Olivocochlear reflex effect on human distortion product otoacoustic emissions is largest at frequencies with distinct fine structure dips

Hearing Research

Hearing Research 223 (2007) 83–92

www.elsevier.com/locate/heares

Research paper

Olivocochlear reflex effect on human distortion product otoacoustic emissions is largest at frequencies with distinct fine structure dips W. Wagner a

a,*

, G. Heppelmann a, J. Mu¨ller b, T. Janssen b, H.-P. Zenner

a

Tu¨bingen Hearing Research Center, Department of Otorhinolaryngology, University of Tu¨bingen, Elfriede-Aulhorn-Str.5, 72076 Tu¨bingen, Germany b Department of Otorhinolaryngology, Technical University of Munich, Germany Received 9 June 2006; received in revised form 18 October 2006; accepted 19 October 2006 Available online 29 November 2006

Abstract Activity of the medial olivocochlear efferents can be inferred by measuring the change of the level of distortion product otoacoustic emissions (DPOAE) during ipsilateral or contralateral acoustic stimulation, the so-called medial olivocochlear reflex (MOCR). A limitation of this measurement strategy, however, is the distinct variability of MOCR values depending on DPOAE primary tone levels and frequency, which makes selection of the stimulus parameters difficult. The objective of this study was to evaluate the dependence of MOCR values on DPOAE fine structure in humans. MOCR during contralateral acoustic stimulation was measured at frequencies with distinct non-monotonicity (‘‘dip’’) in the DPOAE fine structure, and in frequencies with flat fine structure. One hundred and twenty one different primary tone level combinations were used (L1 = 50–60 dB SPL, L2 = 35–45 dB SPL, 1 dB steps). The measurement was repeated on another day. The major findings were: (1) Largest MOCR effects can be found in frequencies which exhibit a distinct dip in DPOAE fine structure. (2) Primary tone levels have a critical influence on the magnitude of the MOCR effect. MOCR changes of up to 23 dB following a L1 change of only 1 dB were observed. Averages of the maximum MOCR change per 1 dB step were in the 3–5 dB-range. Both findings can be interpreted in the light of the DPOAE two-generator model [Heitmann, J., Waldmann, B., Schnitzler, H.U., Plinkert, P.K., Zenner, H.P. 1998. Suppression of distortion product otoacoustic emissions (DPOAE) near 2f1  f2 removes DP-gram fine structure – evidence for a secondary generator. Journal of the Acoustical Society of America 103, 1527–1531]. According to the present results we propose, that assessing MOCR specifically at frequencies with a distinct dip in the DPOAE fine structure, in combination with fine variation of the stimulus tone levels, allows for a more targeted search for maximum MOCR effects. Future studies must show if this approach can contribute to the further clarification of the physiological roles of the olivocochlear efferents.  2006 Elsevier B.V. All rights reserved. Keywords: Olivocochlear efferents; Olivocochlear bundle; Olivocochlear reflex; Contralateral suppression; Otoacoustic emissions; Distortion products

1. Introduction Abbreviations: CAS, contralateral acoustic stimulation; CS, contralateral suppression (of OAE); DPOAE, distortion product otoacoustic emissions; IA, ipsilateral adaptation; MOCB, medial olivocochlear bundle; MOCR, medial olivocochlear reflex; OAE, otoacoustic emissions; OCB, olivocochlear bundle; OHC, outer hair cell; SNR, signal to noise ratio; SR, stapedial reflex * Corresponding author. Tel.: +49 7071 2988088; fax: +49 7071 2983311. E-mail address: [email protected] (W. Wagner). 0378-5955/$ - see front matter  2006 Elsevier B.V. All rights reserved. doi:10.1016/j.heares.2006.10.001

The olivocochlear bundle (OCB) consists of neurons projecting from the upper olivary region in the brainstem to the cochlea. It is comprised of a lateral and a medial subsystem, which both contain crossed and uncrossed fibers (Warr and Guinan, 1979). The lateral OCB projects to dendrites which insert at the inner hair cells. Its exact function is yet unknown. The medial part (MOCB) projects

84

W. Wagner et al. / Hearing Research 223 (2007) 83–92

primarily to outer hair cells (OHCs). To date there is evidence for two possible effects of the MOCB: (1) improvement of low tone detection and sound discrimination in background noise (Winslow and Sachs, 1987; Liberman and Guinan, 1998; Kumar and Vanaja, 2004), including selective attention to frequencies (Scharf et al., 1997) (2) a protective effect against acoustic trauma of the cochlea (e.g. Rajan, 1995; Maison and Liberman, 2000), which is, however, still controversially discussed. An extensive neurophysiological literature demonstrates that the MOCB system attenuates cochlear response to sound by reducing the gain of the OHC mechanical response to stimulation (Galambos, 1956; Wiederhold, 1970; Murugasu and Russell, 1996). Activation of the MOCB by acoustic stimulation leads to amplitude changes of otoacoustic emissions (OAE), the so called MOC-reflex (MOCR) (Puel and Rebillard, 1990; Veuillet et al., 1991).The effect can be measured either by just stimulating the ipsilateral ear (referred to as ipsilateral adaptation, IA), or by also stimulating the contralateral ear (contralateral suppression, CS). Both CS and IA are regarded as a measure for MOCR strength (Collet et al., 1990; Maison and Liberman, 2000). This view is supported by the fact that CS and IA effects, although different in magnitude, are identical in frequency, level dependence and form in guinea pigs (Kujawa and Liberman, 2001). CS was found to be more commonly a decreasing effect (suppression) than an increasing effect (enhancement) and typically amounted to some dB, depending on stimulus parameters. In humans, IA effects are generally smaller than CS values and mostly range below 1 dB (Kim et al., 2001; Bassim et al., 2003). Furthermore, CS was demonstrated to have a higher measurable incidence and a better reproducibility than IA in humans, which recommends CS as the method of choice for the measurement of human efferent activity (Mu¨ller et al., 2005). Recent studies in guinea pigs demonstrated that the measured MOCR effect strongly depends on measured frequency and on DPOAE primary tone levels. Maison and Liberman (2000), who used a matrix of 176 different primary tone level combinations, found that variation of the primary tone level by only 1 dB could result in changes of IA of DPOAE of more than 30 dB, including a change in sign of the amplitude change (bipolar effect). Variation of f2 by 1/4 octave led to changes of the DPOAE adaptation induced by bilateral acoustic stimulation of up to 10 dB (Luebke and Foster, 2002). One consequence of these findings would be the requirement to vary both stimulus frequency and primary tone levels extensively in high resolution, in order to completely describe an individual’s MOCR strength. This, however, would result in impracticably long measurement times in humans. In 2001, Kujawa and Liberman reported, that they found the largest IA effects preferably in frequencies which exhibited non-monotonicities (‘‘notches’’) in the DPOAE input/output functions. After section of the MOCB of the guinea pigs both the notches and the IA effects largely

