Acta Oto-Laryngologica, 2008; 128: 53
60

ORIGINAL ARTICLE

Speech-in-noise intelligibility does not correlate with efferent

olivocochlear reflex in humans with normal hearing

WOLFGANG WAGNER, KATHRIN FREY, GUIDO HEPPELMANN,

STEFAN K. PLONTKE & HANS-PETER ZENNER

Tübingen Hearing Research Center, Department of Otolaryngology/Head and Neck Surgery, University of Tübingen,
Germany

Abstract
Conclusion. According to the presented data, speech-in-noise intelligibility (SI) does not correlate with olivocochlear efferent
activity
as measured by contralateral suppression (CS) of distortion product otoacoustic emissions (DPOAE)
in humans
with normal auditory threshold. Objectives. Literature data indicate a possible role of the medial olivocochlear efferents in
speech intelligibility, especially in background noise. The objective of this study was to investigate this relationship. Materials
and methods. SI was evaluated in three independent sessions by determining the ratio speech level/noise level, at which 50%
of the words are understood (i.e. speech reception threshold, SRT). Efferent activity was inferred measuring CS of DPOAE,
using two different paradigms with extensive variation of stimulus parameters and duplicate measurements. Results. For
optimum measurement of CS, the study was restricted to subjects (n 49) with valid DPOAE down to primary tone levels
L147/L220 dB SPL. Average SRT was 6.66 dB ( 4.50 to 7.65 dB, SD 0.63 dB). CS increased with decreasing
primary tone levels, with mean absolute CS values in the range of 0.6
6 dB SPL. Test
retest repeatability was good.
Statistical evaluation revealed no significant relationship between SI and CS of DPOAE.

Keywords: Olivocochlear efferents, medial olivocochlear reflex, speech intelligibility, discrimination, background noise,
distortion product otoacoustic emissions

Introduction on stimulus parameters. In humans, ipsilateral

The olivocochlear bundle (OCB) consists of efferent
neurons projecting from the upper olivary region in
the brainstem to the cochlea. It is composed of a
lateral and a medial part, which both possess crossed
and uncrossed fibers [1]. While the lateral neurons
synapse with afferent dendrites of the inner hair
cells, the medial olivocochlear (MOC) efferents
terminate directly upon outer hair cells. Therefore,
activation of the efferent bundle by acoustic stimula-
tion leads to amplitude changes of otoacoustic
emissions, the so-called MOC reflex [2]. The effect
can be measured on the side of acoustical stimula-
tion (referred to as ipsilateral adaptation), or con-
tralaterally (contralateral suppression, CS). CS was
found to be more commonly a decreasing effect
(suppression) than an increasing effect (enhance-
ment) and typically amounts to some dB, depending
adaptation effects are generally smaller than CS
values and mostly range below 1 dB [3]. Both CS
and ipsilateral adaptation are regarded as metrics for
efferent activity [4]. In humans, however, CS was
demonstrated to have a higher measurable incidence
and a better reproducibility than ipsilateral adapta-
tion, which recommends CS as the method of choice
for measurement of the MOC reflex in humans [5].
An extensive neurophysiological literature demon-
strates that the MOC efferents attenuate cochlear
response to sound [6,7], presumably by reducing
the gain of the outer hair cell mechanical res-
ponse to acoustic stimulation [8]. It has been
shown that experimental activation of the efferents
via acoustic stimulation can influence the ability
of frequency discrimination in the cat [9], and
sound intensity discrimination in the cat [10
13]

Correspondence: Wolfgang Wagner, MD, Tübingen Hearing Research Center, Department of Otolaryngology/Head and Neck Surgery, University of
Tübingen, Elfriede-Aulhorn-Str. 5, 72076 Tübingen, Germany. Tel: 49 7071 2988088. Fax: 49 7071 293311. E-mail: w.wagner@med.uni-tuebingen.de

(Received 16 January 2007; accepted 28 February 2007)

