Christofer W. Bester, BSc (Hons) Supervised by: Emeritus Professor Don Robertson 1 Emeritus Professor Geoff Hammond 2 This thesis is presented for the degree of Doctor of Philosophy of the University of Western Australia 1 School of Anatomy, Physiology and Human Biology 2 School of Psychology 2015 Mechanisms of auditory attention in normal and hearing impaired listeners
194
Embed
Mechanisms of auditory attention in normal and hearing impaired listeners
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Christofer W. Bester, BSc (Hons)
Supervised by: Emeritus Professor Don Robertson1
Emeritus Professor Geoff Hammond2
This thesis is presented for the degree of Doctor of Philosophy of the University of Western Australia
1 School of Anatomy, Physiology and Human Biology 2 School of Psychology
2015
Mechanisms of auditory attention in normal and hearing impaired listeners
i
Abstract.
Near-threshold tones presented in background noise are detected at a
relatively higher rate if presented more frequently than other tones,
and/or if preceded by a clearly audible cue tone of the same or similar
frequency. A potential candidate that has been suggested to underlie the
formation of this so-called attentional filter is the medial olivocochlear
system (MOCS). This thesis addresses the extent and nature of the
potential MOCS role in forming the attentional filter.
Three sets of experiments are included in the present work. The first set
of experiments correlated the depth of the attentional filter with the
strength of a single MOCS process, the MOCS acoustic reflex, using the
suppression of otoacoustic emissions in normal -hearing participants. The
second set of experiments explored the depth of the attentional filter
using the difference in detection rate of the more frequently presented
tones and the infrequently presented tones, in participants with a loss of
the MOCS efferent t argets due to sensorineural hearing loss (SNHL).
Sensorineural hearing loss participants were recruited with a range of
severities of hearing loss, from mild to moderately -severe, so that fi lter
depth could be correlated with the level of hearing loss. The final set of
experiments measured the attentional filter in cochlear implant
recipients, as a group presumed to have no remaining MOCS action on
the cochlea, but who had undergone a period of auditory relearning. The
findings of the thesis are:
ii
1. In normal-hearing participants there was no evidence for increasing
depth of the attentional fil ter with increasing strength of the MOCS
acoustic reflex, as assessed by the contralateral suppression of
otoacoustic emissions.
2. The depth of the attentional filter was found to decrease slightly
with increasing MOCS acoustic reflex strength, although this was a
weak effect which was observed primarily on the low -frequency
side of the attentional filter.
3. In these clinically normal hearing participants, there was a range of
auditory thresholds from -5 to 10 dB HL. A negative correlation
was found between the depth of the low -frequency side of the
attentional filter and hearing level. The depth of the low -frequency
side of the attentional filter decreased to zero over this sm all range
of subclinical hearing levels.
4. Individuals with SNHL and a loss of otoacoustic emissions had
decreased depth of the attentional fil ter. At the lowest level of
SNHL, the low-frequency side of the attentional filter was no
longer suppressed in comparison with the more-frequently
presented centre frequency. The depth of the high -frequency side of
the fil ter decreased progressively with increasing SNHL, and was
near zero at 60 dB HL.
5. Two participants with conductive hearing loss, who were assumed
to have a hearing loss without an associated loss of MOCS targets
in the cochlea, showed a similar decrease in filter depth as the
SNHL group.
iii
6. Five of the six cochlear implant recipients, who were presumed to
have no remaining MOCS targets in the cochlea, show ed no
evidence of an attentional filter when the stimuli were presented
with free-field acoustics. However, one implant recipient showed a
normal attentional fi lter.
7. Four implant recipients were re -tested with a programmed, direct
stimulation that had no acoustic stimulus, to eliminate the
potentially unwanted effects of the commercial speech processor
used previously. Two of the cochlear implant recipients tested with
programmed, direct stimulation showed an attentional filter.
The decreased depth of the attentional filter with increasing MOCS
acoustic reflex strength in normal -hearing participants is not consistent
with a role for the MOCS in forming the attentional filter. In participants
with SNHL, the decreased depth of the high -frequency side of the f ilter
occurred at levels of hearing loss physiologically relevant to the
impairment of the MOCS efferent targets. In contrast, the loss of the
low-frequency side of the attentional fil ter occurred before clinical
SNHL is classified, and at a hearing level that is not typically associated
with appreciable impairment to the MOCS efferent targets, the outer hair
cells. The subclinical changes in hearing level may be associated with
recent research into so-called “hidden” hearing losses, which have been
identif ied in individuals with normal auditory thresholds. These hidden
hearing losses have been associated with auditory neuropathy, and this
neuropathy may be selective for the fibres that form the afferent input to
the MOCS. The loss of the low-frequency side of the filter may then be
due to reduced input to the MOCS with auditory neuropathy, or it may be
iv
due an alternative, non-MOCS mechanism whose impairment is
associated with the low-level elevation in auditory thresholds. These
results are consistent with a t least a partial role for the MOCS in
forming the filter. However, the loss of the attentional fil ter in two
participants with conductive hearing loss may suggest that the hearing
loss alone can result in the loss of the attentional filter. Finally, the
finding that at least one cochlear implant recipient who had profound
SNHL prior to implantation was able to form the attentional filter
suggests that there must be an alternative mechanism that is able to form
the fil ter in some cochlear implant recipients .
The attentional filter is thought to improve the detectability of signals of
interest in noisy environments. The impaired formation of the filter with
SNHL may then contribute to the poor speech -in-noise perception
associated with this hearing impairment . In addition, the filter decreased
in depth at subclinical hearing levels, which may indicate an important
“hidden” detriment to auditory ability before clinical hearing loss is
classified. Finally, most of the cochlear implant recipients did not show
a typical attentional filter, which may contribute to the difficulties with
speech in noise perception associated with the implants. However, the
apparent presence of the attentional filter in at least one cochlear
implant recipient indicates that the filter can be formed in some
individuals with no remaining MOCS action on the cochlea. Variations
in the ability to form an attentional filter may be a contributing factor to
the wide range in cochlear implant outcomes, in particular when
discriminating signals in competing background noise.
v
vi
Table of contents
Abstract. ............................................................................................................................. i Table of contents .............................................................................................................. vi List of figures ................................................................................................................. viii Abbreviations .................................................................................................................... x Acknowledgements .......................................................................................................... xi Candidate Contributions ................................................................................................. xii Chapter 1. General Introduction ................................................................................... 1
1.1 The attentional filter: History ............................................................................. 3 1.2 The attentional filter: Cues ................................................................................. 5 1.3 The medial olivocochlear system: Proposed role ............................................... 7 1.4 The medial olivocochlear system: Anatomy ...................................................... 8 1.5 The medial olivocochlear system: Physiology ................................................. 10 1.6 The medial olivocochlear system: Effects on hearing...................................... 11 1.7 The medial olivocochlear system: In humans .................................................. 15 1.8 The medial olivocochlear system: Forming the attentional filter..................... 19 1.9 The medial olivocochlear system: Related structures ...................................... 25 1.10 Central mechanisms ......................................................................................... 27 1.11 Structure and aims of the thesis ........................................................................ 28
4.4 Discussion ........................................................................................................ 90 Chapter 5. Formation of the attentional filter in cochlear implant recipients using acoustic presentation ..................................................................................................... 101
5.4 Discussion ...................................................................................................... 120 Chapter 6. Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation .................................................................................... 127
6.4 Discussion ...................................................................................................... 141 Chapter 7. General Discussion.................................................................................. 148
7.1 Implications & Future Directions ................................................................... 159 7.1.1 Formation of the filter in normal hearing individuals ............................. 160 7.1.2 Formation of the attentional filter with conductive hearing loss ............ 163 7.1.3 An alternative mechanism able to form the attentional filter .................. 164
8.1 Depth of the attentional filter as a function of OAE suppression................... 175 8.1.1 TEOAE suppression ................................................................................ 175 8.1.2 DPOAE suppression ............................................................................... 177
viii
List of figures
Chapter 1
Figure 1.1: The attentional filter measured by Greenberg and Larkin (1968) .................. 4
Figure 1.2: Schematic diagram of the MOCS anatomy .................................................... 9
Figure 1.3: Auditory nerve fibre firing rate with MOCS activation in quiet .................. 12
Figure 1.4: Auditory nerve fibre firing rate with MOCS activation in noise .................. 14
Figure 1.5: Schematic of proposed MOCS role in filter formation ................................ 21
Figure 1.6: Change in filter depth after a vestibular neurectomy ................................... 22
Chapter 2
Figure 2.1: Sound spectrum of broadband background noise ......................................... 32
Figure 2.2: Structure of the two interval forced choice procedure ................................. 33
Chapter 3
Figure 3.1: Examples of OAE data, including DPOAE spectrum, DPOAE fine structure and a DPOAE I/O function, as well as the TEOAE waveform ................................. 42-43
Figure 3.2: TEOAE response with and without noise for 3 participants ........................ 47
Figure 3.3: DPOAE response with and without noise for 3 participants ........................ 49
Figure 3.4: Attentional filters for 15 normal hearing participants .................................. 52
Figure 3.5: Attentional filters for 15 normal hearing participants, for the three sessions used to measure the filter ................................................................................................ 54
Figure 3.6: Attentional filters across three sessions for two participants ....................... 55
Figure 3.7: Correlation between TEOAE suppression and depth of the attentional filter ..
Figure 4.1: Audiograms for normal hearing and SNHL participants ........................ 80-81
Figure 4.2: Attentional filters for the normal hearing and SNHL groups ....................... 83
Figure 4.3: Depth of the attentional filter as a function of hearing loss for the normal hearing and SNHL groups .............................................................................................. 85
Figure 4.4: Schematics for attentional filters with normal hearing and SNHL .............. 86
Figure 4.5: OAE suppression as a function of hearing loss for the normal hearing group
Figure 4.6: Audiograms for two conductive hearing loss participants ........................... 88
ix
Figure 4.7: Attentional filters for two conductive hearing loss participants ................... 89
Chapter 5
Figure 5.1: Schematic of stimuli used for cochlear implant simulation ....................... 107
Figure 5.2: Response of the cochlear implant during simulations at threshold ............ 111
Figure 5.3: Response of the cochlear implant during simulations at threshold ............ 113
Figure 5.4: Tracking threshold in two individuals with normal hearing and two cochlear implant recipients using acoustic presentation .............................................................. 114
Figure 5.5: Attentional filters for two normal hearing participants and two cochlear implant recipients in a preliminary study using acoustic presentation ......................... 115
Figure 5.6: Grouped attentional filters for normal hearing individuals and cochlear implant recipients using acoustic presentation .............................................................. 117
Figure 5.7: Attentional filters for individual cochlear implant recipients using acoustic presentation ................................................................................................................... 119
Chapter 6
Figure 6.1: Structure of the programmed, direct stimuli ............................................... 132
Figure 6.2: Response of cochlear implant to programmed stimulus ............................. 135
Figure 6.3: Grouped attentional filters for the cochlear implant recipients using programmed, direct stimulation ................................................................................... 137
Figure 6.4: Attentional filters for individual cochlear implant recipients using both acoustic presentation as well as programmed, direct stimulation ................................. 139
Figure 6.5: Shifted target attentional filters for cochlear implant users using programmed, direct stimulation .................................................................................... 140
Appendix
Figure 8.1: Relationship between filter depth and TEOAE suppression ..................... 176
Figure 8.2: Relationship between filter depth and DPOAE suppression ............... 177-178
Figure 8.3: Frequency allocation table (FAT) for Cochlear™ implants ....................... 179
Figure 8.4: Audiograms for the contralateral ear in the SNHL participants ................ 180
x
Abbreviations
2IFC Two interval forced choice
AC Air conduction
ANOVA Analysis of variance
BC Bone conduction
CHL Conductive hearing loss
CI Cochlear implant
DPOAE Distortion product otoacoustic emission
FFT Fast Fourier transform
HL Hearing level
I/O Input / output
IHC Inner hair cell
LOCS Lateral olivocochlear system
MOCS Medial olivocochlear system
NH Normal hearing
OAE Otoacoustic emission
OHC Outer hair cell
SFOAE Stimulus frequency otoacoustic emission
SNHL Sensorineural hearing loss
SPL Sound pressure level
TEOAE Transiently-evoked otoacoustic emission
xi
Acknowledgements
My first acknowledgements and thanks, must go to my supervisors. Don & Geoff have
worked with (nearly) infinite patience as I have progressed in my studies, and this
document would be a pale shadow of what it is without their careful guidance.
From the Ear Science Institute of Australia, Marcus Atlas, Dunay Taljaard, and
especially Dona Jayakody have helped tremendously with my understanding of hearing
research, and without their and the Institute‟s help I would not have had the
opportunities to work directly with the hearing impaired individuals required for the
work. Thanks for the help, both inside and outside of the soundproof booths!
The Auditory lab has been an excellent base for learning how to perform high-quality
research. The guidance provided by Don, Rob, and Helmy as well as the students I‟ve
worked with, Darryl, Nathaniel, Ahmaed, Kristin, and Kerry made the years I‟ve
worked with lab enjoyable, and informative beyond a simple scientific education. My
sincere thanks to everyone involved in this institution.
The village required to raise this particular child extends beyond the university. Without
the support of my parents this PhD would simply not exist, even though this career path
was third (at most) on their list of approved jobs. Mom, dad, thanks for the help, and the
ability to make hard decisions for all of us. To the rest, Ashleigh, Scott, and Nick,
thanks for providing a context that post-graduate studies are not my whole life.
Audrey Bester, getting married during a PhD was a mad move, and I would do it again
in a heartbeat. My thanks for you would take more than the 180 pages I‟ve written here,
so I will be brief: Every day I‟m with you is the best day I‟ve ever had.
To everyone else who has helped throughout this crazy process, you have my thanks.
To my friends, please pretend that I listed your name here first in a long line of friends,
go you! Thanks to my extended family, in Australia, South Africa, London and now
Malaysia and Singapore!
“Don't Panic.” ― Douglas Adams, The Hitchhiker's Guide to the Galaxy
xii
Candidate Contributions
This thesis does not contain work that I have published, or work that is currently under
review for publication.
All experiments were conducted by the author, Christo Bester, with some assistance.
Many of the LabVIEW programs included in the work are based on versions that were
initially programmed by Geoff Hammond. I was aided by audiologists from the Ear
Science Institute of Australia. Dunay Taljaard aided in the audiometric testing of some
hearing impaired individuals in chapter 4. Dona Jayakody aided in the same audiometric
testing, and in the setup of cochlear implant recipients in chapters 5, and 6. Dr. Peter
Busby and Kieran Reed from CochlearTM provided advice and a proofread of the Python
programming included in chapter 6.
Otherwise, the content of this thesis is of my own composition, and all relevant sources
are acknowledged. This thesis has never been submitted for any other degree in this, or
another institution.
Geoff Hammond
Co-ordinating supervisor
in psychology
6th September 2015
Don Robertson
Co-ordinating supervisor
in physiology
6th September 2015
Christo Bester
PhD Student
6th September 2015
Chapter 1. General Introduction
General introduction
2
Detecting a signal of interest in competing noise, such as understanding the content of
one person‟s speech while others are talking, is a complex auditory task. A healthy
auditory system is surprisingly adept at this task, which forms the basis of Cherry‟s
“Cocktail Party Problem”, a research question that investigates the possible mechanisms
that enable us to separate sounds of interest from background noise (Cherry, 1953).
There are a number of stages within the auditory system that are involved in this ability,
and much of the processing, such as auditory scene analysis (Bregman, 1990), is located
in central auditory structures. However, this thesis is concerned only with a subset of
auditory processing that is achieved by the efferent control of the peripheral auditory
system.
The cochlea acts as the first detailed frequency analyser in the auditory system, but it is
not a passive receiver. The cochlea‟s fine frequency sensitivity and specificity are due to
the active cochlear amplifier, which is powered by the outer hair cells (OHCs). The
OHCs are the targets of an efferent projection from the olivocochlear system.
Consequently, this efferent projection is able to modify the auditory system‟s sensitivity
to incoming sound via the cochlear amplifier. Thus, the efferent control of the
peripheral cochlear amplifier is an ideal location to modify the detectability of incoming
auditory signals.
The first experiments to investigate the effects of selective attention on the detectability
of auditory stimuli used dichotic listening tasks that test the processing of competing
speech signals (Cherry, 1953). However, when we consider the basic physiology of the
peripheral auditory system, it is more useful to consider simple stimuli. Tanner and
Norman (1954) began these experiments by testing the detection rate of pure tones at an
attended frequency („target‟ tones), compared to pure tones presented at an unattended
frequency („probe‟ tones). In their study, participants were trained to detect a 1-kHz
target tone in noise, which was presented at a 65% detection threshold. After the
General introduction
3
training period, the stimulus was switched to a 1.3-kHz probe tone, without informing
the participants. The tone presented at the attended, target, frequency was detected at the
65% threshold level, while the tone presented at the unattended, probe, frequency was
not detected above chance. In fact, the participants reported that they had failed to
switch on the signal generator! The results of this experiment, and a similar study by
Greenberg (1962), were taken as evidence that the auditory system behaves, in part, as
a narrow-band receiver that allows the input of attended stimuli while attenuating the
input of unattended stimuli.
1.1 The attentional filter: History
In 1968, Greenberg and Larkin took this research a step further with their probe-target
procedure (or probe-signal procedure in the original nomenclature). This procedure was
developed to measure the detection rate of target and probe tones as a function of
frequency. The original probe-target procedure used a similar structure to Tanner and
Norman‟s 1954 procedure described above. The participants were trained to detect a
1.0-kHz target tone, and led to only expect this tone during the experiment. After this
training period the procedure continued, but now with the target tone presented on only
70% of all trials, while the remaining 30% contained a probe tone. Each 100-trial run of
the procedure used only a single probe frequency, and multiple runs were done to
measure a range of probe frequencies. The data from multiple runs were combined to
show an apparent narrow band filter of detection rates, with a peak of detection rates at
the target frequency and a progressive reduction of the detection rates of probes with
increasing separation from the target frequency (Greenberg and Larkin, 1968). This is
known as the attentional filter (Botte, 1995). This attentional filter is a separate process
to the more general auditory filter (Fletcher, 1940), as demonstrated by the auditory
filter‟s slightly narrower bandwidth, and tendency to become skewed with increasing
masker level, whereas the attentional filter remains symmetrical (Glasberg and Moore,
General introduction
4
1990; Botte, 1995). An example of the attentional filter is shown in Figure 1.1, from
Greenberg and Larkin‟s original 1968 research. The presence of the attentional filter
supports the notion that auditory signals presented in noise vary in detectability based
on the expectation built by previous trials.
Figure 1.1. Detection rate of the target tone, at 1.0 kHz and the probe tones, all others,
as a function of frequency, demonstrating the attentional filter. The tone at the target
frequency was presented on 70% of all trials, and the participants were only informed of
the presence of the target tone. Figure adapted from Greenberg and Larkin (1968).
Initially, there were concerns that the attentional filter was produced not due to the
attenuation of the probes, but because the participants were ignoring the probe tones that
they were not informed of, even when they were readily detectable. This is known as
the „heard but not heeded‟ hypothesis (Scharf et al., 1987). To address this concern,
Scharf et al. repeated the probe-target procedure, but informed the participants of the
presence of probe tones different from the target tone during the experiment, and found
no appreciable differences in the shape or depth of the filter, which indicates that prior
knowledge of the presence of probes does not affect the formation of the attentional
filter.
General introduction
5
The probe-target procedure up to this point presented the target tone and only a single
probe tone during each run of the experiment, and therefore required several
experimental sessions, over 3 to 5 days, to measure the attentional filter. Dai et al.
(1991) changed this by modifying the procedure to include multiple probe tones during
each run. This was termed the „multi-probe‟ procedure. The multi-probe procedure still
used the 70% target presentation, but the remaining probe frequencies were randomly
dispersed throughout 30% of each run, in equal number. This allowed a more rapid
measurement of the attentional filter. Dai et al.‟s work further estimated the amount of
attenuation which, if applied to the distant probe tones during an auditory threshold
measurement, would result in the reduced detection rates seen in the attentional filter.
The probe suppression equated to roughly 7 dB of suppression at half a critical band
from the target frequency.
1.2 The attentional filter: Cues
A 70% or higher presentation rate of the target was the first method used to build the
expectation of the target tone. However, prior research had demonstrated that a clearly
audible cue tone, that preceded a single to-be-detected tone, increased the tone‟s
detection rate (Greenberg, 1962). In 1972, Penner extended this cueing effect to the
formation of the attentional filter (Penner, 1972). The experiment followed the same
structure as the previous probe-target experiments, but the to-be-detected target and 8
probe tones were presented with equal probability, and all tones were preceded by a
clearly audible cue tone at the target frequency. This cued probe-target procedure
formed attentional filters that were not substantially different from those in the previous
research, showing that the attentional filter can be produced by preceding the to-be-
detected tones with a cue tone. Penner explained the formation of the attentional filter in
response to this cue-effect as an alternative method of leading the subjects to
consciously expect the target tone at the cue frequency (Penner, 1972).
General introduction
6
Research has shown that cues do not have to be presented at the target frequency to
produce the attentional filter. Ebata et al. (2001) used complex tones as cues, which
were constructed to have a missing fundamental frequency at 250 Hz. The probe-target
procedure was then used to measure the attentional filter around this missing
fundamental frequency, at which there was no real acoustic energy in the cue tone. The
normal attentional filters that were produced demonstrated that acoustic energy at the
cue frequency in the cochlea is not required for an effective cue tone. Borra et al. (2013)
supported this notion by demonstrating normal attentional filters were produced at
octave-related frequencies both higher and lower than the cue tone, even when the cue
tone was imagined (Borra et al., 2013). Therefore, the attentional filter can be formed
using cues that require complex frequency-extraction, or are imagined, which implicates
the involvement of high-order auditory processing.
However, the role of cue-effect on the formation of the attentional filter may not be
entirely dependent on high-level processing. In contrast to the cues discussed so far, Tan
et al. (2008) used a cued probe-target procedure which attempted to remove the
expectation building effects of the cue tones (Tan et al., 2008). In this experiment, the
cue tones were randomly selected from a set of 5, and the target and 4 probe tones were
selected with equal probability. Thus, the to-be-detected tone matched the cue tone on
only 20% of all trials, which resulted in a cue tone that was misleading in the majority
of trials, and should not have led the participants to consciously expect the target tone at
the cue frequency. Nevertheless, the on-cue target tone was detected at a higher rate
than any of the probe tones, and the subjects showed attentional filters of approximately
half the depth of those formed with a constant frequency cue and 75% target
presentation. This result demonstrated the formation of the attentional filter in response
to cue tones which did not build-up the expectation of a single target frequency,
although the filters were reduced in depth. While it seems reasonable to suggest that
General introduction
7
these filters were produced using the same high-order auditory processing discussed in
the previous chapter, prior research supports a reflexive mechanism for the cue effect by
demonstrating that the emergence of the cue benefit can be extremely rapid. Pure-tone
cues were found to reduce the threshold of a transient tone in noise within 52 ms
(Scharf et al., 2007). The formation of the attentional filter in response to apparently
non-informative cues, which have been shown to have an effect on detection rates when
they preceded a to-be-detected tone by only 52 ms suggests, but does not necessarily
prove, that the cue-effect is partially provided by a subcortical reflex.