disappeared. We hypothesized that this association would also hold true for non-monotonicities in the DPOAE level vs. frequency function (‘‘dips’’), as both kinds of nonmonotonicities presumably arise from the same phenomenon, i.e. acoustic interferences between two cochlear sources which comprise the DPOAE signal. The scope of this study was to examine the dependence of the MOCR effect – as measured by CS of DPOAE – on DPOAE fine structure and primary tone level in humans. For this we measured MOCR at frequencies in different specific areas of the DPOAE fine structure (dip and flat area), and varied primary tone levels extensively (121 level combinations). The overall aim was to suggest a measurement paradigm for a MOCR assay in humans, which meets the following criteria: (1) registration of the entire range of MOCR strength of an individual including maximum effects, (2) practicability with regard to measurement time. 2. Materials and methods 2.1. Subjects Twenty-six subjects participated in the study. Inclusion criteria were absence of ear disease in the medical history, normal finding in ear microscopy, normal tympanometry, stapedial reflex (SR) thresholds P80 dB SPL (measured with tone elicitors, both ipsi- and contralaterally), audiometric hearing thresholds of 20 dB HL or better at audiometric frequencies between 0.5 and 8 kHz in both ears, and bilaterally recordable DPOAE between 1 and 6 kHz with a minimum signal to noise ratio (SNR) of 6 dB at stimulus levels L2 = 60, 50, 40, 35, 30, 25 and 20 dB SPL, L1 = L2 · 0.4 + 39 dB SPL (scissor paradigm). Four subjects were excluded from the study because of insufficient DPOAE presence according to the inclusion criteria. The remaining 22 subjects (15 female, 7 male) were aged 19– 47 years. 2.2. General measurement procedures Auditory threshold was measured at audiometric frequencies between 0.125 and 8 kHz in both ears (audiometer CA 540, Hortmann, Germany). For tympanometry and SR registration the MADSEN Zodiac 901 Middle-Ear Analyzer (GN Otometrics A/S, Denmark) was used. The 2f1  f2 DPOAE were recorded using a PCMCIA DSP card (Starkey, USA) with an ER-10c (Etymotic Research, USA) ear probe. Contralateral acoustic stimulation (CAS) was delivered via a FZ-PRC1 (Fischer-Zoth, Germany) sound probe, with a priori calibration in an ear simulator (Bru¨el & Kjær, Denmark, Type 4157) and no correction accounting for the individual ear canal volume. Measurement and evaluation software was custom-made, using MATLAB (Mathworks, USA). Ipsilateral stimuli were adjusted according to an in-the-ear-calibration strategy. The frequency ratio was set to f2/f1 = 1.2 for all experimental conditions. Tympanometry and audiometric

W. Wagner et al. / Hearing Research 223 (2007) 83–92

threshold measurement were performed in a sound proof chamber, while during all DPOAE recordings the subjects were seated in a comfortable recliner in a quiet room. The quiet room provided similar low noise floors as the sound proof chamber, as verified by exemplary measurements. The investigations were performed in accordance with the principles of the Declaration of Helsinki and were approved by the human subject review committee of the University of Tu¨bingen. 2.3. Measurement of DPOAE fine structure DPOAE fine structure was measured bilaterally at f2 = 2–6 kHz with a frequency resolution of 47 Hz. Primary tone levels of L2 = 20, 30 and 40 dB SPL and L1 = 0.4L2 + 39 dB SPL were used (Fig. 1). This level setting takes into account the nonlinear interaction of the two primaries at the DPOAE generation site at the f2 place (Boege and Janssen, 2002). The averaging time for DPOAE recording was set to 4 s and the pause time between two measurement settings to 1 s. Total measurement time was approx. 25 min per ear. 2.4. Selection of frequencies fdip and fcontrol As second step, the frequencies for the following CS measurements were selected. A frequency with an especially distinct amplitude dip (‘‘dip-frequency’’, fdip) and a frequency without dip (‘‘control-frequency’’, fcontrol) were selected from the fine structure. One fdip from one ear of each subject was selected according to the following criteria: (a) presence of the dip at all three primary tone level settings, (b) signal to noise ratio of P10 dB at fdip at all three primary tone level settings, (c) maximum depth of the dip (‘‘ddip’’).

Fig. 1. Example of measurement of DPOAE fine structure. Three primary tone level combinations with L2 = 20, 30 and 40 dB SPL and L1 = L2 · 0.4 + 39 dB were used. The three bottom lines represent the three noise floor levels.

