- Email: [email protected]
Effects of effusion in the middle ear and perforation of the tympanic membrane on otoacoustic emissions in guinea pigs
Hearing Research 122 (1998) 41^46
E¡ects of e¡usion in the middle ear and perforation of the tympanic membrane on otoacoustic emissions in guinea pigs Hiromi Ueda a; *, Seiichi Nakata b , Michitaka Hoshino a
a
Department of Otolaryngology, Nagoya University, School of Medicine, 65 Tsurumai-Cho, Showa-Ku, Nagoya 466-8550, Japan b Department of Otolaryngology, Nagoya Red Cross Hospital, Nagoya, Japan Received 9 June 1997; revised 13 April 1998; accepted 21 April 1998
Abstract The influence of fluid in the middle ear and of perforation and closure of the tympanic membrane (TM) on otoacoustic emissions (OAEs) was evaluated in guinea pigs. Click-evoked otoacoustic emissions (CEOAEs) and distortion product otoacoustic emissions (DPOAEs) were measured after the auditory bulla was opened. Neither OAE level changed significantly when fluid filled only half the space in the bulla, but both OAE levels disappeared when fluid completely filled the bulla. These changes were reversible. Thus, the presence of fluid in the bulla influenced CEOAE and DPOAE levels only when its volume filled more than half the space of the bulla. Changes in both CEOAE and DPOAE levels were affected by the size of the perforation of the tympanic membrane. For the smallest perforation, the reduction in both CEOAE and DPOAE levels was restricted to the lower frequencies. However, as the size of the perforation increased, a decrease in high-frequency function occurred. Thus, the results indicate that the magnitude of both OAEs was proportional to the size of the perforation. Both OAE levels improved after the perforation was closed. Because CEOAEs were more sensitive than DPOAEs to perforation and closure of the TM, DPOAEs may be better suited for OAE measurement in ears with perforated TMs. z 1998 Elsevier Science B.V. All rights reserved. Key words: Otoacoustic emission; Tympanic membrane; Perforation; Otitis media with e¡usion; Guinea pig
1. Introduction Evaluation of evoked otoacoustic emissions (OAEs) in man is highly useful in audiometric screening to evaluate cochlear function (Kemp et al., 1990; LonsburyMartin and Martin, 1990). It is well known that middle ear problems, such as the presence of £uid in the middle ear or perforation of the tympanic membrane (TM), can a¡ect OAEs. Bray (1989) examined the e¡ect of £uid loading on the TM and showed that click-evoked otoacoustic emissions (CEOAEs) were almost completely absent after the introduction of three droplets of water onto the TM. Wiederhold (1990) reported the e¡ects of TM modi¢cation on distortion product otoacoustic emissions (DPOAEs) in the cat external auditory meatus by causing a leak with a small perforation * Corresponding author. Tel.: +81 (52) 744-2324; Fax: +81 (52) 744-2325; E-mail: [email protected]
and showed that DPOAEs by primary tones at 4.0 and 5.2 kHz reduced these amplitudes at all measured sound pressures. Owens et al. (1993) reported that ears with type B or C tympanogram patterns showed an absence or a marked reduction of both CEOAEs and DPOAEs. They also observed that ears with ventilating tubes exhibited the presence of DPOAEs, although their amplitudes at mid to high frequencies were lower than those of healthy ears. Amedee (1995) reported that neither a middle ear e¡usion nor an abnormal tympanogram necessarily caused CEOAEs to be absent, and that the type of e¡usion in the middle ear a¡ected the presence or absence of emissions. He also reported that the presence of ventilation tubes resulted in reducing the amplitude of CEOAEs to below 4 kHz. These somewhat diverse ¢ndings emphasized the need for more systematic information about the e¡ects of TM perforation and middle ear £uid on the recordability of OAEs. In this study, the in£uence of £uid in the middle ear
0378-5955 / 98 / $19.00 ß 1998 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 5 9 5 5 ( 9 8 ) 0 0 0 8 4 - 7
HEARES 3050 23-7-98
42
H. Ueda et al. / Hearing Research 122 (1998) 41^46
and of perforation and closure of the TM on both CEOAE and DPOAE levels in guinea pigs was examined. 2. Materials and methods Seven healthy white guinea pigs, 200 g, were anesthetized with ketamine hydrochloride (10 mg/kg, i.m.), placed in a sound-proof chamber, and restrained with a head holder. They were tracheotomized while breathing spontaneously. Body temperature was maintained at 37³C throughout the experiment with a heating pad. After the cartilaginous external ear canal had been cut away, a customized metal tube was inserted into the ear canal. Measurement of CEOAEs utilized the ILO88 OAE analyzer (ver. 3.91; Otodynamics Ltd.). An Etype neonate OAE probe for the ILO88 was attached to the metal tube. Transient responses to non-linear clicks at 73 þ 2 dBpeSPL were averaged 260 times and analyzed during the ¢rst 20 ms after stimulus onset (Kemp et al., 1990). Usually, CEOAEs were distinguished visually from acoustic ringing in individual ear canals between 1 and 2 ms after stimulus onset (Ueda et al., 1992). The time window was then set from the appearance time of CEOAEs to 10 ms, the original waveform was subsequently reconstructed and the CEOAE waveform was displayed. From total echo energy (total echo power, TEP) in the cross-power spectrum of the two independent echo responses, a more precise analysis was conducted after its conversion into ¢ve frequency bands from 0.5 to 5.5 kHz was analyzed. Speci¢cally, the ¢ve bands were as follows : FEP (¢ltered echo power) 1 = 0.5^1.5 kHz; FEP2 = 1.5^2.5 kHz ; FEP3 = 2.5^3.5 kHz; FEP4 = 3.5^4.5 kHz ; and FEP5 = 4.5^5.5 kHz. Noise levels were also calculated from the non-cross-power spectrum of the responses. DPOAEs were measured by use of the CUBe DIS1 system (ver. 3.00-C) packaged by Mimosa Acoustics, including an Ariel-DSP signal-processing board and the ER 10C probe system. Stimuli consisted of two sinusoids (f1 and f2; f2/f1 = 1.2) presented simultaneously, with the lower frequency sinusoid presented at 65 dBSPL and the higher frequency sinusoid at 55 dBSPL. Each of the two tones of the stimulus was placed in a 16 bit by 1024 point integer array in the memory of the DSP-16. These arrays, which at a 50 kHz sampling rate have a duration of T = 20.48 ms, were played out repetitively, one on channel A and one on channel B. After the signal was ramped on (T ms) and the transients had died out (in T ms), the microphone response was periodically time-averaged for a total of 200 averages (200T = 4.096 s) into a third 32 bit by 1024 point integer array. The response bu¡er was copied back to the PC, where it was converted to £oating point and Fourier-transformed, resulting in a
