a dedicated digital signal processor and CODEC (analog-to-digital and digital-to-analog encoder/decoder) to generate and process each measurement was designed and built. The updated system also records both the lower side-band, 2f 1 -f 2 , and upper side-band, 2f 2 -f 1 , emissions simultaneously. This resulted in a time reduction of almost 50% as a single stimulus captures two points in equal level maps. This improved system developed by Creare Inc., (Hanover, NH) can compute a map at a rate of 0.24 s/stimulus, which is 140% of the theoretical limit. This is a marked improvement over the ~0.48 s/stimulus in the system used at the Jerry Pettis Memorial Veterans Medical Center (JPMVMC) for previously published maps, and the 0.3 s/stimulus for the system used to generate the maps in the present study.

This reduction brings the time to complete a map from approximately 25 min with the research prototype used for the present study, to approximately 12 min for an equal level stimulus (e.g., L1,L2 = 75,75 dB SPL) and 16 min for an unequal stimulus (e.g., L1,L2 = 65,55 dB SPL). (Unequal stimulus maps require additional time because the 2f 2 -f 1 emissions are measured with L1,L2 = 55,65 dB SPL and therefore cannot be measured at the same time as the 2f 1 -f 2 .)


DPOAE level mapping offers potential for detecting and tracking NIHL, but the appropriate parameters to use and the areas of the DPOAE level map that change the most with noise exposure have yet to be defined. In this study, we performed DPOAE level maps on 17 individuals without a history of noise exposure and compared those results to maps from 19 individuals with mild-moderate NIHL. The results show that: (a) both DPOAE levels, and the range of ratios where DPOAEs are detected, are reduced in the NIHL group for both the 2f 1 -f 2 and 2f 2 -f 1 emissions, (b) higher level primaries (L1,L2 = 75,75 dB SPL) increase the number of measureable emissions with an SNR of 3, 6 or 9 dB or greater in both groups, (c) the DPOAE ratios and frequencies that elicited the largest DPOAE SNRs were also generally the same ones that were significantly different between the NHNN and NIHL groups, (d) the 2f 1 -f 2 emission is reduced significantly for both L1,L2 = 65,55 dB SPL and 75,75 dB SPL primary levels for the NIHL group, (e) the 2f 2 -f 1 emission is reduced significantly in NIHL, but the changes are evident only when using L1,L2 = 75,75 primary tone levels, (f) the optimal ratio for the 2f 2 -f 1 emission at L1,L2 = 65,55 dB SPL is smaller than has been described previously and (g) displaying DPOAE level map data plotted as a function of SNR, rather than as DPOAE level, improves map interpretation at lower DPOAE frequencies (<2000 Hz).

Overall DPOAE level mapping results

As expected, DPOAE levels were significantly reduced in the NIHL group for both experimental protocols. The overall DPOAE levels, as well as the range of f 2 /f 1 ratios where emissions were detectable, were reduced. This effect was likely due predominately to the history of noise exposure in the NIHL group. Although the NIHL group was statistically significantly older than the NHNN group; the mean age differed by only 7 years. Currently, there is no definitive support for the premise that age alone significantly affects DPOAEs amplitude. [8] There are no age-corrections applied to the clinical interpretation off DPOAE magnitude measurements.

Although the reduction in DPOAE levels in subjects with NIHL is not surprising, these data suggest that there may be potential benefit in examining more than just the optimal DPOAE level at a single ratio when monitoring individuals for changes over time. Low-level 2f 1 -f 2 emissions and the 2f 2 -f 1 emission may disappear early on due to noise exposure. This hypothesis needs to be more fully explored in a longitudinal study.

One benefit of the DPOAE level map is that it presents a large amount of DPOAE data in a color pattern, so that the map image provides a visual snapshot of cochlear function between the DP frequencies of 0.5 kHz and 6.0 kHz across a wide range of stimulus ratios. This form of data presentation allows for more detailed data collection and analyses than standard DP-Grams, and also allows for a quick visual comparison of maps from a given individual. The results from this study show that the DPOAE SNR map, shown in combination with a DPOAE level map, improves the interpretation of the measurements, revealing areas where higher noise levels dominate the response.

