" 12.15.33
,.. 3. .r v, . {:35
.wfiflmmmifi .. . . 7 ... .2
3 ‘-

83....
"trig.-
..

.9 .I .30.. .

71:: 1 .2 :L!

x.
.c i. 1' I.
L. .3. . .i .o
1.

gig

 

«.3, .. z

.. z,

0. 2.4:?!
.. s; it

my

a: . .4.

..._.,,.‘,,‘..%..‘
as.
\- "

155...; i h.

5 . .. .

t ,
iv. an
avg: gun.
enssvwfihh
l.§, flan
«Fa.

434w {mam ¢ _

.u.....;.fi.. . 323$ .1.

:Fu.u.%3h. fiflflmfl§cw
NF

5.2.. ‘
.-
4. 35?. :5
3.122;: #33: .9...
)Is Sculls? .fi
:3 , ha
t $ II . VII.
,‘ I “J 9 - 3".“3‘5:
ivtaisl .- vlql ‘ 3%...
E5103! I... it
V ’I,
‘1’-
I'iII —
.10.!" t‘\:‘
3.2.42.3: .3}... 1..
s . . [9.1.2 .5»
‘3’? u ’ A
3 u.
. s: .

n
5..- a"...
435.11

‘16-
. .619... 3...:

53...... \Izaui’i..51i I)
133132 3...: $2.11.

LIDHHHY

MM ?~ 9,. oiate
University

‘w

This is to certify that the
dissertation entitled

COGNITIVE EFFORT AND PERCEPTION OF SPEECH BY
POSTLINGUALLY DEAFENED ADULT USERS OF COCHLEAR
IMPLANTS

presented by

RANA ANAS ALKHAMRA

has been accepted towards fulfillment
of the requirements for the
Doctorate degree in Communicative Sciences and
Disorders

 

Major Professcf’s Signature ’
8/5/2010

Date

MSU is an Affinnative Action/Equal Opportunity Employer

 

 

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE
f " ‘Gi
1535’ 33 ‘j 5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KIProj/Achres/ClRC/Dateouejndd

COGNITIVE EFFORT AND PERCEPTION OF SPEECH BY POSTLINGUALLY
DEAF ENED ADULT USERS OF COCHLEAR IMPLANTS

By

Rana Anas Alkhamra

A DISSERTATION

Submitted to
Michigan Sate University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Communicative Sciences and Disorders

2010

ABSTRACT

COGNITIVE EFFORT AND PERCEPTION OF SPEECH BY
POSTLINGUALLY DEAFENED ADULT USERS OF COCHLEAR IMPLANTS

By
Rana Anas Alkhamra

Cochlear implants (Cls) are remarkable devices that make functional hearing possible
for users who would otherwise have only minimal access to sound. Postlingually-deafened
adults who receive a Cl often recover the ability to perceive speech. At the same time, the Cl-
processed speech signal that they receive lacks a number of acoustical features that are
available to normally-hearing listeners. The present dissertation study tested the hypothesis
that postlingually-deafened Cl users must therefore make a greater-than-normal commitment
of cognitive effort in order to achieve perceptual success when listening to speech. This
hypothesis was tested in two ways.

In part one of the dissertation, a substantial group of postlingually deafened adult Cl
users (n=127) were surveyed. Survey participants rated the cognitive effort that they must
expend when listening to speech in a wide variety of everyday circumstances. They also rated
the effort needed at various points in time, both pre- and post-implant. Finally, they provided
demographic information and information about hearing history, both of which were
examined in relation to the cognitive effort ratings. Key findings included the following: (a)
CI users reported experiencing a sharp reduction in the cognitive effort needed to perceive
speech during the first days and weeks post-implant, and lesser but continuing reductions for
several months thereafter; (b) the cognitive effort needed to perceive speech was reported to
increase significantly whenever the level of background noise increased, when listening to an

unfamiliar topic of conversation, when listening to an unfamiliar talker (or talkers), and when

listening to a non-native speaker or to a rapidly speaking talker; and (0) CI users who
regularly wear a hearing aid in the opposite ear reported exerting significantly less effort than
those who do not use a hearing aid.