While the features of the attentional filter continued to be a focus for research, the
systems underlying the formation of the filter remained unclear. In the research
described above, the filter was shown to be formed in conditions consistent with high-
order, central processes, as well as in conditions that may be restricted to the effects of a
subcortical reflex. This thesis addresses the efferent control of the cochlear amplifier,
which will be shown below to fulfil both of these conditions, as a potential mechanism
for the formation of the attentional filter.
1.3 The medial olivocochlear system: Proposed role
In 1987, Scharf proposed that efferent control of the cochlear amplifier by the
olivocochlear system plays a role in forming the attentional filter (Scharf et al., 1987).
The cochlear amplifier, powered by the OHCs, was known to have substantial control of
auditory sensitivity (Dallos and Harris, 1978), by providing up to 50 dB of amplification
for incoming sound (Patuzzi et al., 1989b). The OHCs are the targets of the largest
efferent projection to the cochlea from the medial olivocochlear efferents. These
efferents have cell bodies located in the periolivary region of the superior olivary
complex, and both crossed and uncrossed projections to both cochleae (Moore et al.,
1999). The efferents synapse at the base of OHCs with cholinergic synapses (Fuchs,
General introduction
8
1996). The synaptic release of acetylcholine hyperpolarises the OHCs, suppressing their
active processes and reducing the gain of the cochlear amplifier (Housley and Ashmore,
1991). Therefore, each cochlea is under efferent control from both ipsilateral and
contralateral medial olivocochlear efferents, and these efferents act to suppress the
cochlear amplifier gain. The medial olivocochlear efferents and their afferent inputs,
discussed below, form the medial olivocochlear system (MOCS).
1.4 The medial olivocochlear system: Anatomy
The MOCS has both ascending input from the auditory afferents, as well as descending
input from higher auditory structures, summarised in Figure 1.2.
General introduction
9
Figure 1.2. Schematic diagram showing the inputs and output of one medial
olivocochlear system. Ascending afferent input is shown as dotted lines. The MOCS
acoustic reflex, with a feedback loop between the inner hair cells (IHCs) to the OHCs
via the MOCS is shown as dashed lines. Descending control of the MOCS by higher
auditory systems is shown by solid lines. Not shown is the uncrossed afferent input to
the MOCS required for binaural facilitation.
The ascending afferent input to the MOCS is through the primary auditory afferents,
which cross to the contralateral MOCS via interneurons in the cochlear nucleus
(Liberman and Brown, 1986; Robertson and Winter, 1988; De Venecia et al., 2005). As
this afferent input is crossed, it activates the contralateral olivocochlear efferents;
however, it is important to note that there are both crossed and uncrossed efferents, thus
afferent auditory input from a single cochlea will result in efferent suppression of both
cochleae. This creates an olivocochlear feedback system, where afferent auditory input
General introduction
10
results in efferent suppression of the cochlea via the contralateral MOCS, which is
known as the MOCS acoustic reflex. The descending, top-down input to the MOCS
arises from both the auditory cortex and the inferior colliculus (Vetter et al., 1993;
Mulders and Robertson, 2000b; Xiao and Suga, 2001; Perrot et al., 2006; Schofield and
Coomes, 2006). Therefore, the MOCS can respond to auditory input via the auditory
afferents as the MOCS acoustic reflex, as well as through top-down input from higher-
order structures, which is consistent with MOCS activation based on experience and
auditory context. This anatomy, with both a subcortical reflexive component and
higher-order control, is consistent with the conditions under which the attentional filter
is formed.
1.5 The medial olivocochlear system: Physiology
Single MOCS efferent fibres have been shown to respond to auditory input in guinea
pigs and cats, and were as sensitive and sharply tuned as auditory afferents with the
same characteristic frequency (Robertson and Gummer, 1985; Liberman and Brown,
1986). These recordings from single MOCS efferent fibres demonstrated that MOCS
efferent fibres synapse on OHCs close to the cochlear regions which activated them.
Subsequent research supports these findings, with the extension that some of the single
efferents project to broad areas of the cochlea (Brown, 2014). An important
consequence of this result is that the MOCS acoustic reflex, which is activated by sound
and forms a feedback loop to the OHCs, will suppress the cochlear amplifier in the
cochlear regions which activated it.
The activation of the reflex in response to acoustic elicitors has been measured with
single olivocochlear neuron recordings in guinea pigs. Three groups of olivocochlear
neurons were found, those that responded to ipsilateral or contralateral input were
roughly equal in proportion, and made up the majority of the neurons, with a small
General introduction
11
proportion responding to either-ear input (Robertson and Gummer, 1985). A strong
effect of binaural facilitation was found, such that even among the neurons that
responded to only one ear there was an increase in firing rate when binaural stimulation
was used (Robertson and Gummer, 1985; Liberman, 1988; Brown et al., 1998). The
stimuli that resulted in the strongest facilitation depended on the characteristic
frequency of the olivocochlear neuron. The facilitation was found to be strongest when
the facilitating stimulus was a pure tone in neurons that responded best to low-
frequency stimuli, but broadband noise was more effective when the neurons responded
best to high-frequency stimuli (Robertson and Gummer, 1985; Liberman, 1988; Brown
et al., 1998). The response of olivocochlear neurons to a continuous broadband noise
resulted in little to no adaptation when the noise was sustained, and this response was
active without significant change for the duration of long periods of noise (Brown,
2001). Therefore, during the measurement of the attentional filter, which includes
transient tones presented in a continuous background noise, there will be consistent
activation of the MOCS acoustic reflex, and the clearly audible cue tones may cause
additional activation of the reflex at the target frequency.
1.6 The medial olivocochlear system: Effects on hearing
The effect of MOCS activation on afferent nerve fibres‟ response to incoming sound
was first tested with electrical and then with acoustic activation of the system. Electrical
stimulation of MOCS efferent fibres in quiet resulted in a rightward level shift in the
firing rate of auditory-nerve fibres in cats, as shown in Figure 1.3. This rightward shift
indicated an increase in the sound level required to reach the same firing rate after
MOCS activation suppressed the cochlear amplifier (Gifford and Guinan, 1983; Gifford
and Guinan, 1987; Winslow and Sachs, 1987). The suppression of the cochlear
amplifier was not accompanied by a reduction in maximum firing rate in almost all the
auditory afferents, although a decrease was apparent in a small subgroup of afferents
General introduction
12
(Brown et al., 2003). Measuring the effect of acoustic activation of the MOCS took
advantage of the bilateral efferent output of the system by measuring the response of
single auditory nerve fibres during the presentation of noise or tones to the contralateral
ear (Warren and Liberman, 1989). Auditory stimulation of the MOCS in quiet resulted
in rightward shift in the firing rate of auditory afferent fibres in a similar manner to the
electrical stimulation (Kawase et al., 1993).
Figure 1.3. Firing rate of a single auditory nerve fibre in response to pure tone elicitors
with and without additional electrical activation of the MOCS in quiet. The solid line is
for no MOCS activation and the dashed line for with MOCS activation. Figure adapted
from Winslow and Sachs (1987).
This decrease in auditory afferent firing rates was frequency-specific, as suggested by
the anatomy, with contralateral pure tones suppressing auditory fibres at the same
frequency (Liberman and Brown, 1986). Thus, the anatomy and physiology of the
MOCS acoustic reflex, at least as activated in quiet, suggests that the system acts as a
frequency-specific suppressor of the auditory system‟s response to sound via its
suppression of the cochlear amplifier.
General introduction
13
More important to the present research question is the effect of MOCS activation on
transient pure tones that are presented in background noise, as these are the conditions
used to measure the attentional filter. The effects of MOCS activation on the reception
of transient signals in noise has been tested in experiments on single auditory afferent
fibre responses, which first used electrical stimulation of the MOCS (Winslow and
Sachs, 1987) and then acoustic activation by contralateral broadband noise (Kawase et
al., 1993). In low to moderate levels of background noise the activation of the MOCS
increased auditory afferent fibre thresholds, representing a poorer threshold, but
increased the fibres dynamic firing range, as shown in Figure 1.4. In high levels of
background noise, the activation of the MOCS did not detectably decrease auditory
thresholds, but the increase in dynamic firing range remained. The increase in dynamic
firing range is significant, as a higher dynamic firing range can improve the detectability
of near-threshold signals in noise. In support of this notion, Kawase et al. (1993)
additionally interpreted the firing rates of single auditory nerve fibres in terms of
psychophysical detection thresholds for pure tones in background noise, and showed
that MOCS activation could improve detection thresholds for pure tones presented in
background noise. The increases in dynamic range, and the subsequent improvement in
detection thresholds, were suggested to be due to antimasking. Antimasking, sometimes
known as unmasking, occurs when there is a release from the masking caused by
neuronal adaptation to continuous noise (Smith, 1979). Antimasking is thought to occur
by activation of the MOCS inhibiting the cochlear amplifier, which reduces the cochlear
response to background noise, and releases adaptation at the synapse between the inner
hair cells and the auditory nerve fibres (Smith, 1979; Winslow and Sachs, 1987;
Mulders et al., 2008). The release from adaptation provides the restoration of dynamic
range, which can be seen in Figure 1.4. A higher dynamic range increases
General introduction
14
discriminability and signal detection in noise, which is consistent with an enhancement
to detection rates at frequency regions under MOCS activation.
Figure 1.4. Firing rate of a single auditory nerve fibre in response to pure tone elicitors
in noise with and without additional electrical activation of the MOCS. The solid line is
for no MOCS activation and the dashed line for with MOCS activation. Figure adapted
from Winslow and Sachs (1987).
Antimasking is not a guaranteed effect of suppression of the cochlear amplifier, and
Kawase et al. (1993) demonstrated its occurrence in guinea pigs only in certain
conditions. When the to-be-detected pure tone was presented at an auditory-nerve
fibre‟s characteristic frequency, antimasking only occurred with relatively high masker
levels, whereas low masker levels increase detection thresholds. When the tone was
presented below the auditory-nerve fibre‟s characteristic frequency, the activation of the
olivocochlear system always resulted in antimasking, regardless of masker level.
Therefore, the results of Kawase et al. (1993) show that antimasking is unlikely to occur
at low masker levels, whereas at higher masker levels antimasking is likely to increase
the detectability of transient signals in a continuous background noise.
General introduction
15
The effects of antimasking have been shown to persist at higher levels of the auditory
system in guinea pigs. Mulders et al. (2008) measured the effect of electrical
stimulation of the MOCS on single inferior colliculus neurons‟ responses to pure tones
presented in background noise. Eighty-five percent of the inferior colliculus neurons
showed an increase in dynamic firing range that was accompanied by either no change,
or a slight deterioration in auditory threshold, which is similar to the responses of single
auditory afferents discussed above. However, a subset of the neurons showed a
surprising leftward shift in firing pattern with MOCS activation, which indicates
improved auditory thresholds. These responses from neurons in the inferior colliculus
demonstrate the improvement in auditory thresholds that can occur with antimasking,
and support the notion that the activation of an inhibitory system, like the MOCS, can
improve the detectability of near-threshold tones in noise.
1.7 The medial olivocochlear system: In humans
In human research, the activation of the MOCS can be estimated by exploiting the
system‟s effect on otoacoustic emissions (OAEs). OAEs are acoustic signals recorded in
the ear canal that are classically evoked in response to an auditory stimulus. There are
several types of OAEs, which are classified by the stimulus (or lack thereof in
spontaneous OAEs) used to evoke the emission. The most commonly used emissions
are transiently evoked (TEOAE) and distortion product (DPOAE) OAEs (see Shera and
Guinan (1999) for a review). TEOAEs are produced in response to a transient stimulus,
typically a broadband click, and are not considered frequency specific because of the
broadband stimulus. DPOAEs are produced by two pure tone stimuli presented near to
each other in frequency and give a more frequency-restricted measurement of cochlear
amplifier activity. Previous research has demonstrated two, fundamentally different
mechanisms that underlie the generation of otoacoustic emissions (Shera and Guinan,
General introduction
16
1999), however as the emissions are currently measured it is unclear how to separate
these mechanisms, or what the significance of this separation would be (Shera, 2004).
The magnitude of the OAE emission is used as a measure of health of the cochlear
amplifier because it is reduced with mild sensorineural hearing loss (SNHL) and lost
with moderate SNHL (Kim et al., 1996), although the raw OAE magnitude cannot be
used as a measure of strength of the cochlear amplifier (Kemp, 2002). However, more
relevant to the current research question is the suppression of the OAE response by the
activation of the MOCS. Otoacoustic emissions are typically suppressed using
contralateral broadband noise in a manner consistent with a reduction of gain of the
outer hair cell amplifier by the activation of the MOCS (Veuillet et al., 1991). The
contralateral noise is a potent elicitor of the MOCS acoustic reflex, which then acts to
suppress the cochlear amplifier, and thus reduce the magnitude of the OAE response
(Ryan et al., 1991). This enables the use of contralateral noise suppression of OAEs as
an index of the function of the MOCS in humans: a greater suppression of the OAE by
contralateral noise suggests a greater magnitude of cochlear amplifier suppression by
MOCS activation in that individual (Berlin et al., 1994; Giraud et al., 1995). In support
of OAE suppression as a useful measure of MOCS action, there exists considerable
variation in the magnitudes of OAE suppression in normal hearing individuals (Backus
and Guinan, 2007; Collet et al., 1992), and OAE suppression has been successfully
correlated with speech perception measures that are thought to require the functioning
of the MOCS (Kumar and Vanaja, 2004).
The suppression of the OAE response by MOCS activation has been used to test the
frequency-tuning of the MOCS acoustic reflex in humans. The tuning of the reflex was
tested using pure tones and narrow-band noises that were presumed to activate the
reflex („elicitors‟), and measuring the suppression of a frequency specific type of OAE,
called stimulus frequency OAEs, in the contralateral ear (Lilaonitkul and Guinan,
General introduction
17
2009b; Lilaonitkul and Guinan, 2012). Pure tone and narrow-band noise elicitors were
found to suppress the contralateral OAEs of similar frequency, which is a result
consistent with the previously discussed anatomy of the MOCS in guinea pigs and cats.
However, this suppression was not restricted to the elicitor frequency, and the
maximum suppression of the contralateral OAEs was achieved when the MOCS
elicitors were skewed up to half an octave from the tested OAE frequency. The
direction of the skew depended on the frequency tested, with a skew towards higher-
frequency elicitors at 0.5 kHz, lower-frequency elicitors at 1 kHz, and frequency-
matched elicitors at 4 kHz. Thus, while the human research supports a suppression of
the cochlear amplifier at the same frequencies used to elicit MOCS activity, the
maximum suppression of the cochlear amplifier by MOCS activation is suggested to be
skewed up to half an octave from the elicitor frequency, depending on the frequency
used.
The suppression of OAEs has been used to test the function of the top-down MOCS
control in humans. The existence of top-down control was confirmed in humans by the
electrical activation of the auditory cortex in humans during pre-surgical brain mapping
for epilepsy (Perrot et al., 2006). During this procedure, the auditory cortex was
electrically stimulated while OAEs were measured in the contralateral ear. There was a
significant reduction in OAE amplitude during the electrical stimulation, consistent with
the presence of the top-down input to the MOCS from the auditory cortex in humans.
Furthermore, this activation has been demonstrated to occur in humans by measuring
OAE amplitude during different task conditions. This is evident in a slight reduction in
OAE amplitude during a visual task as compared with an auditory task (Puel et al.,
1988; Meric and Collet, 1992), although a later study found no change in the
suppression of OAEs by contralateral noise between a visual task and an auditory task
(de Boer and Thornton, 2007). A frequency-specific effect of the top-down MOCS
General introduction
18
control was suggested by Maison et al. (2001). This study used a variant of TEOAEs
evoked by pure tone pips, and measured the contralateral acoustic noise suppression of
OAEs that were evoked by these pips at two different frequencies during an auditory
task. The tone pips were presented while the participants were simultaneously tasked
with detecting infrequent, near threshold pure tones in noise in the opposite ear, which
matched the frequencies of the tone pips on only some presentations. The results
demonstrated an increased magnitude of suppression of OAEs that were evoked by tone
pips that matched the frequency the participants were attending to, but not of the OAEs
evoked by the tone pips at an unattended frequency (Maison et al., 2001). This suggests
an increase in the suppression of the cochlear amplifier at frequencies under focused
auditory attention. In contrast however, a similar experiment by de Boer and Thornton
(2007), which measured OAE suppression when a participant attended to stimuli
presented in only one ear with dichotic noise, found less OAE suppression in the
attended ear. Therefore, while the actions of the higher-order inputs to the MOCS are
apparently similar to the actions of the MOCS acoustic reflex, in that they would be
active at the target frequency during the attentional filter measurement, there are
contradictory studies that find either no such evidence, or the opposite relationship.
Thus, while the top-down input to the MOCS from higher-order structures may act in a
similar manner to the MOCS acoustic reflex to cause additional suppression of the
cochlear amplifier at the attended target frequency, the evidence supporting this
conclusion is inconsistent.
The above research demonstrates that MOCS activation in humans may have a similar
inhibitory effect on the cochlear amplifier to what was identified in animals in the
previous section. Furthermore, human research into the effects of MOCS activation
supported the presence of an antimasking effect brought about by the system‟s
activation, which can improve the detectability of transient pure tones in noise. The
General introduction
19
activation of the MOCS has been shown to generally enhance the detectability of
transient tones presented in noise, although this effect depends on the participant and
the conditions used (Micheyl and Collet, 1996). Detection thresholds for 1- and 2-kHz
pure tones presented monotically in broadband noise were measured in the same ear as
TEOAEs, which was done successively with and without a contralateral broadband
noise thought to activate the MOCS. The detection thresholds for the 2-kHz tone
improved with the addition of the contralateral broadband noise, and the magnitude of
this improvement correlated with the magnitude of TEOAE suppression. This result is
consistent with an improvement in signal-in-noise detection rates associated with the
activation of the MOCS, in this case by the contralateral broadband noise, and this
improvement scaled with the index of MOCS function. No such relationship was
present for the 1-kHz pure tone. The frequency-specific effects may be due to the
broadband clicks used to test the OAE suppression. Similar research used DPOAEs to
assess MOCS strength, and again correlated this strength with the detectability of
masked tones at 1 and 2 kHz (Bhagat and Carter, 2010). Using DPOAEs, there was a
positive correlation between the magnitude of suppression of DPOAEs and reduced
masking of only 1-kHz pure tones.
1.8 The medial olivocochlear system: Forming the attentional filter
Overall, the above research demonstrates that the activation of the MOCS can cause
frequency specific suppression of the cochlear amplifier, and the system can be
activated in response to pure-tones, like the cues used to form the attentional filter, or in
response to top-down control, consistent with the filter‟s formation in response to
complex cues. Furthermore, the effect of MOCS activation has been demonstrated to
cause antimasking-mediated enhancements of pure-tones in noise in animals, and in
humans there is limited evidence that the system can improve the detectability of
General introduction
20
transient tones. Therefore, the MOCS has an anatomy, physiology and function
consistent with a role in generating the attentional filter.
The specific role that the MOCS has in generating the attentional filter is likely to be an
antimasking-based enhancement at the target frequency. The antimasking effect,
schematized in Figure 1.5 would rely on the activation of the MOCS at the target
frequency, either by the clearly-audible cue tone activating the MOCS acoustic reflex,
or by more-frequent target presentations activating the MOCS by its top-down inputs.
The resulting suppression of the cochlear amplifier at the target frequency then causes
an antimasking effect to enhance the target‟s detection rate, with a rapid drop-off of this
effect with increasing frequency separation from the target. Previous research by Tan et
al. (2008) has suggested that the filter is formed both by an enhancement at the target
frequency as well as a suppression at the probe frequencies. It is unclear how the MOCS
might be involved in this suppression, when the anatomy and physiology discussed so
far has suggested a tonotopic tuning of the MOCS, in which the system suppresses the
cochlear amplifier at, or very near to , the frequencies at which it is active. Therefore,
the present thesis is concerned primarily with MOCS-mediated enhancement at the
target frequency, although probe suppression is not discounted.
General introduction
21
Figure 1.5. Schematic of the proposed role of the MOCS in forming the attentional
filter, demonstrating the suggested antimasking effect at the target frequency.
Support for a causal role for the MOCS specifically forming the attentional filter is
provided by Scharf et al. (1997), who measured auditory function before and after a
vestibular neurectomy. The vestibular neurectomy is a surgical procedure to relieve
severe vertigo, generally associated with Ménière‟s disease, and all of the participants
included in Scharf‟s study except one were diagnosed with Ménière‟s disease. In
addition to relieving severe vertigo, the vestibular neurectomy sections the MOCS
efferent connection to the cochlea as the system‟s efferent output travels with the
inferior vestibular nerve (Bodian and Gucer, 1980). Scharf and his colleagues measured
the auditory capacities of 16 individuals before and after undergoing the surgery,
including measurements of pure tone thresholds, intensity discrimination, loudness
adaptation, and, of relevance to the current work, the difference in detection rate
between expected and unexpected near-threshold tones presented in background noise.
General introduction
22
The only significant change was shallower attentional filters in participants who
previously showed normal filters, as shown in Figure 1.6. The loss or reduction in the
formation of the attentional filter with the presumed total loss of MOCS efferent output
was taken as support that the MOCS has a role in the generation of the attentional filter.
Figure 1.6. Mean difference (± SD) in detection rate between target tones and lower-
(LF) or higher- (HF) frequency probe tones in noise either before the surgery (n = 8), or
in the contralateral healthy ear (n = 3) (left panel) and after (right panel) a vestibular
neurectomy. Prior to the surgery participants showed a larger decrease in detection rate
between the target and the probes relative to after the surgery. This demonstrates the
decrease in depth of the attentional filter with the sectioning of the MOCS output,
however many participants retain a target-probe difference at or greater than 10%. The
figure was composed with estimates from figures published in Scharf et al. (1997).