spectrum sampled every 48.8 Hz. Two points per octave were measured from the minimum f2 frequency of 500 Hz to the maximum f2 frequency of 10 000 Hz. Levels of the 2f13f2 DPOAE and noise £oor were swept as a function of f2. The auditory bulla was opened via a ventral approach, and the bone of the bulla was removed to reveal the lower half of the TM. Care was taken not to injure the TM. After several measurements of CEOAEs and DPOAEs, physiological saline was instilled into the bulla until the umbo was submerged. In this state, both the oval window and the round window were covered with physiological saline. The CEOAEs and DPOAEs were then measured twice after which the bulla was ¢lled with the physiological saline. After the OAEs were again measured twice, all the saline was removed by suction, and the emissions measured again twice. In the next series, the lower portion of the TM was punctured with a 25-gauge needle. The same animals were used for the TM perforation studies. Three sizes of perforation of TM were created. First, a small perforation was created by puncturing the center of the lower potion of the TM. CEOAEs and DPOAEs were measured, the perforation was closed by application of oiled paper, and CEOAEs and DPOAEs were measured again. Next, a larger perforation was made by removing the oiled paper and moving the needle laterally and vertically from within the small perforation. OAEs were measured both before and after application of oiled paper. Finally, the largest perforation was created in the same manner so that the lower portion of the TM was perforated maximally. The tympanic membrane and perforated area were visualized endoscopically from the extended external ear canal in one animal. These pictures were analyzed by NIH image software (ver. 1.55), and the area of perforation was calculated. Two-way repeated-measure ANOVA was used to compare the extent of OAE changes. Independent variables consisted of CEOAEs and DPOAEs before or after change, and of the frequencies used. Dependent variables were the CEOAE or DPOAE levels. The means of the resulting ANOVA values were then compared with the mean values of baseline using a contrast method with the Macintosh software (Super ANOVA ver. 1.11, Abacus Concepts, Berkeley, CA, USA). A level of P 6 0.01 was accepted as statistically signi¢cant. These investigations were performed in accordance with the principles of the Declaration of Helsinki. 3. Results 3.1. In£uence of £uid in the bulla Fig. 1 shows changes in CEOAE and DPOAE levels before and after instillation of physiological saline into
HEARES 3050 23-7-98
H. Ueda et al. / Hearing Research 122 (1998) 41^46
43
Fig. 1. Changes in CEOAE (a) and DPOAE levels (b) before (control) and after the instillation of physiological saline into the bulla (half, Half or completely ¢lled, Full) and after its removal. N = 7.
the bulla. When physiological saline half ¢lled the bulla, CEOAE levels were decreased across the test frequency range and DPOAE levels were decreased slightly at under 1 kHz. However, these changes were not statistically signi¢cant (Table 1). On the other hand, both CEOAE and DPOAE levels essentially equalled those of the noise £oor after the bulla had been completely ¢lled with physiological saline. This tendency was evident at all frequencies. When physiological saline had been totally removed from the bulla, both OAEs recovered completely. 3.2. In£uence of TM perforation 3.2.1. Small perforation The ¢rst perforation constituted 1% of the total TM area, as determined by computer measurement. As shown in Fig. 2a, CEOAE levels were reduced primarily in the low-frequency range with the reduction at 1 kHz (FEP 1) being statistically signi¢cant (P 6 0.001, Table 2). When the perforation was closed, CEOAE levels essentially returned to control levels. Changes in
DPOAEs were similar to those in CEOAEs (Fig. 2b). The DPOAEs below 1 kHz were reduced signi¢cantly after perforation of TM (Table 2), but DPOAEs above 1 kHz were not signi¢cantly changed. When oiled paper was applied to the perforation to close the defect, DPOAE levels at all frequencies recovered to their control levels. 3.2.2. Medium perforation The second perforation of 5% of the TM area reduced CEOAEs signi¢cantly at all frequencies as shown in Fig. 2a with the reduction on CEOAE levels being greatest over the low-frequency range. After closure of TM perforations with oiled paper, CEOAEs improved (Fig. 3a). Recovery appeared to be slightly less than complete, but di¡erences between control values and values after closure of TM perforations were not statistically signi¢cant (Table 2). The DPOAEs were also decreased after the 5% TM perforation with the reductions under 1 kHz and at 3 and 10 kHz being statistically signi¢cantly di¡erent from control values (Table 2). After repairing the perforations with the oiled paper,
Table 1 Changes in CEOAE and DPOAE levels (dB) before and after instillation of £uid into the bulla and after its removal CEOAE FEP1
FEP2
FEP3
FEP4
FEP5
Half 2.2 þ 1.4 2.3 þ 1.6 1.8 þ 1.2 2.8 þ 1.0 4.4 þ 1.1 Full 9.8 þ 2.7*** 18.9 þ 3.8*** 20.3 þ 4.2*** 18.9 þ 4.7*** 15.5 þ 4.1*** Removal 30.8 þ 1.8 0.0 þ 1.3 0.4 þ 1.9 0.1 þ 1.5 30.1 þ 1.0 DPOAE 0.5k Half Full
0.7k
1k
1.5k
2k
3k
4k
6k
8k
10k
1.7 þ 4.7 8.7 þ 8.5*
3.1 þ 4.6 4.5 þ 3.8 0.8 þ 6.9 30.5 þ 6.0 1.3 þ 5.1 30.2 þ 6.6 2.9 þ 3.5 30.8 þ 3.2 31.4 þ 2.7 14.7 þ 15.6* 14.7 þ 10.8** 23.1 þ 5.3*** 25.0 þ 8.9*** 31.3 þ 6.7*** 25.7 þ 12.9**- 30.3 þ 5.7*** 34.6 þ 7.9*** 37.6 þ 7.3*** * Removal 30.3 þ 3.8 33.9 þ 7.5 33.4 þ 4.4 3.2 þ 6.5 31.4 þ 5.7 32.2 þ 3.4 31.5 þ 8.3 1.0 þ 4.5 34.1 þ 5.0 30.9 þ 4.2 Values indicate means þ 1 S.D. (dB). Values with one or more asterisks di¡er signi¢cantly from control values (*P 6 0.01; **P 6 0.001; ***P = 0.0001). N = 7.