Speech in background noise is widely reported to present particular difficulty for CI
users. In the second part of this dissertation, an experimental study was therefore conducted
to measure the cognitive effort needed to perceive speech depending on the level of
competing background noise. Two groups of listeners were tested: Group C1 were n=12
postlingually-deafened adults who had worn a CI for at least a year and who were rated as
very good speech perceivers by their clinicians; Group NH were a comparison group of n=1 8
normally-hearing listeners. Results indicated that when listening in quiet and at a favorable
signal-to-noise ratio (SW) of +1 5 dB, the CI users were capable of maintaining substantial
accuracy in their speech perception performance, nearly as high as that of NH listeners.
However, significantly greater cognitive effort was required in order for the CI group to
perceive speech .in these conditions. When the background noise level increased to S/N=+5
dB, speech perception errors made by both groups increased, but they increased much more ‘
rapidly for the CI group. Also there was evidence of a substantially greater than normal
commitment of cognitive effort to speech processing by members of the CI group. Finally, a
comparison of results for a group of NH listeners tested at S/N = 0 dB and a group of Cl
users tested at S/N = +1 0 dB showed similarities in both their speech perception accuracy and
their commitments of cognitive effort. Hence, the two groups could be roughly equated by a

+10 dB change in S/N.

Copyright by
RANA ANAS ALKHAMRA
2010

DEDICATION
T o my father, mother and brothers
who have always believed

ACKNOWLEDGEMENTS

I would like to thank the members of my committee, Terry Zwolan, Jerry Punch,
and Ester Onaga for their support, encouragement and guidance throughout my graduate
studies. I would also like to extend my special thanks to the University of Michigan
Cochlear Implant Program for all the support and for facilitating collecting the study data.
Lots of gratitude also goes to Dr. Jill Elfenbein, may her spirit rest in peace, who
participated in many ways in my arrival at this milestone. In addition, my appreciation
goes to my adviser Dr. Brad Rakerd, whose intelligence and integrity has made a huge
influence on the development of my passion for research. The completion of this work
would not have been possible without him.

I would also like to thank my beloved family (mother and brothers: Hassan,
Bashar and Hatem). I owe it all to them and to a great father who always believed in me
and guided me towards success. I wish that he was still alive to ripe the fruit of the plant
that he seeded.

Last, I thank all my friends for their love, encouragement, and unfailing support. I
love you all very much.

This degree was partly supported by the University of Jordan scholarship office.

All gratitude for their support.

vi

TABLE OF CONTENTS
List of Tables ....................................................................................... ix

List of Figures ....................................................................................... x

CHAPTER 1 ......................................................................................... 1

Introduction and Review ofthe Literature 1

Cochlear Implants
What is a Cochlear Implant? ................................................................................. 3
History ........................................................................................ 4
Components .................................................................................. 5
Candidacy Criteria .......................................................................... 5
Speech Perception in Individuals with Cochlear Implants ........................... 6
Survey Study ............................................................................... 17
Experiment Study ......................................................................... l7

CHAPTERZ ........................................................................................ 18
Survey 18
Methods 18
Survey Instrument19
Results and D13cussron27
Part1c1pant528
Cognitive Effort QuestronsBl
Data Analysis ofCognitive Effort35
Additional Factor Questions48
Data Analysis ofAdditional Factors49
Interaction354
Conclusion ................................................................................. 63

CHAPTER 3 ....................................................................................... 64
Experiment 64
Methods ............................................................................................. 66
Participants ................................................................................. 66
Dual-tasks ................................................................................... 69
Testing Procedures ........................................................................ 72
Tests .......................................................................................... 73
Results and Discussion ............................................................................ 76
Favorable Listening Conditions ......................................................... 77
More Challenging Listening conditions ................................................ 89
Additional Dual-Task Tests .............................................................. 93

CHAPTER 4 ...................................................................................... 104

Clinical Implications ..................................................................... 104
Study Limitations ......................................................................... 106
Suggestions for Further Research ...................................................... 107

vii

 

Summary .................................................................................. 108