Similar research has used dysfunction of the MOCS, presumed from hearing loss, to test
the proposed role for the system in forming the attentional filter. Moore et al. (1996)
measured the attentional filter in two individuals with moderately-severe SNHL. SNHL
is typically, but not necessarily, caused by the loss or damage of the OHCs, which are
the MOCS efferent targets. Individuals with a moderately-severe SNHL may have a
General introduction
23
near-complete loss of OHC function (Stebbins et al., 1979; Hamernik et al., 1989).
Moore et al. (1996) found no evidence for the attentional filter in the two individuals
with SNHL, consistent with the loss of the filter with impairment to the MOCS action
on the cochlea. In this study, Moore et al. suggest that the loss of the filter in SNHL
participants is due to the broadening of auditory filters, and thus a loss of frequency
selectivity. This loss of frequency selectivity may have prevented the cue tone from
producing a distinct pitch cue for the auditory system, and therefore lead the SNHL
participants to adopt a “broadband” listening strategy that prevented the formation of
the attentional filter. However, the study did not address the potential involvement of
the MOCS in the formation of the attentional filter, and so the possible contributions of
the loss of the MOCS efferent targets, the OHCs, was not addressed.
In similar research Tan (2008) measured the benefit of a cue tone to the detection of
near-threshold tones presented in noise at a randomly selected frequency in eight SNHL
participants with moderately-severe to profound SNHL. Tan (2008) additionally tested
for the absence of detectable OAEs as an indicator of OHC impairment. The eight
SNHL participants showed a loss of the cue-benefit, which equated to a 3-dB
improvement in detection thresholds in a group of normal hearing individuals. Tan had
argued that this cue-benefit was an important part of the filter‟s formation; however, no
direct measurement of the attentional filter was made. The apparent loss of the
attentional filter in Moore et al.‟s SNHL participants, and the loss of a cue-benefit in
Tan‟s SNHL participants is consistent with the results of Scharf et al.‟s vestibular
neurectomy studies in suggesting that the normal formation of the attentional filter, or at
least components of the filter in the loss of the cue-benefit, is dependent on the normal
function of MOCS action on hearing.
Improved speech reception in noise might be expected from a functioning attentional
filter, as the filters presence indicates an ability to attenuate unwanted signals in favour
General introduction
24
of a repeated, or expected signal. Evidence of such an improvement with increasing
OAE suppression by contralateral broadband noise would provide additional support for
the proposed role for the MOCS in forming the attentional filter. In line with this
expectation, some studies have found a positive correlation between the magnitude of
suppression of OAEs by contralateral broadband noise and improvements in speech-in-
noise intelligibility (Giraud et al., 1997, De Boer and Thornton, 2008, Kumar and
Vanaja, 2004). However, conflicting results are present in the literature, as other studies
have either failed to replicate this relationship (Wagner et al., 2008), or found the
opposite relationship (de Boer et al., 2012). The most apparent sources for these
conflicting results are a change in the type of OAE measurement and the speech-in-
noise intelligibility measurement used in the experiment. The experiments that found an
improvement in speech-in-noise intelligibility used the suppression of TEOAEs, which
are evoked in response to broadband click stimuli. In contrast, Wagner et al. (2008)
found no relationship between speech-in-noise intelligibility and the suppression of
DPOAEs, which are measured using 2 pure tones near to each other in frequency. These
two OAE types are thought to be produced by two different mechanisms in the cochlea
(for a review see Shera and Guinan (1999)), although the significance of a result present
with one OAE type but not the other is unclear. The conflicting results may also be due
to different stimuli used; in almost all experiments listed above the speech stimuli were
different. The clearest example of the effects of a change in stimulus comes from de
Boer et al. (2008) and de Boer et al. (2012). These studies used identical procedures and
equipment, with the only changes in the participants and the speech signals used. In the
studies, the suppression of TEOAEs by contralateral noise was correlated with the
discrimination of /bee/ to /dee/ in 2008, and then of /ga/ to /da/ in 2012. Significant
correlations between OAE suppression and phoneme discrimination were found in both
studies, but they were in opposite directions: a positive correlation for improved
General introduction
25
phoneme discrimination with increased OAE suppression was found in 2008, but in
2012 a negative correlation for impaired phoneme discrimination was found. The above
studies demonstrate a complex relationship between the activation of the MOCS and the
detection of signals in noise, in which the type of OAE used to estimate MOCS activity
and the type of signal that is to be detected has important consequences on the direction
of the relationship.
Overall, the above research supports a role for the MOCS in improving the detectability
of transient tones within a background noise. The most direct support that the MOCS
forms the attentional filter is the loss of the filter in Scharf et al.‟s participants after a
vestibular neurectomy. This result occurred without any changes in pure-tone thresholds
or frequency discrimination, and so cannot be explained by any impairment to the
auditory system except for a reduction in MOCS efferent control of the cochlear
amplifier.
1.9 The medial olivocochlear system: Related structures
When considering the involvement of the medial olivocochlear system in the formation
of the attentional filter, it is important to consider related structures which may be
difficult, or impossible, to separate from the medial olivocochlear system during the
experiments.
In addition to the medial olivocochlear efferents, which have myelinated axons that
originate from the medial part of the superior olivary complex, there are lateral
olivocochlear efferents, which have unmyelinated fibres that originate from the lateral
part of the same structure (Rasmussen, 1960). These lateral olivocochlear efferents,
which form the lateral olivocochlear system (LOCS) with their inputs, travel through
the vestibular nerve as the MOCS efferents do, but project to the dendrites of the
auditory afferent fibres at the base of inner hair cells (Warr and Guinan Jr, 1979). The
General introduction
26
synapses of the LOCS efferents contain ACh, although there is evidence of additional
neurotransmitters and neuromodulators, including GABA and CGRP in humans
(Schrott-Fischer et al., 2007). The LOCS receives input from the auditory afferents, but
the efferent fibres have not been directly shown to respond to sound. The LOCS is
presumed to respond to sound, as it receives input from the auditory afferents, and the
lateral superior olivary complex from where the fibres originate has been shown to
respond to ipsilateral auditory input (Adams, 1995). The LOCS cannot be directly
activated by electrical stimulation, as its unmyelinated efferent fibres cannot be
stimulated electrically, but indirect activation is possible by the electrical stimulation of
the inferior colliculus (Groff and Liberman, 2003). This indirect electrical stimulation
resulted in complex changes in auditory nerve firing rates, which took over one minute
to take effect, and lasted up to 5 to 20 minutes after the stimulation ended. Thus, the
function of LOCS activation is unclear, although two proposed roles are the balance of
left to right auditory inputs (Darrow et al., 2006) (although subsequent research by
Larsen and Liberman (2010) did not replicate this result), and the protection of the
auditory system from the effects of aging (Liberman et al., 2014). Importantly for the
work described in this thesis, the MOCS and LOCS systems may interact, and
disentangling the effects of the systems is not always possible. These interactions occur
at a basic level because the systems both receive auditory afferent input, and act to
change the auditory nerve‟s response to input. However, there is evidence for a direct
connection, as the LOCS efferent fibres appear to synapse on MOCS efferents en
passant (Liberman, 1980). Thus, the potential interactions between the LOCS and the
MOCS may prevent a clear separation between the actions of the two systems.
Both the MOCS and the LOCS have collateral branches to the cochlear nucleus in
addition to their efferent outputs to the cochlea (Brown et al., 1988; Ryan et al., 1990).
Collaterals from the two olivocochlear systems innervate specific cell types in the
General introduction
27
cochlear nucleus, and there is evidence of both excitatory and inhibitory effects on
specific cell types (Mulders et al., 2002; Mulders et al., 2003). The function of these
collaterals is unclear, although the MOCS collaterals have been proposed to modulate
the system‟s inhibitory effects on the cochlear amplifier (Benson and Brown, 1990; Kim
et al., 1995; Mulders et al., 2003). Previous research on the formation of the attentional
filter in individuals with altered auditory system function has not tested the potential
influence of the MOCS collaterals on the filter‟s formation. For example, the sectioning
of the vestibular nerve in Scharf et al.‟s research (1997) might not directly change the
collaterals activation in the cochlear nucleus, nor would the loss of the MOCS efferent
targets after SNHL in Moore et al. (1999). It is unclear whether the collaterals have any
role in forming the attentional filter, if indeed the MOCS is responsible for the filter‟s
formation, however the function of the collaterals is likely to be linked to the function
of the MOCS efferents, and their presence must be considered when measuring MOCS
action.
1.10 Central mechanisms
Prior research has argued that auditory attention relies on the formation of perceptual
objects in a manner similar to visual attention, and it is this formation of perceptual
objects that drives the detectability of target stimuli versus off-target stimuli (Shinn-
Cunningham, 2008). Indeed, recent research has demonstrated that when attention is
focused on a single speaker within a background noise, auditory cortical activity
synchronizes with the temporal modulations of the speaker, and that the precision of
this synchronization predicts the speech recognition of the listener (Ding and Simon,
2013). Thus, the detectability of complex speech signals in background noise may be
directly linked to centrally-mediated auditory attention, and the formation of the
attentional filter may be partially a result of these central mechanisms. However, a filter
formed only by central mechanisms is not sufficient to explain the loss of the filter after
General introduction
28
a vestibular neurectomy in Scharf et al‟s studies, in which severing the MOCS efferent
fibers resulted in a loss of the attentional filter, but no other significant impairments to
auditory function (Scharf et al., 1994; Scharf et al., 1997). This strongly suggests a
significant involvement of the efferent output of the MOCS in the formation of the
attentional filter. In the present work, the MOCS is suggested to modulate the
detectability of transient tones prior to these tones becoming perceptual objects, and
therefore these central mechanisms are not a focus of the present work.
1.11 Structure and aims of the thesis
This thesis examines a suggested role for the medial olivocochlear system‟s efferent
control of the cochlear amplifier in forming the attentional filter. Previous research has
established that the MOCS has an anatomy, physiology and function consistent with the
proposed role, and that severely reducing the MOCS action on the cochlea, through a
moderately-severe SNHL or a vestibular neurectomy, reduces the depth of the
attentional filter. The present work aims to correlate the strength of the MOCS with the
depth of the attentional filter, as well as extend the research that shows that the
attentional filter is impaired in conditions presumed to impair the MOCS action on
hearing. Three research questions will be addressed in the thesis. First, is there evidence
that the MOCS acts to form the attentional filter in normal hearing individuals? Second,
does the reduction in depth of the attentional filter with presumed MOCS impairment
scale with the severity of that impairment? Third, can the attentional filter be formed in
conditions where no MOCS action on hearing is possible? The structure of the thesis is
as follows:
Chapter two details the psychophysical methods common to all experiments.
Chapter three focuses on the contralateral noise suppression of OAEs as an index of the
strength the MOCS acoustic reflex in normal-hearing participants, and whether this
General introduction
29
correlates with the depth of the attentional filter. The aim of these experiments was to
investigate whether individuals with relatively strong MOCS acoustic reflexes had
deeper attentional filters when compared with individuals with relatively weak reflexes.
Chapter four continues the use of auditory dysfunction to test the MOCS involvement in
forming the attentional filter, using individuals with SNHL and conductive hearing loss.
SNHL with established OHC damage, by the loss of detectable OAEs, are considered to
have reduced MOCS function due to a loss of the MOCS efferent targets, and there is
evidence that the reduction in MOCS function scales progressively with increasing
hearing loss. This is an opportunity to investigate the formation of the attentional filter
in individuals with decreasing MOCS function as hearing loss increases. Conductive
hearing loss participants were used as a control group presumed to have intact cochlear
amplifiers, but long-standing hearing impairments. The aim was to investigate whether
individuals with SNHL show similarly impaired attentional filters to Scharf et al.‟s
(1997) vestibular neurectomy participants, and if this impairment scaled with the
severity of the hearing impairment.
In chapter five, the attentional filter is measured in individuals presumed to have no
remaining MOCS action on hearing, due to profound SNHL and the use of a cochlear
implant (CI) that bypasses the cochlear amplifier. The experiments contained within this
chapter test for the presence of the attentional filter in six individuals with CIs using
acoustic presentation of the stimuli with a loudspeaker. CI recipients give a unique
perspective on selective attention because the cochlear amplifier and much of the
peripheral auditory system are bypassed, and this is combined with near-normal
auditory thresholds, a substantially different electrical form of hearing, and a period of
auditory relearning.
General introduction
30
In chapter six, attentional filters are again measured in CI recipients; however, the
stimuli are presented using a programmed, direct stimulation mode. This stimulation
mode does not include an acoustic stimulus, and so there is no possible involvement of
the MOCS efferent control of the cochlear amplifier on the reception of the signals. In
addition, the programmed, direct stimulation mode enables a specific set of stimuli to be
presented to each participant, without the use of the commercial speech processor that
was used in the experiments of chapter five.
Chapter 2. General Methods
General methods
32
2.1 Acoustic Stimuli
All experiments were conducted in a sound-attenuating room using a Windows PC with
an ASUS Xonar STX soundcard located in an adjacent room. Except when otherwise
specified, the continuous broadband noise was generated on a separate Windows laptop
running SoundForge XP v4.5. The amplitude of this noise was calibrated to 60 dB(A)
SPL with a Brüel & Kjær 2260 Sound Level Meter combined with a B&K Artificial Ear
Type 4152. Figure 2.1 shows the spectrum of the background noise, with a flat response
from 1 to 3 kHz, and less than 1-dB variation in power. The output of the stimuli-
generating Windows PC and the noise-generating Windows laptop were mixed by a
Behringer Eurorack MX 802, and presented diotically in all experiments except those in
chapter 5, with a pair of Sennheiser HD-280 PRO headphones.
Figure 2.1. Sound spectrum of the broadband noise used during the experiments. Sound
spectrum recorded using the same equipment used to calibrate the 60 dB(A) background
noise, with the output of the sound level meter to a Powerlab 4ST by ADInstruments in
10 second samples. Units are reported in dB relative to 0 mV recorded by the meter,
with an overall amplitude of 60 dB(A) SPL.
General methods
33
2.2 Psychophysical procedures
The psychophysical tasks described in this thesis use a 2 interval forced choice design
(2IFC), the design of which is shown in Figure 2.2. The 2IFC structure presents two
detection intervals, only one of which contains a to-be-detected signal, and requires the
participant to select the interval with the signal. This avoids criterion effects because it
forces the participant to choose which detection interval contained a signal, rather than
whether the signal was presented. Because there are two intervals to choose from, when
a participant is unable to detect the signal their detection rate will be at 50%.
Figure 2.2. Structure of the 2IFC procedure. Each trial began with a 600-ms period with
a blank interval presentation box. The first 300-ms held a cue in some experiments, but
otherwise contained only noise. The first detection interval began at 600 ms, as
indicated by a “1” in the interval presentation box, which remained for the duration of
the interval. The interval presentation box was blank for the 300 ms between the two
detection intervals, until the second detection interval at 1200 ms. Participants were
prompted to respond 300 ms after the completion of the second detection interval, and
no response input was recorded prior to this request. The next trial began 1 s after the
participant‟s response. In almost all the experiments, broadband noise was present
continuously throughout the experiment.
There are two procedures that are common to each experiment in this thesis. The first is
a threshold estimation procedure, which was used at the beginning of each session. The
second is the probe-target procedure, which was used to measure the attentional filter.
Small alterations are made to the procedures in each chapter, and these will be defined
in the appropriate chapters. The psychophysical procedures were programmed in
General methods
34
LabVIEW 7 or 12 and presented on a Windows 7 PC with an ASUS XONAR STX
Soundcard, at a 44100 Hz sample rate.
2.3 Auditory thresholds
Auditory thresholds were estimated in the presence of continuous background noise
using a three-down one-up adaptive staircase procedure that produced a threshold
corresponding to a 79% detection rate (Levitt, 1971). The threshold procedure used the
2IFC structure described above. One of the 300-ms long intervals was randomly
assigned to contain a stimulus, chosen with a 50% probability and no constraints.
Subjects signalled their response by left clicking if they thought they heard the signal in
interval 1, or right clicking for interval 2. Correct responses were signalled to the
participant with a green visual indicator and incorrect responses by a red visual
indicator displayed on the monitor. Initially, a correct response changed tone amplitude
in 5-dB steps, which changed to 1-dB steps after the first incorrect response. Eighty
trials were presented for each threshold procedure, and the mean of the last eight
reversals, i.e. a change from correct-correct-correct-incorrect or incorrect-correct, was
taken as the threshold estimate.
2.4 Cued probe-target procedure to measure the attentional filter
The attentional filter was measured over at least 4 one hour long experimental sessions.
The first session contained 3 practice runs each of the threshold and probe-target
procedures. Subsequent sessions, which were held at least one hour but not more than
one week apart, began with threshold estimation, and then 3 runs of the probe-target
procedure. The probe-target procedure is based on those previously reported in the
literature (Greenberg, 1962; Dai et al., 1991; Schlauch and Hafter, 1991; Botte, 1995;
Tan et al., 2008).
General methods
35
The probe-target procedure used the same 2IFC structure to measure the attentional
filter, with the addition of a cue tone in some experiments. Four probes were presented
around the 2.0-kHz target frequency; the specific probe frequencies will be outlined in
the relevant chapters. In the experiments with a cue tone, shown in table 2.1, a 300-ms
2-kHz cue tone preceded the first detection interval by 300 ms, and was presented at 14
dB above the 2-kHz target tone‟s threshold. In the experiments with normal-hearing
participants, the target and all probes were presented at the previously measured 2.0-
kHz threshold. The alternative procedure for hearing impaired participants will be
outlined in the relevant chapters. Each run of the experiment contained 192 trials, with
the order of the to-be-detected tones set at the beginning by randomizing blocks of 48
trial tones plus 4 of each of the probe tones equating to 75% target presentation and
equal numbers of each of the probe tones over the remaining 25%. The 192 trial runs
were repeated 3 times in each session, with a 5 minute break between each run. In total
the target was presented 1296 times, and each probe 108 times for each participant.
Table 2.1. Inclusion of a cue tone during the attentional filter experiment.
Experiment Cue?
Chapter 3: Normal hearing participants Yes
Chapter 4: Hearing impaired participants Yes
Chapter 5: Cochlear Implant recipients using
acoustic presentation No
Chapter 5: Cochlear Implant recipients using
programmed, direct presentation Yes
General methods
36
Chapter 3. Formation of the attentional filter in normal-hearing
participants
Formation of the attentional filter in normal hearing participants
38
3.1 Introduction
The formation of the attentional filter is impaired when the efferent fibres of the MOCS
are sectioned (Scharf et al., 1994; Scharf et al., 1997). Currently, this is the only direct
evidence that the function of the MOCS affects the formation of the attentional filter,
and it relies on an abnormal, surgical dysfunction of the MOCS in individuals with pre-
existing medical conditions, mainly Ménière‟s disease. At present, no research has
demonstrated a relationship between the function of the MOCS and the formation of the
attentional filter in individuals with normal MOCS function and normal auditory system
function. Such a relationship is expected if the MOCS does act to form a significant
portion of the attentional filter.
To study the relationship between the function of the MOCS and the formation of the
attentional filter, the experiments included in this chapter used the suppression of OAEs
by contralateral broadband noise as an index of MOCS function (Berlin et al., 1994).
This measurement of MOCS strength was then correlated with specific features of the
attentional filter, to test whether a stronger MOCS was associated with a deeper
attentional filter. In past literature the typical range of OAE suppression has been
between -1 to 3 dB, which has correlated with up to 10% improvements in phoneme
recognition (Giraud, 1997), or a 5 dB improvement in the detection threshold of a multi-
tone complex in noise (Micheyl et al., 1995). Previous research has suggested that the
attentional filter may be formed by up to 7 dB of effective suppression at the distant
probes (Dai et al., 1991), which provides an adequate size of change to be detectable
with OAE suppression, if this relationship is present.
The experiments described in this chapter aimed to correlate the magnitudes of
suppression of both TEOAEs and DPOAEs by contralateral broadband noise with
specific features of the attentional filter. In the primary experiment, a measurement of
Formation of the attentional filter in normal hearing participants
39
OAE suppression was made in each individual in a separate session, which was held
after three sessions used to measurement the attentional filter. This is a commonly used
procedure in the literature, and is based on the assumption that the magnitude of OAE
suppression by contralateral broadband noise is a stable measurement in an individual
over time. A previous study supported the assumption by showing that the suppression
of TEOAEs is stable in individuals across sessions, with a within-subject variability of
0.01 to 0.07 dB, and a Cronbach‟s alpha of 0.8 (Mishra and Lutman, 2013). The same
data do not exist for the suppression of DPOAEs. A preliminary study is included in this
chapter to examine the validity of this assumption in comparison with the previous
research on TEOAEs, and as an extension of this research to DPOAEs.
Formation of the attentional filter in normal hearing participants
40
3.2 Methods
3.2.1 Participants
Six male and 9 female individuals ranging in age from 19 to 25 years participated in this
research (median age = 22). Participants had normal hearing (<20dB HL) from 250 Hz
to 8 kHz as tested using a Grason Stadler GSI 61 Clinical Audiometer. All
psychophysical stimuli were presented diotically through Sennheiser HD280 Pro
headphones.
3.2.2 MOCS acoustic reflex strength measurement
The strength of MOCS acoustic reflex in each subject was assessed using contralateral
noise suppression of both DPOAEs and TEOAEs using an Otodynamics ILO292 with
an Otodynamics DPOAE Probe. Probe fit was tested at the beginning of each trial using
the procedure specified by Otodynamics, and the probe was left in position between
trials unless a refit was required, as indicated by the Otodynamics ILO v6 software.
Each OAE waveform was averaged into one of two alternating buffers, A and B. The
response magnitude was measured from the average of the A and B buffers, while noise
was estimated from an A minus B waveform. Waveform reproducibility was measured
by the cross-correlation of the averaged A and B buffers, and OAEs were only included
when reproducibility was above 80%. All normal hearing participants had DPOAE
amplitudes greater than 5 dB tested for L1 = L2 = 65 dB elicitors, with f2 = 2 kHz and
with f1/ f2 = 1.22
DPOAE fine structure was assessed by measuring the DPOAE responses for L1 = L2 =
55 dB elicitors, with f1/f2 = 1.22 and f2 frequencies of 220 Hz above and below 2 kHz
measured in small increments (initially 50 Hz followed by 20 Hz in the vicinity of
identified peaks in DPOAE amplitude). Following the method described by Abdala et
al., (2009) DPOAE suppression was measured at an individual‟s peak, or maximum,
Formation of the attentional filter in normal hearing participants
41
DPOAE response in their fine structure (Abdala et al., 2009). In cases where two peaks
of identical amplitude were found, the peak closer to 2 kHz or the peak at a higher
frequency was used. For suppression measurements, a DPOAE I/O function was
measured for L1 = L2 = 45 to L1 = L2 = 55 dB SPL in 1 dB increments, with f2 equal to
each individual‟s peak of DPOAE amplitude and with f1/f2 = 1.22. For each
measurement, DPOAE amplitude was averaged over at least 60 seconds. A WO
condition that did not include contralateral noise was interleaved with a WN condition
that included a contralateral 60 dB SPL broadband noise, presented with the Sennheiser
HD-280 PRO headphones. The conditions were separated by 60 seconds. Three
measurements each of the WO and WN conditions were made, and DPOAE suppression
is expressed as a decrement from the averaged WO amplitude minus the averaged WN
amplitude at either the L1 = 45 or L1 = 55 dB SPL amplitudes. Figure 3.1 shows
examples of the OAE data collected during the experiments.