HEARES 3050 23-7-98
44
H. Ueda et al. / Hearing Research 122 (1998) 41^46
Fig. 2. Di¡erences in changes in CEOAE (a) and DPOAE levels (b) before (control) and after creation of a small TM perforation (Small perf.), a medium-sized TM perforation (Medium perf.) and a large TM perforation (Large perf.). N = 7.
DPOAEs at all frequencies recovered to their control levels. 3.2.3. Large perforation The third perforation at 30% of the TM area reduced CEOAEs to noise £oor levels as indicated in Fig. 2a. When TM perforations were closed with oiled paper, CEOAEs recovered partially. Although the post-closure recovery of CEOAEs at 3, 4, and 5 kHz was incomplete, that at 1 and 2 kHz was almost total, and showed no signi¢cant di¡erence from control values (Table 2). The DPOAEs were also decreased signi¢cantly at all frequencies following the 30% perforation, with DPOAEs 6 2 kHz being reduced to noise £oor levels (Fig. 2b). However, DPOAEs above 4 kHz, although reduced in level, were still present. After closure of the
TM perforations, DPOAEs tended to recover except for those at 10 kHz, which remained signi¢cantly lower in level than their baseline values (Table 2). 3.2.4. In£uence of perforation size The data of Fig. 2 shows how changes in CEOAEs and DPOAEs di¡ered according to the size of the perforation. With a 1% perforation, both OAEs were reduced at low frequencies. Reductions in both OAEs at low frequencies increased with increases in the size of the perforation, with an accompanying loss of the highfrequency response. With the 30% perforation, CEOAEs at all frequencies were reduced to the noise level, whereas the DPOAEs at higher frequencies persisted.
Fig. 3. Di¡erences in changes in CEOAE (a) and DPOAE levels (b) before (control) and after a closure of three kinds of TM perforations (Small perf., Medium perf., and Large perf.) with oiled paper. N = 7.
HEARES 3050 23-7-98
H. Ueda et al. / Hearing Research 122 (1998) 41^46
45
Table 2 Changes in CEOAE and DPOAE levels (dB) before and after TM perforation and upon perforation closure with oiled paper CEOAE
FEP1
FEP2
FEP3
FEP4
FEP5
2.3 þ 1.0 1.8 þ 1.7 7.2 þ 2.3*
1.5 þ 0.8 1.8 þ 1.4 5.9 þ 2.4*
Small perf. 5.8 þ 2.9** Close 0.9 þ 2.6 Medium 11.9 þ 2.4*** perf. Close 1.3 þ 1.9 Large perf. 11.9 þ 4.2*** Close 4.9 þ 5.1
3.5 þ 2.3 3.3 þ 1.7 0.5 þ 1.6 0.9 þ 1.6 10.5 þ 2.7*** 9.3 þ 1.8**
1.1 þ 1.5 2.4 þ 1.7 2.9 þ 1.5 17.7 þ 4.8*** 16.7 þ 4.0*** 14.4 þ 3.3 3.5 þ 6.3 6.0 þ 5.6* 8.3 þ 3.8***
3.5 þ 1.2 12.0 þ 4.0*** 7.9 þ 2.0***
DPOAE
0.7k
2k
3k
32.0 þ 6.0 0.7 þ 4.3 4.0 þ 7.8
2.1 þ 2.1 3.0 þ 3.2 1.7 þ 2.3 2.7 þ 7.6 11.1 þ 5.3* 8.3 þ 6.2
0.5k
Small perf. 9.3 þ 5.1* Close 0.9 þ 1.6 Medium 13.9 þ 5.0*** perf. Close 2.0 þ 2.5 Large perf. 19.2 þ 8.5*** Close 7.8 þ 8.0
1k
1.5k
12.9 þ 6.6** 5.7 þ 4.2 33.3 þ 6.0 30.3 þ 1.0 30.8 þ 4.8 30.6 þ 4.4 23.1 þ 3.8*** 15.0 þ 7.3*** 2.9 þ 2.9
4k
6k
8k
30.9 þ 4.1 2.4 þ 1.6 30.3 þ 3.7 1.3 þ 3.7 4.9 þ 4.0 5.5 þ 3.9
10k 2.4 þ 1.2 1.1 þ 1.7 7.4 þ 1.8**
0.5 þ 1.2 32.1 þ 5.6 31.6 þ 7.2 0.4 þ 5.6 3.6 þ 3.1 2.8 þ 8.3 30.1 þ 4.3 2.6 þ 4.8 2.7 þ 1.4 29.6 þ 4.8*** 22.7 þ 6.8*** 18.9 þ 5.0*** 19.6 þ 10.8*** 20.1 þ 9.4** 21.6 þ 9.5*** 11.9 þ 7.2** 12.6 þ 3.6** 16.8 þ 4.7*** 8.8 þ 6.7 1.8 þ 8.4 32.8 þ 9.4 1.1 þ 8.2 4.7 þ 9.6 8.1 þ 4.5 5.0 þ 4.5 5.6 þ 5.0 6.5 þ 2.5**
Values indicate means þ 1 S.D. (dB). Values with one or more asterisks indicate that these values di¡ered signi¢cantly from control values (*P 6 0.01; **P 6 0.001; ***P = 0.0001). N = 7.