85

The noise floor level was computed by averaging the levels at six frequencies located around the DPOAE frequency. Whereas dip-frequencies were selected from the entire measured frequency range 2–6 kHz, fcontrol was set at 4 kHz. If 4 kHz exhibited a dip, or fdip was 4 kHz in the particular ear, a frequency with flat fine structure closest possible to 4 kHz was chosen as fcontrol. ddip was defined as the arithmetic mean of the differences between fdip and the adjacent lower and higher frequencies at all three primary tone level settings (Fig. 2), according to the equation: d dip ðf2 ; L2 Þ ¼ ½Ldp ðf2  47 Hz;L2 Þ  Ldp ðf2 ; L2 Þ þ ½Ldp ðf2 þ 47 Hz;L2 Þ  Ldp ðf2 ; L2 Þ d dip ðf2 Þ ¼ fd dip ðf2 ; L2 ¼ 20 dBÞ þ d dip ðf2 ; L2 ¼ 30 dBÞ þ d dip ðf2 ; L2 ¼ 40 dBÞg=3 Accordingly, the SNR was calculated considering the arithmetic mean of the three SNR values at the three primary tone level settings. 2.5. Measurement of MOCR CS of DPOAE was measured at fdip and fcontrol using 121 different primary tone level combinations. L1 was varied from 50 to 60 dB SPL in 1 dB steps, L2 was varied from 35 to 45 dB SPL in 1 dB steps (11L1 · 11L2 = 121). CAS consisted of broadband noise of 60 dB SPL with a frequency range of 100 Hz to 10 kHz. In more detail, DPOAE measurement started with the first L1/L2 combination (L1/ L2 = 50/35 dB SPL) without CAS. This measurement was repeated with the same L1/L2 combination but with CAS application. The procedure was conducted for all 121 L1/ L2 combinations (change of L2 in 1 dB steps) with fixed L1; after measurements with all 11 L2 levels change of L1 by 1 dB and again sweeping of L2, and so forth. An automatic re-calibration was performed before every single measurement, whether with or without CAS. The primary tone levels were not changed during re-calibrations but

Fig. 2. Calculation of depth of fine structure ddip (section of DPOAE fine structure example taken from Fig. 1).

86

W. Wagner et al. / Hearing Research 223 (2007) 83–92

were kept constant, according to the first calibration. The re-calibrations were intended to give the examiner a control if the ear canal probe was still in the same correct position. Total individual measurement time for both frequencies was approx. 70 min. CS measurement was repeated in an identical way on a second day, 1–15 days (mean 5.5 days) after the first measurement session. In the following the CS effect is referred to as DLdp. The difference between maximum and minimum DLdp across the 121 primary tone level combinations was termed DLdp range. 3. Results 3.1. DPOAE fine structure DPOAE fine structure was measured in all 44 ears of 22 subjects. In total, 10,312 out of 10,692 data points were valid (96.5%). Three hundred and four data points were rejected because of an SNR < 6 dB, in 76 data points no DPOAE amplitude was measurable. Mean DPOAE amplitudes were in the range of ca. 3 dB for the highest primary tone level combination (L2 = 40 dB SPL, L1 = 55 dB SPL) as compared to ca. 14 dB for the lowest (L2 = 20 dB SPL, L1 = 47 dB SPL). For evaluation of dip distribution over frequency, the five deepest dips with valid values for all three level combinations (SNR P 6 dB, no artefacts) were selected from every ear. Overall dip incidence was higher in the frequency range below 4.2 kHz than in the higher frequencies. This finding was caused by a group of 14 ears, which showed a bimodal distribution pattern with high dip incidence below ca. 4 kHz and low dip incidence above ca. 4 kHz. In the other 30 ears the dips were about evenly distributed over the entire frequency range. Comparing the fine structure in both ears of the subjects, we did not find any significant differences between right and left ears. Within the fine structure, mean DP amplitude slope in the five deepest dips was as great as 4.5, 6.4 and 6.8 dB per 47 Hz step (for L2 = 40, 30 and 20 dB), with maximum gradients in one individual of 19.9, 21.5 and 21.6 dB per D47 Hz. Alternative evaluation by octave steps revealed mean slopes in the dips of 222, 316, 348 dB per octave (for L2 = 40, 30 and 20 dB), with maximum individual gradients of 1146, 1269 and 1351 dB per octave. 3.2. Distribution of dip- and control-frequencies The selection criteria for dip- and control-frequencies and the calculation method for the depth of the dip (ddip) have been described in Section 2. fdip belonged to the right ear in 12 cases and to the left ear in 10 cases. fcontrol was selected from the same side as fdip. The mean frequency of fdip was 3672 Hz ± 725 Hz (2016–5063 Hz), with a mean ddip of 16 ± 4.9 dB (8.2–24 dB). For fcontrol, 4000 Hz could be selected 13 times, in nine subjects another fcontrol had to be chosen (five subjects with other local dips at 4000 Hz, four subjects with fdip at 4000 Hz). Mean frequency of fcontrol

Fig. 3. Distribution of dip frequencies fdip (circles) and control-frequencies fcontrol (asterisks). The horizontal lines show the SD of frequency, the vertical lines show the SD of the depth of fine structure.