Formation of the attentional filter in normal hearing participants
42
Formation of the attentional filter in normal hearing participants
43
Figure 3.1. OAE responses for two subjects included in chapter 3. A: DPOAE
amplitude spectrum for a single measurement, including the elicitors used to evoke the
emission, the emission itself, and the noise included in the recording. B: DPOAE fine
structure measured for a single participant, showing DPOAE measurements that were
taken from 1850 to 2150 Hz at 50-Hz intervals, with a further 4 measurements taken
around the highest DPOAE response at ±10 Hz and ±20 Hz intervals. The arrow
indicates this individual‟s peak DPOAE response magnitude. C: DPOAE input/output
function for a single participant. WO represents the without noise condition, and WN
the with contralateral broadband noise condition. Subsequent DPOAE suppression is
measured as the decrement between the WO and WN conditions at the L1 = 45 or L1 =
55 dB amplitude. D: TEOAE response waveform for the WO condition (black line) and
WN condition (grey line) from a single participant.
The TEAOE threshold was determined following the procedure described by De
Ceulaer et al. (2001). TEOAEs were evoked by brief clicks (80-µs duration) in the
presence of 60-dB SPL broadband noise in the contralateral ear. Clicks were presented
at 50 Hz, in the linear presentation mode which presents all clicks at an equal
magnitude. Although this presentation mode does not remove the presence of the initial
click from the recording, known as the click artefact, like the non-linear mode, which
presents 3 clicks equal in magnitude and a fourth click with inversed polarity and tripled
in magnitude, it does keep the entire emission intact, and preserves useful emission
information. In the following data, the click artefact is limited by the relatively low click
amplitude. The ILO292 software uses a fast Fourier transform (FFT) to measure the
magnitude of the TEOAEs over a window beginning 20 ms after the click, with a cosine
window applied to the first 2.5ms to attenuate the ringing of the click stimulus. Initially
the clicks were presented at 60 dB SPL, which was decremented in 3-dB steps until the
TEOAE was no longer detectable above the noise floor. Subsequent TEOAE testing
was performed 12 dB above this threshold in each participant, which previous research
has demonstrated to be the most effective elicitor amplitude with a 40 dB SPL
contralateral noise suppressor (De Ceulaer et al., 2001). Each trial ran for 260
Formation of the attentional filter in normal hearing participants
44
presentations of the click train. A WO condition that included no contralateral noise was
interleaved with a WN condition that included a contralateral 60 dB SPL broadband
noise, presented with the Sennheiser HD-280 PRO headphones. TEOAE suppression
was expressed as the decrement in TEOAE amplitude between the averaged WO and
the averaged WN conditions.
In addition to activating the MOCS acoustic reflex, contralateral broadband noise has
been shown to activate the middle-ear muscle reflex, and this activation can confound
the effect of MOCS activation. Ten participants were tested for the activation of the
reflex by testing for an observable change in admittance of the tympanic membrane
caused by the presentation of up to 95 dB SPL contralateral broadband noise using the
same equipment as used in the later experiments. The activation of the reflex was
monitored using a Grason-Stadler GSI-38 Auto Tymp.
3.2.3 Preliminary study
A preliminary study was used to establish typical OAE magnitudes, and the variability
and test-retest repeatability of the suppression of the OAEs. The OAE measurements
were paired with a simplified selective attention procedure on each session, which was
used to ensure that the detection rate of the target and a representative probe were
within a range which would allow an appropriate comparison in the following
experiment. Reliability of the OAEs was measured using Cronbach‟s α, which was
calculated using the following equation:
where n is the number of sessions (4), c is the average of the covariance‟s for each
session, and v is the average variance across the sessions. The interpretation of the
subsequent α is for: α ≥ 0.9 indicates excellent reliability, α ≥ 0.8 is good reliability, α ≥
Formation of the attentional filter in normal hearing participants
45
0.7 is acceptable, α ≥ 0.6 is questionable, α ≥ 0.5 is poor, and less than 0.5 is
unacceptable. Cronbach‟s α calculation and interpretation was taken from the protocol
outlined by Mishra and Lutman (2013).
Three of the fifteen subjects participated in the preliminary study. The study included 4
sessions over two days, with a morning and afternoon session each day, which were
matched in time the following day. Each session consisted of a measurement of DPOAE
and TEOAE suppression, followed by a threshold estimation and then a shortened cue-
probe procedure. The shortened cued probe-target procedure used the same 2IFC
structure as the probe-target procedure described in section 2.3. Only a single 2.08-kHz
probe was included in the measurement to accompany the 2-kHz target tone. The target
and the probe were presented with equal probability, with 192 presentations of each in
every session.
3.2.4 Primary experiment: Attentional filters and the suppression of OAEs
This experiment evaluated the attentional filter and its relationship to the contralateral
broadband suppression of OAEs measured in all 15 subjects. OAE suppression was
measured in each subject in a single separate session after all the psychophysical
measurements had been completed. Participants attended 5 sessions, the first of which
was to conduct initial hearing tests, perform initial OAE calibration as described above,
and two training runs each of the below experiments. Sessions two to four were used to
measure the attentional filter, and then DPOAE and TEOAE suppression were
measured on session 5.
Sessions used to measure the attentional filter began with a threshold measurement
(described in section 2.2) for the 2-kHz target tone in noise. After the threshold
measurement there were 3 runs of the cued probe-target procedure. The cued probe-
target procedure (described in section 2.4) used a 75% presentation of the 2-kHz target
Formation of the attentional filter in normal hearing participants
46
and 25% presentation of the probes at 1.8, 1.92, 2.08 and 2.2 kHz, which were
presented with equal number.
3.3 Results
3.3.1 Activation of the middle-ear muscle reflex
The threshold for activation of the middle-ear muscle reflex was monitored in 10 of the
15 subjects, using broadband noise of up to 95 dB SPL. Six of the subjects showed
activation of the reflex, with one subject showing activation at 75 dB SPL, four at 80 dB
SPL, and one at 85 dB SPL. The other 4 subjects showed no measureable middle-ear
reflex activation at contralateral noise levels of up to 95 dB SPL. The amplitude of
broadband noise used in the later experiments was 60 dB SPL, and so not expected to
activate the reflex.
3.3.2 Preliminary study: TEOAEs
Contralateral broadband noise reliably suppressed TEOAEs, with the three subjects
included in the preliminary study all showing a suppression of the TEOAE magnitude in
each session, as shown in Table 3.1, and as measured by the decrement from the
without noise emission amplitude from the with noise emission amplitude. Table 3.1
also shows the variations in the magnitudes of TEOAE suppression within individuals,
with the coefficient of variation for subjects 1, 2 and 3 equal to 0.26, 0.08 and 0.25
respectively. Figure 3.2 shows the raw TEOAE magnitudes for both the without noise
and with contralateral noise conditions. A concern was whether the magnitude of the
TEOAE suppression was dependent on the initial magnitude of the TEOAE in quiet,
however plotting the raw TEOAE magnitude for with and without noise shows an
approximately linear relationship that is parallel to y = x, which demonstrates that the
magnitude of TEOAE suppression was independent of the initial emission magnitude.
Formation of the attentional filter in normal hearing participants
47
Table 3.1. Magnitudes of TEOAE suppression with mean and standard deviations for
the three subjects on each session in the preliminary study.
Figure 3.2. The relationship between TEOAE response without (WO) and with noise
(WN) for 3 participants over 4 experimental sessions. The dotted line shows y = x.
TEOAE suppression shows a relatively small amount of variability within individuals
when averaged across the four sessions,
The TEOAE measurement was taken twice in the 15 individuals for the experiment
below, to calculate the reliability of the magnitude of TEOAE suppression in the form
of Chronbach‟s alpha. For the suppression of TEOAEs, α = 0.79. TEOAEs were
measured at +12 dB SL, and for the 15 participants the median TEOAE amplitude was
62 dB SPL ± 6.1, which equates to a median TEOAE threshold of 40 dB SPL.
Session 1 Session 2 Session 3 Session 4 Mean SD
Subject 1 0.70 0.75 1.10 1.30 0.96 0.25
Subject 2 1.70 1.45 1.45 1.40 1.50 0.12
Subject 3 0.75 1.30 1.25 1.60 1.23 0.31
Formation of the attentional filter in normal hearing participants
48
3.3.3 Preliminary study: DPOAEs
Contralateral broadband noise resulted in an overall suppression of DPOAEs in all
participants, although as shown in Table 3.2, in Subject 2 there was one session which
had a small increase in emission amplitude. The magnitude of DPOAE suppression
shows more variability when compared to that of TEOAEs, with the coefficient of
variation for subjects 1, 2 and 3 equal to 0.44, 0.92 and 0.19 for L1 = 45 dB DPOAEs,
and 0.57, 0.31 and 0.39 for L1 = 55 dB DPOAEs respectively. Figure 3.3 shows the raw
DPOAE magnitudes for the without noise and with contralateral noise conditions. As
reported for the TEOAEs, the raw TEOAE magnitudes for both the with noise and the
without noise conditions show an approximately linear relationship that was parallel to
y = x, which demonstrates that the magnitude of DPOAE suppression is independent of
the initial emission magnitude.
Table 3.2. L1 = 45 and L1 = 55 DPOAE suppression values on each of the 4
experimental session as well as the mean and standard deviation across the sessions for
Figure 5.4 shows the result of three runs of the threshold tracking procedure at the target
frequency for two normal-hearing participants and two CI recipients. The threshold
tracking procedure typically shows a rapid drop to the threshold, and then a near-flat
track around it. This pattern is present in both the normal-hearing participants and the
CI recipients.
Formation of the attentional filter in cochlear implant recipients using acoustic presentation
114
Figure 5.4. Amplitude of the to-be-detected signal in sound card units as a function of
trial number for two normal-hearing participants (NH#1 and NH#2) and two cochlear
implant recipients (CI#1 and CI#2). The tracking is shown for three sessions using a
1.938 kHz pure tone, with an initial 5-dB step size until the first incorrect response, and
a 1-dB step size thereafter.
The mean detection rates of the target and probes as a function of frequency for the two
CI recipients and the two normal-hearing participants, calculated from the attentional
filter measurement, are shown in Figure 5.5. Although slightly below the 79% threshold
estimate, the detection rates for the CI recipients are sufficiently above the 50% chance-
detection level to demonstrate that the CI recipients were able to detect the tones. The
normal-hearing participants had prior experience in the attentional filter task, and
showed the typical attentional filter shape, with an increase in detection rate at the target
frequency.
Formation of the attentional filter in cochlear implant recipients using acoustic presentation
115
Figure 5.5. The detection rates of the target and probes as a function of frequency for
the two CI recipients and the two NH participants in the pilot study. The error bars show
SEM, where the SEM is above 1%.
5.3.3 Primary experiment: Measuring the attentional filter
The estimated thresholds for the 1.938-kHz tone were averaged for each of the six CI
recipients and are reported in Table 5.2 in decibels relative to the estimated threshold
for a normal hearing individual. While all CI recipients required louder tones to reach
the threshold, CI#2 and CI#4 were within 5 dB of the normal hearing thresholds. CI#7
had the highest threshold, with a tone amplitude 11.8 dB higher than that of a normal
hearing individual. With a severe to profound hearing loss without the cochlear implant,
Formation of the attentional filter in cochlear implant recipients using acoustic presentation
116
this suggests that the tones will be too quiet to activate any residual hearing in the
profoundly deafened CI recipients.
Table 5.2. Estimated thresholds for a 79% detection rate of the 1.938-kHz target tone,
reported in decibels relative to the 79% detection threshold estimated for a normal
hearing participant.
CI Recipient Estimated threshold in
dB re: normal hearing
CI#2 0.9
CI#3 9.7
CI#4 3.7
CI#5 9.1
CI#6 6.8
CI#7 11.8
Attentional filters were measured with acoustic presentation in six CI recipients, and in
three individuals with normal hearing. Figure 5.6 shows the mean detection rates of the
target and the probes as a function of frequency in the CI group and the normal-hearing
group. The attentional filters in the normal-hearing group show an increase in detection
rate at the target frequency in the attentional filter condition compared to the detection
rate of the target in the equal-likelihood condition. These are important indications of
the typical attentional filter, and demonstrate that the filter can be formed using the
larger frequency separation, no cue tone, and monotic presentation in normal hearing
individuals. In contrast however, the CI recipients do not show a clear increase in the
detection rate of the target tone compared to that of the probe tones. Importantly, there
is no clear difference in detection rate between the attentional filter and equal-likelihood
conditions.
Formation of the attentional filter in cochlear implant recipients using acoustic presentation
117
Figure 5.6. Average detection rates of the target and probe tones for the 3 normal-
hearing participants and the 6 cochlear implant recipients. Two conditions were
included, the attentional filter condition with a target at 1.938 kHz, and probes at 1.125,
1.438, 2.5 and 3.313 kHz, and an equal-likelihood condition, with the same target and
probe frequencies, but all tones were presented in equal number. The error bars show
SEM, where the SEM is higher than 1%.
Figure 5.7 shows the mean detection rates of the target and probes for the individual CI
recipients in the attentional filter and equal-likelihood conditions. CI#5, CI#6 and CI#7
do not show clear differences in detection rate between the attentional filter and equal-
likelihood conditions. CI#4 shows a higher detection rate of the target tone compared to
that of the 1.125-, 2.5- and 3.313-kHz probes in the attentional filter condition, and the
target‟s detection rate is higher in the attentional filter condition relative to the equal-
likelihood condition. Thus, CI#4 shows an increased detection rate for the more-
frequently presented target tone which was only present at this frequency when the
target was presented more frequently than the probes, which is consistent with the
normal formation of the attentional filter. CI#2 shows a peak of detection rates at the
target frequency in the attentional filter condition, apparently consistent with the normal
formation of the filter, however this pattern of result is present in the equal-likelihood
condition. This suggests that the apparent attentional filter shown in CI#2 is not due to
Formation of the attentional filter in cochlear implant recipients using acoustic presentation
118
the more frequent presentations of the target tone, and therefore not truly representative
of the presence of the attentional filter. However, this is the only participant to attend
the previous, preliminary study, which included an attentional filter measurement task
with a 75% presentation rate of the 1.938-kHz target frequency. CI#3 demonstrated a
higher detection rate of the two high frequency probes in the attentional filter condition
than in the equal-likelihood condition, whereas detection rates of the target and low-
frequency probes were the same in the two conditions. Thus, the more-frequent
presentations of the target tone failed to increase its detection rate but may have
enhanced the detection rates of higher-frequency tones.
Formation of the attentional filter in cochlear implant recipients using acoustic presentation
119
Figure 5.7. Mean detection rate of the target and probes for the individual CI recipients.
Two conditions are included, the attentional filter condition (solid lines and filled
circles) with a target at 1.938 kHz, and probes at 1.125, 1.438, 2.5 and 3.313 kHz, as
well as an equal-likelihood condition (dashed lines, and open circles), with the same
target and probe frequencies, but all tones were presented in equal number. The error
bars show 99% CIs and * indicates a significant difference in detection rate from that
probe compared to the target tone in only the attentional filter condition.
Formation of the attentional filter in cochlear implant recipients using acoustic presentation
120
5.4 Discussion
The present chapter investigated the attentional filter in six CI recipients, all of whom
had profound SNHL prior to implantation and were presumed to have no remaining
MOCS action on hearing, due to the loss of the MOCS efferent targets (Stebbins et al.,
1979; Hamernik et al., 1989). To support the presence of the attentional filter, the CI
recipients would have to show both a clear increase in the detection rate of the target
tone in the attentional filter condition, and an increased detection rate of the target in the
attentional filter condition compared to its detection rate in the equiprobable condition.
Five of the six CI recipients failed to meet these criteria, and so failed to show the
attentional filter. This result is consistent with the results of chapter 4, and reaffirms the
dependence of the attentional filter on the normal function of MOCS action on hearing.
However, in contrast to the absence of the filter in these participants, CI#4 shows a clear
increase in the detection rate of the target tone compared to the detection rate of the
probes in the attentional filter condition, and an increase in the target‟s detection rate in
the attentional filter condition compared to the equal-likelihood condition. Therefore,
CI#4 fulfils the typical features of the attentional filter, and demonstrates the filter‟s
formation in an individual presumed to have no possible MOCS action on hearing.
Thus, there are two major results in the present chapter. First is the absence of the
attentional filter in five of the six tested CI recipients, which is consistent with the
results of chapter 4 and with previous research by Scharf et al. (1994, 1997) and Moore
et al. (1997). Second is the apparent formation of the attentional filter in CI#4, which
suggests the presence of an alternative mechanism able to form the attentional filter in
some CI recipients.
The absence of the attentional filter in five of the six CI recipients is consistent with
previous research on the formation of the filter in individuals with reduced function of
the MOCS. However, is it possible that the formation of the attentional filter was
Formation of the attentional filter in cochlear implant recipients using acoustic presentation
121
eliminated by the function of the cochlear implant, rather than the impairment to MOCS
function, such as the current spread associated with the cochlear implant? Firstly, is it
possible that the settings of the cochlear implant affected the filter's formation? During
the task, each CI recipient used their individual MAP, and the available dynamic range
on each electrode (i.e. the difference between the comfort level and the threshold level)
varied across the participants. However, to avoid an effect of varying dynamic ranges,
the detection thresholds were estimated for the target and each of the probes, which
prevents inter-subject MAP differences from altering the detection thresholds.
Secondly, current spread may affect the experiment, and there were two types of current
spread that may have been involved. The first is a software-based spread of electrode
activation based on the workings of the speech processor, and the second is an
intracochlear spread of current from the electrodes. The effect of software-based spread
of electrode activation was estimated in the simulations prior to running the full
experiment, with the conclusion that a spread of activation existed towards the
numerically lower or more basal electrodes with increasing stimulus amplitude. This
software-based spread was limited to the electrode immediately adjacent to the targeted
electrode, from the target electrode 11 onto the adjacent electrode 10, and did not occur
on the other tested electrodes, electrodes 9, 12, or 13. Therefore, the software-based
current spread did not affect the electrodes responsible for the to-be-detected tones in
the attentional filter measurement, for the 2.5 kHz tone on electrode 9 or the 1.48 kHz
tone on electrode 13. This was true for simulations of up to +15 dB to an estimated
normal hearing individual‟s threshold, and the CI recipients‟ thresholds did not exceed
+12 dB above the threshold, and so the simulations remain valid. However, no direct
measurement of the stimulation patterns could be made in the individual CI recipients.
It remains possible that the speech processor caused a wider spread of activation during
the attentional filter measurement, as the processor regularly received five different pure
Formation of the attentional filter in cochlear implant recipients using acoustic presentation
122
tones, the target and four probes, which were presented at the centre frequencies of 5
different implant channels. This condition was not tested by the simulations, during
which only one tone was presented in the background noise, and so it is not clear
whether there was a software-based spread of stimulation during the attentional filter
measurement. The second type of current spread is an intracochlear spread due to the
electrodes being immersed in conductive perilymph. This intracochlear current spread
has been shown in previous research to travel preferentially towards the basal side of the
cochlea, and may be substantial with the monopolar stimulation used in the present
work (Cohen et al., 2003), although there is evidence that other stimulation modes, like
the commonly used bipolar stimulation, cause similar levels of spread during a signal-
in-noise task, and so changing stimulation modes would not have addressed this concern
(Kwon and van den Honert, 2006; Snyder et al., 2008). The present work was unable to
measure this intracochlear current spread, and the spread is likely to vary between CI
recipients, as it increases with a roughly linear relationship with stimulus amplitude
(Cohen et al., 2003). If this intracochlear current spread was sufficient to prevent the CI
recipients from discriminating the target and probes, it may have eliminated the
attentional filter. However, this level of current bleed would be detectable as an increase
in the frequency difference limens, and, as discussed below, individuals with cochlear
implants typically have difference limens that are typically much smaller than the
frequency range used in to measure the attentional filter.
A second potential cause of the loss of the attentional filter, which is unrelated to the
loss of MOCS action on hearing, is the reduced peripheral bandwidth caused by
decreased frequency discriminability in implant users. If a decrease in frequency
discrimination was sufficient to make the target and probes indistinguishable, it may
prevent the formation of the attentional filter, similar to the spread of intracochlear
current discussed above. A prior study measured the frequency difference limens for a
Formation of the attentional filter in cochlear implant recipients using acoustic presentation
123
group of CI recipients, of which twenty-one out of forty-nine had CochlearTM Nucleus
CI22M or CI24M implants that are comparable to those used by the CI recipients in the
present work (Gfeller et al., 2002b). At 2 kHz, the median just noticeable difference for
the forty-nine CI recipients was 134 Hz, with a maximum of 314 Hz. This result
suggests that the probes nearest the target in the present work, at 500 Hz below and 562
Hz above the target frequency will be distinguishable from the target. Although this
places the probes nearest the target within 200 Hz of the just noticeable difference, there
are still the probes distant to the target, at 813 Hz below and 1375 Hz above the target
frequency. These more distant probes were outside the just noticeable limits and so
would be expected to be discriminable from the target tone, and thus show the presence
of the attentional filter, if formed in these CI recipients. Thus, although it is possible
that an inability to discriminate the target and the probes nearest the target affects the
worst performers in the present work, this is unlikely to render the target and distant
probes indistinguishable, and therefore is not expected to cause the absence of the
attentional filter in the CI recipients.
CI#2 showed one of the two conditions required for the presence of the attentional filter,
the increase in the detection rate of the target compared to that of the probes in the
attentional filter condition; however, this bias towards the target frequency was also
present in the equal-likelihood condition. This may have happened due to the variations
in detection rate in the equal-likelihood condition that were present in all CI recipients.
These variations in detection rate occurred even though each tone was presented at an
amplitude measured to be detectable on 79% of all presentations for each participant.