Effect of primary tone level

Stover et al. [9] suggested that moderate-level primaries in the 50-60 dB SPL range are optimal for distinguishing normal-hearing versus mild to moderately hearing impaired ears when comparing to conventional pure-tone audiometry. The data from this study show that higher primary tones produce higher amplitude emissions in both the noise-exposed and non-noise exposed group. Since DPOAE level mapping takes time, testing at multiple primary tone levels may be feasible for a research application, but impractical for field use in hearing conservation programs. Helleman et al. [10] also suggested that higher level primary tones may be advantageous in terms of minimizing the effect of background noise when DPOAE monitoring was implemented as part of an occupational hearing conservation program. The data from this study show that higher level primaries provide higher SNRs in both the NHNN and NIHL groups, and are especially important for the measurement of DPOAEs in a mild-moderately NIHL impaired ear. Therefore, the higher-level primaries may be more useful for DPOAE level mapping studies in these populations. Although, since lower-level primaries are generally believed to be more sensitive to cochlear damage compared to higher level primaries, lower-level primaries may still be useful in individuals with normal DPOAE levels at baseline.

2f 2 -f 1 emissions

DPOAE level maps provide data on both 2f 1 -f 2 and 2f 2 -f 1 emissions; however, it is not clearly established if measuring the 2f 2 -f 1 emission adds value in detecting NIHL. Results from this study show significant changes in the 2f 2 -f 1 region when comparing the NHNN to the NIHL group. This difference was only apparent at the higher primary tone levels. These data suggest that when using the mapping for longitudinal tracking, the 2f 2 -f 1 emission area may be useful in detecting early cochlear changes due to hazardous noise exposure. This finding also provides further evidence that "coherent linear reflection" emission sources are useful in the detection of NIHL. [3],[4]

The data also revealed that the optimal 2f 2 -f 1 emission occurred at an f 2 /f 1 ratio of 1.025 at L1,L2 = 65,55 dB SPL, which is a smaller ratio value than has been seen in other studies. This may be because none of the cited works other than Fitzgerald and Prieve, [11] tested at such a low ratio and their results yielded the lowest ratio at which they did test. Although Fitzgerald and Prieve tested ratios as low as 1.025, they selected a single ratio that maximized the emission over a range of primary level combinations, and not specifically L1,L2 = 65,55. In addition, previous studies used DPOAE level to determine the optimal ratio, [11],[12],[13],[14] whereas the present study relies on both DPOAE levels and SNR to determine the optimal ratio in part because of the very short averaging time used in the protocol.

Technical factors for DPOAE level mapping

One factor limiting the use of DPOAE level maps as a clinical tool is the amount of time each map takes. As part of this study, the technology was developed to minimize the time required for each individual map. Testing time can be shortened further by using data from this study to reduce the number of points in each map. The data obtained in this study established the outlines of a smaller-sized map containing the most useful information [Figure 7]. Another approach to reducing test time is to use a flexible map protocol that allows for longer averaging times at some frequencies and ratios, and shorter times at frequencies not contaminated by noise. This could lead to an adaptive clinical DPOAE mapping protocol wherein frequencies, ratios and averaging time are optimized to maximize the response of the ear while maintaining a reasonable clinical testing time of about 5-10 min. Further, research and development will be required to design an efficient and clinically relevant DPOAE mapping protocol.

Analytical considerations

Using a point-by-point ANOVA after median filtering has limitations when applied to these data (e.g., multiple comparisons). As mentioned previously, this approach was undertaken in order to determine the areas with the highest likelihood of illustrating differences in DPOAE levels between the NHNN and NIHL groups as the mapping technology and protocol are being developed. This serves a different purpose than quantifying exact differences between groups. It is also recognized that DPOAE level maps are highly individualized, and consequently group comparisons are further confounded. As the mapping technology is advanced, more sophisticated statistical analysis can be undertaken to longitudinally monitor for individual changes in DPOAE level maps over time using fewer data points.