4. Discussion It is readily accepted that OAEs are a¡ected by the transfer function of the middle ear, because sound energy passes through the middle ear twice to re£ect to the external canal. The transfer function of the middle ear is proportional to the input admittance of the middle ear. Friction, mass, and compliance are all factors contributing to input admittance. The middle ear space is one of the most important contributors to compliance with the tympanic membrane (Relkin, 1988). If the bulla is opened, the cavity of middle ear is considered to be large. In such a case, for a large middle ear space, the total compliance approximates the compliance of the tympanic membrane. If the middle ear is partially ¢lled with a low viscosity £uid, compliance of the tympanic membrane is not much changed. It is, therefore, understandable that CEOAE and DPOAE levels did not change signi¢cantly after half the space in the bulla had been ¢lled with £uid. Amedee (1995) reported that, in humans, the presence of e¡usion in the middle ear per se did not necessarily cause an absence of otoacoustic emissions. The present results are consistent with these ¢ndings. However, as that author pointed out, increased viscosity of the £uid in the middle ear might cause the OAEs to disappear. The presence of viscous £uid in the middle ear would be expected to reduce the compliance of the TM and to increase both friction and mass. If even a low viscosity £uid completely ¢lls the middle ear cavity, the compliance of the middle ear will decrease and the friction and mass will increase. In our experiment, CEOAEs and DPOAEs disappeared after the bulla was completely ¢lled with physiological saline. Moreover, the phenomenon was reversible. Clinically, these conditions resemble only the serous type of chronic otitis media with e¡usion. In the future, ad-
ditional investigations into the e¡ect on OAEs of changes of e¡usion £uid viscosity in the bulla will be needed. In patients, the e¡ect of the perforation of the TM on the transmission of sound is associated with a lowfrequency hearing loss (Anthony and Harrison, 1972; Austin, 1978). The frequency range of hearing loss has been less clear in animal studies. McArdle and Tonndorf (1968) reported that losses in transfer function, as judged by cochlear microphonics (CMs), were relatively uniform with respect to frequency when there was a wide opening in the bulla, but were large in the lowfrequency range when the bulla was closed. Wada et al. (1995) reported that middle ear sti¡ness was reduced by tympanic bulla perforation in guinea pigs, and that there was a drastic increase in the DPOAE level at primary f2 frequency lower than 1 kHz after tympanic bulla perforation. These authors also indicated that the e¡ect of tympanic bulla perforation on the DPOAE level at primary f2 frequency higher than 7 kHz was small. Bigelow et al. (1996), who measured the e¡ect of the size of the perforation of the TM on umbo velocity in the bulla with a small opening in the rat, reported that, in the low-frequency range, velocity was reduced with a small perforation, and was accompanied by a decrease in the velocity of high-frequency responses in the case of a large perforation. In our experiments, the bulla was opened widely, because it was otherwise dif¢cult to perforate the TM and then close it from the external ear canal. Results showed clearly that the loss of the CEOAE and DPOAE levels was restricted to low frequencies in the case of a small perforation, and that a loss of high-frequency function emerged as the size of the perforation increased. These ¢ndings agree with those of Bigelow et al. (1996), although the changes in OAEs exceeded those of the umbo velocities. OAEs
HEARES 3050 23-7-98
46
H. Ueda et al. / Hearing Research 122 (1998) 41^46
may thus be more sensitive than umbo velocities to small changes in perforation size. The CEOAE and DPOAE levels recovered almost completely after the small or medium TM perforation was closed, but both OAE levels recovered only incompletely after a large perforation was closed. The improvement at high frequencies was especially poor. This incomplete recovery at high frequencies may be ascribed to an increase in the mass of the middle ear, since application of an oiled paper to the tympanic membrane increases the mass component of the middle ear. Owens et al. (1993) reported that the amplitude of mid- to high-frequency DPOAEs in children with ventilation tubes was diminished as compared with that of children with normal ears, perhaps attributable to an increase in the mass component of the middle ear caused by the weight of the ventilation tube in the ear. It seems that CEOAEs are more sensitive than DPOAEs to perforation and closure of the TM. CEOAEs were signi¢cantly reduced at all frequencies after a medium-sized perforation of the TM, whereas DPOAEs were signi¢cantly reduced only at limited frequencies. In addition, the improvement of DPOAEs after closure of a large perforation exceeded that of CEOAEs. Thus, DPOAEs may be better suited for use in measurement of ears with a perforated TM. 5. Conclusions The CEOAE and DPOAE levels did not change signi¢cantly after half the space in the bulla was ¢lled with £uid, but both OAEs disappeared completely when the bulla was ¢lled entirely. These results indicate that the presence of £uid in the bulla does not in£uence the CEOAE or DPOAE levels if the £uid volume occupies less than half the total bulla space. Changes in both CEOAE and DPOAE levels were a¡ected by the size of the perforation of the tympanic membrane. With a small perforation, loss of both CEOAEs and DPOAEs was restricted to low frequen-
cies. As the size of the perforation increased, a loss of high-frequency function emerged. Results indicate that the magnitude of both OAE levels was proportional to the size of the perforation. Both OAE levels improved after the perforation was closed. References Amedee, R.G., 1995. The e¡ects of chronic otitis media with e¡usion on the measurement of transiently evoked otoacoustic emissions. Laryngoscope 105, 589^595. Anthony, W.P., Harrison, C.W., 1972. Tympanic membrane perforation: E¡ect on audiogram. Arch. Otolaryngol. 95, 506^510. Austin, D.F., 1978. Sound conduction of the diseased ear. J. Laryngol. Otol. 92, 367^372. Bigelow, D.C., Swanson, P.B., Saunders, J.C., 1996. The e¡ect of tympanic membrane perforation size on umbo velocity in the rat. Laryngoscope 106, 71^76. Bray, P., 1989. Click Evoked Otoacoustic Emissions and the Development of a Clinical Otoacoustic Hearing Test Instrument, Thesis, University of London. Kemp, D.T., Ryan, S., Bray, P., 1990. A guide to the e¡ective use of otoacoustic emissions. Ear Hear. 11, 93^105. Lonsbury-Martin, B.L., Martin, G.K., 1990. The clinical utility of distortion-product otoacoustic emissions. Ear Hear. 11, 144^154. McArdle, F.E., Tonndorf, J., 1968. Perforation of the tympanic membrane and their e¡ects upon middle-ear transmission. Arch. Klin. Exp. ONK Heilk. 192, 145^152. Owens, J.J., McCoy, M.J., Lonsbury-Martin, B.L., Martin, G.K., 1993. Otoacoustic emissions in children with normal ears, middle ear dysfunction, and ventilating tubes. Am. J. Otol. 14, 34^40. Relkin, E.M., 1988. Introduction to the analysis of middle-ear function. In: Jahn, A.F., Santos-Sacchi, J. (Eds.), Physiology of the Ear, Raven Press, New York, pp. 103^123. Ueda, H., Hattori, T., Sawaki, M., Niwa, H., Yanagita, N., 1992. The e¡ect of furosemide on evoked otoacoustic emissions in guinea pigs. Hear. Res. 62, 199^205. Wada, H., Ohyama, K., Kobayashi, T., Koike, T., Noguchi, S., 1995. E¡ect of middle ear on otoacoustic emissions. Audiology 34, 161^ 176. Wiederhold, M.L., 1990. E¡ects of tympanic membrane modi¢cation on distortion product otoacoustic emissions in the cat ear canal. In: Dallos, P., Geisler, C.D., Matthews, J.W., Ruggero, M., Steele, C.R. (Eds.), The Mechanics and Biophysics of Hearing, Springer, New York, pp. 251^258.