ISSN 0001-6489 print/ISSN 1651-2551 online # 2008 Taylor & Francis
DOI: 10.1080/00016480701361954
54 W. Wagner et al.
and human [14,15]. Consistently, the influence of
acoustic stimulation on discrimination was dimin-
ished through surgical section of the human OCB
[16]. It was proposed that efferent activity supports
the ability of sound discrimination and low tone
detection in background noise [12,17,18], including
selective attention to frequencies [19].
However, discrepant data exist according to the
psychoacoustic abilities after OCB section. While
frequency and intensity discrimination in back-
ground noise deteriorated in surgically de-efferented
cats [9,20], no such impairment was observed in
human patients whose OCB was sectioned in
vestibular neurotomy [19,21]. Accordingly, conclu-
sive evidence for a correlation between efferent
activity and sound discrimination in noise, which is
closely related to speech-in-noise intelligibility (SI),
is still pending in humans. The goal of the present
study was to further evaluate this relationship.
For assessment of efferent activity, we measured
the change of DPOAE level during acoustic stimula-
tion of the contralateral ear (i.e. CS of DPOAE).
The contralateral stimulus of 60 dB SPL broadband
noise proved to be well below the subjects’ stapedial
reflex thresholds. In 13 of the 49 subjects, CS
was registered using stimulus parameters that are
comparable to those used in previously described
audiologic paradigms (f2 1
6 kHz in 1 kHz steps,
L2 60, 50, 40, 35, 30, 25, and 20 dB SPL,
L1 L2 0.439 dB). Such a paradigm covers a
relatively wide range of stimulus frequencies and
primary tone levels, however, accepting relatively
wide intervals of stimulus frequency and level.
Thereby large efferent effects may eventually be
overlooked. It was recently demonstrated that ex-
treme CS values occur preferably at frequencies at
which DPOAE level exhibits a pronounced sharp-
edged minimum (‘dip’) in a DP gram measured in
fine frequency steps [22]. The phenomenon of
maximum CS values in ‘dip frequencies’ can be
explained by interference of two separate sources
along the basilar membrane generating the DPOAE
signal (see, for example, the study by Heitmann et al.
[23]). According to these findings, in 36 of the
49 subjects, CS of DPOAE was not measured over a
wide frequency range, but at dip frequencies that
were individually selected for each subject. Within
these frequencies, primary tone levels were varied in
very fine resolution (121 primary tone level combi-
nations with L1 50
60 dB SPL in 1 dB steps, L2  measurements. The investigations were performed in
35
45 dB SPL in 1 dB steps). In both paradigms,
CS measurements were repeated on average 5.5 days
later.
Materials and methods
Subjects
Forty-nine subjects participated in the study (30
female, 19 male, age 19.7
41.7 years, mean 25.2
years). Inclusion criteria comprised absence of ear
disease in the medical history, normal findings in ear
microscopy, normal tympanometry, stapedial reflex
thresholds of ]80 dB SPL (measured with tones),
and bilaterally recordable DPOAE between f2 1
and 6 kHz with a minimum signal to noise ratio
(SNR) of 6 dB at primary tone levels L2 60, 50,
40, 35, 30, 25, and 20 dB SPL, L1L2*0.4
39 dB SPL. This primary tone level setting (‘scissor
paradigm’) takes into account the nonlinear interac-
tion of the two primary tones at the DPOAE
generation site at the f2 place [24].
General measurement procedures
Auditory threshold was measured at 0.125, 0.25,
0.5, 0.75, 1, 1.5, 2, 3, 4, 6, and 8 kHz in both ears
(audiometer AT 900, Auritec GmbH, Hamburg,
Germany). For tympanometry and stapedial reflex
registration the MADSEN Zodiac 901 Middle-
Ear Analyzer (GN Otometrics A/S, Copenhagen,
Denmark) was used. The 2f1-f2 DPOAE were
recorded using a PCMCIA DSP card (Starkey,
Minnesota, USA) with an ER-10c ear probe
(Etymotic Research, Illinois, USA). Contralateral
acoustical stimulation (CAS) was delivered via a
FZ-PRC1 sound probe (Fischer-Zoth, Germering,
Germany), with a priori calibration in an ear
simulator (Type 4157, Brüel & Kjær, Nærum,
Denmark) and no correction accounting for the
individual ear canal volume. Measurement and
evaluation software was custom-made in cooperation
with Prof. Janssen (Department of Otorhinolaryn-
gology, Technical University of Munich, Germany),
using MATLAB (Mathworks, Natick, MA, 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 threshold measure-
ment were conducted 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 to the sound proof chamber, as verified by test
accordance with the principles of the Declaration of
Helsinki and were approved by the human subject
review committee of the University of Tübingen.