was 4244 Hz ± 434 Hz (3656–5156 Hz), with a mean dip of 0.2 ± 3.8 dB (7.4 to 9.2 dB) (Fig. 3). 3.3. MOCR General remarks: Both negative and positive DLdp values – i.e. DPOAE suppressions and enhancements – could be observed, however suppression effects occurred more frequently and were mostly greater in amplitude. Both suppressions and enhancements are supposed to reflect efferent activity, and both effects can occur in one individual at different frequencies or primaries, leading to eventual arithmetical extinction in averaging operations. Therefore we considered DLdp regardless of the sign, resulting in an absolute DLdp termed ‘‘DLdp abs’’. MOCR was measured for 121 primary tone level combinations in each ear, and in the following the term DLdp abs describes the arithmetic mean of the absolute values of the 121 measurement points in one individual. Test–retest-repeatability: Test–retest-repeatability of DLdp between measurement day 1 and 2 was generally good (Fig. 4), with a Pearson correlation coefficient of r = 0.8. For further evaluation of both DLdp abs and DLdp range, the arithmetic mean from the two measurement days was used. If only one valid measurement point from one measurement day was available, this value was used. Correlation between DLdp abs and DLdp range: Regression analysis showed a distinct positive correlation between the two measures which can be used for description of the MOCR (r = 0.88, p = 0.000), indicating that comparable overall findings should be found using either variable (Fig. 5). Influence of DPOAE fine structure on MOCR: Generally, MOCR values were greater in frequencies with distinct fine structure dips(Figs. 6 and 7). In detail, the mean of DLdp abs over all subjects and level combinations was 2.17 dB in fdip as compared to 1.29 dB in fcontrol. Similarly, DLdp range was higher in fdip (mean of 9.2 dB) than in fcontrol (mean of 5.7 dB) (Fig. 8). Individual DLdp abs was greater in fdip than

W. Wagner et al. / Hearing Research 223 (2007) 83–92

87

Fig. 7. Mean value and standard deviation of MOCR in fdip and fcontrol: (a) evaluation by DLdp abs; (b) evaluation by DLdp range. Fig. 4. Test–retest-repeatability of the MOCR effect (r = 0.94). Circles: measurement points at fdip; asterisks: fcontrol.

Fig. 5. Relationship between DLdp abs and DLdp range (r = 0.88). Circles: measurement points at fdip; asterisks: fcontrol.

in fcontrol in 14/22 (63.6%) of the subjects, and individual DLdp range was greater in fdip than in fcontrol in 17/22 subjects (77.3%). For statistical analysis, oneway variance analysis

was used, by grouping MOCR values both according to L1 and according to L2. The difference between MOCR values in fdip and fcontrol proved to be statistically significant, both with consideration of DLdp abs and DLdp range (a < 0.05). In line with these results, MOCR magnitude was positively correlated with the depth of the DPOAE fine structure in the corresponding frequency. Regression analysis demonstrated a weak however significant positive correlation between DLdp abs and ddip (r = 0.41, p = 0.005), and between DLdp range and ddip (r = 0.44, p = 0.003) (Fig. 8). Influence of primary tone levels on MOCR: The overall finding was that the primary tone levels had a critical influence on the measured MOCR effect (example in Fig. 6). The change of DLdp with variation of L1 by 1 dB averaged 1.3 dB in fdip, and 0.8 dB in fcontrol, respectively, with maximum DLdp changes over the 121 level combinations of averagely 5.7 dB (fdip) and 3.7 dB (fcontrol). L2 variation by 1 dB changed DLdp by averagely 0.9 dB in fdip as opposed to 0.6 dB in fcontrol, with maximum DLdp changes of averagely 4.4 dB (fdip) and 3.0 dB (fcontrol). The highest individual DLdp change found after a 1 dB step of L1 amounted to 23 dB, the largest individual DLdp range was 25.2 dB.

Fig. 6. Measurement example (subject #1) of MOCR with 121 primary tone level combinations. Left: fdip, right: fcontrol.

88

W. Wagner et al. / Hearing Research 223 (2007) 83–92

Fig. 8. Relationship between depth of DPOAE fine structure (ddip) and MOCR effect: (a) evaluation by DLdp abs (r = 0.41); (b) evaluation by DLdp range (r = 0.44). Circles: measurement points at fdip; asterisks: fcontrol.

A change from suppression to enhancement or vice versa over the 121 level combinations was found in 100% of the subjects, both in fdip and fcontrol. In some subjects the change of sign occurred even several times within the level matrix. In four ears (18%) the change from maximum to minimum DLdp occurred within a few dB steps (so called bipolar effects, Maison and Liberman, 2000). On average, DLdp abs was slightly greater at lower primary tone levels L1 and L2, as expected. However, this relationship was more clearly apparent for fdip than for fcontrol, and, as the primaries were varied only in a range of 10 dB in the CAS measurements, the primary tone level dependent differences were small and inconstant (Fig. 9). In general, we found no primary tone level constellations which were typically associated with extreme DLdp values. Side-related effects: The mean of DLdp abs was larger in the dip frequencies of the left ears (2.31 dB) as compared to the right ears (2.0 dB). This difference was significant in ANOVA-testing (p = 0.000). ddip in the dip frequencies, which positively correlates with DLdp abs, was also greater on the left side (mean of 16.5 dB left, 15.6 dB on the right). The latter side difference, however, was not statistically sig-

nificant. In the control frequencies we observed no siderelated differences. 4. Discussion There are two major findings of the present study: (1) Largest MOCR effects can be found in frequencies which exhibit a distinct dip in DPOAE fine structure in humans. The positive correlation appeared both when considering the single suppression/enhancement values (DLdp), and when evaluating the range between the extrema of the effect (DLdp range). (2) The levels of the primary tones have a critical influence on the magnitude of the MOCR effect, as measured by CS of DPOAE. MOCR changed up to 23 dB when varying L1 by only 1 dB. The largest individual DLdp range was as high as 25.2 dB. To our knowledge, these are the largest primary tone level dependent CS changes reported in humans so far. Averages of the maximum CS change per 1 dB step in our subjects were still in the 3– 5 dB-range. Explanations for the large effects we found may be the extensive stimulus matrix with 121 primary tone level combinations which makes it more likely to detect

Fig. 9. Mean values of DLdp over the 121 primary tone level combinations: (a) fdip; (b) fcontrol.