APPENDIX A .................................................................................... 110
Survey ............................................................................................. 111
APPENDIX B .................................................................................... 122
BKB Test Lists .................................................................................... 123
REFERENCES .................................................................................... 125

viii

LIST OF TABLES

Table 2-1 the etiologies of hearing loss across participants .................................. 29
Table 2-2. Summary ofcognitive effort response rates 32
Table 2-3. Summary of “No Opinion” reports .................................................. 33
Table 2-4. Pairwise comparisons between adjacent time since implant means ............. 38
Table 2-5. Pairwise comparisons between adjacent noise level means .......................... 41
Table 2-6. “Other Factors” subgroups sizes ....................................................... 49
Table 2-7. Mean cognitive effort rating when listening under the “Best .................... 51
Table 2-8. Mean cognitive effort rating when listening under the “Worst Conditions”. . .53
Table 2-9. Interactions for which there was statistical evidence of a significant trend ...55
Table 3-1. Demographics and hearing history of the individuals in the CI group ......... 67
Table 3-2. Visual reaction time score for the NH group81
Table 3-3 Comparisons of the CI and NH listener group means on several measures of
speech perception performance ....................................................................... 85
Table 3-4. Visual reaction time scores for the CI group .......86
Table 3-5. Comparisons of the CI and NH listener group means on several measures of
visual reaction time performance .................................................................. 88
Table 3-6. Comparing speech test scores at: 0 dB (NH) and 10 dB (CI) ............... 101
Table 3-7. Comparing RT scores at: 0 dB (NH) and 10 dB (CI) .......................... 103

LIST OF FIGURES
Figure 2-1. Percentage of “no opinion” responses ........................................... 34

Figure 2-2. Cognitive effort before receipt of the cochlear implant and one week post-
implant ............................................................................................... 35

Figure 2-3. Cognitive effort and time since implant .......................................... 37

Figure 2-4. Cognitive effort ratings for listening in quiet and in three different levels of
background noise ................................................................................... 40

Figure 2-5. Cognitive effort for listening in environmental vs. speech background
noise... ............................................................................................................................... 42

Figure 2-6. Cognitive effort and level of familiarity of the conversation t0pic ............... .43

Figure 2-7. Cognitive effort results when the talkers were known (familiar) or
unfamiliar ........................................................................................................................... 45

Figure 2-8. Cognitive effort results regarding familiarity of talker on the phone ............. 46
Figure 2-9. Graph of cognitive effort results according to the talker characteristics ........ 47
Figure 2-10. Interaction between Cl users’ gender and talkers’ native language ........... 56
Figure 2-1 1. Interaction between CI users’ age and telephone partners’ familiarity. . ......57
Figure 2-12. Interaction between hearing loss onset type and talkers’ native language....57
Figure 2-13. Interaction between hearing loss onset type and speaking rate. . . . . . . . .5....8
Figure 2-14. Interaction between C I users’ gender and background noise level ............ 59

Figure 2-15. Interaction between using a hearing aid in the opposite ear and talkers’
familiarity .............................................................................................. 60

Figure 2-16. Interaction between using a hearing aid in the opposite ear and speaking
rate .................................................................................................... 60

Figure 2-17. Interaction between using a hearing aid in the opposite ear and telephone
partners’ familiarity ................................................................................. 61

Figure 2-18. Interaction between using a hearing aid in the opposite ear and background
noise level61

Figure 2-19. Interaction between CI users’ age and time since implant. . . . .. ................... 62

Figure 3-1a. Normally-hearing (NH) listeners’ raw score results for the speech test
presented at +15 dB S/N .. ............................................................................................ 78

Figure 3-1b. NH listeners’ normalized score results for the speech test. . . . . . . . . . . ...7.9

Figure 3-2a. NH listeners’ raw score results for the visual reaction time test presented at

+15dB ............................................................................................... 80
Figure 3-2b. NH listeners’ raw score results for the visual reaction time test .................. 82
Figure 3-3a. CI users’ (n=12) raw score results for the speech test. ................................ 83
Figure 3-3b. CI users’ normalized score results for the speech test. ........................ 84
Figure 3-4a. CI users’ raw score results for the visual reaction time test .................... 86
Figure 3-4b. CI users’ normalized score results for the visual reaction time test ........... 87
Figure 3-5a. Speech test raw scores for the dual-task test at S/N=+5 dB. 90

Figure 3-5b. Speech test normalized scores for participants in both groups when tested at
+5dB S/N ............................................................................................... 91