Assessment of Speech-in-noise intelligibility
SI was measured using the Oldenburg Sentence
Test, which is equal to the Danish DANTALE II
test (for detailed description see Wagener et al.
[25,26]). In short, the test uses a base list consisting
of 10 sentences with 5 words each (name, verb,
numeral, adjective, object, for example: ‘Peter kauft
achtzehn nasse Schuhe’ [‘Peter buys eighteen wet
shoes’]). The sentences do not contain meaningful
information, therefore they cannot be easily memor-
ized and there is no benefit from a context. The
background noise in the Oldenburg Sentence Test is
generated by 30-fold overlapping of temporal shifted
components of the entire test speech material. Thus,
the long time spectrum of this noise equals the
speech material, creating an optimum masking
effect. After presentation of each five-word sentence,
the subject is asked to repeat as many of the words as
possible. Noise level is held constant at 65 dB SPL.
Speech level, starting at 65 dB SPL, is raised if fewer
than two of the five words of the sentence are
repeated correctly (raising by 2 dB with none of
five words correct, and by 1 dB with one of five
words correct), and lowered if more than three of
five words are understood (by 2 dB with five of five
words correct, and by 1 dB with four of five words
correct). With two of five or three of five correctly
repeated words, speech level is kept unchanged. By
this method, the so-called speech reception thresh-
old (SRT) can be determined according to the
definition: SRTdifference (dB) speech level minus
noise level, at which 50% of the words are under-
stood.
In the present study, speech and background
noise (played from the commercially available CD
‘Oldenburger Satztest’) were presented synchro-
nously by one loudspeaker (AT 966-1, Auritec
GmbH, Hamburg, Germany) in free-field condi-
tion in a sound-proof chamber. The loudspeaker
was located in front of the subject at a distance of
1 m. The test was performed three times on three
different days, using different randomly selected
sentences out of the Oldenburg Sentence Test
battery for each session. The reason for the
triplicate measurements is the known training
effect during the first measurements, which is
due to a general familiarization with the measure-
ment procedure and the limited number of words
in the test material. However, it has been demon-
strated that there is no strong further training
effect after the third measurement [26]. Thus, the
SRT value emerging from the third test session
(SRT3) is considered an appropriate estimate of a
subject’s SI.
Speech intelligibility and efferent olivocochlear reflex 55
Assessment of MOC efferent activity
For evaluation of MOC efferent activity, CS of
DPOAE was measured under CAS of 60 dB SPL
broadband noise (frequency range of 100 Hz to
10 kHz), which was started 1 s before onset of the
primary tones and ended 0.5 s after their termina-
tion. Two different paradigms were used, as follows.
Paradigm A (DP growth function): in the first
13 subjects, CS of DPOAE was registered at
the frequencies f2 6, 5, 4, 3, 2, and 1 kHz,
with L2 60, 50, 40, 35, 30, 25, and 20 dB SPL,
L1 L2 0.439 dB. The averaging time for
DPOAE recording was set to 4 s and the pause
time between two measurement settings to 1 s. The
measurement was repeated on another day (on
average after 5.4 days, range 1
15 days), resulting
in 168 measurement values per subject (7 primary
tone level combinations 6 frequencies 2 ears 2
measurements). Paradigm B (dip frequency): in the
following 36 subjects, the paradigm was modified
due to recent findings of our group, indicating that
optimum conditions for MOC reflex measurements
are preferably found in frequencies with distinct
non-monotonocities (so-called dips) in DPOAE fine
structure [22]. DPOAE fine structure (i.e. DP level
versus frequency function, measured in high resolu-
tion of frequency) was measured bilaterally in 47 Hz
frequency steps at 2
6 kHz (Figure 1). From this, a
frequency with distinct dip (fdip) was selected from
one ear of each subject according to the following
criteria: (i) presence of the dip at three different
primary tone level settings (L2 40, 30, 20 dB SPL,
L1 L2 0.439 dB SPL); (ii) SNR ]10 dB at
10
L2 = 40 dB SPL
DPOAE level [dB SPL]
0
L2 = 30 dB SPL
-10
-20 L2 = 20 dB SPL
-30
-40
2 3 4 5 6
frequency f2 [kHz]
Figure 1. Example of a DP gram (DPOAE level over stimulus
frequency) measured in fine frequency steps of 47 Hz, revealing
distinct DP level maxima and minima, the so-called DPOAE fine
structure. DPOAE fine structure was measured at f2 2
6 kHz,
and with three primary tone level combinations, resulting in
the three DP level curves (squares, DP level at L2 40 dB
SPL; triangles, L2 30 dB SPL; diamonds, L220 dB SPL;
L1L20.439 dB SPL for all L2). The dashed lines on the
bottom depict the according three noise floor levels.
56 W. Wagner et al.
10
DPOAE level [dB SPL]
Ldp(2675 Hz) Ldp(2770 Hz)
5 -3
Ldp(2715 Hz)
0
L2=40 dB SPL
2.60 2.65 2.70 2.75 2.80
frequency f2 [kHz]
Figure 2. Section from DPOAE fine structure measure-
ment example in Figure 1, showing the frequency range
f2 2.6
2.9 kHz in higher magnification. A DP level minimum
(dip) at f2 2715 Hz, L2 40 dB SPL can be seen. The depth
of the dip is calculated from the differences in DP level bet-
ween the dip frequency and the adjacent higher and lower
frequencies, according to the equation: depth of the dip
(ddip) at f2 2715 Hz/L240 dB SPL [Ldp(f2 2675 Hz/
L240 dB SPL)Ldp(2715 Hz/40 dB)]dp(2770 Hz/40 dB)
Ldp(2715 Hz/40 dB)].
fdip at all three primary tone level settings; (iii)
maximum depth of the dip (ddip) (Figure 2). fdip
belonged to the right ear in 21 subjects and to the
left ear in 15 subjects. The mean frequency of fdip
was 3.37 kHz (range 2.02
5.63 kHz), with a mean
depth of the dip (arithmetic mean of the differ-
ences between fdip and the neighboring lower and
higher frequencies at all three primary tone level
settings, see Figure 2) of 16.9 dB (range 8.9

