Sanford, Chris A.1; Hunter, Lisa L.2; Feeney, M. Patrick3,4; Nakajima, Hideko Heidi5,6
Investigations by Keefe et al. (1993) and Voss and Allen (1994) contributed to early work describing a technique to measure wideband, power-based response functions of acoustic impedance, admittance, and reflectance (i.e., wideband acoustic immittance [WAI]) in the human ear canal. This technique varied from conventional admittance tympanometry in several significant ways including freedom from pressurizing the ear canal (see Rosowski et al., this issue, pp. 9S–16S). The ability to obtain wideband acoustic estimates of middle ear function under ambient ear-canal conditions was considered especially significant for measurements in newborns and infants (see Kei et al., this issue pp. 17S–26S; and Hunter et al., this issue, pp. 36S–42S). After the application of WAI in human ears by Keefe and colleagues (1993), numerous studies in individuals across the age span have demonstrated the potential usefulness of these measurements obtained at ambient ear-canal pressure (see Hunter et al., this issue, pp. 36S–42S; Kei et al., this issue, pp. 17S–26S; Nakajima et al., this issue, pp. 48S–53S; Prieve et al., this issue, pp. 54S–59S). However, just as Terkildsen and Thompson (1959) introduced the concept of gathering immittance measurements in pressurized ear canals in a clinical manner (e.g., tympanometry), researchers hypothesized that valuable diagnostic information would be available from wideband measurements obtained over a range of ear-canal pressures (e.g., wideband tympanometry [WT]) (Keefe & Levi 1996; Margolis et al. 1999). A number of studies have examined this hypothesis via investigations of WT in individuals with normal middle ears and individuals with middle ear dysfunction. The purpose of this article is to review and summarize the results of studies using WT and propose directions for future research and potential clinical utility. The authors note, as Rosowski et al. (this issue, pp. 9S–16S) describe, that power or energy reflectance is the square of pressure reflectance. However, in the present article, the term reflectance will be used for simplicity when referring to previous work where either energy or power reflectance is referenced. As discussed in the consensus statement (see Feeney et al., this issue, pp. 78S–79S). it is recommended that more uniform and consistent use of terminology be used in future WAI research and clinical applications.
Several studies have suggested that WT measurements may be more sensitive to middle ear disorders and the effects of middle ear development than ambient WAI tests (Margolis et al. 1999; Keefe & Simmons 2003; Sanford & Feeney 2008). Early work focused on analyses of reflectance data derived from single- and multifrequency admittance tympanograms (Keefe & Levi 1996). A reflectance tympanogram (a term first used by Keefe and Levi) was represented by plotting reflectance (scaling from 0 to 1), as a joint function of pressure (daPa) and frequency (0.25 to 8.0 kHz). Later work, for ease of interpretation, plotted absorbance (absorbance = 1 − energy reflectance) as a function of pressure and frequency; absorbance has morphological properties similar to conventional tympanometry with peaked maxima (see Fig. 1 from Liu et al. 2008). Relative to the overall body of literature surrounding the family of WAI measurements, few studies have investigated WT. The relative ease with which ambient WAI tests can be performed, the perceived difficulties of WT, such as technological constraints (more so in the early studies), and questions regarding analysis techniques, are perhaps the most likely reasons why WT measurements have not received more attention.
WT data collection techniques have evolved from a process of deriving power-based response functions from traditional admittance tympanometry data via mathematical calculations (Keefe & Levi 1996) to a research system composed of a combination of PC-controlled software and a conventional tympanometer (Liu et al. 2008; Sanford & Feeney 2008), to a hand-held device (e.g., Interacoustics Titan Instrument, Assens, Denmark) capable of making both WT and traditional tympanometry measurements with dynamic pressure sweeps. Technological advances, more accessibility to equipment, and evolving data analysis techniques should encourage more interest in this area. The authors have chosen to review the relevant literature in a chronological fashion, highlighting technological advancements and knowledge gained regarding, (1) ear-canal/middle ear maturation, (2) middle ear disorders, and (3) test performance in classifying normal and disordered middle ears.