DPOAE level mapping provides a comprehensive overview of the changes in DPOAE magnitude for both the 2f 1 -f 2 and the 2f 2 -f 1 emissions in NIHL. DPOAE level mapping using higher level, equi-level primaries (75,75) provides more measureable DPOAEs in both the NHNN and NIHL groups than the lower level, unequal-level primaries (65,55). On average, the changes in DPOAE level maps in the NIHL group are concentrated in particular map areas and this knowledge could be used to design more efficient mapping protocols. Using SNR to display map results provides additional information for an improved interpretation of the mapping data, particularly at lower DPOAE frequencies in humans. Advances in the technology for acquiring DPOAE level maps have been made, which allow DPOAE level maps to be collected in clinically-acceptable times (5-10 min). DPOAE level mapping may offer a novel approach for differentiating and monitoring changes in cochlear function due to hazardous noise exposures.


We thank Dr. Brenda Lonsbury-Martin, Dr. Glen Martin and Barden Stagner for their support and assistance in developing and validating the instrumentation and software used in this research. We appreciate the work of Abigail Fellows, Ashley Huerta, Amber Powner, Kathy Pritzl and numerous personnel at Creare Inc. who supported the project.[15]


1Knight 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.
2Knight RD, Kemp DT. Wave and place fixed DPOAE maps of the human ear. J Acoust Soc Am 2001;109:1513-25.
3Martin 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
4Meinke DK, Stagner BB, Lonsbury-Martin BL, Martin GK. Detailed DPOAE level/phase maps in normal and noise-damaged human ears. Am Audit Soc Bull 2006;31:31.
5American National Standards Institute. Specifications for Audiometers (ANSI) S3.6 [R2010]. New York.
6American National Standards Institute. Maximum permissible ambient noise for audiometric test rooms (ANSI S3.1-1999[2008]). New York.
7Harris FP, Lonsbury-Martin BL, Stagner BB, Coats AC, Martin GK. Acoustic distortion products in humans: Systematic changes in amplitudes as a function of f2/f1 ratio. J Acoust Soc Am 1989;85:220-9.
8Strouse AL, Ochs MT, Hall JW 3 rd . Evidence against the influence of aging on distortion-product otoacoustic emissions. J Am Acad Audiol 1996;7:339-45.
9Stover L, Gorga MP, Neely ST, Montoya D. Toward optimizing the clinical utility of distortion product otoacoustic emission measurements. J Acoust Soc Am 1996;100:956-67.
10Helleman HW, Jansen EJ, Dreschler WA. Otoacoustic emissions in a hearing conservation program: General applicability in longitudinal monitoring and the relation to changes in pure-tone thresholds. Int J Audiol 2010;49:410-9.
11Fitzgerald TS, Prieve BA. Characteristics of the acoustic distortion product 2f 2 -f 1 from the normal hearing ear. J Speech Lang Hear Res 2005;48:21.
12Moulin A, Collet L, Veuillet E, Morgon A. Interrelations between transiently evoked otoacoustic emissions, spontaneous otoacoustic emissions and acoustic distortion products in normally hearing subjects. Hear Res 1993;65:216-33.
13Knight RD, Kemp DT. Relationships between DPOAE and TEOAE amplitude and phase characteristics. J Acoust Soc Am 1999;106:1420-35.
14Erminy M, Avan P, Bonfils P. Characteristics of the acoustic distortion product 2f 2 -f 1 from the normal human ear. Acta Otolaryngol 1998;118:32-6.
15Martin GK, Stagner BB, Chung YS, Lonsbury-Martin BL. Characterizing distortion-product otoacoustic emission components across four species. J Acoust Soc Am 2011;129:3090-103.