An alternative explanation for the apparent bias towards the target frequency in CI#2 is
the prior experience this participant had in an attentional filter measurement task. CI#2
attended the preliminary study, which included three runs of the attentional filter
condition and was held prior to running the equal-likelihood condition. This prior
Formation of the attentional filter in cochlear implant recipients using acoustic presentation
124
training may have caused the participant to form the shape of the attentional filter
during the very similar equal-likelihood condition, and CI#2 was the only participant
included in both the preliminary study as well as the primary experiment. However,
there is no direct evidence to support this interpretation, and so the participant is
considered to show no formation of the attentional filter in the present chapter.
CI#3 showed an off-target enhancement of the high-frequency probes in the attentional
filter condition compared with the equal-likelihood condition, and no change in
detection rate of the target tone. This off-target increase in detectability of the high-
frequency probes may be due to a mismatch between the cochlear region stimulated by
the 1.938-kHz target tone and the non-MOCS mechanism that may have been
responsible for forming the attentional filter in CI#4, discussed below. There is
precedent for such a mismatch, as the electrode array typically stimulates regions of the
cochlea at a more basal location of the normal pitch-place map in the cochlea, due to
limitations in fitting an electrode array in the cochlear spiral (Ketten et al., 1998). As a
result, the electrode-array stimulates at locations on the cochlea typically associated
with a higher pitch percept than the speech processor assigns. Over one to two years, the
pitch brought about by activating each electrode shifts, from a relatively high-frequency
pitch that is reflective of the pitch-place map, to a lower frequency pitch that reflects the
pitches assigned by the speech processor (Reiss et al., 2007). The results of Reiss et al.
(2007) were observed with relatively short electrode arrays, with the Iowa/Nucleus
hybrid implant, that is implanted only in the basal regions of the cochlea. This result is
not seen with all cochlear implants, for example the effect was a non-significant trend in
research that used a longer 31-mm FLEXSOFT MED-EL electrode, perhaps due to the
smaller difference between the pitch-place and assigned frequency on each electrode
with the longer electrode (Vermeire et al., 2015). CI#3‟s abnormal filter may be a result
of this frequency mismatch between the pitch-place regions of the cochlea stimulated by
Formation of the attentional filter in cochlear implant recipients using acoustic presentation
125
the implant and the central cochleotopic map, which may be in the process of
undergoing a plastic adaptation to better reflect the higher-than-expected stimulation
from the cochlear implant.
CI#4 shows the typical signs of the attentional filter, with a detection bias towards the
target over the probes, and an enhancement of the target in the attentional filter
condition compared with its detection rate in the equal-likelihood condition. CI#4 was,
therefore, able to bias sensitivity towards specific frequencies in noise based only on the
history of presentation, to form a normal attentional filter. This participant had a severe
SNHL prior to receiving the cochlear implant, and had a detection threshold at 1.938
kHz that was 5.8 dB higher than a normal hearing individual, and so the to-be-detected
target tone would not be detectable with the participant‟s residual hearing. Thus, CI#4
would not be expected to have any MOCS efferent action on hearing during the
experiment. It is unclear what enables CI#4 to form the filter, but not the other CI
recipients. CI#4 did not have the shortest duration of profound deafness (longer than
CI#3 and equal to CI#2), nor did he have the longest duration of experience with his
current CI. The apparent formation of filter in CI#4 suggests the presence of an
alternative mechanism that is able to form the attentional filter in at least one CI
recipient, and this mechanism must be unrelated to MOCS efferent control of the
cochlear amplifier. Potential sources for the alternative mechanism will be discussed in
chapter 6.
The formation of the attentional filter in the present work may have been affected by the
speech processor. All user-selectable noise-processing strategies were switched off for
the duration of the experiment, and to limit any hardware or software related variations
in the processing of the stimuli the same speech-processor was used for each participant.
However, it is unclear whether the pattern of stimulation in response to the tones was
consistent, both between individual participants and between participants‟ sessions. Any
Formation of the attentional filter in cochlear implant recipients using acoustic presentation
126
variability in the presentation of the stimuli may have resulted in, or prevented, the
appearance of the attentional filter. Thus, the results obtained in the present chapter are
limited by the unknown and potentially unwanted effects of the speech processor used
to receive the acoustic presentation.
The results of this chapter showed the loss of the attentional filter in five of the six
tested CI recipients, all of whom were presumed to have no remaining MOCS action on
the cochlea. This finding is consistent with the results of chapter 4, and supports a role
for the MOCS in generating the attentional filter. However, CI#4 showed the apparent
formation of a normal attentional filter, which suggests the presence of an alternative
mechanism that is able to form the filter in some CI recipients. As yet, it is unclear
whether this mechanism is unique to CI recipients, or if it functions in the normal
formation of the attentional filter. No evidence of the alternative mechanism was found
in chapter 4, with all individuals with at least moderate SNHL showing a substantial
loss of the attentional filter. However, the use of acoustic presentation and a commercial
speech processor in the present chapter may have caused or prevented the formation of
the attentional filter in the present experiments. The next chapter, chapter 6, describes a
similar set of experiments, but using a programmed, direct stimulation mode that uses
no acoustic presentation and a known stimulation pattern, to complement the results
obtained with the acoustic presentation in the present chapter.
Chapter 6. Formation of the attentional filter in cochlear implant
recipients using programmed, direct stimulation
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
128
6.1 Introduction
The attentional filter has been shown to be impaired in conditions presumed to reduce
the MOCS action on the cochlea in previous research (Moore et al., 1996; Scharf et al.,
1997), and in chapters 4 and 5 of this thesis. Consistent with these studies, the results of
chapter 5 showed the absence of the attentional filter in five of six tested CI recipients.
One CI recipient, however, showed a normal attentional filter. The presence of the filter
in this CI recipient, who would not be expected to have any remaining MOCS efferent
targets (Stebbins et al., 1979; Hamernik et al., 1989), suggests the presence of an
alternative, non-MOCS mechanism that is able to form an attentional filter. However,
the experiments described in chapter 5 may have included unknown effects of the
commercial CP810 speech processor which was used during the experiments. While all
user-selectable sound processing strategies were disabled (e.g. Adaptive Dynamic
Range Optimization and AutoSensitivity Control), the effect of the speech processor on
the reception of the signals is unknown. Therefore, it is unclear whether the results of
chapter 5 were physiological phenomena or due to the effects of the speech processor.
In the present chapter, attentional filters were again measured in CI recipients; however,
the stimulation on each electrode was directly programmed during the procedure, and
presented to the participant using the Cochlear Ltd. Nucleus Implant Communicator™
software and an L34 research processor. The L34 research processor differs
significantly from the commercially available CP810 speech processor used in the
previous chapter. The software-based speech processing can be disabled on the L34
research processor, and a known pattern of stimuli can be programmed to be presented
to the electrode array. This eliminates the potentially unwanted, unknown effects of the
speech processor which may have been present in chapter 5. Therefore, the experiments
of the present chapter complement those of the previous chapter, as the attentional filter
is measured in a group presumed to have no possible MOCS action on the reception of
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
129
the signals, with the additional elimination of the potentially unwanted effects of the
commercial speech processor.
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
130
6.2 Methods
6.2.1 Participants
Four of the six CI recipients that had participated in the experiments of chapter 5, CI#2,
CI#3, CI4 and CI#5, participated in the direct stimulation experiments. The durations
and aetiologies of each subject‟s hearing loss are stated in Table 5.1. The CI recipients
used a Cochlear CP810 speech processor in day-to-day living, and had at least one
year’s experience with their current processor. Each of the CI recipients had bilateral
implants, however only the ear with the greatest dynamic range at electrode 11 was
used. All of the CI recipients had severe to profound hearing loss prior to implantation
with pre-operative CUNY (spoken sentences in quiet) scores of 0%. Post implantation
CUNY scores ranged from 90 to 100% (mode = 100%).
6.2.2 Constructing the stimuli
There were no acoustic signals used in the experiments described in the present chapter;
all stimuli were presented directly through the L34 body-worn research processor which
had its microphones disabled for the duration of the experiments. The stimuli were
programmed using the Python programming language (Rossum, 2007) and presented to
the implants using the Cochlear™ Nucleus Implant Communicator (NIC) software and
a Cochlear™ L34 body-worn research processor. By programming the stimuli, the
software-based spread of stimulation that was produced by the speech processor in
response to the cue tone (that was identified as a concern in chapter 5) was avoided,
allowing the use of a cue tone in this experiment.
The construction of the stimuli followed the ACE processing strategy used day-to-day
by the CI recipients (Patrick et al., 2006). According to this strategy the Cochlear™
implants used by the participants in this chapter stimulate on one electrode at a time,
with a biphasic pulse using a 25 µs pulse width and an 8 µs interphase gap. The
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
131
stimulation rate is variable among CI recipients, but a 900 Hz rate per electrode, the
default setting, was used by all participants in the present experiment. The ACE
processing strategy chooses 8 maxima, based on the frequency ranges with the most
acoustic energy, and stimulates on 8 corresponding electrodes in a sequential, ascending
pattern. These 8 electrode cycles must contain 8 non-repeated electrodes. An important
consideration is each CI recipient‟s comfort level on each electrode, as any stimulation
above this level can cause physical discomfort. If, at any point, a desired current level
exceeded the recipient‟s comfort level for that electrode the program stopped before
sending this to the implant, and aborted the test.
All stimuli were programmed on-the-fly; however, a hardware memory limitation
restricted the length of the stimuli to 2.7 seconds. This prevented the use of a continuous
background noise as used in the previous experiments. Instead, the stimuli were
constructed to be equivalent to a 2.7 second noise stimulus, with a 300 ms cue stimulus
beginning 500 ms after the beginning of the noise, and a 300 ms to-be-detected stimulus
in one of two intervals either 300 or 600 ms after the end of the cue stimulus.
Programming the noise was done by stimulating randomly selected electrodes
throughout the noise period. To follow the ACE processing strategy, the noise was
programmed by stimulating 8 randomly selected electrodes during each cycle, from
electrodes 3 to 19. An initial current level at an equivalent of 25% of each electrode‟s
dynamic range was used, with an additional jitter of plus or minus up to 3 current levels
using a Gaussian distribution. The noise was calibrated in amplitude and low to high
frequency balance at the beginning of each experiment as described in Figure 6.1 below.
To insert a cue, target, or probe stimulus, a stimulation of the desired electrode was
substituted into the existing noise array on the 8-electrode cycles. To do this, one of the
electrodes that had been chosen to present noise on one cycle was randomly chosen to
be substituted with stimulation on the desired cue, target, or probe electrode. This was
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
132
done for each cycle over the desired length of stimulation. The target and probes were
presented at the current level previously estimated for a 79% detection threshold, while
the cue was presented at +5 current levels relative to this threshold. An outline of the
structure with an example is shown below in Figure 6.1.
Figure 6.1. Structure of the Python program used to construct the 2.7 second stimuli,
including both the noise and the tone, for the direct stimulation research, including an
example of the program‟s output for a single cycle.
6.2.3 Measuring the attentional filter
Before measuring the attentional filter, the background noise was calibrated to produce
equal loudness from the low to the high frequency electrodes. To achieve this, the
participants were given controls in a graphical user interface to increase or decrease the
amplitude of the stimulus or to replay it. The participants were then presented with a 1
second noise stimulus in the structure described in section 2.2 above, with the amplitude
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
133
on each electrode set at 25% of each electrode‟s dynamic range, the range of amplitudes
between the lowest detectable stimulus and the highest stimulus that does not cause
physical discomfort. Participants were asked to make the stimulus of comparable
loudness to a comfortably spoken conversation. Once this level was reached, the low to
high frequency balance of the noise was calibrated. Using the same graphical user
interface, the participants were able to increase or decrease the amplitude of the “Low”
(electrodes 12, 13, 14, 15, 16, 17, 18, and 19) or the “High” (electrodes 4, 5, 6, 7, 8, 9,
10 and 11) frequencies. The participants were asked to equalise any low or high
frequency imbalances, so that the noise sounded equally loud across the electrode array.
Any response changed the amplitude in current level increments equal to 5% of the
dynamic range on each electrode, and the participants were free to make as many
changes as they saw fit. In practice, very few participants altered the balance of the
noise after setting it to a comfortable loudness. The noise was described by the
participants as a “Hiss”, “Fuzz”, or a “TV set to the wrong channel”.
The threshold measuring procedure followed the same structure as used in the previous
experiments, with an initial change in stimulus amplitude of 5 current levels until the
first incorrect response, after which 1 current level steps were used. Thresholds were
measured at the beginning of each session for the target electrode 11 and each of the
probe electrodes, with the order randomized on each session. All thresholds were
measured in the presence of the previously adjusted background noise stimulation.
The attentional filter measurement procedure used the same structure described in
chapter 2, Section 2, except for the use of 2.7 second long bursts of noise stimuli, rather
than a continuous background noise. The cue stimulus was presented at +5 current
levels relative to the threshold on the target electrode 11, which made it clearly audible,
and the probe stimuli were presented on electrodes 9, 10, 12 and 13. The target was
presented on 75% of all trials, with the remaining 25% equally spread across the probes.
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
134
Participants‟ hearing aids or cochlear implants contralateral to the test ear were switched
off for the duration of the experiment.
6.2.4 Shifted target experiments
A second experiment was conducted using the attentional filter measurement procedure
described above, but the cue and target were shifted to electrode 14. Electrode 14 was
chosen as it was outside the electrode range used in the previous attentional filter
measurement task (electrodes 9 to 13). Therefore, the detectability of the new target
electrode 14 would not have been affected by prior experience, which the results of
chapter 5, and specifically the apparent target bias in CI#2 suggested may affect the
formation of the filter. The probe electrodes were 12, 13, 15 and 16. Only CI#2 and
CI#4 were able to participate in the shifted target experiment.
6.2.5 Statistics
As in the previous chapter, detection rates are reported as a proportion across the three
experimental sessions used to measure the filter, although this is converted to
percentage to be reported in the figures. The confidence intervals for the proportion of
correct responses at each frequency were calculated using the Newcombe-Wilson
method without continuity correction (Method 10 from (Newcombe, 1998), calculated
using Herbert (2013). The Bonferroni correction for multiple comparisons was applied,
and as a result, 99% confidence intervals were used in each subject.
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
135
6.3 Results
6.3.1 Stimulation details
Figure 6.2 shows an example stimulation pattern for participant CI#4 before, during and
after the 300-ms „tone‟ stimulus. These results show a close similarity to those in Figure
5.2, which were a result of simulations for the acoustic stimuli.
Figure 6.2. Mean amplitude, in current level (panel A) and number of stimulations
(panel B) during the programmed, direct stimulation of the experimental stimulus. Top
panel shows mean ± SD.
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
136
6.3.2 Attentional Filter measurements
Figure 6.3 shows the mean detection rates of the target and each of the probes with 99%
CIs. In a similar result to chapter 5, CI#4 shows an increase in the detection rate of the
target compared to the probes, although only significant between the target and the
probes on electrodes 9 and 13. CI#2 shows a similar, apparently normal attentional filter
shape of detection rates, with the detection rate of the target significantly higher than
that of the probes. CI#3 shows no significant differences in detection rate across the
tested frequency range, consistent with showing an absence of an attentional filter. CI#5
showed a detection rate of the target near 50%, which suggests that this participant did
not detect the target, or the probes on electrodes 12 and 13. However, CI#5 was able to
complete the task, supported by a consistent detection of the probe on electrode 9
throughout the task.
It is also interesting to note that the detection rate of the target stimulus was greater than
the estimated 79% detection threshold for CI#4, elevated to approximately 84%, and for
CI#2, greatly elevated to near-to 100%. In CI#4 there was a decrease of the detection
rate of the probe on electrode 9 from its 79% detection threshold, whereas CI#2 had a
detection rate of all probes near to the 79% threshold. This result may represent an
underestimation of the target‟s threshold in CI#2.
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
137
Figure 6.3. Detection rate of the target, on electrode 11, and the probe stimuli on
electrodes 9, 10, 12 and 13. Electrodes arrayed in descending order, as a higher
electrode equates to a lower frequency percept. Error bars show the 99% CIs and *
indicates a significant difference in detection rate from that of the target tone.
The results of the present chapter can be compared with those of the previous chapter, in
which the attentional filter was measured in the same participants, but without a cue and
using acoustic presentation. To make this comparison possible, the pure tones used as
the target and probe stimuli in the previous chapter can be converted to electrode
numbers using the standard frequency allocation table (see Appendix 8.3).
Figure 6.4 shows the detection rates of the acoustically presented pure tone target and
probes from chapter 5 as well as the detection rates of the programmed, direct stimuli
used in the present chapter. Overall, a relatively close match was found for the detection
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
138
rates between the two stimulation modes, which was most notable for CI#2 and CI#4.
CI#2 consistently showed an increase in detection rates of the target tone compared to
that of the probes, consistent with the presence of the attentional filter. However, CI#2
showed a substantial increase in the detection rate of the target stimulus, from a mean
detection rate of 81% with acoustic presentation to a mean of 98% with programmed,
direct presentation. Similarly, CI#4 showed a consistent bias towards the target
frequency over the detection rate of the probes, although there were no apparent
differences in the detection rates between the stimulation modes. CI#3 showed small
changes in detection rates using programmed, direct stimulation compared with acoustic
stimulation, although both results were generally consistent with a flat detection rate
across the tested frequency range. CI#5 had a substantial change in detection rates with
programmed, direct presentation compared to acoustic presentation. With acoustic
presentation, CI#5 showed relatively high detection rates of the target electrode and the
nearby low-frequency probe. With programmed, direct stimulation however, the
detection rates of the target electrode and both electrodes programmed for lower
frequencies were near to 50%, which is consistent with the participant guessing for the
majority of these presentations.
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
139
Figure 6.4. The detection rates of the target and the probes as a function of electrode for
both the present chapter, with programmed, direct stimulation, and the previous chapter,
with acoustic presentation. The centre frequencies are used to convert the target and
probe frequencies with acoustic presentation to electrode number, using the standard
FAT shown in the appendix 8.3. Error bars show the 99% CIs.
An additional “Shifted Target” experiment was performed with CI#2 and CI#4, to test
whether the formation of the filter was a physiological phenomenon, or due to a chance
bias towards the detection of stimulation of the target electrode. In this experiment, the
target and cue electrode was shifted to electrode 14, and the probe electrodes were
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
140
shifted to 12, 13, 15 and 16. Figure 6.5 shows the detection rates of the new shifted
target and probes in CI#2 and CI#4, in comparison with their earlier results. As in the
previous experiments, CI#2 and CI#4 showed a higher detection rate of the target
electrode compared to the probe electrodes, although the new, shifted filter was reduced
in depth compared with the previous filter with a target at electrode 11.
Figure 6.5. The detection rates of the target, on electrode 14, and of the probes on each
electrode in a shifted target experiment, with a comparison to the previous experiment
using a target on electrode 11. Electrodes are arrayed in descending order, as a higher
electrode equates to a lower frequency percept. The error bars show 99% CIs, and *
indicates significant differences in the shifted target experiment, relative to the target
electrode 14.
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
141
6.4 Discussion
Previous work has consistently shown reduced depth of the attentional filter in
conditions presumed to impair the MOCS action on the cochlea, as shown in chapter 4
of the present thesis, and in previous research by Scharf et al. (1997) and Moore et al.
(1999). The results of the chapter 5 were broadly consistent with these findings, as the
filter was absent in five of six tested CI recipients who were presumed to have no
appreciable function of the MOCS efferent targets, the outer hair cells, because of their
severe to profound SNHL (Hamernik et al., 1989). However, in contrast to the earlier
findings, CI#4 satisfied both conditions required to support the presence of the
attentional filter, with an enhanced detection rate of the target frequency over the probe
frequencies that was not present in an equal-likelihood condition. A possible
explanation for the presence of the attentional filter in CI#4 was the use of acoustic
presentation to a commercial speech processor, as the output of the speech processor
was unknown, and may have affected the detectability of particular frequencies. To
address this potential issue, the present chapter used programmed, direct stimulation
which allowed a known, controlled set of stimuli to be presented to the participants.
With the programmed, direct stimulation, two of the four participants, CI#3 and CI#5,
did not show an attentional filter, which was consistent with their results in chapter 5,
and supports the loss of the attentional filter in conditions presumed to impair MOCS
action on hearing. However, two participants, CI#2 and CI#4, showed a higher
detection rate of the target compared with the probes, which was consistent with the
formation of the attentional filter. Importantly, when the target was shifted to a different
electrode, the formation of the attentional filter was replicated at this new target, which
supports a physiological mechanism forming the attentional filter. The presence of a
second, shifted attentional filter mirrors what would be expected in a normal hearing
participant, in whom filters can be shown at multiple frequencies, for example Dai et al.
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
142
(1991) showed the filters formation at five different target frequencies from 0.25 to 4
kHz. This result demonstrates the formation of the attentional filter in two CI recipients,
in a condition with no possible effect of the MOCS efferent control of the cochlear
amplifier on the reception of the signals, and with no influence of the commercial
speech processor that was present in chapter 5.
The absence of the attentional filter in CI#3 and CI#5 supports the results of chapters 4
and 5 in the present work, by the absence of the filter in conditions presumed to have no
MOCS action on the cochlea. In the previous chapters, there was a concern that the
filter was not present due to the reduced peripheral bandwidth associated with SNHL or
a cochlear implant, rather than specifically the loss of MOCS action. This reduction in
frequency discrimination is of considerable concern in the present chapter, as the CI
recipients were tasked with discriminating between the target electrode and electrodes
that were either adjacent or separated by only one other electrode. In previous research
including 9 CI recipients, only 2 were able to perfectly discriminate every electrode,
although these recipients used a different electrode type to those used by those in the
present work (Zwolan et al., 1997). This may have rendered the target and probes
indistinguishable when using programmed, direct stimulation, which would have
prevented the formation of the attentional filter. However, CI#3 and CI#5 showed no
evidence of the attentional filter in chapter 5 with acoustic presentation, in which the
distant probes were well outside the maximum frequency difference limens reported for
forty-nine CI recipients (Gfeller et al., 2002a). In addition, the CI recipients used in the
present work had excellent speech recognition in quiet, which would not be expected of
CI recipients unable to discriminate over 8 electrodes. Taken together, the absence of
the attentional filter in CI#3 and CI#5 with acoustic presentation as well as with
programmed, direct stimulation demonstrates that this absence was not due to the
effects of the speech processor, and suggests that the absence of the filters with direct
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
143
stimulation was not due to reduced frequency discrimination. The replicable absence of
the filter in these CI recipients is consistent with the dependence of the filter on the
normal function of MOCS action on hearing.
CI#4 demonstrated the normal formation of an attentional filter in chapter 5 with
acoustic presentation and in the present chapter with programmed, direct stimulation.