HEARES 3050 23-7-98
E¡ects of e¡usion in the middle ear and perforation of the tympanic membrane on otoacoustic emissions in guinea pigs Hiromi Ueda a; *, Seiichi Nakata b , Michitaka Hoshino a
a
Department of Otolaryngology, Nagoya University, School of Medicine, 65 Tsurumai-Cho, Showa-Ku, Nagoya 466-8550, Japan b Department of Otolaryngology, Nagoya Red Cross Hospital, Nagoya, Japan Received 9 June 1997; revised 13 April 1998; accepted 21 April 1998
Abstract The influence of fluid in the middle ear and of perforation and closure of the tympanic membrane (TM) on otoacoustic emissions (OAEs) was evaluated in guinea pigs. Click-evoked otoacoustic emissions (CEOAEs) and distortion product otoacoustic emissions (DPOAEs) were measured after the auditory bulla was opened. Neither OAE level changed significantly when fluid filled only half the space in the bulla, but both OAE levels disappeared when fluid completely filled the bulla. These changes were reversible. Thus, the presence of fluid in the bulla influenced CEOAE and DPOAE levels only when its volume filled more than half the space of the bulla. Changes in both CEOAE and DPOAE levels were affected by the size of the perforation of the tympanic membrane. For the smallest perforation, the reduction in both CEOAE and DPOAE levels was restricted to the lower frequencies. However, as the size of the perforation increased, a decrease in high-frequency function occurred. Thus, the results indicate that the magnitude of both OAEs was proportional to the size of the perforation. Both OAE levels improved after the perforation was closed. Because CEOAEs were more sensitive than DPOAEs to perforation and closure of the TM, DPOAEs may be better suited for OAE measurement in ears with perforated TMs. z 1998 Elsevier Science B.V. All rights reserved. Key words: Otoacoustic emission; Tympanic membrane; Perforation; Otitis media with e¡usion; Guinea pig
1. Introduction Evaluation of evoked otoacoustic emissions (OAEs) in man is highly useful in audiometric screening to evaluate cochlear function (Kemp et al., 1990; LonsburyMartin and Martin, 1990). It is well known that middle ear problems, such as the presence of £uid in the middle ear or perforation of the tympanic membrane (TM), can a¡ect OAEs. Bray (1989) examined the e¡ect of £uid loading on the TM and showed that click-evoked otoacoustic emissions (CEOAEs) were almost completely absent after the introduction of three droplets of water onto the TM. Wiederhold (1990) reported the e¡ects of TM modi¢cation on distortion product otoacoustic emissions (DPOAEs) in the cat external auditory meatus by causing a leak with a small perforation * Corresponding author. Tel.: +81 (52) 744-2324; Fax: +81 (52) 744-2325; E-mail: [email protected]
and showed that DPOAEs by primary tones at 4.0 and 5.2 kHz reduced these amplitudes at all measured sound pressures. Owens et al. (1993) reported that ears with type B or C tympanogram patterns showed an absence or a marked reduction of both CEOAEs and DPOAEs. They also observed that ears with ventilating tubes exhibited the presence of DPOAEs, although their amplitudes at mid to high frequencies were lower than those of healthy ears. Amedee (1995) reported that neither a middle ear e¡usion nor an abnormal tympanogram necessarily caused CEOAEs to be absent, and that the type of e¡usion in the middle ear a¡ected the presence or absence of emissions. He also reported that the presence of ventilation tubes resulted in reducing the amplitude of CEOAEs to below 4 kHz. These somewhat diverse ¢ndings emphasized the need for more systematic information about the e¡ects of TM perforation and middle ear £uid on the recordability of OAEs. In this study, the in£uence of £uid in the middle ear
0378-5955 / 98 / $19.00 ß 1998 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 5 9 5 5 ( 9 8 ) 0 0 0 8 4 - 7
HEARES 3050 23-7-98
42
H. Ueda et al. / Hearing Research 122 (1998) 41^46
and of perforation and closure of the TM on both CEOAE and DPOAE levels in guinea pigs was examined. 2. Materials and methods Seven healthy white guinea pigs, 200 g, were anesthetized with ketamine hydrochloride (10 mg/kg, i.m.), placed in a sound-proof chamber, and restrained with a head holder. They were tracheotomized while breathing spontaneously. Body temperature was maintained at 37³C throughout the experiment with a heating pad. After the cartilaginous external ear canal had been cut away, a customized metal tube was inserted into the ear canal. Measurement of CEOAEs utilized the ILO88 OAE analyzer (ver. 3.91; Otodynamics Ltd.). An Etype neonate OAE probe for the ILO88 was attached to the metal tube. Transient responses to non-linear clicks at 73 þ 2 dBpeSPL were averaged 260 times and analyzed during the ¢rst 20 ms after stimulus onset (Kemp et al., 1990). Usually, CEOAEs were distinguished visually from acoustic ringing in individual ear canals between 1 and 2 ms after stimulus onset (Ueda et al., 1992). The time window was then set from the appearance time of CEOAEs to 10 ms, the original waveform was subsequently reconstructed and the CEOAE waveform was displayed. From total echo energy (total echo power, TEP) in the cross-power spectrum of the two independent echo responses, a more precise analysis was conducted after its conversion into ¢ve frequency bands from 0.5 to 5.5 kHz was analyzed. Speci¢cally, the ¢ve bands were as follows : FEP (¢ltered echo power) 1 = 0.5^1.5 kHz; FEP2 = 1.5^2.5 kHz ; FEP3 = 2.5^3.5 kHz; FEP4 = 3.5^4.5 kHz ; and FEP5 = 4.5^5.5 kHz. Noise levels were also calculated from the non-cross-power spectrum of the responses. DPOAEs were measured by use of the CUBe DIS1 system (ver. 3.00-C) packaged by Mimosa Acoustics, including an Ariel-DSP signal-processing board and the ER 10C probe system. Stimuli consisted of two sinusoids (f1 and f2; f2/f1 = 1.2) presented simultaneously, with the lower frequency sinusoid presented at 65 dBSPL and the higher frequency sinusoid at 55 dBSPL. Each of the two tones of the stimulus was placed in a 16 bit by 1024 point integer array in the memory of the DSP-16. These arrays, which at a 50 kHz sampling rate have a duration of T = 20.48 ms, were played out repetitively, one on channel A and one on channel B. After the signal was ramped on (T ms) and the transients had died out (in T ms), the microphone response was periodically time-averaged for a total of 200 averages (200T = 4.096 s) into a third 32 bit by 1024 point integer array. The response bu¡er was copied back to the PC, where it was converted to £oating point and Fourier-transformed, resulting in a