Figure 3-6a. Visual reaction time raw scores in both groups when tested at +5dB S/N...92
Figure 3-6b. Visual reaction time difference scores in both groups in +5dB S/N ............ 93
Figure 3-7a. Speech test raw scores for the CI group listening at +10 dB S/N ............. 94
Figure 3-7b. Speech test proportion scores for the CI group listening at +10 dB S/N. . ...94
Figure 3-8a. Visual reaction time raw scores in the CI group at +10 dB S/N. . . . . ............95
Figure 3-8b.Visual reaction time proportion score in the CI groups at +10 dB S/N ....... 96
Figure 3-9a. Speech test raw scores for the NH group at 0 dB S/N .......................... 97
Figure 3-9b. Speech test proportion score for the NH groups at 0 dB S/N. . . . . . . . . . . ........97
Figure 3-10a. Visual reaction time raw scores for the NH group tested at 0 dB S/N ...... 98

Figure 3-10b. Visual reaction time difference score for the NH group at 0 dB S/N ....... 99

xi

Figure 3-11a Speech test raw scores for the NH group tested at 0 dB S/N compared to
scores for the CI group tested at +10 dB S/N... ..100

Figure 3-1 lb. Speech test normalized scores for the NH group tested at 0 dB S/N
compared to scores for the CI group tested at +10 dB S/N .................................. 100

Figure 3-12a. Visual reaction time test raw scores for the NH group tested at 0 dB S/N
compared to scores for the CI group tested at +10 dB S/N .................................. 102

Figure 3-12b. Visual reaction time test normalized scores for the NH group tested at 0 dB
S/N compared to scores for the CI group tested at +10 dB S/N102

xii

CHAPTER 1
Introduction and Review of the Cochlear Implant Literature
In years past, hearing aids and tactile devices were the only treatment options
available for persons with sensorineural hearing loss (SNHL). Today, many people with
severe-to-profound hearing loss are able to receive benefit fiom cochlear implants (CIs).
Many CI users perform well on speech perception tests especially when listening in quiet.
They may also have the ability to use the telephone (Anderson, Baumgartner, Boheim,
Nahler, Arnoldner, & D'Haese, 2006; Wilson & Dorman, 2008). However, there is
considerable variation in treatment outcomes. This variation can be partially explained by
differences among patients in cortical or auditory pathway function (Sharma, Dorman, &
Spahr, 2002; Fallon, Irvine & Shepherd, 2008). The longer the period of auditory
deprivation, the more challenging speech perception becomes for CI users (Blarney et al.,
1996; Sharma et al., 2002; Summerfield & Marshall, 1995). Many studies have
demonstrated that other factors such as age at the time of surgery, gender, duration of
hearing loss, aetiology of hearing loss, residual hearing, implant type, speech processor
strategy,and number of active implanted electrodes may also account for some but not all
of the variability seen in adults and children (Blamey et al., 1993; Dowell, Dettman,
Blarney, Barker, & Clark, 2002; Dowell, Hollow, & Winton, 2004; Gantz, Woodworth,
Knutson, Abbas, & Tyler,1993; Geers, Brenner, & Davidson, 2003; Green e1al., 2007;
Rubinstein, Parkinson, Tyler, & Gantz, 1999; Summerfield, & Marshall,1993; Turner,

Gantz, Vidal, Behrens, & Henry, 2004).

While many multichannel CI users perform very well when listening in quiet,
listening to speech in a competing background noise remains a challenge. Speech
performance can deteriorate rapidly with increasing levels of background noise, even for
the most successful CI users (e. g., Fu & Nogaki, 2005; Friesen, Shannon, Baskent, &
Wang, 2001).

Most studies that have investigated Us to date have used standard audiometric
tests of speech intelligibility to assess the listening abilities of children and adults with
C15. For example, they have measured the intelligibility of individual words or whole
sentences (Beijen, Mylanus, Leeuw, & Snik, 2008; Cullington & Zeng, 2008; Riss,
Arnoldner, Baumgartner, Kaider, & Hamzavi, 2008).

The present study was conducted with a different focus. Specifically, the
cognitive effort required to achieve listening success was examined here. A number of
previous studies have found that the cognitive effort needed to perceive speech can be
affected by a listener’s hearing status (Choi, 2005; Downs, 1982; Feuerstein, 1988;
Hicks &Tharpe, 2002; Moore, 2003; McFadden & Pittman, 2008; Rakerd, Seitz, &
Whearty, 1996; Rossiter, Stevens, & Walker, 2006). The present investigation

extended this perspective to the assessment of persons with Us

This section provides background information about CIs. It discusses: what a CI
is, the history of CI’s, components of C1 systems, candidacy criteria for receiving a CI,
and speech perception findings regarding individuals with C13. It also reviews factors that
can influence speech perception by C1 users such as time since implant, level of

background noise, and talker-related variability.