There was no indication that CI#4 has remaining MOCS function, or was otherwise
substantially different from the other CI recipients. CI#4 had severe SNHL prior to
implantation, which is consistent with a total loss of function of the MOCS efferent
targets (Hamernik et al., 1989). CI#4 did not have an unusual duration of severe SNHL
prior to implantation, as his duration was matched by CI#2, longer than that of CI#3,
but shorter than CI#5, nor did CI#4 have an extended experience with his implant
compared with the other CI recipients. Indeed, the formation of the attentional filter in
the present chapter demonstrates the filter‟s presence in response to purely electric
signals, as no acoustic signals were used in the programmed, direct stimulation.
Therefore, the cochlear amplifier could not have affected the reception of the stimuli, if
there were remaining MOCS efferent targets, which conclusively eliminates the efferent
output of the MOCS as a possible cause for the formation of the attentional filter in this
experiment. This consistent, replicable formation of the filter in CI#4, without the
influence of the MOCS action or the commercial speech processor used in chapter 5,
supports the presence of an alternative mechanism that is able to form the attentional
filter in some CI recipients.
The second participant to show the attentional filter in the present chapter, CI#2, did not
satisfy the conditions for the filter‟s presence in chapter 5, because the enhancement of
the target was also present in an equal-likelihood condition. In chapter 5, it was
suggested that this enhancement was due to the participant‟s prior training in the
attentional filter condition, during the preliminary study. In these sessions, the target
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
144
frequency was presented on 75% of all trials, and these sessions were held before
running the equal-likelihood condition. This prior experience may have created a
persistent attentional bias towards the more-frequently presented target frequency, and
this bias may have affected his performance in the equal-likelihood procedure, which
used the same probe frequencies. The formation of the attentional filter with the shifted
target in the present chapter supports the real formation of the attentional filter in CI#2,
although it remains possible that the participant had a chance bias towards the target
electrodes, which would have caused the apparent formation of the filter. In retrospect,
it would have been useful to run a shifted target experiment that included the previous
target stimulus, but presented rarely as a probe stimulus. This method may have shown
a reduction in the previous target‟s detection rate, which would have provided stronger
support for the physiological presence of the attentional filter.
The source of the alternative mechanism that formed the attentional filter in CI#4, and
perhaps in CI#2, is not necessarily independent of the MOCS, although it cannot
involve the MOCS efferent control of the cochlear amplifier. The MOCS has another
output, as collateral fibres that project to the cochlear nucleus (Rasmussen, 1960;
Rasmussen, 1967; White and Warr, 1983; Brown et al., 1988; Winter et al., 1989). The
MOCS collaterals have been shown to have both excitatory and inhibitory effects on
cochlear nucleus neurons‟ spontaneous firing rates and responses to sound (Mulders et
al., 2002; Mulders et al., 2003). The function of these collaterals is unclear, although
they have been proposed to compensate for the MOCS‟ inhibitory effect on the cochlear
amplifier (Benson and Brown, 1990; Kim et al., 1995). It may be that the collaterals can
have a role in the formation of the attentional filter in some CI recipients, although this
role is speculative.
An alternative mechanism for the filter‟s formation may be located in central auditory
structures. Previous research has demonstrated that auditory cortical activity correlates
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
145
with performance in a speech recognition in noise task, which was argued to
demonstrate a selective coding of speech in noise mediated by the auditory cortex (Ding
and Simon, 2013). This central auditory attention may be sufficient to form the
attentional filter in the CI recipients. In support of a centrally located mechanism, the
only CI recipients who showed evidence for the filter‟s presence had relatively short
durations of profound SNHL prior to receiving their cochlear implants, with 1 year each
for CI#4 and CI#2, compared to 4 years for CI#5. A prolonged duration of profound
deafness prior to implantation is well documented to have negative effects on CI
outcomes (Dowell et al., 1985; Dorman et al., 1989; Blamey et al., 2013; Holden et al.,
2013). These poor outcomes are typically attributed to both the degeneration of
peripheral auditory structures, including the spiral ganglion cells that cochlear implants
are thought to communicate with (Nadol et al., 1989; Miura et al., 2002), as well as
central plasticity related to the degeneration of central auditory structures, which causes
the degradation of the fine cochleotopic maps (Robertson and Irvine, 1989; Raggio and
Schreiner, 1999) and/or the invasion of cross-modal plasticity (Rebillard and Rebillard,
1980; Lee et al., 2001). The relationship between deafness-induced plasticity and poor
speech perception outcomes in CI recipients is consistent with a role for central auditory
structures in the forming attentional filter, if the presumed role for the filter in
improving the detectability of signals in noise is true. However, no specific
measurement has been made on the filter‟s formation in relation to deafness-induced
plasticity in CI recipients in the present work or in previous research.
If the central auditory system is the basis for the alternative mechanism forming the
filter in CI#4, how might this mechanism function? Previous research by Moore et al.
(1996) suggested a role for central template matching, as described by Dau et al. (1996),
in forming the attentional filter. The template-matching mechanism is consistent with
some of the known features of the attentional filter, as prior research found that
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
146
providing a more effective template improves detection rates in a manner consistent
with the formation of the attentional filter, such as the benefit provided by imaginary
cues (Borra et al., 2013), as well as the increased detection rates of probes when they
match a cue in duration (Wright and Dai, 1994). Thus, the attentional filter in CI#4 may
have been formed, at least partially, by an effective template-matching mechanism
which may have remained or emerged in this participant due to his relatively short
duration of profound deafness prior to implantation.
In summary, the present chapter replicated the formation of the attentional filter in
CI#4, both without the influence of the speech processor, or any possible effect of the
MOCS control of the cochlear amplifier on the reception of the electrical stimuli.
Therefore, CI#4 has demonstrated a consistent and replicable formation of the filter in
this thesis. This finding supports the presence of an alternative mechanism that is able to
form the filter in some CI recipients. In contrast, participants CI#3 and CI#5
demonstrate a consistent absence of the filter in conditions with impaired MOCS action
on hearing, which is consistent with the MOCS role in the normal formation of the
attentional filter and previous research described above. However, with a combined
total of 6 CI recipients in chapters 5 and 6, further studies are needed to elucidate why
CI#4 was the only participant to consistently show the formation of the attentional filter.
Formation of the attentional filter in cochlear implant recipients using programmed, direct stimulation
147
Chapter 7. General Discussion
General Discussion
149
The main aim of this thesis was to better understand the mechanisms that underlie the
formation of the attentional filter, with a focus on the role of the MOCS. The effect of
MOCS function on the formation of the attentional filter was measured in four sets of
experiments. The first set of experiments, in chapter 3, measured the effect of the
strength of the MOCS acoustic reflex using the suppression of OAEs in normal-hearing
participants. The second set of experiments, in chapter 4, investigated the effect of
reduced MOCS action on the cochlea in individuals with SNHL and no detectable
OAEs. The third and fourth sets of experiments measured the attentional filter in CI
recipients, as a group presumed to have no remaining MOCS action on the cochlea. In
CI recipients, the attentional filter was initially measured using acoustic presentation in
chapter 5, and then with programmed, direct stimulation in chapter 6. The findings of
the present thesis suggest a complex relationship between the MOCS and the attentional
filter, in which the typical formation of the attentional filter requires the normal function
of the MOCS action on the cochlea. However, the results also indicate the presence of
an alternative mechanism that is able to form the filter in some CI recipients, which
does not involve the MOCS efferent control of the cochlear amplifier.
In the normal hearing subjects of chapter 3, the strength of a single process of the
MOCS, the MOCS acoustic reflex, was measured using the suppression of OAEs by
contralateral broadband noise. This index of MOCS strength was then correlated with
the depth of the attentional filter. A positive correlation between the strength of the
MOCS acoustic reflex and the depth of the attentional filter would have supported a role
for the reflex in the filter‟s formation. No evidence for this relationship was found,
indeed, a stronger reflex correlated with reduced depth of the filter, and specifically
with a small elevation in the detection rate of the probes on the low-frequency side of
the filter. This result does not support a role for the MOCS acoustic reflex in forming
the attentional filter, although the absence of the expected relationship may have been
General Discussion
150
due to an unsuitable temporal separation between the OAE measurement and the
attentional filter measurement.
The index of the MOCS acoustic reflex‟s strength used in chapter 3 was measured on a
separate session to the measurement of the attentional filter, and this may limit the
index‟s suitability for assessing the MOCS role in forming the attentional filter. The
magnitude of OAE suppression by a contralateral broadband noise has been shown in
previous research to be altered by the task conditions used during the measurement,
which was first shown in a comparison between an auditory-task and a visual-task (Puel
et al., 1988). More relevant to conditions used to measure the attentional filter is
research that showed that the OAE suppression can be increased during a task that
requires focused auditory attention, and that this increase is located specifically at the
frequencies that are under focused attention (Maison et al., 2001). In addition, recent
research has suggested that OAE suppression associated with an auditory task alters
systematically with attention and experience in the task (de Boer and Thornton, 2008;
de Boer et al., 2012). Taken together, these studies suggest that a single measurement of
OAE suppression by contralateral broadband noise that is taken separately from the task
of interest may not be an appropriate measurement of MOCS activity during the task.
Indeed, the studies suggest that an effective measurement of MOCS activity may
require OAE suppression to be measured both during each session with the task
(Maison et al., 2001), as well as across multiple sessions as a subject trains with the task
(de Boer and Thornton, 2008). Therefore, the OAE suppression results in the present
work may not have captured the MOCS activity relevant to the system‟s role in forming
the attentional filter, if such activity exists.
Although the measurement of OAE suppression used in this thesis may not have been
appropriate to test the MOCS role in forming the filter, the measurement may be
informative of the effect of the MOCS acoustic reflex on the reception of near-threshold
General Discussion
151
tones in noise. The magnitude of DPOAE suppression showed consistent, although
weak, positive correlations with the detection rate of the probe tones on the low-
frequency side of the attentional filter, which suggests a role for the MOCS in
enhancing the detectability of transient tones in noise at specific frequencies. The
frequency-tuning of this enhancement is consistent with previous research that
measured the frequency-tuning of the MOCS acoustic reflex. This previous research
used stimulus frequency OAEs, a type of highly frequency specific OAE that uses a
single pure tone probe (Lilaonitkul and Guinan, 2009b, Lilaonitkul and Guinan, 2009a),
or spontaneous OAEs (Zhao and Dhar, 2012), and correlated the magnitude of OAE
suppression with the type and bandwidth of various MOCS elicitors. Lilaonitkul and
Guinan (2009a) demonstrated that the magnitude of OAE suppression elicited by a
narrow-band noise increased with the elicitor‟s bandwidth, and that broadband noise
was a highly effective MOCS elicitor. However, the suppression of the cochlear
amplifier elicited by the broadband noise was not uniform in magnitude across the
frequencies tested (Lilaonitkul and Guinan, 2009a). The maximum suppression of the
cochlear amplifier by broadband noise was located at 1 kHz; with the minimum
suppression located at 4 kHz (2 kHz was not tested). While this is a poor frequency
resolution, it suggests that the broadband noise used in the experiments of chapter 3
would maximally suppress the cochlear amplifier below the 2-kHz target frequency.
The increased suppression of the cochlear amplifier can lead to improvements in the
detectability of transient tones in noise by antimasking, in which the auditory system‟s
adaptation to a sustained noise is suppressed, which leads to an increase in dynamic
firing ranges and improvements in auditory thresholds (Winslow and Sachs, 1987;
Mulders and Robertson, 2000a). In the results of chapter 3, the increased detection rate
of the low-frequency probes is consistent with an antimasking effect, and this increase
occurred for probes located at the frequencies which would receive the maximum
General Discussion
152
suppression by the MOCS acoustic reflex, at least as activated by broadband noise.
Thus, the positive correlation between the detection rates of the low-frequency probes
with DPOAE suppression may have been due to an antimasking benefit provided by the
activation of the MOCS by broadband noise.
It is important to note that the enhancement of the low-frequency probes was consistent
with the activation of the MOCS by the broadband noise, and not by the 2-kHz cue
tone. The studies listed above suggested that a pure tone MOCS elicitor at 2 kHz would
result in the maximum suppression of the cochlear amplifier at frequencies up to half an
octave higher than 2 kHz (Lilaonitkul and Guinan, 2009b). This frequency tuning
suggests that the activation of the reflex by the cue tone was not involved in the increase
in detection rates of the low-frequency probe tones. Any contribution of the cue tone to
changes in detection rates may not be included in the estimation of MOCS strength,
because the suppression of OAEs used to estimate the strength of the MOCS used
contralateral suppression in a separate task to the measurement of the attentional filter.
As discussed in the previous paragraph, the methods of chapter 3 were best placed to
capture the effects of tonic activation of the MOCS acoustic reflex by broadband noise,
as these were the conditions used to suppress the OAEs. Overall, the results of chapter 3
do not support a role for the MOCS acoustic reflex in forming the attentional filter,
although they are consistent with the reflex enhancing the detectability of near-threshold
tones in an unrelated manner.
In chapter 3 the index of MOCS function was limited to the action of the MOCS
acoustic reflex as activated by a contralateral broadband noise, and on a separate session
to the measurement of the attentional filter. This index may only include a subset of
MOCS action, and may not have captured the system‟s activity relating to its suggested
role in forming the attentional filter. In chapter 4 this concern was addressed, by
measuring the attentional filter in individuals with a presumed reduction in MOCS
General Discussion
153
function due to a loss of the MOCs efferent targets, the OHCs, that was association with
SNHL and a loss of detectable OAEs. In these participants the MOCS action on the
cochlea is thought to be impaired, which would affect the actions of the MOCS as a
whole, rather than a subset of the system‟s action. Thus, the results of chapter 4 may be
more representative of the function of the MOCS as a whole, and will include any
effects of the system on the formation of the attentional filter.
The results of chapter 4 demonstrated reduced depth of the attentional filter with
increasing SNHL, and a complete loss of the filter in individuals with at least a
moderate SNHL. This reduction in filter depth, which was associated with impaired
MOCS function, is consistent with previous research into the filter‟s formation in
similar conditions (Scharf et al., 1994; Moore et al., 1996; Scharf et al., 1997), although
the relationship in the present work was more complex than expected, perhaps due to
the considerably larger sample size (n = 14 in the present work, n = 2 for Moore et al.
1996) and range of HLs included. This earlier research included individuals with either
a section of the MOCS efferent fibres that was presumed to be complete (Scharf et al.,
1997) or with a SNHL consistent with a total loss of the MOCS efferent targets (Moore
et al., 1996), and both studies demonstrated reductions in filter depth that were either
incomplete but generally symmetric, or showed a complete absence of the filter. In
contrast, the present results demonstrated an asymmetric reduction in filter depth as a
function of the magnitude of SNHL. The low-frequency side of the filter was found to
be highly sensitive to hearing loss, with this side of the filter reduced in depth at
subclinical hearing levels. On the other hand, the high-frequency side of the attentional
filter reduced in depth beginning with mild SNHL, and continued in a graded manner
with increasing hearing loss at levels of SNHL that are physiologically relevant to
reduced function of the MOCS efferent targets. This graded reduction in the depth of
the high-frequency side of the attentional filter, which occurred in line with presumed
General Discussion
154
reduction of MOCS efferent target function, supports a role for the MOCS in the
generation of the filter.
In chapter 4 it was also shown that the low-frequency side of the attentional filter
reduced in depth with increasing HL in the normal hearing group. This was the same set
of participants that showed a reduction in the depth of the filter on its low-frequency
side with increasing DPOAE suppression in chapter 3. As discussed above, in chapter 3
the decrease in filter depth was attributed to an increase in the strength of the MOCS
acoustic reflex, as activated by broadband noise. Can this conclusion be reconciled with
the apparently similar relationship for decreasing filter depth with increasing HL? As
discussed in chapter 4, there was some evidence that the relationships were linked, with
a near-significant positive correlation between the magnitude of DPOAE suppression
and HL. This potential relationship may have been due to elevations in central auditory
system activity shown to occur after hearing loss, as a plastic response to the decreased
afferent input (Schaette and McAlpine, 2011; Hébert et al., 2013). The increased
activity may have subsequently increased the efferent drive of the MOCS through its
top-down inputs (Knudson et al., 2014), and caused a positive relationship between
subclinical hearing losses and the magnitude of DPOAE suppression. However, this is
an unsatisfying explanation for the reduction in depth of the low-frequency side of the
attentional filter, as this impairment remained through-out the SNHL group, who were
likely to have significant impairments to MOCS action on the cochlea. Instead, the loss
of the low-frequency side of the filter may have been due to an auditory neuropathy
associated with hidden hearing losses (Kujawa and Liberman, 2009; Schaette and
McAlpine, 2011), which are selective for the fibres that form the afferent input to the
MOCS (Furman et al., 2013). This reduction in MOCS afferent drive may have
eliminated the formation of the low-frequency side of the attentional filter, and explains
the initial loss with subclinical hearing losses, and continued absence of the low-
General Discussion
155
frequency side of the filter with SNHL. However, the present work does not include a
measurement of central auditory structure activity, or a sensitive measurement of
auditory neuropathy, and so this interpretation remains speculative.
Overall, chapter 4 presents an argument that the attentional filter is at least partially
formed by MOCS action on the cochlea, as the filter reduced in depth at HLs that are
physiologically relevant to reduced function of the MOCS efferent targets. The
progressive loss of the high-frequency side of the attentional filter with increasing HL
has not been previously reported. Similarly, the loss of the low-frequency side of the
attentional filter at subclinical hearing levels has not been previously reported, and
demonstrates a sensitivity to hearing loss that may have negative effects on an
individual‟s ability to process signals in noise before a clinical level of hearing loss is
reached.
The results of Chapter 4, and the conclusions drawn from them, rely on presumed
impairments to the MOCS due to the loss or damage of its efferent targets, for which no
direct measurement could be made. While a SNHL and the loss of detectable OAEs is
strongly associated with impairment to OHC function (Abdala, 2000, Kim et al., 1996),
the loss of OAEs cannot be used to grade the severity of this impairment. Otoacoustic
emissions are lost with a relatively small impairment to OHC function (Attias et al.,
1995, Kim et al., 1996). Thus, an increasing impairment to MOCS function must be
inferred from the increasing SNHL, a conclusion which is supported by research that
shows that the percentage of damaged or lost OHCs increases with increasing hearing
loss, at least in guinea pigs (Stebbins et al., 1979; Hamernik et al., 1989). However,
SNHL can also be caused by the loss or damage of IHCs or auditory neuropathy, neither
of which would directly reduce MOCS action on the cochlea. The present work also has
no direct or indirect measurements of how IHC damage or auditory neuropathy may
have contributed to each individual‟s hearing loss. Therefore, while a SNHL and a loss
General Discussion
156
of detectable OAEs is associated with reduced MOCS function, and this reduction is
likely to scale with increasing hearing impairment, it is possible that other causes of
SNHL were present that have no relationship with MOCS function. This potential issue
was addressed in chapters 5 and 6, by measuring the attentional filter in individuals in
whom MOCS action on the cochlea was presumed to be impossible, rather than
suffering a degree of impairment.
In chapter 5 the attentional filter was measured in six CI recipients using acoustic
presentation to a commercial CP810 speech processor. The CI recipients had profound
SNHL prior to implantation, and so would not be expected to have a significant inner or
outer hair cell function remaining (Stebbins et al., 1979; Hamernik et al., 1989). In
addition to the presumed loss of hair cell function, the acoustic stimuli were presented at
an amplitude that was within 12 dB of a normal hearing listener‟s threshold, and so
would be too quiet to be affect any remaining OHC function in the severe to profoundly
deafened CI recipients. The attentional filter was measured using comparisons between
detection rates in an equal-likelihood condition, in which the to-be-detected tones were
presented with equal probability, and in an attentional filter condition, which used a
75% presentation rate of the target tone. To support the presence of the attention filter,
there would have to be a clear preference for the target tone that occurred only in the
attentional filter condition, which was demonstrated in a group of normal hearing
individuals. The results showed five of six CI recipients did not detect the target at a
higher rate in only the attentional filter condition, which demonstrates an inability to
bias sensitivity towards specific frequencies based on the history of occurrence, and
suggests a total loss of the attentional filter. The majority of CI recipients, therefore,
showed a result consistent with the results of chapter 4, with the loss of the attentional
filter in a condition presumed to remove possible MOCS action on the cochlea.
General Discussion
157
In contrast to the apparent absence of the filter in five of the six tested CI recipients,
CI#4 showed a clear enhancement of the detection rate of the target above the probes
that was only present in the attentional filter condition, and not the equal-likelihood
condition. Thus, CI#4 satisfied both conditions required to support the presence of the
attentional filter, and this occurred in an individual presumed to have no remaining
MOCS action on the cochlea. This result suggests the existence of an alternative
mechanism, which does not involve the MOCS control of the cochlear amplifier, and is
able to form the attentional filter in at least one CI recipient.
The validity of the results in chapter 5 may be limited due to the use of acoustic
presentation to a commercial speech processor. While every effort was made to disable
the speech processor‟s noise processing strategies (such as Adaptive Dynamic Range
Optimization and AutoSensitivity Control), the processor introduces unknowns which
may have caused or prevented the formation of the attentional filters in chapter 5.
The experiments described in chapter 6 eliminated the potentially unwanted effects of
the speech processor by using programmed, direct stimulation. This programmed, direct
stimulation did not use an acoustic stimulus, which ensures that there would be no way
for the MOCS efferent control of the cochlear amplifier to influence the reception of the
target and probes. The results showed a lack of formation of the attentional filter in two
out of four CI recipients, CI#3 and CI#5, who had shown the absence of the filter in
chapter 5. However, CI#2 and CI#4 had significantly increased detection rates of the
target stimulus over the probe tones. Importantly, the increased detection rate of the
target stimulus was repeated when the target electrode was shifted from electrode 11 to
electrode 14 in both of these CI recipients. These results support those of chapter 5, with
the loss of the attentional filter in some CI recipients, but there was evidence for the
formation of the filter in others. CI#4 in particular, demonstrated the consistent,
replicable formation of the attentional filter in every relevant test of the present work.
General Discussion
158
Thus, the results of chapter 6 support the loss of the attentional filter in conditions of
impaired MOCS action on hearing; however, there is additional support for an
alternative mechanism that is able to form the filter in some CI recipients.
The alternative mechanism responsible for the formation of the attentional filter in CI#4
may be located in central auditory structures. Moore et al. (1996) suggested that
template-matching in central auditory structures aids in the formation of the attentional
filter, based on earlier work by Dau et al. (1996) which has been recently updated
(Jepsen and Dau, 2011). A role for central template-matching in, at least partially,
forming the attentional filter is consistent with the filter‟s formation in response to
complex cues. Previous research has demonstrated the filter‟s formation in response to
cues that require complex frequency extraction (Ebata et al., 2001), harmonic
complexes, or even when the cue was imagined (Borra et al., 2013). The formation of
the attentional filter in response to complex cues may rely on the production of an
effective template in central auditory structures. Template-matching may by the method
used by some of the individuals with sectioned MOCS efferent fibres to form partial
attentional filters in Scharf et al.‟s vestibular neurectomy studies (1994, 1997).