spectrum sampled every 48.8 Hz. Two points per octave were measured from the minimum f2 frequency of 500 Hz to the maximum f2 frequency of 10 000 Hz. Levels of the 2f13f2 DPOAE and noise £oor were swept as a function of f2. The auditory bulla was opened via a ventral approach, and the bone of the bulla was removed to reveal the lower half of the TM. Care was taken not to injure the TM. After several measurements of CEOAEs and DPOAEs, physiological saline was instilled into the bulla until the umbo was submerged. In this state, both the oval window and the round window were covered with physiological saline. The CEOAEs and DPOAEs were then measured twice after which the bulla was ¢lled with the physiological saline. After the OAEs were again measured twice, all the saline was removed by suction, and the emissions measured again twice. In the next series, the lower portion of the TM was punctured with a 25-gauge needle. The same animals were used for the TM perforation studies. Three sizes of perforation of TM were created. First, a small perforation was created by puncturing the center of the lower potion of the TM. CEOAEs and DPOAEs were measured, the perforation was closed by application of oiled paper, and CEOAEs and DPOAEs were measured again. Next, a larger perforation was made by removing the oiled paper and moving the needle laterally and vertically from within the small perforation. OAEs were measured both before and after application of oiled paper. Finally, the largest perforation was created in the same manner so that the lower portion of the TM was perforated maximally. The tympanic membrane and perforated area were visualized endoscopically from the extended external ear canal in one animal. These pictures were analyzed by NIH image software (ver. 1.55), and the area of perforation was calculated. Two-way repeated-measure ANOVA was used to compare the extent of OAE changes. Independent variables consisted of CEOAEs and DPOAEs before or after change, and of the frequencies used. Dependent variables were the CEOAE or DPOAE levels. The means of the resulting ANOVA values were then compared with the mean values of baseline using a contrast method with the Macintosh software (Super ANOVA ver. 1.11, Abacus Concepts, Berkeley, CA, USA). A level of P 6 0.01 was accepted as statistically signi¢cant. These investigations were performed in accordance with the principles of the Declaration of Helsinki. 3. Results 3.1. In£uence of £uid in the bulla Fig. 1 shows changes in CEOAE and DPOAE levels before and after instillation of physiological saline into
HEARES 3050 23-7-98
H. Ueda et al. / Hearing Research 122 (1998) 41^46
43
Fig. 1. Changes in CEOAE (a) and DPOAE levels (b) before (control) and after the instillation of physiological saline into the bulla (half, Half or completely ¢lled, Full) and after its removal. N = 7.
the bulla. When physiological saline half ¢lled the bulla, CEOAE levels were decreased across the test frequency range and DPOAE levels were decreased slightly at under 1 kHz. However, these changes were not statistically signi¢cant (Table 1). On the other hand, both CEOAE and DPOAE levels essentially equalled those of the noise £oor after the bulla had been completely ¢lled with physiological saline. This tendency was evident at all frequencies. When physiological saline had been totally removed from the bulla, both OAEs recovered completely. 3.2. In£uence of TM perforation 3.2.1. Small perforation The ¢rst perforation constituted 1% of the total TM area, as determined by computer measurement. As shown in Fig. 2a, CEOAE levels were reduced primarily in the low-frequency range with the reduction at 1 kHz (FEP 1) being statistically signi¢cant (P 6 0.001, Table 2). When the perforation was closed, CEOAE levels essentially returned to control levels. Changes in
DPOAEs were similar to those in CEOAEs (Fig. 2b). The DPOAEs below 1 kHz were reduced signi¢cantly after perforation of TM (Table 2), but DPOAEs above 1 kHz were not signi¢cantly changed. When oiled paper was applied to the perforation to close the defect, DPOAE levels at all frequencies recovered to their control levels. 3.2.2. Medium perforation The second perforation of 5% of the TM area reduced CEOAEs signi¢cantly at all frequencies as shown in Fig. 2a with the reduction on CEOAE levels being greatest over the low-frequency range. After closure of TM perforations with oiled paper, CEOAEs improved (Fig. 3a). Recovery appeared to be slightly less than complete, but di¡erences between control values and values after closure of TM perforations were not statistically signi¢cant (Table 2). The DPOAEs were also decreased after the 5% TM perforation with the reductions under 1 kHz and at 3 and 10 kHz being statistically signi¢cantly di¡erent from control values (Table 2). After repairing the perforations with the oiled paper,
Table 1 Changes in CEOAE and DPOAE levels (dB) before and after instillation of £uid into the bulla and after its removal CEOAE FEP1
FEP2
FEP3
FEP4
FEP5
Half 2.2 þ 1.4 2.3 þ 1.6 1.8 þ 1.2 2.8 þ 1.0 4.4 þ 1.1 Full 9.8 þ 2.7*** 18.9 þ 3.8*** 20.3 þ 4.2*** 18.9 þ 4.7*** 15.5 þ 4.1*** Removal 30.8 þ 1.8 0.0 þ 1.3 0.4 þ 1.9 0.1 þ 1.5 30.1 þ 1.0 DPOAE 0.5k Half Full
0.7k
1k
1.5k
2k
3k
4k
6k
8k
10k
1.7 þ 4.7 8.7 þ 8.5*
3.1 þ 4.6 4.5 þ 3.8 0.8 þ 6.9 30.5 þ 6.0 1.3 þ 5.1 30.2 þ 6.6 2.9 þ 3.5 30.8 þ 3.2 31.4 þ 2.7 14.7 þ 15.6* 14.7 þ 10.8** 23.1 þ 5.3*** 25.0 þ 8.9*** 31.3 þ 6.7*** 25.7 þ 12.9**- 30.3 þ 5.7*** 34.6 þ 7.9*** 37.6 þ 7.3*** * Removal 30.3 þ 3.8 33.9 þ 7.5 33.4 þ 4.4 3.2 þ 6.5 31.4 þ 5.7 32.2 þ 3.4 31.5 þ 8.3 1.0 þ 4.5 34.1 þ 5.0 30.9 þ 4.2 Values indicate means þ 1 S.D. (dB). Values with one or more asterisks di¡er signi¢cantly from control values (*P 6 0.01; **P 6 0.001; ***P = 0.0001). N = 7.