W. Wagner et al. / Hearing Research 223 (2007) 83–92

extreme measurement points, and, most importantly, the fact that we specifically selected dip-frequencies for CS measurement. The effect magnitude resembles the data from guinea pigs (Maison and Liberman, 2000; DLdp range > 30 dB, as shown in an example in Fig. 6), which, however, refer to IA and not to CS, and are therefore not directly comparable with our findings. For IA in guinea pigs, enhancements were preferably observed when L2 was smaller than L1, while the effect changed to suppression with L2 moving towards L1 level (Maison and Liberman, 2000). In our data, however, no such rule became apparent (see Fig. 9). Occurrence of both suppression and enhancement effects as response to CAS, as we found it, has been previously described by different groups (Lisowska et al., 2002; Janssen et al., 2003; Wagner et al., 2005). Beyond this observation, variation of the primary tone levels led to a change from suppression to enhancement or vice versa in 100% of the subjects, in dip frequencies as well as in frequencies with flat fine structure. The same effect has been reported for IA in humans (Meinke et al., 2005). In some of our subjects the change of sign occurred even several times within the level matrix. In 18% of the ears the change from maximum to minimum DLdp occurred within a few dB steps, resulting in a DLdp/DL-function which was not sinusoidal but exhibited a point of reversal with a steep outline edge. These bipolar effects are similar to the ones which have been found for IA in guinea pigs (Maison and Liberman, 2000; Halsey et al., 2005). The actual findings are well compatible with a recent study of our group (Mu¨ller et al., 2005), where a different protocol with greater stimulus tone level intervals (L2 varied in 5 dB steps, L1 in 2 dB steps) within a broader level range (L2 = 20–60 dB SPL) also demonstrated a distinct dependence of MOCR on stimulus tone level and greatest CS effects in frequencies with distinct dips in DPOAE fine structure. The current paper, however, presents new data using a higher stimulus tone level resolution and a larger number of subjects. DPOAE comprise the sum of two intracochlear sources (Kim, 1980; Heitmann et al., 1998; Shera and Guinan, 1999), which interact constructively and destructively depending on DPOAE level and phase, and may be differentially influenced by MOCB-activity. Both dips in DPOAE fine structure and in DPOAE input–output-functions are thought to reflect this phenomenon (Maison and Liberman, 2000; Luebke and Foster, 2002). Accordingly, we propose the two-source generation of DPOAE to be an underlying mechanism for both major findings of our study as described above. The observations could be explained in the following way: in dips of the DPOAE fine structure the two DPOAE sources are almost cancelled due to similar magnitude and different phase polarity (e.g. source #1 amplitude: 22 lPa [=0.8 dB SPL], source #2 amplitude: 20 lPa [=0.0 dB SPL]). This results in a very small absolute amplitude of the resulting composite DPOAE signal (22  20 lPa= 2 lPa [=20.0 dB SPL]).

89

If one of the two DPOAE sources is now influenced by the MOCB-activity in a slightly different magnitude than the other one (e.g. #1CAS: 24 lPa [=1.6 dB SPL], #2CAS: 20 lPa [=0.0 dB SPL]), this results in a high absolute change in DPOAE level (24  20 lPa = 4 lPa [=14.0 dB SPL]; DLdp = 6.0 dB). Granted that the second DPOAE source amplitude is constant, an increase (decrease) of the first source DPOAE amplitude due to MOCB-activity leads to enhancement (suppression) of the DPOAE level. The magnification is highest in the notch and is much less present in peaks or high level regions of the DPOAE fine structure where signals of the DPOAE sources add and therefore result in a high absolute amplitude of the composite DPOAE signal (with the same values as for the example above but with constructive interference of the two sources: 22 lPa + 20 lPa = 42 lPa [=6.4 dB SPL]). In this case a small MOCB-induced difference in change of amplitude of the two DPOAE sources only results in a small absolute change in DPOAE level (24 lPa + 20 lPa = 44 lPa [=6.8 dB SPL], DLdp = 0.4 dB). However, we propose that medial efferent activity does not differ as much between different frequency and amplitude regions as the high variation of the MOCR values seems to indicate. Instead, maximum MOCR measurement points may represent sections in which MOCBactivity exerts the maximum measurable effect. Two different groups (Maison and Liberman, 2000; Luebke and Foster, 2002) found a relationship between MOCR strength as measured by maximum DLdp values, and another physiological parameter, namely noise vulnerability. This is a strong argument in favour of the view, that the MOCR extrema represent a useful metric for individual MOCB activity. To further clarify the influence of the two-source generation of DPOAE on MOCR measurement, CS measurements with simultaneous suppression of the second source (Heitmann et al., 1998) are planned for the future. In the literature we found 16 studies on CS of DPOAE in humans, starting in 1992. In 12 of these, stimulus frequency was varied whereas primary tone levels were held constant (Bassim et al., 2003; James et al., 2002; Kim et al., 2002; Lisowska et al., 2002; Di Girolamo et al., 2001; Sasaki et al., 2000; Timpe-Syverson and Decker, 1999; Williams and Brown, 1995; Moulin et al., 1992; Nieschalk et al., 1997; Chery-Croze et al., 1993; Moulin and Carrier, 1998). Reported suppression effects ranged from <1 to a few dB. Generally, larger MOCR effects were observed with lower primary tone levels, higher CAS levels, with broad band noise as CAS as compared to narrow band noise or sinusoidal tones, and in the frequency range 62 kHz as compared to the higher frequencies. Four groups applied systematic primary tone level variation: 5-dB steps were used by Sliwinska-Kowalska and Kotylo (2002) (L1 = L2 = 55–70 dB SPL, f2 = 4 kHz) and Moulin et al. (1993) (L1 = L2 = 30–80 dB SPL, f2 = 1.4 kHz). Janssen et al. (2003) varied L2 in 10 dB steps from 20 to