Derived WT Data
Early work by Keefe and Levi (1996) presented reflectance tympanograms that were obtained by deriving power-based response functions of reflectance from conventional single- and multifrequency admittance tympanograms and estimates of ear-canal diameter; participants were normal-hearing (NH) adults and infants. With 226 Hz tympanometry, it is assumed that when the tympanic membrane is exposed to a high degree of pressure (e.g., +200 daPa), that the admittance measured in the ear canal represents only the admittance of the volume of air in the enclosed canal. This assumption forms the basis for the procedure used to “compensate” for the effects of the volume of air in the ear canal between the measurement probe and tympanic membrane. Keefe and Levi suggested that this assumption could be tested via reflectance tympanometry measurements because the reflectance measured at the probe is approximately equal to the reflectance measured along the ear canal and at the eardrum. Their results suggested that the assumption was valid for adults, but not for infants in the 3 to 6 months age range, and that the differences for infants may be due to maturational effects of the infant ear. Results from Keefe and Levi demonstrated the feasibility of calculating reflectance tympanograms from conventional admittance tympanometry data and that measurements made in the presence of ear-canal pressure provide insight into the maturational changes in the infant ear. A limitation of this approach is the reliance on conventional tympanometry technologies, which have an upper frequency limit of 0.7 kHz.
Static Pressure Methods
Building on work by Keefe and Levi (1996), Margolis et al. (1999) obtained WAI data (reflectance and admittance from ambient and tympanometric conditions) using a PC-based data-acquisition system similar to that used by Keefe et al. (1993). Ear-canal pressure variations were accomplished with a microsyringe and manometer; with this method, WAI data acquisition (e.g., wideband reflectance/admittance tympanometry) across a broad range of discrete ear-canal pressures (+300 to −300 daPA) was possible. Wideband patterns, presented in Figure 2, show increased reflectance from 0.25 through 3.0 kHz in response to positive and negative ear-canal pressures (−300 and +300 daPa) relative to measures at ambient pressure (0 daPa). This increase in reflectance was expected, due to the stiffening of the tympanic membrane. At higher frequencies, from 3.0 kHz up to about 8.0 kHz, both positive and negative pressure caused a decrease in reflectance relative to results at ambient pressure.
Margolis et al. (1999) hypothesized that helpful diagnostic information could be obtained from WAI measures obtained in the presence of static pressure changes in the ear canal. Specifically, if an individual presented with negative middle ear pressure, it may be advantageous to assess the middle ear system at ambient and tympanometric peak pressure (TPP). For example, the authors presented a case study with audiometric and reflectance tympanometry results for a 10-year-old boy with recurrent otitis media. The boy presented with a 35 dB HL conductive hearing loss (CHL) and negative TPP (−250 daPa, as determined with 226 Hz tympanometry) for his left ear. Otoscopy revealed a retracted tympanic membrane, but no evidence of middle ear effusion. The reflectance pattern at ambient pressure resembled a pattern typically observed in normal ears under pressurized conditions (similar to data obtained at −300 daPa in Fig. 2). Margolis and colleagues speculated that if the ear canal was pressurized to match TPP, a normal reflectance measurement would be observed. However, when the ear canal was pressurized to match TPP, abnormal reflectance patterns were observed, indicating the possibility of a middle ear disorder, concurrent with negative middle ear pressure, but otherwise undetected; it is possible that the effects of negative middle ear pressure “overshadowed” the other middle ear dysfunction.
Margolis and colleagues (1999) also obtained conventional multifrequency (0.5 to 2.0 kHz) tympanometry data for comparison with the WAI data; due to known limitations, multifrequency tympanometry data were not obtained above 2.0 kHz. A comparison of multifrequency and WT data from 20 adults revealed similar tympanometric patterns and estimates of resonant frequency for frequencies from 0.5 to 2.0 kHz. WT data, (e.g., admittance and reflectance) from 2.0 through 11.3 kHz, were also examined. Interestingly, wideband admittance tympanograms did not follow an orderly pattern, but wideband reflectance tympanograms did progress in an orderly fashion as frequency increased from 2.0 through 11.3 kHz. Margolis and colleagues concluded that the orderly behavior of wideband reflectance tympanometry measurements and their apparent sensitivity to middle ear disorders might make them useful for clinical middle ear assessment.