HEARES 3050 23-7-98
44
H. Ueda et al. / Hearing Research 122 (1998) 41^46
Fig. 2. Di¡erences in changes in CEOAE (a) and DPOAE levels (b) before (control) and after creation of a small TM perforation (Small perf.), a medium-sized TM perforation (Medium perf.) and a large TM perforation (Large perf.). N = 7.
DPOAEs at all frequencies recovered to their control levels. 3.2.3. Large perforation The third perforation at 30% of the TM area reduced CEOAEs to noise £oor levels as indicated in Fig. 2a. When TM perforations were closed with oiled paper, CEOAEs recovered partially. Although the post-closure recovery of CEOAEs at 3, 4, and 5 kHz was incomplete, that at 1 and 2 kHz was almost total, and showed no signi¢cant di¡erence from control values (Table 2). The DPOAEs were also decreased signi¢cantly at all frequencies following the 30% perforation, with DPOAEs 6 2 kHz being reduced to noise £oor levels (Fig. 2b). However, DPOAEs above 4 kHz, although reduced in level, were still present. After closure of the
TM perforations, DPOAEs tended to recover except for those at 10 kHz, which remained signi¢cantly lower in level than their baseline values (Table 2). 3.2.4. In£uence of perforation size The data of Fig. 2 shows how changes in CEOAEs and DPOAEs di¡ered according to the size of the perforation. With a 1% perforation, both OAEs were reduced at low frequencies. Reductions in both OAEs at low frequencies increased with increases in the size of the perforation, with an accompanying loss of the highfrequency response. With the 30% perforation, CEOAEs at all frequencies were reduced to the noise level, whereas the DPOAEs at higher frequencies persisted.
Fig. 3. Di¡erences in changes in CEOAE (a) and DPOAE levels (b) before (control) and after a closure of three kinds of TM perforations (Small perf., Medium perf., and Large perf.) with oiled paper. N = 7.
HEARES 3050 23-7-98
H. Ueda et al. / Hearing Research 122 (1998) 41^46
45
Table 2 Changes in CEOAE and DPOAE levels (dB) before and after TM perforation and upon perforation closure with oiled paper CEOAE
FEP1
FEP2
FEP3
FEP4
FEP5
2.3 þ 1.0 1.8 þ 1.7 7.2 þ 2.3*
1.5 þ 0.8 1.8 þ 1.4 5.9 þ 2.4*
Small perf. 5.8 þ 2.9** Close 0.9 þ 2.6 Medium 11.9 þ 2.4*** perf. Close 1.3 þ 1.9 Large perf. 11.9 þ 4.2*** Close 4.9 þ 5.1
3.5 þ 2.3 3.3 þ 1.7 0.5 þ 1.6 0.9 þ 1.6 10.5 þ 2.7*** 9.3 þ 1.8**
1.1 þ 1.5 2.4 þ 1.7 2.9 þ 1.5 17.7 þ 4.8*** 16.7 þ 4.0*** 14.4 þ 3.3 3.5 þ 6.3 6.0 þ 5.6* 8.3 þ 3.8***
3.5 þ 1.2 12.0 þ 4.0*** 7.9 þ 2.0***
DPOAE
0.7k
2k
3k
32.0 þ 6.0 0.7 þ 4.3 4.0 þ 7.8
2.1 þ 2.1 3.0 þ 3.2 1.7 þ 2.3 2.7 þ 7.6 11.1 þ 5.3* 8.3 þ 6.2
0.5k
Small perf. 9.3 þ 5.1* Close 0.9 þ 1.6 Medium 13.9 þ 5.0*** perf. Close 2.0 þ 2.5 Large perf. 19.2 þ 8.5*** Close 7.8 þ 8.0
1k
1.5k
12.9 þ 6.6** 5.7 þ 4.2 33.3 þ 6.0 30.3 þ 1.0 30.8 þ 4.8 30.6 þ 4.4 23.1 þ 3.8*** 15.0 þ 7.3*** 2.9 þ 2.9
4k
6k
8k
30.9 þ 4.1 2.4 þ 1.6 30.3 þ 3.7 1.3 þ 3.7 4.9 þ 4.0 5.5 þ 3.9
10k 2.4 þ 1.2 1.1 þ 1.7 7.4 þ 1.8**
0.5 þ 1.2 32.1 þ 5.6 31.6 þ 7.2 0.4 þ 5.6 3.6 þ 3.1 2.8 þ 8.3 30.1 þ 4.3 2.6 þ 4.8 2.7 þ 1.4 29.6 þ 4.8*** 22.7 þ 6.8*** 18.9 þ 5.0*** 19.6 þ 10.8*** 20.1 þ 9.4** 21.6 þ 9.5*** 11.9 þ 7.2** 12.6 þ 3.6** 16.8 þ 4.7*** 8.8 þ 6.7 1.8 þ 8.4 32.8 þ 9.4 1.1 þ 8.2 4.7 þ 9.6 8.1 þ 4.5 5.0 þ 4.5 5.6 þ 5.0 6.5 þ 2.5**
Values indicate means þ 1 S.D. (dB). Values with one or more asterisks indicate that these values di¡ered signi¢cantly from control values (*P 6 0.01; **P 6 0.001; ***P = 0.0001). N = 7.