Cochlear Implants

What is a cochlear implant?

Cochlear implants (CIs) are surgically implanted devices that can be used
effectively by both prelingually-deafened children and postlingually-deafened children
and adults. For persons with normal hearing, the outer and middle ears act as a bridge to
transfer acoustic pressure changes present in the surrounding atmosphere to the fluids in
the cochlea. Vibrations set up in the fluids then cause other structures within the cochlea
to vibrate and to stimulate the sensory receptors known as hair cells (Martin & Clark,
2003).

The cochlea is organized so that the highest frequency sounds cause maximum
stimulation of the hair cells at its base. The lowest frequency sounds cause maximum
stimulation of the sensory cells at the other end, the apex of the cochlea (Moller, 2006;
Von Bekesy, 1989). Thus, the information about the incoming sound is detected and
passed onto the auditory nerve in a frequency-selected way. This organization of the hair
cells and the auditory nerve is referred to as tonotopic organization (Rauschecker &
Shannon, 2002). lnforrnation about sound is coded by the place in the cochlea and also by
the timing patterns of how the nerve fibers respond to sounds (Estabrooks, 1998; Moore
& Teagle, 2002; Tye-Murray, 2004).

People with hearing losses in the severe- to-profound range often have absent or
malfunctioning sensory cells in the cochlea. Cls are designed to substitute for the
functions of the outer, middle, and inner ears, transforming sound energy into electrical

energy that directly stimulates the auditory nerve tonotopically. Specifically, electrical

signals transmitted through implanted electrodes activate the auditory nerve tonotopically
which begins the signal propagation in the nervous system. The neural signals are first
passed in to the brain stem and then up to higher centers of the brain, ultimately reaching

the auditory cortex (Allum, 1996; Rauschecker & Shannon, 2002).

History

The history of CIs began with Alessandro Volta in 1790, when he inserted metal
rods into each of his ear canals and experienced a “boom within the head,” followed by a
sensation of sound similar to that of “boiling, thick soup.” The more recent history
began with the work of Djoumo and Eyries in Paris in 1957 (N iparko, 2009). They
stimulated the hearing of a deaf adult by placing an electrode on the auditory nerve. The
patient reported sensing the presence of environmental sounds but could not understand
speech or discriminate among speakers or many sounds. Shortly after Djoumo and Eyries
work, William House in Los Angeles (1961) was inspired by their work and initiated an
effort to develop a practical and reliable way to treat deafness using electrical stimulation
of the cochlea. By the 19805, single-channel implants were in widespread use among
adults with severe-to-profound hearing loss. By 1988, multichannel systems with
multiplechannels of processing and with multiple sites of stimulation in the cochlea had
come into use (Wilson & Dorman, 2008). In 1990, the Food and Drug Administration
(FDA) approved multichannel CIs for children (Tye-Murray, 2004; Wilson & Dorman,

2008).

 

Components of implant systems

CI systems consist of both internal and external components. The external
components include: a directional microphone, which picks up sound from the
environment; a speech processor that filters, analyzes, and digitizes the sound signal into
coded electrical signals; and a transmitter coil that transmits the signals across the skin to
the implanted receiver via an F M radio signal. The principal internal component is a
receiver that is surgically placed under the skin and is inductively coupled with the
wearable transmitter (Owens & Kessler, 1998, Tye-Murray, 2004; Wilson & Dorman,
2008). This device receives the signal from the externally worn coil and delivers it to the
electrodes in the implanted array.

The electrode array is a small wire that is inserted into the cochlea. Direct
stimulation of the auditory nerve is produced by currents delivering the correct amount of
electrical stimulation to one or more of the electrodes in the scalatympani. Most CIs
today are multichannel devices, which mean that the electrode pairs in the electrode array
present different information to different regions of the cochlea (Tye-Murray, 2004;

Wilson & Dorman, 2008).