Based on the work of Piskorski et al. (1999), suggesting that ambient WAI measurements may be a good predictor of CHL, Keefe and Simmons (2003) evaluated the test performance of ambient and tympanometric WAI tests in predicting CHL. While 226 Hz tympanometry is useful for detecting middle ear dysfunction, clear predictive relationships between tympanometric measures and CHL have not been identified (Dempster & MacKenzie 1991). Therefore, an objective test of CHL would be of interest and especially useful for pediatric populations. Keefe and Simmons described their WAI measurements as acoustic transfer functions, choosing a more generic name that could be used to describe many different acoustic ear-canal measurements. They also used transmittance as an alternative term to reflectance. Transmittance, as defined by Keefe and Simmons, is analogous to wideband absorbance described earlier in this article and in the work by Rosowski etal. (this issue, pp. 9S–16S). To deal with the large amount of data generated with WAI tympanograms, Keefe and Simmons reduced the two-dimensional wideband transmittance tympanograms into one-dimensional functions, or moments (see Cohen 1995). Specifically, frequency moments were calculated by averaging pressure as a function of frequency, and frequency moments were calculated by averaging frequency as a function of pressure. Moment analysis was chosen as a means of reducing a relatively large data set (re: ambient wideband data) into a small set of variables or predictor criteria. Keefe and Simmons hypothesized that wideband transmittance tympanometry data, organized by moment analysis, would provide more useful, quantitative predictors of CHL, relative to ambient wideband measurements. Keefe and Simmons used a prototype wideband diagnostic middle ear (DME) system, consisting of a computer with a two-channel digital signal processor and GSI-33 middle ear analyzer (Grason-Stadler Inc., Eden Prairie, MN USA); the GSI-33 was modified so the pressure pump was under computer control. WAI data was obtained for 17 static pressure intervals, at approximately eight “steps” in both positive-negative directions relative to TPP. Adult subjects, ranging in age from 10 to 55 years, provided 42 normal ears and 18 impaired ears with conductive or mixed losses. Clinical decision theory analysis was performed and the area under the receiver operating characteristic (AROC) curve was calculated for both wideband and 226 Hz tympanometry measurements. Test performance in predicting the presence of CHL, based on an air–bone gap of ≥20 dB, was evaluated and the sensitivity of each test was compared at a fixed specificity of 0.90. Results showed AROC values of 0.28 for static acoustic admittance (226 Hz), 0.72 for ambient wideband transmittance, and 0.94 for transmittance tympanometry. These results suggest that WAI tympanometry may be a good tool for the detection of middle ear disorders and an objective measure of CHL in children and adults. While computer-controlled changes in static pressure are an improvement over the system used by Margolis et al. (1999, 2001), the time to complete a WAI test (40 seconds) may preclude its use in some pediatric populations. One significant aspect of the work by Keefe and Simmons included a moment analysis approach. This approach was used to translate the large amount of data obtained across both frequency and pressure domains into univariate parameters, which were more suitable for clinical decision analysis. The ability to use a quantitative analysis technique to classify middle ears as either normal or abnormal would enhance the clinical utility of WAI tests, and would be especially useful in screening situations or other circumstances where testing is performed by individuals less skilled in interpretation of WAI results.