90

W. Wagner et al. / Hearing Research 223 (2007) 83–92

60 dB, while changing L1 in 4 dB steps (according to the scissor paradigm L2 · 0.4 + 39 dB; f2 = 2 kHz). In none of these studies large CS changes resulting from small primary tone level changes, as observed in this study, are reported. However, the cited authors did not specifically search for primary tone level dependent CS changes as it was not within the scope of these studies, and the level steps were substantially larger than the 1-dB steps used in the present study. Thus, level dependent CS changes may have been missed. The study with the highest primary tone level resolution was reported by Williams and Brown (1997), who varied L2 in 2 dB steps from 20 to 70 dB SPL with L1 = 55 dB SPL (trial 1), and L1 in 2 dB steps from 30 to 70 dB SPL with L2 = 40 dB SPL (trial 2). Additionally, four frequencies between 1587 and 6350 Hz were measured. The authors observed a maximum CS change of ca. 3 dB following a L1 change of 2 dB (at L1/L2 = 62/40 dB SPL, estimated from a figure in the manuscript), which is considerably smaller than the effects found in our measurements. It is important to note, that other types of OAE (e.g. click-evoked OAE, stimulus frequency OAE, for an overview see Guinan et al., 2003) are also established tools for assessing medial efferent activity, however comparison between the advantages and disadvantages of the different OAE types for this purpose was not within the scope of this paper. The measurements in this study were comparatively extensive with regard to the fine variation of primary tone levels in 1 dB steps with 121 level combinations, the complete duplicate measurement on a second day, and the number of 22 subjects. As a tribute to practicability, CS was measured only in one ear of each subject, after selection of the largest and most valid fine structure dip of both ears. In unilateral measurements, the MOCR effect is exerted by activity of the ipsilaterally-responding crossed bundle (activated by the primary tones) and the contralaterally-responding uncrossed bundle (activated by CAS). Hence, even unilateral measurements provide – albeit limited – bilateral information on the MOCB. With CAS held constant, the contribution of the ipsilateral bundle to the overall effect grows with decreasing primary tone level. To judge the diagnostic value of a unilateral approach, it is important to know to what extent the MOCR in both ears of one individual do correlate. Few data are available on this question, however two groups (Khalfa et al., 1997; Kumar and Vanaja, 2004) reported slightly greater values of CS of TEOAE in right than in left ears in human subjects. The authors interpreted their results as arguments in favour of peripheral auditory lateralization especially at OHC and MOCB levels. We, on the contrary, observed slightly higher CS values in the left ears as compared to the right ears, which was in line with the slightly greater depth of the dip in fdip of the left ears. Our data, however, did not allow analysis of intraindividual correlation between right and left ears, as CS was measured only in one ear of each subject. Anyhow, because of the close intraindividual correlation between activity of the crossed and the uncrossed

system demonstrated in guinea pigs (Kujawa and Liberman, 2001), we regard the unilateral approach as an acceptable strategy. Test–retest-repeatability was generally good, especially considering the condition of re-positioning the ear probe on another measurement day, and the low minimum stimulus tone levels of L1/L2 = 47/20 dB SPL. This confirms the significance of the MOCR effect as a well repeatable parameter which is characteristic for each individual. In all assays using acoustic stimulation, the possibility of elicitation of the SR with subsequent interference of efferent effects and middle-ear mediated DPOAE suppression must be taken into account. Regarding this aspect substantial differences between species exist. On the one hand there is little doubt that the suppression effects on otoacoustic emissions are mediated by olivocochlear efferent activity in guinea pigs (Puel and Rebillard, 1990; Kujawa and Liberman, 2001) and cats (Liberman et al., 1996). On the other hand, similar effects might be largely due to middle-ear reflex activity in rabbits (Whitehead et al., 1991) and rats (Relkin et al., 2005). Concerning humans, however, it has been demonstrated that the SR is not involved in CS of OAE, when CAS levels similar to the one used in the present study are applied (e.g. Janssen et al., 2003). Furthermore, elicitation of the SR would be expected to result in a general reduction of the DPOAE level for all ipsilateral measurement conditions, while we observed not only DPOAE suppression but also enhancements. Therefore we propose that the contribution of middle-ear reflex to CS effects can be largely excluded in the present study. As an interesting additional finding, we found remarkably high amplitude/frequency gradients within the DPOAE fine structure, with mean DP amplitude changes in the greatest dips in the range of ca. 220–350 dB/octave (ca. 4–7 dB/47 Hz step), and a maximum gradient in one individual of 1351 dB/octave (21.6 dB/47 Hz step). Only Dhar and Shaffer (2004) reported even greater gradients in the range 800–2700 dB/octave with application of an extreme high fine structure resolution of 4–12 Hz (values estimated from Fig. 1 of the manuscript as data not given in text). Gaskill and Brown (1990) found slopes of 30– 420 dB/octave using a lower frequency resolution of 1/3 octave. Obviously the observed gradients in DPOAE amplitude vs. frequency increase with finer frequency resolution and with decreasing stimulus levels. Regarding the latter aspect it is important to note, that in the present study the individual fine structure remained qualitatively unchanged when the stimulus level was altered, only the magnitude of the dips changed (see Fig. 1). DLdp and DLdp range showed a close relationship, obviously reflecting the same phenomena. For this reason either approach of data evaluation appears appropriate. In our data, however, the difference between the efferent effects was somewhat larger when considering DLdp range, which may suggest to use rather this parameter in an effort to characterize an individual’s maximum MOCR strength.