Candidacy criteria

Late-deafened adults are generally good candidates for CIs. Current Food and
Drug Administration (FDA) candidacy criteria for adults 18 years and over are (a)
severe-to-profound sensorineural hearing loss (SNHL) in both ears (>70 dBHL); (b)
receiving little or no useful benefit from hearing aids; (c) scoring, with a hearing aid,

50% or less on sentence recognition tests in the ear to be implanted and 60% or less in the

non-implanted ear or bilaterally; (d) desire to improve hearing and realistic expectations;
and having (e) no medical contraindications (Cohen , 2004 ; “candidacy criteria”, 2010;

Copeland & Pillsbury, 2004; Zwolan , 2001).

Speech perception in individuals with cochlear implants

There is good evidence that both adults and children with severe-to-profound
hearing loss can receive substantial benefits from CIs and attain high levels of open-set
speech recognition (Moore & Teagle, 2002; Wilson & Dorman, 2008). Particularly
relevant to the present study are the speech perception abilities of postlingually-deafened
adults who receive a CI. A majority of these users can understand speech well (Lassaletta
et al., 2005; Zwolan, 2001). Many implantees successfully return to their old jobs
(Hamzavi, Baumgartner, Pok, Franz, & Gstoettner, 2003). However, despite all the
benefits of C18, many individuals still experience difficulty hearing in group situations
and in background noise (Cullington & Zeng, 2008; Fu & Nogaki, 2005; Zhao, Stephens,
Sim, & Meredith, 1997). A number of factors can affect the benefit and performance of
adults using a unilateral CI. Several of these factors are examined in detail below.

Time since implant.

A number of studies have evaluated the impact of length of use of multichannel
implants. Hamzavi et al. (2003) tested 66 late-deafened adults (36 females, 30 males;
mean age 53.8 years; age range 19—78 years) using CIs. Tests were carried out at first
fitting, after 3 months, 6 months, 12 months, and then annually for up to 6 years
following implantation. Three speech perception tests were used in conducting the study.

The Freiburger Numbers test is a test that is composed of four groups of 10 four-syllable

numbers. The Freiburger Monosyllables test is composed of three groups of 20 everyday
monosyllables. Both tests were presented to listeners at 70 dB SPL in free field, using a
recorded voice of a male speaker. In the Innsbrucker Sentence Test (Hamzavi,
Baumgartner, Adunka, Franz, & Gstoettner, 2000), one list of 10 three- to eight-word
sentences was presented by the examiner himself at 65 dB SPL. Visual clues for lip-
reading were not given. Repetition was not permitted in the three tests.

Results indicated that all participants showed a steady improvement over time on
every test. Scores obtained after one month showed significant improvement over scores
obtained at time of fitting for all tests (p<0.05). There was also significant progress on
one or more of the tests between 1 and 3 months, 3 and 6 months and 6, and 12 months
(p<0.05). After 12 months of use, testing showed some continuing but statistically
insignificant improvement for the group as a whole. Many patients reached a plateau
within the first 12 months.

Ruffin and colleagues (2007) found that 'speech perception abilities stabilize over
a period of 24 months post-implant. All of the participants in their study (n=31) had
postlingual profound bilateral SNHL and received no benefit from hearing aids prior to
receiving a CI (the mean years of profound deafiiess were 8.8, and the mean age at
implantation was 51).

Participants were evaluated at approximately 3, 6, 9, 12, and 18 months after
initial stimulation, and annually thereafter. Participants were initially tested with the NU-
6 Word lists and then later with the consonant-nuclens-consonant (CNC) word test as
speech discrimination testing measures evolved. This study treated both word lists as

identical test items because no significant differences between NU-6 and CNC words

were found. Data used in analysis were fi‘om one of three possible sources: the CNC
score when this was the only test given to the subject at that time point; the NU-6 score
when this was the only test given; the average of the two test scores when both tests were
given to the subject at the same point in time.

The most significant performance gains occurred in the first 9 months with slower
improvement from 9 to 24 months. After 24 months post-implantation, there was no
significant growth or decline in speech perception unless adverse medical events were
experienced. Age at implantation was significantly and substantially negatively correlated
with most scores over all tested periods. There were no age-related declines in
performance.

Another study by Oh et al. (2003) targeted evaluation of long-term speech
perception of cochlear implantees and compared the developing auditory performance
patterns of prelingually-deafened children and postlingually-deafened adults. The study
included 29 children and 17 adults who had been followed for four years. Speech
perception ability was assessed using tests of vowel and consonant confusion and the
Korean version of the Central Institute of Deafness test (K-CID) (performed without
visual cues). The K-CID test assesses understanding of speech in everyday conversational
situations. Test results were analyzed at three and six months after implantation and then
annually. In the prelingually-deafened children, the average results continued to improve
over the four year period. In postlingually-deafened adults, improvement did not extend
beyond the first two years post-implantation. Adults with < 5 years of deafiiess had a
faster rate of recovery of speech perception than those who had been deaf for > 5 years.