While the potential benefits of WT measurements in adults and older children had been suggested (re: the studies reviewed earlier), questions with regard to their value in infants had not been investigated. Sanford and Feeney (2008) shed light on these questions by investigating the maturational effects of the infant ear on ambient and tympanometric WAI measurements in 4-, 12-, and 24-week-old infants. A DME system similar to that used by Keefe and Simmons (2003) was used to gather WAI data. Results from Sanford and Feeney showed developmental effects for ambient reflectance measurements (see Fig. 3), which varied as a function of frequency; ambient reflectance results were very similar to reflectance at TPP (results at TPP are represented as solid lines for each age group in Fig. 3). The WAI tympanometry measurements, at frequencies from 0.25 to 0.75 kHz, showed as much as a 30% change in mean reflectance with variation in static ear-canal pressure from +200 to −200 daPa (see Fig. 3). However, the effects of ear-canal pressure resulted in minimal differences in reflectance across age at frequencies from 0.75 to 2.0 kHz. For 4-week-old infants, at high frequencies from 2.0 to 6.0 kHz, negative pressures were associated with increased reflectance and positive pressures were associated with decreased reflectance; these effects observed at high frequencies were not observed for either group of older infants. While a specific timeline of infant ear-canal development has not been fully described, some age-related effects observed in WT responses may be the result of a more compliant ear-canal wall (i.e., more susceptible to collapse) or less rigid coupling of the ossicles (Saunders et al. 1983; Kei et al., this issue, pp. 17S–26S), which may become more resistant to changes in pressure with age. The developmental effects observed with WT may explain some of the variability reported in 226 Hz tympanometry data in infants. Data from Sanford and Feeney shed light on the effects of ear-canal pressure on WAI measurements in infant ears and suggest that by 12 weeks of age, the effects of ear-canal pressure on WAI measurements are similar to those in adult ears.
Dynamic Pressure Methods
The development of a system capable of more rapid acquisition of both ambient and tympanometric WAI data was reported by Liu et al. (2008). The system described by Liu et al. obtained responses to multiple click stimuli as pressure was swept in the ear canal (+200 to −300 daPa). The system consisted of a Windows-based computer with a CardDeluxe (Digital Audio Labs, Chanhassen, MN) internal soundcard and custom software for control of stimulus generation and data acquisition; an Interacoustics AT235 tympanometer (Assens, Denmark) was modified and functioned as the pressure control. Advantages of this prototype system over the DME system used in previous studies include: (1) changes in pressure that were more similar to 226 Hz tympanometry (e.g., sweep versus discrete steps), (2) a finer description of WAI data across the pressure continuum, and (3) test time was reduced dramatically (e.g., from 40 to as little as 4 seconds). To assess the performance of the system, Liu et al. obtained WT data from 48 adult subjects (92 ears) with NH. The authors reported normative WT data, a measure of bandpass TPP (absorbance averaged over frequency from 0.38 to 2 kHz), and extractions of single-frequency (226 and 1000 Hz) tympanograms, which resembled the shape of single-frequency tympanometry measurements. Effects of sweep speed and direction were also observed in WAI tympanograms and were similar to those observed in single-frequency tympanometry measurements. Results from Liu et al. showed that wideband tympanograms obtained using dynamic pressure sweeps were similar to those obtained using a series of static pressures (Margolis et al. 1999; Keefe & Simmons 2003; Sanford & Feeney 2008) and demonstrated the feasibility of a system using both ambient and swept-pressure WAI measurements. The ability of the system used by Liu et al. to quickly obtain wideband and traditional tympanometry data may be useful in both research and clinical settings.
Using a system similar to that described by Liu et al. (2008), Sanford et al. (2009) evaluated test performance of WAI measurements (ambient and tympanometric) and 1 kHz tympanometry in relation to outcomes on a distortion-product otoacoustic emission (DPOAE)–based Newborn Hearing Screening (NHS) test. DPOAE testing was used to determine the NHS status of 455 infant ears (375 passed and 80 referred); infants were less than 2 days old. In a continued effort to identify appropriate measurement criteria, and reduce the multivariate WAI tympanometry data to univariate predictors, log likelihood ratio calculations (using means and standard deviations of WAI responses across frequency) were performed (Van Trees 1967). The purpose of using this method was to calculate likelihood criteria that indicated whether a response from an individual ear was from either the pass or refer group. The entire wideband frequency response range (0.226 to 8.0 kHz) was included, and the calculations produced larger relative weightings for frequencies where the differences between the distributions for pass and refer groups were greatest. Likelihood ratios were computed for each WAI test (ambient and tympanometric absorbance, admittance, etc). The 1 kHz tympanometry data analyses were based on methods used by Margolis et al. (2003), Kei et al. (2003), and Baldwin (2006).