However, if a template-matching mechanism is able to form the attentional filter, it is
unclear why CI#4 was able to use this mechanism, but the other five CI recipients were
not, and it is unclear why the mechanism was not engaged in the SNHL participants
included in chapter 4. Recent research has demonstrated that the model proposed by
Dau et al. (1996) is still unable to successfully predict the deficits in signal-in-noise
detection in some SNHL participants, and it may be that an impaired template-matching
ability is involved in this deficit (Jepsen and Dau, 2011).
What conclusions can be made on the overarching research question, on whether the
MOCS is responsible for the formation of the attentional filter? The attentional filter
was reduced in depth in individuals with SNHL, and was not present in the majority of
General Discussion
159
CI recipients, which supports and extends previous research that demonstrates the
dependence of the attentional filter on the normal function of MOCS action on hearing.
These findings are consistent with a major role for the MOCS in forming the attentional
filter, as there is evidence of a complete loss of the filter with a presumed complete loss
of MOCS action on the cochlea. However, the constant requirement for presumed loss
of MOCS action on the cochlea, both in the present work and in previous research due
to the lack of a direct measurement of the function of the MOCS, requires a cautious
interpretation of the results. In addition, the failure to demonstrate a connection between
the normal function of the MOCS and the formation of the attentional filter in the
normal hearing individuals of chapter 3 prevents a stronger conclusion for the role of
the MOCS in forming the filter. Finally, the consistent, replicable formation of the filter
in CI#4 demonstrates the presence of an alternative mechanism that is able to form the
attentional filter, but it unrelated to the MOCS efferent control of the cochlear amplifier.
Nonetheless, some aspects of the present work are consistent with a role for the MOCS
in the normal formation of the attentional filter.
7.1 Implications & Future Directions
The present thesis provides some support for the involvement of the MOCS in the
formation of the attentional filter, with evidence of an additional mechanism that is
unrelated to the MOCS‟ efferent output. There are three major research questions still to
be addressed, which are pertinent to the main research question, as well as the
implications raised by the results of the thesis. First is the formation of the attentional
filter in normal hearing individuals, both in relation to the absence of the expected
positive relationship between the estimated strength of the MOCS and the depth of the
filter, as well as the apparent deficits in filter depth with subclinical elevations in
auditory thresholds. Second is the apparent deficit in the attentional filter in the two
individuals with a clinically-pure conductive hearing loss in chapter 4. Third is the
General Discussion
160
suggestion of an alternative mechanism that was able to form the attentional filter in at
least one CI recipient in chapters 5 and 6.
7.1.1 Formation of the filter in normal hearing individuals
In chapter 3, the effect of MOCS strength on the depth of the attentional filter was
measured using the suppression of OAEs in a separate session to the measurement of
the attentional filter. This measurement of the strength of the MOCS acoustic reflex
may have been inappropriate to test the system‟s suggested relationship with the
attentional filter due to the temporal separation between the measurement of OAE
suppression and that of the attentional filter. As discussed earlier, research has
demonstrated substantial changes in the magnitude of OAE suppression depending on
the task conditions and throughout training with a single task (Puel et al., 1988; Maison
et al., 2001; de Boer and Thornton, 2008), and so the MOCS action relevant to the
measure of the attentional filter may not have been measured in the experiments of
chapter 3. Further research must include the potentially task-relevant top-down MOCS
action, by eliminating the temporal separation between the measurement of OAE
suppression and the attentional filter. Ideally, this would take place using a
measurement of OAE suppression that occurred during the attentional filter task,
although this task would have to overcome the standard use of continuous broadband
noise during the attentional filter measurements.
The results of chapter 3 revealed substantial variations in the depth of the attentional
filter in the normal hearing group, although this variation was not substantially different
from that reported in the first study of the attentional filter (Greenberg and Larkin,
1968). In chapter 4 it was shown that the variation in the present work was better
explained by hearing loss rather than OAE suppression. Thus, subclinical elevations in
hearing levels may represent an important confounding factor when measuring the
General Discussion
161
effect of MOCS strength on the filter, and further research into the formation of the
attentional filter in normal hearing individuals must control for slight hearing losses.
Importantly, elevations in pure-tone audiometric thresholds may not be a sufficient
indicator for these hearing losses. As discussed in chapter 4, these levels of hearing loss
are not associated with appreciable losses of function of the MOCS efferent targets, but
may instead be associated with „hidden‟ hearing losses that have been linked with
auditory neuropathy (Kujawa and Liberman, 2009). No sensitive measurement of
auditory neuropathy was included in the present work. Future work could employ the
evoked auditory brainstem response as a sensitive measurement of auditory neuropathy.
A strong correlation between the magnitude of the auditory brainstem response and the
population of surviving auditory nerve fibres has been demonstrated in animal models
(Goldstein and Kiang, 1958; Hall, 1990). On the basis of this animal research, the
brainstem response has been used in humans to estimate the surviving population of
spiral ganglion neurons (Fifer and Novak, 1991); however the relationship between the
response magnitude and the surviving neuronal population has not been directly
confirmed in humans (Miller et al., 2008). In support of the proposed relationship, the
auditory brainstem response has been successfully used to predict CI outcomes (e.g.
(Walton et al., 2008)), and it is known that these outcomes are heavily influenced by
surviving spiral ganglion neuron counts (Khan et al., 2005). Recent studies have
proposed a rapid testing procedure to measure the response in humans (Bharadwaj and
Shinn-Cunningham, 2014), and this measure has been used to correlate the estimate of
„hidden‟ hearing loss with various auditory performance measures, such as detection
thresholds for frequency-modulated tones in noise (Bharadwaj et al., 2014). Therefore,
the auditory brainstem response may be used to test for hearing impairment that would
not be detected with pure-tone audiometry, and could then be used to control for
General Discussion
162
potential auditory neuropathy prior to experiments that aim to correlate OAE
suppression with features of the attentional filter.
The deficit in the attentional filter that was identified in individuals with normal hearing
indicates a diminished ability to bias auditory sensitivity towards expected or cued
signals in noise. This may result in an elevation of speech reception thresholds in noise
in individuals with clinically normal hearing. Presently, speech reception thresholds in
noise have not been correlated with elevations in subclinical hearing levels, nor with
direct measurements of hidden hearing losses that may be related to these impairments,
such as the evoked brainstem response discussed above. However, previous research
has indirectly linked impaired speech reception in noise thresholds in individuals who
would be expected to have subclinical elevations in HL or hidden hearing losses. These
hearing impairments are associated with temporary threshold shifts brought on by noise
exposure (Alvord, 1983; Kujawa and Liberman, 2009). Previous research has shown
that human subjects with a history of noise exposure, but with normal pure-tone
thresholds, do have the predicted deficit in speech reception thresholds in noise when
compared with individuals with the same pure-tone thresholds but no history of noise
exposure (Alvord, 1983; Kujala et al., 2004; Kumar et al., 2012). It is currently unclear
whether the impairment of the attentional filter plays a role in these deficits, or whether
they are due more simply to the reduction in afferent input associated with auditory
neuropathy. Future work could link auditory neuropathy in individuals with normal
pure-tone thresholds, using the evoked auditory brainstem response, discussed above,
with a reduction in the depth of the attentional filter, and then correlate the impairment
of the attentional filter with any deficits in speech reception in noise. This would first
identify whether the attentional filter is impaired specifically by the auditory neuropathy
associated with hidden hearing losses, and second, quantify the effects of impairment to
the attentional filter on speech reception in noise thresholds.
General Discussion
163
7.1.2 Formation of the attentional filter with conductive hearing loss
An important follow up to the present work is a measurement of the attentional filter in
a larger sample of individuals with conductive hearing loss. Chapter 4 included a
measurement of the filter in two individuals with a clinically pure conductive hearing
loss, however there was a subclinical sensorineural component to their hearing loss that
may have contributed to their apparent lack of the attentional filter. The absence of the
filter in the conductive hearing loss participants is significant, as it suggests that hearing
loss alone can eliminate the formation of the attentional filter. This might indicate that
the loss of the attentional filter in the SNHL participants was due to hearing loss, rather
than a specific impairment to the MOCS efferent targets. Only two conductive hearing
loss participants were included in the present work, as the recruitment of individuals
with a pure conductive hearing loss and no sensorineural component to this loss was
difficult. While a large number of individuals with appropriate conductive hearing
losses were contacted, there was an extremely low rate of reply. It was speculated that
this unwillingness to attend the research was prompted by the restoration of near-normal
auditory perception by a hearing aid that is possible with a conductive hearing loss, but
individuals with SNHL still suffer considerable difficulties in noisy environments (Hol
et al., 2004). If this is correct, and the conductive hearing loss individuals who refused
to attend the research have effectively normal hearing, then there is a greater issue of
selection bias for the two conductive hearing loss participants who attended the
research, as these participants were not able to form the attentional filter and so may be
expected to have difficulty hearing in noise. Still, whether the attentional filter is
impaired by a general hearing loss, rather than a specific impairment to the MOCS
efferent targets is a significant question.
General Discussion
164
7.1.3 An alternative mechanism able to form the attentional filter
The implications for the hearing impaired group, both for individuals with mild SNHL
up to profound SNHL and those with a cochlear implant, rest on the apparent formation
of the attentional filter in an individual presumed to have no possible MOCS efferent
action on hearing. This result suggests the presence of an alternative mechanism that is
able to form a typical attentional filter in individuals with reduced MOCS action on
hearing. Presently, the source of this alternative mechanism is not known, but there is
no clear reason to suspect that CI#4, who showed the filter, is unique among all
individuals with reduced MOCS action on hearing in possessing the alternative
mechanism. Thus, depending on the mechanism, it may be possible to activate or train
the alternative mechanism to form the attentional filter, at least in some individuals.
This may be significant, as speech reception thresholds in noise are impaired in
individuals with even a mild SNHL, and considerably impaired in CI recipients. The
formation of the attentional filter represents an ability to bias sensitivity towards certain
frequencies, such as the frequency components of a speaker‟s voice, and away from
unwanted noise that may distract from attending to the speaker. Therefore, the
formation of the attentional filter may improve speech reception thresholds in noise, and
the identification of the alternative mechanism may enable a method for improving
speech reception thresholds in some hearing impaired individuals.
In chapter 6 it was suggested that a central, template-matching, or perceptual object
encoding mechanism may have been an alternative mechanism able to form the
attentional filter. The focus of the present work was the efferent control of peripheral
auditory structures, as a potential source of the attentional filter. However, there is a
great deal of auditory processing in central auditory structures, which forms the basis of
auditory scene analysis (Bregman, 1990), which this thesis has not directly addressed. It
is likely that this central processing has a role in forming the attentional filter in
General Discussion
165
normally hearing individuals, for example recent research has demonstrated a benefit
from an informative speech cue on following a single speaker with competing speech
signals (Woods and McDermott, 2015), and these mechanisms, particularly stream
segregation (Noorden, 1975; Moore and Gockel, 2012), may be involved in forming the
filter in the absence of MOCS action on the cochlea. Indeed, the presence of the
attentional filter in individuals with cochlear implants strongly suggests that central
auditory structures can form the entirety of the attentional filter in certain conditions.
7.2 Conclusions
The expected relationship for increasing depth of the attentional filter with increasing
strength of the MOCS was not found. Instead, a consistent finding was that even small
reductions in auditory sensitivity reduce the formation of the attentional filter. In
participants with SNHL, the filter reduced in depth at levels of hearing loss that are
associated with a loss of the function of the MOCS efferent targets, which supports the
notion that the MOCS in involved in the formation of the filter. However, there was a
reduction in the depth of the filter even in normal hearing individuals at subclinical
levels of hearing loss, and it is unclear whether this reduction in depth is associated with
MOCS action. Finally, the formation of the attentional filter in at least one cochlear
implant recipient demonstrates that the attentional filter can be formed with a restoration
of auditory sensitivity, even with a complete absence of MOCS action on the cochlea.
This result strongly suggests the presence of central mechanisms in the formation of the
attentional filter, however it is unclear why this restoration of the filter was only present
in cochlear implant recipients, and not in individuals with aided SNHL, nor is it clear to
what degree these central mechanisms may be involved in the filter‟s formation in
normal hearing individuals.
References
References
ABDALA, C., MISHRA, S. K. & WILLIAMS, T. L. 2009. Considering distortion product otoacoustic emission fine structure in measurements of the medial olivocochlear reflex. The Journal of the Acoustical Society of America, 125, 1584.
ADAMS, J. C. 1995. Sound stimulation induces Fos‐related antigens in cells with common morphological properties throughout the auditory brainstem. Journal of Comparative Neurology, 361, 645-668.
ALVORD, L. S. 1983. Cochlear dysfunction in "normal-hearing" patients with history of noise exposure. Ear and Hearing, 4, 247-250.
ATTIAS, J., FURST, M., FURMAN, V., RESHEF, I., HOROWITZ, G. & BRESLOFF, I. 1995. Noise-induced otoacoustic emission loss with or without hearing loss. Ear and Hearing, 16, 612-618.
BACKUS, B. C. & GUINAN JR, J. J. 2007. Measurement of the distribution of medial olivocochlear acoustic reflex strengths across normal-hearing individuals via otoacoustic emissions. Journal of the Association for Research in Otolaryngology, 8, 484-496.
BENSON, T. & BROWN, M. 1990. Synapses formed by olivocochlear axon branches in the mouse cochlear nucleus. Journal of Comparative Neurology, 295, 52-70.
BERLIN, C. I., HOOD, L. J., HURLEY, A. & WEN, H. 1994. Contralateral suppression of otoacoustic emissions: an index of the function of the medial olivocochlear system. Otolaryngology – Head and Neck Surgery, 110, 3-21.
BHAGAT, S. P. & CARTER, P. H. 2010. Efferent-induced change in human cochlear compression and its influence on masking of tones. Neuroscience Letters, 485, 94-97.
BHARADWAJ, H. M., VERHULST, S., SHAHEEN, L., LIBERMAN, M. C. & SHINN-CUNNINGHAM, B. G. 2014. Cochlear neuropathy and the coding of supra-threshold sound. Frontiers in Systems Neuroscience, 8, 26.
BLAMEY, P., ARTIERES, F., BAŞKENT, D., BERGERON, F., BEYNON, A., BURKE, E., DILLIER, N., DOWELL, R., FRAYSSE, B. & GALLÉGO, S. 2013. Factors affecting auditory performance of postlinguistically deaf adults using cochlear implants: an update with 2251 patients. Audiology & neuro-otology, 18, 36-47.
BODIAN, D. & GUCER, G. 1980. Denervation study of synapses of organ of Corti of old world monkeys. Journal of Comparative Neurology, 192, 785-796.
BOOTHROYD, A., HANIN, L. & HNATH, T. 1985. A sentence test of speech perception: Reliability, set equivalence, and short term learning. New York, NY: City University of New York. Internal Report RCI 10: Speech and Sciences Research Center.
BORRA, T., VERSNEL, H., KEMNER, C., VAN OPSTAL, A. J. & VAN EE, R. 2013. Octave effect in auditory attention. Proceedings of the National Academy of Sciences, 110, 15225-15230.
BOTTE, M. C. 1995. Auditory attentional bandwidth: Effect of level and frequency range. The Journal of the Acoustical Society of America, 98, 2475.
BREGMAN, A. S. 1990. Auditory scene analysis: The perceptual organization of sound, MIT press.
BROWN, M. C. 2001. Response Adaptation of Medial Olivocochlear Neurons Is Minimal.
References
167
BROWN, M. C. 2014. Single-unit labeling of medial olivocochlear neurons: the cochlear frequency map for efferent axons. Journal of Neurophysiology, 111, 2177-2186.
BROWN, M. C., DE VENECIA, R. K. & GUINAN, J. J. 2003. Responses of medial olivocochlear neurons. Experimental Brain Research, 153, 491-498.
BROWN, M. C., KUJAWA, S. G. & DUCA, M. L. 1998. Single Olivocochlear Neurons in the Guinea Pig. I. Binaural Facilitation of Responses to High-Level Noise.
CHERRY, E. C. 1953. Some Experiments on the Recognition of Speech, with One and with Two Ears. The Journal of the Acoustical Society of America, 25, 975-979.
COHEN, L. T., RICHARDSON, L. M., SAUNDERS, E. & COWAN, R. S. 2003. Spatial spread of neural excitation in cochlear implant recipients: comparison of improved ECAP method and psychophysical forward masking. Hearing Research, 179, 72-87.
COLLET, L., VEUILLET, E., BENE, J. & MORGAN, A. 1992. Effects of contralateral white noise on click-evoked emissions in normal and sensorineural ears: towards an exploration of the medial olivocochlear system. International Journal of Audiology, 31, 1-7.
DAI, H., SCHARF, B. & BUUS, S. 1991. Effective attenuation of signals in noise under focused attention. The Journal of the Acoustical Society of America, 89, 2837.
DALLOS, P. & HARRIS, D. 1978. Properties of auditory nerve responses in absence of outer hair cells. Journal of Neurophysiology, 41, 365-383.
DARROW, K. N., MAISON, S. F. & LIBERMAN, M. C. 2006. Cochlear efferent feedback balances interaural sensitivity. Nature Neuroscience, 9, 1474-1476.
DAU, T., PÜSCHEL, D. & KOHLRAUSCH, A. 1996. A quantitative model of the „„effective‟‟signal processing in the auditory system. I. Model structure. The Journal of the Acoustical Society of America, 99, 3615-3622.
DE BOER, J. & THORNTON, A. R. D. 2007. Effect of subject task on contralateral suppression of click evoked otoacoustic emissions. Hearing Research, 233, 117.
DE CEULAER, G., YPERMAN, M., DAEMERS, K., VAN DRIESSCHE, K., SOMERS, T., OFFECIERS, F. E. & GOVAERTS, P. J. 2001. Contralateral suppression of transient evoked otoacoustic emissions: normative data for a clinical test set-up. Otology & neurotology, 22, 350.
DE VENECIA, R. K., LIBERMAN, M. C., GUINAN, J. J. & BROWN, M. C. 2005. Medial olivocochlear reflex interneurons are located in the posteroventral cochlear nucleus: a kainic acid lesion study in guinea pigs. The Journal of Comparative Neurology, 487, 345.
DING, N. & SIMON, J. Z. 2013. Adaptive temporal encoding leads to a background-insensitive cortical representation of speech. The Journal of Neuroscience, 33, 5728-5735.
DORMAN, M. F., HANNLEY, M. T., DANKOWSKI, K., SMITH, L. & MCCANDLESS, G. 1989. Word recognition by 50 patients fitted with the Symbion multichannel cochlear implant. Ear and Hearing, 10, 44-49.
DOWELL, R., MARTIN, L., CLARK, G. M. & BROWN, M. 1985. Results of a preliminary clinical trial on a multiple channel cochlear prosthesis. Scientific Publications, vol. 3, 1984-1986, no. 176.
EBATA, M., MIYAZONO, H., KUMAMARU, K., CHISAKI, Y. & USAGAWA, T. 2001. The formation of attentional filters for a missing-fundamental complex tone and frequency-gliding tones. Acoustical Science and Technology, 22, 401-406.
References
168
ENGSTRÖM, B. 1984. Fusion of Stereocilia on Inner Hair Cells in Man and in the Rabbit, Rat and Guinea Pig. Scandinavian Audiology, 13, 87-92.
FIFER, R. C. & NOVAK, M. A. 1991. Prediction of auditory nerve survival in humans using the electrical auditory brainstem response. Otology & Neurotology, 12, 350-356.
FLETCHER, H. 1940. Auditory patterns. Reviews of modern physics, 12, 47. FUCHS, P. A. 1996. Synaptic transmission at vertebrate hair cells. Current opinion in
neurobiology, 6, 514-519. FURMAN, A. C., KUJAWA, S. G. & LIBERMAN, M. C. 2013. Noise-induced
cochlear neuropathy is selective for fibers. J Neurophysiol, 110, 577-586. GFELLER, K., TURNER, C., MEHR, M., WOODWORTH, G., FEARN, R.,
KNUTSON, J. F., WITT, S. & STORDAHL, J. 2002a. Recognition of familiar melodies by adult cochlear implant recipients and normal-hearing adults. Cochlear Implants International, 3, 29-53.
GFELLER, K., TURNER, C., MEHR, M., WOODWORTH, G., FEARN, R., KNUTSON, J. F., WITT, S. & STORDAHL, J. 2002b. Recognition of familiar melodies by adult cochlear implant recipients and normal‐hearing adults. Cochlear implants international, 3, 29-53.
GIFFORD, M. L. & GUINAN, J. J. 1983. Effects of crossed‐olivocochlear‐bundle stimulation on cat auditory nerve fiber responses to tones. The Journal of the Acoustical Society of America, 74, 115-123.
GIFFORD, M. L. & GUINAN, J. J. 1987. Effects of electrical stimulation of medial olivocochlear neurons on ipsilateral and contralateral cochlear responses. Hearing research, 29, 179-194.
GIRAUD, A. L., COLLET, L., CHÉRY-CROZE, S., MAGNAN, J. & CHAYS, A. 1995. Evidence of a medial olivocochlear involvement in contralateral suppression of otoacoustic emissions in humans. Brain research, 705, 15-23.
GLASBERG, B. R. & MOORE, B. C. 1990. Derivation of auditory filter shapes from notched-noise data. Hearing research, 47, 103-138.
GOLDSTEIN, M. H. & KIANG, N. Y. S. 1958. Synchrony of Neural Activity in Electric Responses Evoked by Transient Acoustic Stimuli. The Journal of the Acoustical Society of America, 30, 107-114.
GREENBERG, G. Z. 1962. Cueing signals and frequency uncertainty in auditory detection. DTIC Document.
GREENBERG, G. Z. & LARKIN, W. D. 1968. Frequency-response characteristic of auditory observers detecting signals of a single frequency in noise: The probe-signal method. Journal of the Acoustical Society of America; Journal of the Acoustical Society of America.
GROFF, J. A. & LIBERMAN, M. C. 2003. Modulation of cochlear afferent response by the lateral olivocochlear system: activation via electrical stimulation of the inferior colliculus. Journal of neurophysiology, 90, 3178-3200.
GUINAN, 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. JARO-Journal of the Association for Research in Otolaryngology, 4, 521-540.
HALL, R. D. 1990. Estimation of surviving spiral ganglion cells in the deaf rat using the electrically evoked auditory brainstem response. Hearing research, 49, 155-168.