Lastly, in a study by Taina et al. (2000), questionnaires were completed by three

Hearing Centers in Finland. Participants in the study were 67 adults (mean age 47 years).
The mean duration of profound hearing loss prior to CI was 13 years (range 0—51 years).
Data were pooled to present the following evaluation intervals: preoperative, 0-3, 4—6, 7—
11, 12—23, and 24 months post-implantation. After three months of implant use, the CI
users’ mean word recognition score was 54% correct. Scores improved steadily over the
succeeding intervals, and after 24 months the mean word recognition score was 71%
correct. Six months after switch-on, the majority of participants (48/67) were able to
recognize some speech without speechreading, and 26 of the 48 participants were able to
use the telephone with a known speaker, gaining good fiinctional benefit fi'om the CI.

Overall, results from the studies reported here provide evidence of improving
speech perception for the first one to two years post-implant for postlingually-deafened
adults, and little or no further improvements (or declines) thereafter.

Level of background noise.

Gifford et al. (2008 ) noted that hearing with C13 has advanced to the point that it
has become difficult to track changes in patient performance over time because many
patients achieve 90 to 100% scores on standard tests of sentence intelligibility in quiet
(28% of the participants in their study achieved perfect soores). They therefore called for
new, more challenging tests of intelligibility. The need for such tests speaks to the
progress made in CI designs during the past two decades. Despite this progress, the
current devices still do not restore normal speech perception when listening to speech in
competition with noise. Compared to normal-hearing listeners, CI users have tremendous
difficulty understanding speech in noise and their speech perception performance

deteriorates rapidly with increased levels of background noise, even for the best CI users

(Friesen, Shannon, Baskent, & Wang, 2001; Fu & Nogak, 2005; Muller, Schmidt, &
Rubert, 1995; Nelson, Jin, Carney, & Nelson, 2003). Loss of low-frequency detailed
information in the speech signal and limitations on the number of effective channels of
stimulation due to overlap in electric fields of the implanted electrodes are factors that are
believed to lead to this deterioration in performance in noise (Wilson & Dorman, 2008).

When individuals with C15 decide to undergo surgery, the decision comes with a
trade-off of losing any residual low-frequency hearing that may be present in the ear of
implant. The residual acoustic hearing is no often longer usable, and only electric
stimulation is available. Many CI users report after the initial device activation that: (a)
the perception of sound becomes “mechanical’ when compared to the acoustic hearing
they had before the surgery, and (b) many of the aesthetic qualities of sound are
diminished (Turner et al., 2004). This loss of aesthetic quality may be related to a
decrease in the ability to perceive the pitch information of sounds (Gfeller et al., 2002).
The loss of pitch perception is primarily a consequence of the limited spectral resolution
of current Cls (Cullington & Zeng, 2008; Gfeller et al., 2002; Turner, Reiss, & Gantz,
2008).

Present designs and placements of electrodes for CIs supports 8 or slightly more
functional channels regardless of the number of electrodes (Henry & Turner, 2003). In
comparison, the number of channels is about 39 for the full range of fiequencies from 50
Hz to 15 kHz, and is about 28 for the range of frequencies covered by speech sounds
(N iparko, 2009). These numbers are much higher than the number of effective channels

available with a CI.

10

Fu and Nogaki (2005) conducted a study on sentence recognition by C1 users
(n=10) and by normally-hearing (NH) participants (n=6) in the presence of steady or
modulated speech-shaped noise. Participants were native speakers of English. All C1
users were tested using their everyday speech processor and they were instructed to use
their everyday volume and microphone sensitivity settings and disable any no ise-
reduction settings in their speech processors. CI users listened to combined speech and
noise (unprocessed) through their speech processors.

For NH participants, speech and noise were processed by an acoustic simulation
of a CI. In the acoustic simulation, NH listeners were tested for different degrees of
spectral resolution (16, 8, or 4 channels) and different degrees of spectral smearing
(carrier filter slopes of -24 or -6 dB/octave). The signal was introduced at 65 dBA.
During testing, NH participants listened to combined speech and noise processed by CI
simulations. They also listened to unprocessed speech and noise for baseline measures.