Results from Sanford et al. (2009) showed that, for the most part, ears that passed the DPOAE test had higher wideband absorbance compared with ears that were referred, indicating the importance of middle ear status for interpreting DPOAE test data in newborns. Clinical decision theory analysis was used to determine the test performance of the WAI and 1 kHz tests in terms of their ability to classify ears that passed or were referred on the NHS test. The highest AROC values were 0.87 for ambient WAI tests and 0.84 for WT tests compared with 0.75 for 1 kHz tympanometry. Unlike the results from the study by Keefe and Simmons (2003), which demonstrated significantly improved performance for WAI tympanometry relative to ambient WAI measurements with adults and older children, test performance for ambient and tympanometric WAI measurements in the Sanford et al. study were not significantly different. The differences in WT test performance between the two studies may be due to several factors. First, reference standards for middle ear status were different; Keefe and Simmons used an air–bone gap (ABG) of ≥ 20 dB, based on behavioral audiometric assessment. This type of behavioral reference standard is not feasible for 1- to 2-day-old neonates; thus, DPOAE-based Universal Newborn Hearing Screening tests were used as the reference standard. Second, the study by Keefe and Simmons included children (10 years of age) and adults. The differences in age are important in light of studies reporting maturational effects for ambient and tympanometric WAI data in infants and children (Keefe et al. 1993; Sanford & Feeney 2008; Kei et al., this issue, pp. 17S–26S). Additional interesting findings from Sanford et al. are shown in Figure 4. These data show group median wideband absorbance for both ascending and descending pressure sweeps in infants who passed a DPOAE screening. For the ascending sweeps (negative to positive pressure), lower absorbance persists across a larger pressure range, compared with the absorbance over the same pressure range in the descending sweep condition; these effects are more dramatic than those observed in adults (see Liu et al. 2008). Holte et al. (1991) reported similar effects of ascending and descending pressure sweeps on multifrequency admittance tympanograms in infants. It may be that the young infants’ immature and compliant ear-canal walls behave differently than adult ears under negative pressure conditions.
In general, and at least qualitatively, to help interpret the WT shown in Figures 1, 4, it may be helpful to think of the WT response as a series of single-frequency tympanograms obtained at discrete frequencies. One can also envision a single-frequency tympanogram “extracted” from the whole wideband absorbance plot. Additional data analysis utilities developed by researchers at Boys Town National Research Hospital provide means to extract and plot WT data in a variety of ways. For example, Figure 5A shows wideband absorbance plotted as a joint function of frequency and pressure, Figure 5B shows absorbance plotted as a function of pressure, and in Figure 5C absorbance is plotted as a function of frequency. Absorbance in Figure 5C is plotted for wideband responses at ambient ear-canal pressure with a dashed line and at TPP with a solid line. The extent to which a comparison of WAI data at ambient pressure and TPP is diagnostically useful, as shown by Margolis et al. (1999, 2001), will require more research.
A more recent study by Keefe et al. (2012) tested the hypothesis that WAI measurements accurately predict CHL in young children suspected of having otitis media with effusion. The reference standards for CHL were ABGs at octave frequencies from 0.25 to 4 kHz, based on behaviorally measured audiometric thresholds. WAI responses (both ambient and tympanometric) and 226 Hz tympanometry measurements were obtained from 25 children (36 ears; aged 3.5 to 8.2 years) with CHL and 23 children (44 ears; aged 2.6 to 8.2 years) with NH. A system similar to those used in the studies by Liu et al. (2009) and Sanford et al. (2009) was used to gather WAI data. Log likelihood ratio predictors (also used in Sanford et al. 2008) for ambient and tympanometric wideband absorbance tests (0.7 to 5.6 kHz) and single-frequency predictors (wideband data centered at the ABG frequencies) for ambient wideband absorbance tests were calculated; tympanometric width and compensated admittance magnitude were 226 Hz tympanometric test predictors. Overall, for WAI data, the largest differences between ears with and without CHL were observed from approximately 0.7 to 6 kHz. WAI data for both CHL and NH groups were essentially the same for frequencies below 0.7 kHz. Therefore, it would be expected that tests including frequencies above 0.7 kHz would be more accurate predictors of CHL. Specifically, wideband absorbance tests were the best overall predictors of CHL with AROC values ≥ 0.97. Tympanometric width was the best 226 Hz tympanometric predictor (AROC ranging from 0.68 to 0.93), but both WAI tests (ambient and tympanometric) always had better test performance compared with 226 Hz tympanometry. Similar to findings reported by Sanford et al. (2008), no significant differences in test performance were observed for ambient and tympanometric data from Keefe et al. when comparing AROC for both types of WAI tests. While these results support the hypothesis that WAI tests are accurate predictors of CHL in children, offering improved test performance relative to 226 Hz tympanometric measurements, neither WAI test was significantly better than the other.