31.1 dB). CS of DPOAE was measured in fdip,
applying 121 different primary tone level combina-
tions (L1 50
60 dB SPL, L2 35
45 dB SPL,
variation in 1 dB steps). As in paradigm A, the
measurement was repeated on another day (on
average 6.2 days, range 1
35 days), resulting in
242 measurement values per subject.
Results
Audiometric thresholds
As a consequence of the inclusion criterion of valid
DPOAE down to primary tone levels of L1/L2 47/
20 dB SPL, pure tone thresholds were in a normal
range in all subjects. The average thresholds in
the frequency range 0.125
10 kHz were 5.5 dB HL
(0.1 to 22.5 dB HL) in the left ears, and 5.1 dB HL
(0.7 to 14.3 dB HL) in the right ears, respectively.
Speech-in-noise intelligibility
The Speech reception thresholds from the three
Oldenburg Sentence Test sessions are referred to as
SRT1, SRT2, and SRT3 in the following discussion.
0
speech reception thresholds [dB]
-1
-2
-4
-5
-6
-7
-8
2.85 2.90 -9
day 1 day 2 day 3
measurement day
Figure 3. Speech-in-noise intelligibility, as measured by the
Oldenburg Sentence Test. Sentences of five words are presented
in the presence of background noise. Speech reception threshold
(SRT) is defined as the difference in dB between speech level and
noise level at which the subject understands 50% of the presented
words. The test is performed three times on three different days to
take into account the known training effect. The result of the third
measurement is considered primarily. This figure presents the
SRT values of the three measurement days in box plots.
SRT1 was on average 5.62 dB (3.50 to
8.20 dB, SD 0.72 dB), SRT2 amounted to
5.83 dB (
4.05 to
7.05 dB, SD 0.6 dB), and
SRT3 had a mean value of 6.66 dB (4.50 to
7.65 dB, SD 0.63 dB) (Figure 3). As expected,
SRT values improved over measurement days, with
the improvement being statistically significant from
day 1 to day 3, and from day 2 to day 3 (SRT day
1/day 2: p0.118 (not significant), SRT day 2/day
3: pB0.001, SRT day 1/day 3: p B0.001).
No influence of the pure tone thresholds on SI
appeared, however these varied within a normal
range in all subjects. This was true both with
consideration of the thresholds at 0.5
6 kHz (corre-
lation coefficient between pure tone thresholds and
SRT3, r 0.174) (Figure 4), and with consideration
-4
speech reception threshold [dB]
-5
-6
-7
-8
0 2 4 6 8 10 12 14
average pure tone threshold 0.5 – 6 kHz [dB HL]
Figure 4. Scatter plot of the 49 subjects showing speech reception
threshold (SRT) over auditory threshold. No correlation between
SRT and pure tone threshold, which was within normal range in
all subjects, appeared.
 Speech intelligibility and efferent olivocochlear reflex 57