4. Discussion It is readily accepted that OAEs are a¡ected by the transfer function of the middle ear, because sound energy passes through the middle ear twice to re£ect to the external canal. The transfer function of the middle ear is proportional to the input admittance of the middle ear. Friction, mass, and compliance are all factors contributing to input admittance. The middle ear space is one of the most important contributors to compliance with the tympanic membrane (Relkin, 1988). If the bulla is opened, the cavity of middle ear is considered to be large. In such a case, for a large middle ear space, the total compliance approximates the compliance of the tympanic membrane. If the middle ear is partially ¢lled with a low viscosity £uid, compliance of the tympanic membrane is not much changed. It is, therefore, understandable that CEOAE and DPOAE levels did not change signi¢cantly after half the space in the bulla had been ¢lled with £uid. Amedee (1995) reported that, in humans, the presence of e¡usion in the middle ear per se did not necessarily cause an absence of otoacoustic emissions. The present results are consistent with these ¢ndings. However, as that author pointed out, increased viscosity of the £uid in the middle ear might cause the OAEs to disappear. The presence of viscous £uid in the middle ear would be expected to reduce the compliance of the TM and to increase both friction and mass. If even a low viscosity £uid completely ¢lls the middle ear cavity, the compliance of the middle ear will decrease and the friction and mass will increase. In our experiment, CEOAEs and DPOAEs disappeared after the bulla was completely ¢lled with physiological saline. Moreover, the phenomenon was reversible. Clinically, these conditions resemble only the serous type of chronic otitis media with e¡usion. In the future, ad-
ditional investigations into the e¡ect on OAEs of changes of e¡usion £uid viscosity in the bulla will be needed. In patients, the e¡ect of the perforation of the TM on the transmission of sound is associated with a lowfrequency hearing loss (Anthony and Harrison, 1972; Austin, 1978). The frequency range of hearing loss has been less clear in animal studies. McArdle and Tonndorf (1968) reported that losses in transfer function, as judged by cochlear microphonics (CMs), were relatively uniform with respect to frequency when there was a wide opening in the bulla, but were large in the lowfrequency range when the bulla was closed. Wada et al. (1995) reported that middle ear sti¡ness was reduced by tympanic bulla perforation in guinea pigs, and that there was a drastic increase in the DPOAE level at primary f2 frequency lower than 1 kHz after tympanic bulla perforation. These authors also indicated that the e¡ect of tympanic bulla perforation on the DPOAE level at primary f2 frequency higher than 7 kHz was small. Bigelow et al. (1996), who measured the e¡ect of the size of the perforation of the TM on umbo velocity in the bulla with a small opening in the rat, reported that, in the low-frequency range, velocity was reduced with a small perforation, and was accompanied by a decrease in the velocity of high-frequency responses in the case of a large perforation. In our experiments, the bulla was opened widely, because it was otherwise dif¢cult to perforate the TM and then close it from the external ear canal. Results showed clearly that the loss of the CEOAE and DPOAE levels was restricted to low frequencies in the case of a small perforation, and that a loss of high-frequency function emerged as the size of the perforation increased. These ¢ndings agree with those of Bigelow et al. (1996), although the changes in OAEs exceeded those of the umbo velocities. OAEs
HEARES 3050 23-7-98
46
H. Ueda et al. / Hearing Research 122 (1998) 41^46
may thus be more sensitive than umbo velocities to small changes in perforation size. The CEOAE and DPOAE levels recovered almost completely after the small or medium TM perforation was closed, but both OAE levels recovered only incompletely after a large perforation was closed. The improvement at high frequencies was especially poor. This incomplete recovery at high frequencies may be ascribed to an increase in the mass of the middle ear, since application of an oiled paper to the tympanic membrane increases the mass component of the middle ear. Owens et al. (1993) reported that the amplitude of mid- to high-frequency DPOAEs in children with ventilation tubes was diminished as compared with that of children with normal ears, perhaps attributable to an increase in the mass component of the middle ear caused by the weight of the ventilation tube in the ear. It seems that CEOAEs are more sensitive than DPOAEs to perforation and closure of the TM. CEOAEs were signi¢cantly reduced at all frequencies after a medium-sized perforation of the TM, whereas DPOAEs were signi¢cantly reduced only at limited frequencies. In addition, the improvement of DPOAEs after closure of a large perforation exceeded that of CEOAEs. Thus, DPOAEs may be better suited for use in measurement of ears with a perforated TM. 5. Conclusions The CEOAE and DPOAE levels did not change signi¢cantly after half the space in the bulla was ¢lled with £uid, but both OAEs disappeared completely when the bulla was ¢lled entirely. These results indicate that the presence of £uid in the bulla does not in£uence the CEOAE or DPOAE levels if the £uid volume occupies less than half the total bulla space. Changes in both CEOAE and DPOAE levels were a¡ected by the size of the perforation of the tympanic membrane. With a small perforation, loss of both CEOAEs and DPOAEs was restricted to low frequen-
cies. As the size of the perforation increased, a loss of high-frequency function emerged. Results indicate that the magnitude of both OAE levels was proportional to the size of the perforation. Both OAE levels improved after the perforation was closed. References Amedee, R.G., 1995. The e¡ects of chronic otitis media with e¡usion on the measurement of transiently evoked otoacoustic emissions. Laryngoscope 105, 589^595. Anthony, W.P., Harrison, C.W., 1972. Tympanic membrane perforation: E¡ect on audiogram. Arch. Otolaryngol. 95, 506^510. Austin, D.F., 1978. Sound conduction of the diseased ear. J. Laryngol. Otol. 92, 367^372. Bigelow, D.C., Swanson, P.B., Saunders, J.C., 1996. The e¡ect of tympanic membrane perforation size on umbo velocity in the rat. Laryngoscope 106, 71^76. Bray, P., 1989. Click Evoked Otoacoustic Emissions and the Development of a Clinical Otoacoustic Hearing Test Instrument, Thesis, University of London. Kemp, D.T., Ryan, S., Bray, P., 1990. A guide to the e¡ective use of otoacoustic emissions. Ear Hear. 11, 93^105. Lonsbury-Martin, B.L., Martin, G.K., 1990. The clinical utility of distortion-product otoacoustic emissions. Ear Hear. 11, 144^154. McArdle, F.E., Tonndorf, J., 1968. Perforation of the tympanic membrane and their e¡ects upon middle-ear transmission. Arch. Klin. Exp. ONK Heilk. 192, 145^152. Owens, J.J., McCoy, M.J., Lonsbury-Martin, B.L., Martin, G.K., 1993. Otoacoustic emissions in children with normal ears, middle ear dysfunction, and ventilating tubes. Am. J. Otol. 14, 34^40. Relkin, E.M., 1988. Introduction to the analysis of middle-ear function. In: Jahn, A.F., Santos-Sacchi, J. (Eds.), Physiology of the Ear, Raven Press, New York, pp. 103^123. Ueda, H., Hattori, T., Sawaki, M., Niwa, H., Yanagita, N., 1992. The e¡ect of furosemide on evoked otoacoustic emissions in guinea pigs. Hear. Res. 62, 199^205. Wada, H., Ohyama, K., Kobayashi, T., Koike, T., Noguchi, S., 1995. E¡ect of middle ear on otoacoustic emissions. Audiology 34, 161^ 176. Wiederhold, M.L., 1990. E¡ects of tympanic membrane modi¢cation on distortion product otoacoustic emissions in the cat ear canal. In: Dallos, P., Geisler, C.D., Matthews, J.W., Ruggero, M., Steele, C.R. (Eds.), The Mechanics and Biophysics of Hearing, Springer, New York, pp. 251^258.
HEARES 3050 23-7-98