SUMMARY AND FUTURE DIRECTIONS
While results from early wideband tympanometry (WT) work showed improvements over ambient wideband tests in terms of test performance for identifying middle ear dysfunction and conductive hearing loss, more recent studies have shown high, but similar, test performance for ambient and tympanometric WAI tests. Reasons for these contradictory findings may include differences in equipment, participant age, analysis procedures, and the type of pathologies/dysfunctions examined. More recent work, showing that ambient and tympanometric WAI tests demonstrate high test performance, may suggest that neither test can “outperform” the other (e.g., ceiling effects for test performance).
Data from the case study presented by Margolis (1999) in Figure 2 suggest that WAI obtained under variations in ear-canal pressure can provide additional and potentially diagnostically useful information for individual ears (e.g., comparison of WAI at ambient and TPP). These aforementioned individual findings and group results, demonstrating the sensitivity of WAI tympanometry to middle ear dysfunction and maturational changes in the middle ear, are promising and suggest the need for additional investigations in individual subjects and large subject populations. Future studies could include ears with a wider variety of middle ear disorders and from individuals across broader age ranges. A key component of ongoing work in this area is to investigate alternative ways of analyzing the large amount of data obtained with WT. While a qualitative pattern-recognition approach may be informative in individual cases, quantitative analysis techniques would be especially useful in screening situations or when used by individuals less skilled in interpretation of WT results. While strategies to simplify large, multivariate data sets to univariate predictors have shown promising results (e.g., moment analysis and log likelihood ratios), additional analysis approaches may further improve the utility of WT tests.
In addition to the potential of improved middle ear assessment, there are other possible benefits of having a WAI system with capabilities of inducing changes in ear-canal pressure. Such a system could have the capacity of gathering ambient and tympanometric WAI and conventional tympanometry data (e.g., 226 and 1000 Hz tympanograms) during the same test; this would enable users to use a variety of test results for interpretation of middle ear status. If acoustic reflex functionality was possible on such a system, the capability to induce changes in ear-canal pressure would allow for reflexes to be obtained at TPP, as is done with conventional immittance equipment. Another potential benefit could be the ability to obtain otoacoustic emissions at varying ear-canal pressures. While such a technique has not been thoroughly investigated, several studies have examined comparisons of otoacoustic emissions obtained in ambient and pressurized conditions (Trine et al. 1993; Hof et al. 2005; Sun & Shaver 2009). A system with the aforementioned capabilities would provide the user with a comprehensive test battery to evaluate the status of the middle ear and other auditory functions as well.
WT measurements provide a view of the acoustic response properties of the middle ear over a broad range of frequencies, and ear-canal pressures and appear to be good indicators of middle ear status, conductive hearing loss, and the effects of middle ear maturation. However, there are several challenges facing these new and evolving techniques. First, additional work is needed to determine normative data for a variety of age groups because age-related differences in ambient and tympanometric WAI have been observed; included in these groups of normative data would be both normal and disordered ears. Specifically with regard to young infants, developmental effects on WAI measurements should be addressed to assess the potential usefulness of this test in infants. Second, if WT measurements are to be successfully applied in clinical settings, data analysis and interpretation must be relatively straightforward. With the large amount of data that can be obtained with wideband measures, key indicators should be identified in an effort to develop middle ear tests with high sensitivity and specificity. The devices with WT capabilities, used in the studies discussed earlier, were research systems approved only for use in research studies. However, with new WT technology and equipment options becoming available, more widespread access to clinically friendly technology will be an important factor for advancements in research and clinical utility. Because improved assessment of middle ear dysfunction is desirable, additional investigations are needed to shed more light on the utility of WT tests.