of the entire measured range of 0.125
10 kHz (r  approaches, which included calculation of correla-
0.105) (not shown).
MOC efferent activity
In both paradigms, MOC efferent activity was
evaluated according to: (i) the arithmetic mean of
all CS values of each subject; (ii) the arithmetic
mean of all absolute CS values of each subject
(referred to as CSabs); and (iii) the differ-
ence between maximum and minimum CS values
of each subject (CSrange). CS was on average
1.0 dB (2.7 to 0.0 dB, SD 0.8 dB) in paradigm
A, and 0.9 dB (6.0 to 3.1 dB, SD 2.0 dB) in
paradigm B. CSabs had a mean of 1.8 dB (1.1 to
3.0 dB, SD 0.5 dB) in paradigm A, and 2.4 dB (0.6
to 6.0 dB, SD 1.2 dB) in paradigm B. Average
CSrange amounted to 11.6 dB (6.9 to 17.5 dB, SD
3.0 dB) in paradigm A, and 9.1 dB (2.4 to 20.4 dB,
SD 4.4 dB) in paradigm B (Figure 5).
Test
retest repeatability was generally good, as
indicated by the repeatability coefficients, which
define the difference between two measurements,
which is exceeded in only 5% of all cases (CS:
repeatability coefficient of 0.67 dB in paradigm A,
and 1.21 dB in paradigm B; CSabs: 0.6 dB in
paradigm A, and 0.88 in paradigm B; CSrange:
4.84 dB in paradigm B). An exception was the low
repeatability of CSrange within paradigm A group
(repeatability coefficient of 8.23 dB), for which we
do not have an explanation.
Correlation between MOC efferent activity and
Speech-in-noise intelligibility
No correlation between MOC efferent activity and
SI became apparent. This was found in all evaluation
tion between: (i) SRT3 and CS (paradigm A,
r 0.117; paradigm B, r 0.247; common evalua-
tion of both paradigms, r 0.105) (Figure 6a); (ii)
SRT3 and CSabs (paradigm A, r 0.186; paradigm
B, r 0.125; both paradigms, r 0.087) (Figure 6b);
(iii) SRT3 and CSrange (paradigm A, r 0.302;
paradigm B, r 0.065; both paradigms, r 0.065)
(Figure 6c).
To ensure that any unexpected relationships were
not overlooked, the same evaluations were con-
ducted replacing SRT3 by (a) SRT1, (b) SRT2,
and (c) the arithmetic mean of SRT1, SRT2, and
SRT3. Similarly, none of these approaches revealed a
correlation between efferent activity and SI.
Discussion
Our data revealed no correlation between speech
intelligibility in background noise and MOC efferent
activity in 49 normally hearing subjects. Evaluation
of SI was based on measurements on 3 different
days. Within the three sessions the expected con-
tinuous improvement of SRT with comparable
learning curves in nearly all subjects was observed,
indicating the validity and plausibility of the test. It
has to be noted that SRT values varied over a
relatively small range among the subjects, which
makes it difficult to uncover a relationship with
another metric based on a correlation. Test
retest
repeatability of CS of DPOAE was generally good,
especially considering the condition of re-positioning
the ear probe on another measurement day, and
the low minimum primary tone levels of L1/L2 47/
20 dB SPL. This indicates that efferent activity
as a response to acoustical stimulation is a well

Paradigm A (“DP growth function”) Paradigm B (“dip frequency”)

4 8 22 4 8 22 (b)
(a)
contralateral suppression [dB]

20 20
2 18 2 18
6 16 6 16
0 14 0 14
12 12
4 4
10 10
-2 -2
8 8
2 6 2 6
-4 -4
4 4
2 2
-6 0 0 -6 0
CS CSabs CSrange CS CSabs CSrange

Figure 5. Amount of contralateral suppression (CS) of DPOAE, which served as measure of MOC efferent activity. CS of DPOAE was
evaluated using different arithmetical approaches (CS, real values, positive or negative; CSabs, absolute values; CSrange, difference between
maximum and minimum CS values of each subject), and with two different paradigms using different sets of DP primary tone levels and DP
measurement frequencies (for details see text).
58 W. Wagner et al.
(a) and, in consequence, in a large inter-individual CS
-4
range. This facilitates identification of the suppres-
speech reception threshold [dB]

sion effects, distinction from artefacts, and quantita-

-5 tive classification. As in previous studies, both
negative and positive CS values
i.e. DPOAE
suppressions and enhancements
were observed in
-6
the present work. Both effects can occur in one
individual at different frequencies or primary tone
-7 levels, leading to eventual arithmetical extinction in
averaging operations. However, both suppressions
-8
and enhancements are supposed to reflect efferent
-6 -5 -4 -3 -2 -1 0 1 2 3 4 activity. Therefore, the absolute CS values (CSabs) or
contralateral suppression [dB] the range between minimal and maximal CS
(CSrange) should be preferentially considered. So as
(b) not to overlook any effects, however, CS was also
-4
speech reception threshold [dB]

evaluated, without detecting any correlation with SI.

Evidence from a number of animal studies sug-
-5 gests a role of the olivocochlear efferents in sound
discrimination in background noise. Electrical sti-
-6 mulation of the OCB can improve the ability of
auditory neurons to change their firing rate as a
response to sound stimulus changes in background
-7 noise in the cat [12,13]. Background noise can
increase the sound-evoked firing rate of efferent
-8 fibers, thereby modulating the neurons’ dynamic
0 1 2 3 4 5 6 7 range, which influences sound discrimination [10].
contralateral suppression, absolute values [dB] It was concluded that sound discrimination in
background noise is improved by an efferent-
(c)
-4 mediated broadening of the dynamic range of the
auditory nerve fibers [11]. The aspect of a possible
speech reception threshold [dB]