HAMERNIK, R. P., PATTERSON, J. H., TURRENTINE, G. A. & AHROON, W. A. 1989. The quantitative relation between sensory cell loss and hearing thresholds. Hearing Research, 38, 199-211.
References
169
HÉBERT, S., FOURNIER, P. & NOREÑA, A. 2013. The auditory sensitivity is increased in tinnitus ears. The Journal of Neuroscience, 33, 2356-2364.
HELLMAN, R. P. 1999. Cross-modality matching: A tool for measuring loudness in sensorineural impairment. Ear and hearing, 20, 193.
HERBERT, R. 2013. Confidence Interval Calculator. HOL, M., SNIK, A., MYLANUS, E. & CREMERS, C. 2004. Does the bone-anchored
hearing aid have a complementary effect on audiological and subjective outcomes in patients with unilateral conductive hearing loss? Audiology & neuro-otology, 10, 159-168.
HOLDEN, L. K., FINLEY, C. C., FIRSZT, J. B., HOLDEN, T. A., BRENNER, C., POTTS, L. G., GOTTER, B. D., VANDERHOOF, S. S., MISPAGEL, K. & HEYDEBRAND, G. 2013. Factors affecting open-set word recognition in adults with cochlear implants. Ear and hearing, 34, 342.
HOOD, J. & POOLE, J. 1971. Speech audiometry in conductive and sensorineural hearing loss. British Journal of Audiology, 5, 30-38.
HOUSLEY, G. & ASHMORE, J. 1991. Direct measurement of the action of acetylcholine on isolated outer hair cells of the guinea pig cochlea. Proceedings of the Royal Society of London. Series B: Biological Sciences, 244, 161-167.
ISON, J. R., VIRAG, T. M., ALLEN, P. D. & HAMMOND, G. R. 2002. The attention filter for tones in noise has the same shape and effective bandwidth in the elderly as it has in young listeners. The Journal of the Acoustical Society of America, 112, 238-246.
JEPSEN, M. L. & DAU, T. 2011. Characterizing auditory processing and perception in individual listeners with sensorineural hearing loss. The Journal of the Acoustical Society of America, 129, 262-281.
JOHNSSON, L.-G., HAWKINS, J. E. & KINGSLEY, T. C. 1981. Aminoglycoside-induced cochlear pathology in man, Acta oto-laryngologica.
KAWASE, T., DELGUTTE, B. & LIBERMAN, M. C. 1993. Antimasking effects of the olivocochlear reflex. II. Enhancement of auditory-nerve response to masked tones. Journal of neurophysiology, 70, 2533-2549.
KAWASE, T. & TAKASAKA, T. 1995. The effect of contralateral noise on masked compound action potential in humans. Hearing research, 91, 1-6.
KEMP, D. T. 2002. Otoacoustic emissions, their origin in cochlear function, and use. British Medical Bulletin, 63, 223.
KETTEN, D. R., SKINNER, M. W., WANG, G., VANNIER, M. W., GATES, G. A. & NEELY, J. G. 1998. In vivo measures of cochlear length and insertion depth of nucleus cochlear implant electrode arrays. The Annals of otology, rhinology & laryngology. Supplement, 1-16.
KHAN, A. M., WHITEN, D. M., NADOL, J. J. B. & EDDINGTON, D. K. 2005. Histopathology of human cochlear implants: Correlation of psychophysical and anatomical measures. Hearing Research, 205, 83-93.
KIM, D., PAPARELLO, J., JUNG, M., SMURZYNSKI, J. & SUN, X. 1996. Distortion product otoacoustic emission test of sensorineural hearing loss: Performance regarding sensitivity, specificity and receiver operating characteristics. Acta oto-laryngologica, 116, 3-11.
KIM, D., PARHAM, K., ZHAO, H. & GHOSHAL, S. 1995. The olivocochlear feedback gain control subsystem: ascending input from the small cell cap of the cochlear nucleus? WENNER GREN INTERNATIONAL SERIES, 31-31.
KNUDSON, I. M., SHERA, C. A. & MELCHER, J. R. 2014. Increased Contralateral Suppression of Otoacoustic Emissions Indicates a Hyper-responsive Medial Olivocochlear System in Humans with Tinnitus and Hyperacusis. Journal of neurophysiology, jn. 00576.2014.
References
170
KUMAR, U. A. & VANAJA, C. 2004. Functioning of olivocochlear bundle and speech perception in noise. Ear and hearing, 25, 142.
KWON, B. J. & VAN DEN HONERT, C. 2006. Effect of electrode configuration on psychophysical forward masking in cochlear implant listenersa). The Journal of the Acoustical Society of America, 119, 2994-3002.
LARSEN, E. & LIBERMAN, M. C. 2010. Contralateral cochlear effects of ipsilateral damage: no evidence for interaural coupling. Hearing research, 260, 70-80.
LEE, D. S., LEE, J. S., OH, S. H., KIM, S.-K., KIM, J.-W., CHUNG, J.-K., LEE, M. C. & KIM, C. S. 2001. Deafness: cross-modal plasticity and cochlear implants. Nature, 409, 149-150.
LEVITT, H. 1971. Transformed up‐down methods in psychoacoustics. The Journal of the Acoustical society of America, 49, 467-477.
LIBERMAN, M. 1980. Efferent synapses in the inner hair cell area of the cat cochlea: an electron microscopic study of serial sections. Hearing research, 3, 189-204.
LIBERMAN, M. C. 1988. Response properties of cochlear efferent neurons: monaural vs. binaural stimulation and the effects of noise. J. Neurophysiol, 60, 1779-1798.
LIBERMAN, M. C. & BROWN, M. C. 1986. Physiology and anatomy of single olivocochlear neurons in the cat. Hearing Research, 24, 17-36.
LIBERMAN, M. C., LIBERMAN, L. D. & MAISON, S. F. 2014. Efferent feedback slows cochlear aging. The Journal of Neuroscience, 34, 4599-4607.
LILAONITKUL, W. & GUINAN, J. J. 2009a. Human medial olivocochlear reflex: effects as functions of contralateral, ipsilateral, and bilateral elicitor bandwidths. JARO-Journal of the Association for Research in Otolaryngology, 10, 459-470.
LILAONITKUL, W. & GUINAN, J. J. 2009b. Reflex control of the human inner ear: A half-octave offset in medial efferent feedback that is consistent with an efferent role in the control of masking. Journal of neurophysiology, 101, 1394.
LILAONITKUL, W. & GUINAN, J. J. 2012. Frequency tuning of medial-olivocochlear-efferent acoustic reflexes in humans as functions of probe frequency. Journal of Neurophysiology, 107, 1598-1611.
LILAONITKUL, W. & JOHN, J. 2002. Tuning of Ipsilateral, Contralateral and Binaural Medial Efferent Reflexes in Humans.
MAISON, S., MICHEYL, C. & COLLET, L. 2001. Influence of focused auditory attention on cochlear activity in humans. Psychophysiology, 38, 35-40.
MCDERMOTT, H. J., LECH, M., KORNBLUM, M. S. & IRVINE, D. R. 1998. Loudness perception and frequency discrimination in subjects with steeply sloping hearing loss: possible correlates of neural plasticity. The Journal of the Acoustical Society of America, 104, 2314-2325.
MEIKLE, M. B., GRIEST, S. E., STEWART, B. J. & PRESS, L. S. 1995. Measuring the negative impact of tinnitus: a brief severity index. Abstracts of the Association for Research in Otolaryngology, 167.
MERIC, C. & COLLET, L. 1992. Visual attention and evoked otoacoustic emissions: a slight but real effect. International Journal of Psychophysiology, 12, 233-235.
MICHEYL, C. & COLLET, L. 1996. Involvement of the olivocochlear bundle in the detection of tones in noise. The Journal of the Acoustical Society of America, 99, 1604.
MISHRA, S. K. & LUTMAN, M. E. 2013. Repeatability of click-evoked otoacoustic emission-based medial olivocochlear efferent assay. Ear and Hearing, 34, 789-798.
MOORE, B. C. & GLASBERG, B. R. 2004a. A revised model of loudness perception applied to cochlear hearing loss. Hearing Research, 188, 70-88.
References
171
MOORE, B. C. & GOCKEL, H. E. 2012. Properties of auditory stream formation. Philosophical Transactions of the Royal Society B: Biological Sciences, 367, 919-931.
MOORE, B. C. & OXENHAM, A. J. 1998. Psychoacoustic consequences of compression in the peripheral auditory system. Psychological review, 105, 108.
MOORE, B. C. J. 1981. Auditory filter shapes derived in simultaneous and forward masking. The Journal of the Acoustical Society of America, 70, 1003.
MOORE, B. C. J. & GLASBERG, B. R. 2004b. A revised model of loudness perception applied to cochlear hearing loss. Hearing research, 188, 70-88.
MOORE, B. C. J., HAFTER, E. R. & GLASBERG, B. R. 1996. The probe signal method and auditory filter shape: Results from normal and hearing impaired subjects. The Journal of the Acoustical Society of America, 99, 542.
MOORE, J. K., SIMMONS, D. D. & GUAN, Y. L. 1999. The human olivocochlear system: organization and development. Audiology and Neurotology, 4, 311-325.
MULDERS, W., PAOLINI, A., NEEDHAM, K. & ROBERTSON, D. 2003. Olivocochlear collaterals evoke excitatory effects in onset neurones of the rat cochlear nucleus. Hearing Research, 176, 113-121.
MULDERS, W. & ROBERTSON, D. 2000a. Effects on cochlear responses of activation of descending pathways from the inferior colliculus. Hearing Research, 149, 11-23.
MULDERS, W. & ROBERTSON, D. 2000b. Evidence for direct cortical innervation of medial olivocochlear neurones in rats. Hearing Research, 144, 65-72.
MULDERS, W., SELUAKUMARAN, K. & ROBERTSON, D. 2008. Efferent pathways modulate hyperactivity in inferior colliculus. Hearing Research, 243, 35-46.
MULDERS, W., WINTER, I. & ROBERTSON, D. 2002. Dual action of olivocochlear collaterals in the guinea pig cochlear nucleus. Hearing research, 174, 264-280.
NADOL, J. B., YOUNG, Y.-S. & GLYNN, R. J. 1989. Survival of spiral ganglion cells in profound sensorineural hearing loss: implications for cochlear implantation. Annals of Otology, Rhinology & Laryngology, 98, 411-416.
NEWCOMBE, R. G. 1998. Interval estimation for the difference between independent proportions: comparison of eleven methods. Statistics in Medicine, 17, 873-890.
NOORDEN, V. L. 1975. Temporal coherence in the perception of tone sequences. Technische Hogeschool Eindhoven.
PATRICK, J. F., BUSBY, P. A. & GIBSON, P. J. 2006. The development of the Nucleus® Freedom™ cochlear implant system. Trends in Amplification, 10, 175-200.
PATUZZI, R. B., YATES, G. K. & JOHNSTONE, B. M. 1989a. Changes in cochlear microphonic and neural sensitivity produced by acoustic trauma. Hearing Research, 39, 189-202.
PATUZZI, R. B., YATES, G. K. & JOHNSTONE, B. M. 1989b. Outer hair cell receptor current and sensorineural hearing loss. Hearing Research, 42, 47-72.
PENNER, M. 1972. The effect of payoffs and cue tones on detection of sinusoids of uncertain frequency. Perception & Psychophysics, 11, 198-202.
PERROT, X., RYVLIN, P., ISNARD, J., GUÉNOT, M., CATENOIX, H., FISCHER, C., MAUGUIÈRE, F. & COLLET, L. 2006. Evidence for corticofugal modulation of peripheral auditory activity in humans. Cerebral Cortex, 16, 941-948.
PUEL, J. L., BONFILS, P. & PUJOL, R. 1988. Selective attention modifies the active micromechanical properties of the cochlea. Brain research, 447, 380-383.
References
172
RAGGIO, M. W. & SCHREINER, C. E. 1999. Neuronal responses in cat primary auditory cortex to electrical cochlear stimulation. III. Activation patterns in short-and long-term deafness. Journal of Neurophysiology, 82, 3506-3526.
RASMUSSEN, G. 1960. Efferent fibers of the cochlear nerve and cochlear nucleus. Neural Mechanisms of the Auditory and Vestibular Systems, 8, 105-115.
REBILLARD, G. & REBILLARD, M. 1980. Factors affecting the recording of visual-evoked potentials from the deaf cat primary auditory cortex (AI). Brain research, 188, 252-254.
REISS, L. A., TURNER, C. W., ERENBERG, S. R. & GANTZ, B. J. 2007. Changes in pitch with a cochlear implant over time. Journal for the Association for Research in Otolaryngology, 8, 241-257.
ROBERTSON, D. & GUMMER, M. 1985. Physiological and morphological characterization of efferent neurones in the guinea pig cochlea. Hearing Research, 20, 63-77.
ROBERTSON, D. & IRVINE, D. R. 1989. Plasticity of frequency organization in auditory cortex of guinea pigs with partial unilateral deafness. Journal of Comparative Neurology, 282, 456-471.
ROBERTSON, D. & WINTER, I. M. 1988. Cochlear nucleus inputs to olivocochlear neurones revealed by combined anterograde and retrograde labelling in the guinea pig. Brain research, 462, 47-55.
ROSSUM, V. 2007. Python Programming Language. USENIX Annual Technical Conference.
RYAN, S., KEMP, D. & HINCHCLIFFE, R. 1991. The influence of contralateral acoustic stimulation on click-evoked otoacoustic emissions in humans. British Journal of Audiology, 25, 391-397.
SCHAETTE, R. & MCALPINE, D. 2011. Tinnitus with a normal audiogram: physiological evidence for hidden hearing loss and computational model. The Journal of Neuroscience, 31, 13452-13457.
SCHARF, B., MAGNAN, J. & CHAYS, A. 1997. On the role of the olivocochlear bundle in hearing: 16 case studies. Hearing Research, 103, 101-122.
SCHARF, B., MAGNAN, J., COLLET, L., ULMER, E. & CHAYS, A. 1994. On the role of the olivocochlear bundle in hearing: a case study. Hearing Research, 75, 11-26.
SCHARF, B., QUIGLEY, S., AOKI, C., PEACHEY, N. & REEVES, A. 1987. Focused auditory attention and frequency selectivity. Attention, Perception, & Psychophysics, 42, 215-223.
SCHARF, B., REEVES, A. & SUCIU, J. 2007. The time required to focus on a cued signal frequency. The Journal of the Acoustical Society of America, 121, 2149.
SCHLAUCH, R. S. & HAFTER, E. R. 1991. Listening bandwidths and frequency uncertainty in pure‐tone signal detection. The Journal of the Acoustical Society of America, 90, 1332.
SCHOFIELD, B. R. & COOMES, D. L. 2006. Pathways from auditory cortex to the cochlear nucleus in guinea pigs. Hearing Research, 216–217, 81-89.
SCHROTT-FISCHER, A., KAMMEN-JOLLY, K., SCHOLTZ, A., RASK-ANDERSEN, H., GLUECKERT, R. & EYBALIN, M. 2007. Efferent neurotransmitters in the human cochlea and vestibule. Acta Oto-laryngologica, 127, 13-19.
SCHUKNECHT, H. F. & WOELLNER, R. C. 1955. An Experimental and Clinical Study of Deafness from Lesions of the Cochlear Nerve. The Journal of Laryngology & Otology, 69, 75-97.
References
173
SHERA, C. A. 2004. Mechanisms of mammalian otoacoustic emission and their implications for the clinical utility of otoacoustic emissions. Ear and hearing, 25, 86-97.
SHERA, C. A. & GUINAN, J. J. 1999. Evoked otoacoustic emissions arise by two fundamentally different mechanisms: a taxonomy for mammalian OAEs. The Journal of the Acoustical Society of America, 105, 782.
SHINN-CUNNINGHAM, B. G. 2008. Object-based auditory and visual attention. Trends in cognitive sciences, 12, 182-186.
SIMON, H. J. & YUND, E. W. 1993. Frequency discrimination in listeners with sensorineural hearing loss. Ear and Hearing, 14, 190-201.
SMITH, R. L. 1979. Adaptation, saturation, and physiological masking in single auditory‐nerve fibers. The Journal of the Acoustical Society of America, 65, 166-178.
SNYDER, R. L., MIDDLEBROOKS, J. C. & BONHAM, B. H. 2008. Cochlear implant electrode configuration effects on activation threshold and tonotopic selectivity. Hearing Research, 235, 23-38.
STEBBINS, W. C., HAWKINS, J. E., JOHNSSON, L. G. & MOODY, D. B. 1979. Hearing thresholds with outer and inner hair cell loss. American Journal of Otolaryngology, 1, 15-27.
TAN, M. N. 2008. Selective Listening Processes in Humans. Ph.D thesis, University of Western Australia.
TAN, M. N., ROBERTSON, D. & HAMMOND, G. R. 2008. Separate contributions of enhanced and suppressed sensitivity to the auditory attentional filter. Hearing Research, 241, 18-25.
TANNER, W. & NORMAN, R. 1954. The human use of information--II: Signal detection for the case of an unknown signal parameter. Information Theory, Transactions of the IRE Professional Group on, 4, 222-227.
TRAUTWEIN, P., HOFSTETTER, P., WANG, J., SALVI, R. & NOSTRANT, A. 1996. Selective inner hair cell loss does not alter distortion product otoacoustic emissions. Hearing Research, 96, 71-82.
VETTER, D. E., SALDAÑA, E. & MUGNAINI, E. 1993. Input from the inferior colliculus to medial olivocochlear neurons in the rat: a double label study with PHA-L and cholera toxin. Hearing Research, 70, 173-186.
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., FREY, K., HEPPELMANN, G., PLONTKE, S. K. & ZENNER, H. P. 2008. Speech-in-noise intelligibility does not correlate with efferent olivocochlear reflex in humans with normal hearing. Acta oto-laryngologica, 128, 53-60.
WALTON, J., GIBSON, W. P. R., SANLI, H. & PRELOG, K. 2008. Predicting cochlear implant outcomes in children with auditory neuropathy. Otology & Neurotology, 29, 302-309.
WARR, W. B. 1978. The olivocochlear bundle: its origins and terminations in the cat. Evoked electrical activity in the auditory nervous system, Academic Press, New York, 43-63.
WARR, W. B. & GUINAN JR, J. J. 1979. Efferent innervation of the organ of Corti: two separate systems. Brain Research, 173, 152-155.
WARREN, E. H. & LIBERMAN, M. C. 1989. Effects of contralateral sound on auditory-nerve responses. I. Contributions of cochlear efferents. Hearing Research, 37, 89-104.
References
174
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.
WRIGHT, B. A. & DAI, H. 1994. Detection of unexpected tones with short and long durations. The Journal of the Acoustical Society of America, 95, 931-938.
XIAO, Z. & SUGA, N. 2001. Modulation of cochlear hair cells by the auditory cortex in the mustached bat. Nature Neuroscience, 5, 57-63.
ZHAO, W. & DHAR, S. 2012. Frequency tuning of the contralateral medial olivocochlear reflex in humans. Journal of Neurophysiology, 108, 25-30.
ZWOLAN, T. A., COLLINS, L. M. & WAKEFIELD, G. H. 1997. Electrode discrimination and speech recognition in postlingually deafened adult cochlear implant subjects. The Journal of the Acoustical Society of America, 102, 3673-3685.
Appendix
Chapter 8. Appendix
8.1 Depth of the attentional filter as a function of OAE suppression
In chapter 3 the detection rates of the target and probes were correlated with both
TEOAE suppression and DPOAE suppression. Below, instead of the detection rates of
the target and probes, OAE suppression is correlated with the depth of the attentional
filter as measured by subtracting each probes‟ detection rate from the target detection
rate.
8.1.1 TEOAE suppression
The relationships between the depth of the attentional filter at each probe frequency,
calculated by subtracting the mean of each probe‟s detection rate from the mean target
detection rate, is shown in Figure 8.1. Slight trends for increasing filter depth at the
distant probe frequencies existed, however these trends did not near significance, as
shown in Table 8.1.
Appendix
176
Figure 8.1. The relationship between the suppression of TEOAEs and the depth of the
attentional filter at each of the probes (N = 14).
Table 8.1. The correlation coefficients and associated statistics for the detection rate of
the target and probes as a function of the suppression of TEOAEs (N = 14). No
correction for multiple comparisons was made.
r 95% CI
1.8-kHz probe .02 (-.72, .26)
1.92-kHz probe .26 (-.32, .69)
2.08-kHz probe .01 (-.31, .70)
2.2-kHz probe .30 (-.72, .26)
Appendix
8.1.2 DPOAE suppression
The relationship between the depth of the each of the probes tones relative to the target
as a function of the suppression of the L1 = 55 dB DPOAEs, and from 13 participants
for the suppression of the L1 = 55 dB DPOAEs in Figure 8.2A-B, with the statistics of
the relationships shown in Table 8.2. No significant correlations were present.
Figure 8.2. A, the relationship between the suppression of the L1 = 45 dB DPOAEs and
the detection rate of the target and probes. B, the same relationship for the suppression
of L1 = 55 dB DPOAEs. (N = 13 for L1 = 45 dB DPOAEs, n = 15 for L1 = 55 dB
DPOAEs).
Appendix
178
Table 8.2. The correlation coefficients and associated statistics for the relationships
between the detection rate of the target and probes and the suppression of the L1 = 45
dB, and L1 = 55 dB DPOAEs (N = 13 for L1 = 45 dB DPOAEs, n = 15 for L1 = 55 dB
DPOAEs).
L1 r 95% CI
1.8-kHz probe 45 .28 (-.31, .72)
55 -.01 (-.52, .50)
1.92-kHz probe 45 .29 (-.31, .73)
55 .26 (-.29, .68)
2.08-kHz probe 45 .14 (-.64, .44)
55 .01 (-.50, .52)
2.2-kHz probe 45 .44 (-.80, .14)
55 -.29 (-.70, .26)
Appendix
Figure 8.3. Frequency allocation table (FAT) for Cochlear™ implants. This table shows
the default frequency ranges that, after being received and processed by the speech
processor, are allocated to each electrode along the implanted electrode array. The
values shown are the defaults, and they can be modified to suit individual cochlear
implant recipients. The implant recipients included in chapters 5 and 6 all used these
default values. Taken from a screenshot of the Cochlear™ implant program. UF: Upper
frequency. LF: Lower frequency. BW: Bandwidth. AE: Active electrode. SM/IE:
Simulation mode.
Appendix
180
Figure 8.4. Companion figure to Figure 4.1, showing the audiograms for the ears with
worse thresholds at 2 kHz for the SNHL group. This includes air conduction thresholds
(solid lines) and bone conduction thresholds (dashed lines).