Baldwin M.. Choice of probe tone and classification of trace patterns in tympanometry undertaken in early infancy. Int J Audiol. (2006); 45:417–427
Cohen L.. Time-Frequency Analysis. (1995); New York, NY Prentice-Hall
Dempster J. H., MacKenzie K.. Tympanometry in the detection of hearing impairments associated with otitis media with effusion. Clin Otolaryngol Allied Sci. (1991); 16:157–159
Hof J. R., Anteunis L. J., Chenault M. N., et al. Otoacoustic emissions at compensated middle ear pressure in children. Int J Audiol. (2005); 44:317–320
Holte L., Margolis R. H., Cavanaugh R. M. Jr. Developmental changes in multifrequency tympanograms. Audiology. (1991); 30:1–24
Keefe D. H., Bulen J. C., Arehart K. H., et al. Ear-canal impedance and reflection coefficient in human infants and adults. J Acoust Soc Am. (1993); 94:2617–2638
Keefe D. H., Levi E.. Maturation of the middle and external ears: Acoustic power-based responses and reflectance tympanometry. Ear Hear. (1996); 17:361–373
Keefe D. H., Sanford C. A., Ellison J. C., et al. Wideband aural acoustic absorbance predicts conductive hearing loss in children. Int J Audiol. (2012); 51:880–891
Keefe D. H., Simmons J. L.. Energy transmittance predicts conductive hearing loss in older children and adults. J Acoust Soc Am. (2003); 114:(6 Pt 1)3217–3238
Liu Y. W., Sanford C. A., Ellison J. C., et al. Wideband absorbance tympanometry using pressure sweeps: System development and results on adults with normal hearing. J Acoust Soc Am. (2008); 124:3708–3719
Margolis R. H., Bass-Ringdahl S., Hanks W. D., et al. Tympanometry in newborn infants–1 kHz norms. J Am Acad Audiol. (2003); 14:383–392
Margolis R. H., Saly G. L., Keefe D. H.. Wideband reflectance tympanometry in normal adults. J Acoust Soc Am. (1999); 106:265–280
Piskorski P., Keefe D. H., Simmons J. L., et al. Prediction of conductive hearing loss based on acoustic ear-canal response using a multivariate clinical decision theory. J Acoust Soc Am. (1999); 105:1749–1764
Sanford C. A., Feeney M. P.. Effects of maturation on tympanometric wideband acoustic transfer functions in human infants. J Acoust Soc Am. (2008); 124:2106–2122
Sanford C. A., Keefe D. H., Liu Y. W., et al. Sound-conduction effects on distortion-product otoacoustic emission screening outcomes in newborn infants: Test performance of wideband acoustic transfer functions and 1-kHz tympanometry. Ear Hear. (2009); 30:635–652
Saunders J. C., Kaltenbach J. A., Relkin E. M.. R. Romand. The structural and functional development of the outer and middle ear. In: Development of Auditory and Vestibular Systems. (1983); New York, NY Academic Press
Sun X. M., Shaver M. D.. Effects of negative middle ear pressure on distortion product otoacoustic emissions and application of a compensation procedure in humans. Ear Hear. (2009); 30:191–202
Terkildsen K., Thomson K.. The influence of pressure variations on the impedance of the human eardrum. J Laryngol Otol. (1959); 73:409–418
Trine M. B., Hirsch J. E., Margolis R. H.. The effect of middle ear pressure on transient evoked otoacoustic emissions. Ear Hear. (1993); 14:401–407
Van Trees H. L.. Detection, Estimation, and Modulation Theory. (1967); New York, NY Wiley Interscience
Voss S. E., Allen J. B.. Measurement of acoustic impedance and reflectance in the human ear canal. J Acoust Soc Am. (1994); 95:372–384