efferent influence on SI was originally addressed in

-5
-6
-7
-8
0 2 4 6 8 10 12 14 16 18 20
contralateral suppression, range [dB]
Figure 6. Relationship between medial olivocochlear (MOC)
efferent activity, represented by contralateral suppression (CS)
of DPOAE, and speech-in-noise intelligibility, as measured by
speech reception threshold. Solid circles, paradigm A (DP growth
function); open circles, paradigm B (dip frequency). Efferent
activity was evaluated according to real values of CS of DPOAE
(CS) (a), absolute values of CS of DPOAE (CSabs) (b), and range
between minimum and maximum of CS of DPOAE (CSrange) (c).
Statistical evaluation revealed no correlation between the two
variables in either approach.
repeatable parameter, representing a constant indi-
vidual property of each subject.
The strategy of measuring MOC reflex in DPOAE
fine structure dips with high primary tone level
resolution results in especially large CS effects [5,22]
1968 by Dewson [27], who observed impairments in
vowel discrimination in noise in rhesus monkeys with
OCB lesions. In line with this observation, surgical
section of the OCB in cats has been shown to
deteriorate vowel discrimination [9] and intensity
discrimination in background noise [20], albeit only
under certain conditions. Hienz et al. [9] reported
that the effect was only present with SNR of 3 dB,
22 not with 13 or 23 dB; May and McQuone [20]
reported that the effect was only present at 8 kHz,
not at 1 kHz. On the other hand, the absence of
post-synaptic MOC efferent receptors in alpha9
acetylcholine receptor knockout mice does influence
tone detection and intensity discrimination in back-
ground noise [28]. This raised the hypothesis that
the central efferent pathways, whose post-synaptic
targets do not express alpha9 acetylcholine recep-
tors, work in combination with the peripheral
olivocochlear system to enhance hearing in noise,
and may compensate for manipulations of peripheral
feedback in routine testing procedures.
In humans, data from invasive electrophysiological
experiments are not available. Therefore one has
to rely on psychoacoustic studies. Micheyl and
Collet [14] observed a significant positive correlation