W. Wagner et al. / Hearing Research 223 (2007) 83–92

5. Conclusions The present data demonstrate that (1) the magnitude of the MOCR effect, as monitored by CS of DPOAE, is critically dependent on DPOAE fine structure in humans. Largest effects are typically found in frequencies exhibiting a distinct fine structure dip. (2) With fine variation of the primary tone levels, a change from suppression to enhancement or vice versa regularly occurs, resulting in large ranges between maximum and minimum MOCR values. Within the dip frequencies, CS changes of up to 23 dB could be observed after a change of a primary tone level by only 1 dB, which represents the highest Dstimulus tone level vs. DCS gradients reported in humans so far. It is proposed that assessing the MOCR at fixed frequencies and with only a single primary tone level combination will eventually miss maximum MOCR values, and may therefore not always be appropriate. Instead, measuring olivocochlear feedback in individually selected dip frequencies with fine variation of the primary tone levels may allow for a more targeted search for an individual’s entire range of MOCR strength, while at the same time being feasible according to measurement time. Literature data concerning putative physiological effects of the medial efferent bundle, such as improvement of sound discrimination in background noise and noise protection of the cochlea, are partially contradictory. Future studies must show, if the use of the ‘‘dip paradigm’’ proposed here can add new information to yet unclear questions on the role of the OCB. Acknowledgements This work was supported by the Deutsche Forschungsgemeinschaft (Wa 1677/2-1, Ja 597/7). We would like to thank the reviewers for their helpful comments and particularly one reviewer for giving an idea for understanding the influence of the two DPOAE sources on the suppression/ enhancement effect during CAS. References Bassim, M.K., Miller, R.L., Smith, D.W., Buss, E., 2003. Rapid adaptation of the 2f1  f2 DPOAE in humans: binaural and contralateral stimulation effects. Hearing Research 182, 140–152. Boege, P., Janssen, T., 2002. Pure-tone threshold estimation from extrapolated distortion product otoacoustic emission I/O-functions in normal and cochlear hearing loss ears. Journal of the Acoustical Society of America 111, 1810–1818. Chery-Croze, S., Moulin, A., Collet, L., 1993. Effect of contralateral sound stimulation on the distortion product 2f1  f2 in humans: evidence of a frequency specificity. Hearing Research 68, 53–58. Collet, L., Kemp, D.T., Veuillet, E., Duclaux, R., Moulin, A., Morgon, A., 1990. Effect of contralateral auditory stimuli on active cochlear micro-mechanical properties in human subjects. Hearing Research 43, 251–261. Dhar, S., Shaffer, L.A., 2004. Effects of a suppressor tone on distortion product otoacoustic emissions fine structure: why a universal suppressor level is not a practical solution to obtaining single-generator DPgrams. Ear and Hearing 25, 573–585.

91

Di Girolamo, S., D’Ecclesia, A., Quaranta, N., Garozzo, A., Evoli, A., Paludetti, G., 2001. Effects of contralateral white noise stimulation on distortion product otoacoustic emissions in myasthenic patients. Hearing Research 162, 80–84. Galambos, R., 1956. Suppression of auditory nerve activity by stimulation of efferent fibers to cochlea. Journal of Neurophysiology 19, 424–437. Gaskill, S.A., Brown, A.M., 1990. The behavior of the acoustic distortion product, 2f1  f2, from the human ear and its relation to auditory sensitivity. Journal of the Acoustical Society of America 88, 821–840. Guinan, J.J.J., Backus, B.C., Lilaonitkul, W., Aharonson, V., 2003. Medial olivocochlear efferent reflex in humans: otoacoustic emission (OAE) measurement issues and the advantages of stimulus frequency OAEs. Journal of the Association for Research in Otolaryngology (JARO) 4, 521–540. Halsey, K., Dolan, D.F., Skjo¨nsberg, A., Ulfendahl, M., 2005. Efferentmediated adaptation of the DPOAE as a predictor of aminoglycoside toxicity. Hearing Research 201, 99–108. Heitmann, J., Waldmann, B., Schnitzler, H.U., Plinkert, P.K., Zenner, H.P., 1998. Suppression of distortion product otoacoustic emissions (DPOAE) near 2f1  f2 removes DP-gram fine structure – evidence for a secondary generator. Journal of the Acoustical Society of America 103, 1527–1531. James, A.L., Mount, R.J., Harrison, R.V., 2002. Contralateral suppression of DPOAE measured in real time. Clinical Otolaryngology and Allied Sciences 27, 106–112. Janssen, T., Gehr, D., Kevanishvili, Z., 2003. Contralateral DPOAE suppression in humans at very low sound intensities. In: Gummer, E. (Ed.), Biophysics of the Cochlea: From Molecules to Models. World Scientific, New Jersey, pp. 498–505. Khalfa, S., Morlet, T., Micheyl, C., Morgon, A., Collet, L., 1997. Evidence of peripheral hearing asymmetry in humans: clinical implications. Acta Oto-Laryngologica 117, 192–196. Kim, D.O., 1980. Cochlear mechanics: implications of electrophysiological and acoustical observations. Hearing Research 2, 297–317. Kim, D.O., Dorn, P.A., Neely, S.T., Gorga, M.P., 2001. Adaptation of distortion product otoacoustic emissions in humans. Journal of the Association for Research in Otolaryngology (JARO) 2, 31–40. Kim, S., Frisina, D.R., Frisina, R.D., 2002. Effects of age on contralateral suppression of distortion product otoacoustic emissions in human listeners with normal hearing. Audiology and Neuro-Otology 7, 348– 357. Kujawa, S.G., Liberman, M.C., 2001. Effects of olivocochlear feedback on distortion product otoacoustic emissions in guinea pig. Journal of the Association for Research in Otolaryngology (JARO) 2, 268–278. Kumar, U.A., Vanaja, C.S., 2004. Functioning of olivocochlear bundle and speech perception in noise. Ear and Hearing 25, 142–146. Liberman, M.C., Guinan Jr., J.J., 1998. Feedback control of the auditory periphery: anti-masking effects of middle ear muscles vs. olivocochlear efferents. Journal of Communication Disorders 31, 471–483. Liberman, M.C., Puria, S., Guinan Jr., J.J., 1996. The ipsilaterally evoked olivocochlear reflex causes rapid adaptation of the 2f1  f2 distortion product otoacoustic emission. Journal of the Acoustical Society of America 99, 3572–3584. Lisowska, G., Smurzynski, J., Morawski, K., Namyslowski, G., Probst, R., 2002. Influence of contralateral stimulation by two-tone complexes, narrow-band and broad-band noise signals on the 2f1  f2 distortion product otoacoustic emission levels in humans. Acta Oto-Laryngologica 122, 613–619. Luebke, A.E., Foster, P.K., 2002. Variation in inter-animal susceptibility to noise damage is associated with?9 Acetylcholine receptor subunit expression level. Journal of Neuroscience 22, 4241–4247. Maison, S.F., Liberman, M.C., 2000. Predicting vulnerability to acoustic injury with a noninvasive assay of olivocochlear reflex strength. Journal of Neuroscience 20, 4701–4707. Meinke, D.K., Stagner, B.B., Martin, G.K., Lonsbury-Martin, B.L., 2005. Human efferent adaptation of DPOAEs in the L(1), L(2) space. Hearing Research 208, 89–100.