between CS of transiently evoked OAE (CAS  found no significant effect of CAS on loudness
30 dB SL) and detection thresholds for 1 and
2 kHz tones in background noise. In a recent study,
Kim et al. [29] investigated the relationship between
CS of DPOAE (CAS 30 dB SL) and (1) sentence
intelligibility in noise with speech and noise coming
from the same frontal direction (HINT08), and (2)
improvement of sentence perception when speech
and noise were spatially separated and emanated
from different directions (RFM). The authors re-
ported weak correlations between both of the two
parameters and CS of DPOAE; however, the corre-
lations appeared only in certain frequency bands (at
4
6 kHz for HINT08 and at 1
2 kHz for RFM), and
the reason for the finding of a positive correlation
between CS and RFM, as opposed to the reported
negative correlation between CS and HINT08,
remained unclear. In summary, the study by Kim
et al. did not demonstrate convincing evidence for a
relationship between SI and MOC activity among
the tested subjects.
Most of the other human studies refer to the
change of the ipsilateral discrimination ability in
background noise during stimulation of the efferent
system by CAS. It has been demonstrated that CAS
can improve intensity discrimination [15], tone
detection thresholds [14], and speech intelligibility
[16,18] in background noise. In all of the studies
mentioned above, the magnitude of the CAS-
mediated improvement was positively correlated
with CS of transiently evoked OAE. Section of the
OCB during vestibular neurotomy, which is even-
tually performed in severe cases of Ménière’s disease
to alleviate vertigo, abolished the effect [16]. In line
with this view, Zheng et al. [30] reported deteriora-
tion of SI in four of six patients after vestibular
neurotomy. However, tone detection thresholds and
pure tone discrimination in silence remained un-
changed in these subjects. The largest series of
patients with section of the vestibular nerve was
reported by Scharf et al. [19]. In 16 cases, no
impairment of multiple psychoacoustic abilities (de-
tection of tonal signals, intensity discrimination,
frequency selectivity, loudness adaptation, frequency
discrimination within a tonal series, in-head later-
alization) both in quiet, and, where relevant, in
background noise, became apparent. The only
change found after OCB section was an improved
detection threshold for tones of unexpected fre-
quency in noise, which was interpreted by the
authors as an epiphenomenon of a disturbed selec-
tive attention to tones. Notably, the patients re-
ported no subjective deterioration of their overall
hearing ability after the surgical ablation of the
efferents. Morand-Villeneuve et al. [21] evaluated
different aspects of sound intensity perception. They
Speech intelligibility and efferent olivocochlear reflex 59
functions and temporal loudness integration, as
well as no significant difference between patients
after vestibular neurotomy and unoperated subjects.
Only loudness summation was influenced by CAS;
however, this effect was not correlated with CS of
transiently evoked OAE.
Summarizing the literature, a conclusive definition
of the role of the olivocochlear efferents on psychoa-
coustic abilities including sound discrimination in
noise in humans is still pending. In animal experi-
ments, the efferent-mediated effects appeared to be
restricted to specific conditions such as a certain
SNR [9] and frequency [20]. Such a restriction of
detectable efferent effects to specific conditions is
reminiscent of findings in studies on cochlear
vulnerability to noise trauma. While a noise protec-
tion through MOC efferent activity could be demon-
strated at certain noise levels and frequencies, the
effect appeared to be absent at other levels and
frequencies in the same animals (for an overview see
Wagner et al. [31]). When interpreting the current
results and literature data, one must be aware of the
possibility that activation of the efferent system by
acoustic stimulation
which occurs in every hearing
situation
might indeed improve sound discrimina-
tion, but MOC reflex strength may not adequately
reflect this ability. On the one hand, the fact that
sectioning of the OCB abolishes the MOC reflex in
animal experiments [32] and in humans [16] in-
dicates that the MOC reflex at least qualitatively
reflects the efferent-mediated effects in the auditory
periphery. On the other hand, how exactly does the
MOC reflex indicate efferent activity quantitatively?
Maison and Liberman [4] evaluated ipsilateral
adaptation, which is another metric for MOC reflex
magnitude, in guinea pigs. They used a semi-
quantitative classification according to weak, mod-
erate, and strong efferent activity, and in doing so
found a significant relationship between MOC reflex
and a physiological parameter, i.e. cochlear noise
vulnerability. This is an argument in favor of the view
that CS values represent efferent activity not only
qualitatively but also quantitatively.
In conclusion, the presented data did not demon-
strate a correlation between the magnitude of MOC
efferent activity (as measured by CS of DPOAE) and
Speech-in-noise intelligibility. The present study was
restricted to subjects with valid DPOAE down to
primary tone levels as low as L147/L220 dB
SPL in order to grant optimum conditions for
measurement of CS. Future studies might include
individuals with mild to moderate perceptional
hearing loss, however, accepting less optimal mea-
surement conditions for efferent activity. The goal
remains to clarify a potential role of the olivocochlear
60 W. Wagner et al.
efferents in impairment of speech intelligibility,
which is one of the most important and yet poorly
understood deficits in perceptional hearing loss.
Acknowledgement
The study was supported by the grant fortüne 1469-
0-0 of the University of Tübingen, Germany.
References
[1] Warr WB, Guinan JJ Jr. Efferent innervation of the organ of
Corti: two separate systems. Brain Res 1979;173:152
5.
[2] Puel JL, Rebillard G. Effect of contralateral sound stimula-
tion on the distortion product 2F1-F2: evidence that the
medial efferent system is involved. J Acoust Soc Am 1990;
87:1630
5.
/
[3] Kim DO, Dorn PA, Neely ST, Gorga MP. Adaptation of
distortion product otoacoustic emissions in humans. J Assoc
Res Otolaryngol 2001;2:31
40. / /
[4] Maison SF, Liberman MC. Predicting vulnerability to
acoustic injury with a noninvasive assay of olivocochlear
reflex strength. J Neurosci 2000;20:4701
7. / /
[5] Müller J, Janssen T, Heppelmann G, Wagner W. Relation-
ship between fine-structure, contralateral suppression, and
ipsilateral adaptation of distortion product otoacoustic
emissions in humans. J Acoust Soc Am 2005;118:3747
56.
[6] Galambos R. Suppression of auditory nerve activity by
stimulation of efferent fibers to cochlea. J Neurophysiol
1956;19:424
37.
/ /
[7] Wiederhold ML. Variations in the effects of electric stimula-
tion of the crossed olivocochlear bundle on cat single
auditory nerve-fiber responses to tone bursts. J Acoust Soc
Am 1970;48:966
77. / /
[8] Murugasu E, Russell IJ. The effect of efferent stimulation on
basilar membrane displacement in the basal turn of the
guinea pig cochlea. J Neurosci 1996;16:325
32.
[9] Hienz RD, Stiles P, May BJ. Effects of bilateral olivocochlear
lesions on vowel formant discrimination in cats. Hear Res
1998;116:10
20.
/ /
[10] Liberman MC. Auditory-nerve response from cats raised in
a low-noise chamber. J Acoust Soc Am 1978;63:442
55.
[11] Kawase T, Liberman MC. Antimasking effects of the
olivocochlear reflex. I. Enhancement of compound action
potentials to masked tones. J Neurophysiol 1993;70:2519

32.
[12] Winslow RL, Sachs MB. Effect of electrical stimulation of
the crossed olivocochlear bundle on auditory nerve response
to tones in noise. J Neurophysiol 1987;57:1002
21.
[13] Winslow RL, Sachs MB. Single-tone intensity discrimination
based on auditory-nerve rate responses in backgrounds of
quiet, noise, and with stimulation of the crossed olivoco-
chlear bundle. Hear Res 1988;35:165
89. / /
[14] Micheyl C, Collet L. Involvement of the olivocochlear
bundle in the detection of tones in noise. J Acoust Soc Am
1996;99:1604
10.
/ /
[15] Micheyl C, Perrot X, Collet L. Relationship between
auditory intensity discrimination in noise and olivocochlear
efferent system activity in humans. Behav Neurosci 1997;
111:801
7.
/ Am 1996;99:3572
84. /
[16] Giraud AL, Garnier S, Micheyl C, Lina G, Chays A, Chery-
Croze S. Auditory efferents involved in speech-in-noise
intelligibility. NeuroReport 1997;8:1779
83.
/ /
[17] Liberman MC, Guinan JJ Jr. Feedback control of the
auditory periphery: anti-masking effects of middle ear
muscles vs. olivocochlear efferents. J Commun Disord
1998;31:471
83.
/ /
[18] Kumar UA, Vanaja CS. Functioning of olivocochlear bundle
and speech perception in noise. Ear Hear 2004;25:142
6. / /
[19] Scharf B, Magnan J, Chays A. On the role of the olivoco-
chlear bundle in hearing: 16 case studies. Hear Res 1997; /
103:101
21.
/
[20] May BJ, McQuone SJ. Effects of bilateral olivocochlear
lesions on pure-tone intensity discrimination in cats. Audi-
tory Neuroscience 1995;1:385
400. / /
[21] Morand-Villeneuve N, Garnier S, Grimault N, Veuillet E,
Collet L, Micheyl C. Medial olivocochlear bundle activation
and perceived auditory intensity in humans. Physiol Behav
/
2002;77:311
20.
/ /
[22] Wagner W, Heppelmann G, Müller J, Janssen T, Zenner HP.
Olivocochlear reflex effect on human distortion product
otoacoustic emissions is largest at frequencies with distinct
fine structure dips. Hear Res 2007;223:83
92. / /
[23] Heitmann J, Waldmann B, Schnitzler HU, Plinkert PK,
Zenner HP. Suppression of distortion product otoacoustic
emissions (DPOAE) near 2f1-f2 removes DP-gram fine
structure
evidence for a secondary generator. J Acoust
/ /
Soc Am 1998;103:1527
31. / /
[24] Boege P, Janssen T. Pure-tone threshold estimation from
extrapolated distortion product otoacoustic emission I/O-
functions in normal and cochlear hearing loss ears. J Acoust
Soc Am 2002;111:1810
8. / /
[25] Wagener K, Kühnel V, Kollmeier B. Entwicklung und
Evaluation eines Satztestes für die deutsche Sprache 1:
Design des Oldenburger Satztestes. Zeitschrift für Arbeits-
wissenschaft 1999;1:4
15. / /
[26] Wagener K, Brand T, Kollmeier B. Entwicklung und
/ /
Evaluation eines Satztestes für die deutsche Sprache 3:
Evaluation des Oldenburger Satztestes. Zeitschrift für Ar-
beitswissenschaft 1999;3:86
95. / /
[27] Dewson JH 3rd. Efferent olivocochlear bundle: some re-
lationships to stimulus discrimination in noise. J Neurophy-
siol 1968;31:122
30.
/ /
/ /
[28] May BJ, Prosen CA, Weiss D, Vetter D. Behavioral investiga-
tion of some possible effects of the central olivocochlear
pathways in transgenic mice. Hear Res 2002;171:142
57.
/ /
/ /
[29] Kim S, Frisina RD, Frisina DR. Effects of age on speech
understanding in normal hearing listeners: relationship
/ /
between the auditory efferent system and speech intellig-
ibility in noise. Speech Commun 2006;48:855
62. / /
[30] Zheng XY, McFadden SL, Ding DL, Henderson D.
Cochlear de-efferentation and impulse noise-induced acous-
tic trauma in the chinchilla. Hear Res 2000;144:187
95. / /
[31] Wagner W, Heppelmann G, Kuehn M, Tisch M, Vonthein
R, Zenner HP. Olivocochlear activity and TTS-susceptibility
in man. Laryngoscope 2005;115:2021
8. / /
[32] Liberman MC, Puria S, Guinan JJ Jr. The ipsilaterally
evoked olivocochlear reflex causes rapid adaptation of the
/
2f1-f2 distortion product otoacoustic emission. J Acoust Soc
/