92

W. Wagner et al. / Hearing Research 223 (2007) 83–92

Moulin, A., Carrier, S., 1998. Time course of the medial olivocochlear efferent effect on otoacoustic emissions in humans. NeuroReport 9, 3741–3744. Moulin, A., Collet, L., Morgon, A., 1992. Influence of spontaneous otoacoustic emissions (SOAE) on acoustic distortion product input/output functions: does the medial efferent system act differently in the vicinity of an SOAE? Acta Oto-Laryngologica 112, 210–214. Moulin, A., Collet, L., Duclaux, R., 1993. Contralateral auditory stimulation alters acoustic distortion products in humans. Hearing Research 65, 193–210. Mu¨ller, J., Janssen, T., Heppelmann, G., Wagner, W., 2005. Relationship between fine-structure, contralateral suppression, and ipsilateral adaptation of distortion product otoacoustic emissions in humans. Journal of the Acoustical Society of America 118, 3747–3756. Murugasu, E., Russell, I.J., 1996. The effect of efferent stimulation on basilar membrane displacement in the basal turn of the guinea pig cochlea. Journal of Neuroscience 16, 325–332. Nieschalk, M., Beneking, R., Stoll, W., 1997. Die Pegelzunahme von Distorsionsproduktemissionen des Menschen durch kontralaterale Beschallung niedriger Intensitat. HNO 45, 378–384. Puel, J.L., Rebillard, G., 1990. Effect of contralateral sound stimulation on the distortion product 2f1  f2: evidence that the medial efferent system is involved. Journal of the Acoustical Society of America 87, 1630–1635. Rajan, R., 1995. Frequency and loss dependence of the protective effects of the olivocochlear pathways in cats. Journal of Neurophysiology 74, 598–615. Relkin, E.M., Sterns, A., Azeredo, W., Prieve, B.A., Woods, C.I., 2005. Physiological mechanisms of onset adaptation and contralateral suppression of DPOAEs in the rat. Journal of the Association for Research in Otolaryngology (JARO) 6, 119–135. Sasaki, N., Kawase, T., Ogura, M., Takasaka, T., 2000. The time-course of the effects of contralateral sound on the level of distortion product otoacoustic emissions. Tohoku Journal of Experimental Medicine 191, 71–78.

Scharf, B., Magnan, J., Chays, A., 1997. On the role of the olivocochlear bundle in hearing: 16 case studies. Hearing Research 103, 101–121. Shera, C.A., Guinan Jr., J.J., 1999. Evoked otoacoustic emissions arise by two fundamentally different mechanisms: a taxonomy for mammalian OAEs. Journal of the Acoustical Society of America 105, 782–798. Sliwinska-Kowalska, M., Kotylo, P., 2002. Occupational exposure to noise decreases otoacoustic emission efferent suppression. International Journal of Audiology 41, 113–119. Timpe-Syverson, G.K., Decker, T.N., 1999. Attention effects on distortion-product otoacoustic emissions with contralateral speech stimuli. Journal of the American Academy of Audiology 10, 371–378. Veuillet, E., Collet, L., Duclaux, R., 1991. Effect of contralateral acoustic stimulation on active cochlear micromechanical properties in human subjects: dependence on stimulus variables. Journal of Neurophysiology 65, 724–735. Wagner, W., Heppelmann, G., Kuehn, M., Tisch, M., Vonthein, R., Zenner, H.P., 2005. Olivocochlear activity and TTS-susceptibility in man. The Laryngoscope 115, 2021–2028. Warr, W.B., Guinan Jr., J.J., 1979. Efferent innervation of the organ of Corti: two separate systems. Brain Research 173, 152–155. Whitehead, M.L., Martin, G.K., Lonsbury-Martin, B.L., 1991. Effects of the crossed acoustic reflex on distortion-product otoacoustic emissions in awake rabbits. Hearing Research 51, 55–72. Wiederhold, M.L., 1970. Variations in the effects of electric stimulation of the crossed olivocochlear bundle on cat single auditory nerve-fiber responses to tone bursts. Journal of the Acoustical Society of America 48, 966–977. Williams, D.M., Brown, A.M., 1995. Contralateral and ipsilateral suppression of the 2f1  f2 distortion product in human subjects. Journal of the Acoustical Society of America 97, 1130–1140. Williams, D.M., Brown, A.M., 1997. The effect of contralateral broadband noise on acoustic distortion products from the human ear. Hearing Research 104, 127–146. Winslow, R.L., Sachs, M.B., 1987. Effect of electrical stimulation of the crossed olivocochlear bundle on auditory nerve response to tones in noise. Journal of Neurophysiology 57, 1002–1021.