Speech test material consisted of sentences taken from the Hearing in Noise Test
(HINT) sentence set (Nilsson et al. 1994). Speech-shaped noise was used as a masker.
The onset of noise was 500 ms before the sentence presentation and the offset of noise
was 500 ms after the end of the sentence presentation; masker rise and fall times were 5
ms.

Results revealed that NH listeners’ performance was significantly affected by the
noise conditions. The poorest performance was observed with steady-state noise and
improved with gated modulated noise. In CI users, no significant main effect of noise

conditions was noted in their performance.

11

With modulated noise, NH listeners experienced significant release from masking
(i.e., compared to steady-state noise) when the original unprocessed speech was
presented. CI users, on the other hand, experienced no release from masking. As the
spectral resolution was reduced, NH listeners’ release from masking gradually
diminished. Release from masking was further reduced as the degree of spectral smearing
increased even when the spectral resolution was relatively high (eight channels).

In summary, regardless of the type of noise, CI users’ speech recognition was
negatively affected by the loss of spectro-temporal fine structure caused by the limited
number of electrodes and spectral channels in the speech processing strategy. However,
the number of spectral channels may not completely account for the poorer performance
of CI users in noise. Spectral cues transmitted by the limited number of implanted
electrodes are further reduced by the degree of channel interaction between electrodes.
Findings suggest that implant users’ susceptibility to noise may be caused by the reduced
spectral resolution and the high degree of spectral smearing associated with channel
interaction.

To facilitate hearing in noisy situations for C 1 users, combined electric and
acoustic stimulation of the auditory system for patients with significant residual hearing
in the low frequencies has been developed. To provide electric acoustic stimulation
(EAS), an electrode is only partially inserted into the cochlea in order to preserve the
residual acoustic hearing that many patients still have for low frequencies (Gantz &
Turner, 2003; Turner el al., 2004; Wilson, Lawson, Muller, Tyler, & Kiefer, 2003).

Relatively normal hearing ability at low frequencies, with excellent frequency

resolution and other attributes of normal hearing, may be preserved by EAS. EAS can

12

 

 

. LT .

also provide a complementary representation of high frequency sounds with the CI and
electrical stimulation.

In summary, studies agree that speech performance deteriorates rapidly with
increased levels of background noise. Overall, listening in challenging situations, such as
speech presented in competition with noise or other talkers, tend to be challenging to
most CI users. Possible causes for the reduced performance in noise include reduced
spectral resolution from a limited number of channels, interactions between channels, and
loss of low-frequency information.

Talker-related variability.

Conditions that introduce variability into the speech signal include differences
between talkers, changes in speaking rate, and dialectal variations (Pisoni 1993). In
normally-hearing individuals, it is generally assumed that spoken word recognition
includes some type of perceptual normalization process in which listeners compensate for
signal variability and hence derive a more idealized internal representation of speech
sounds (Pisoni, 1993; Pisoni & Luce, 1986; Takayanagi, Dirks, & Moshfegh, 2002).
Whether and how listeners with hearing loss cope with variability is the subject of
continuing investigation.

Kirk et al. (1997) investigated the consequences of talker variability on the word-
recognition performance of listeners with hearing loss. Participants in Kirk’s study (9
males, 8 females) were between the ages of 29 and 66 years (M=55 years). All had mild-
to-moderate bilateral SNHL and showed speech recognition scores of 80% or higher. All

participants had postlingually acquired hearing loss; nine of the participants used

13

amplification. Speech recognition performance was assessed with the NU-6 test and
Modified Rhyme Test (MRT).

Results from the study showed that word recognition performance was
consistently lower in multiple-talker conditions (where the listener had to continually
adjust to between-talker differences), than it was in a single-talker condition. Moreover,
the participants’ performance in the multiple-talker condition correlated more strongly
with self-reports of communication ability than performance in the single-talker
condition. A second finding was that performance was lower for speech produced at a
fast speaking rate than for slower or normal speaking rates.

Results from a study by Takayanagi and Moshfegh (2002) were consistent with
the Kirk et al. (1997) study. This study examined the effects of talker variability on
spoken-word recognition among four groups of listeners: native speakers with normal
hearing, native speakers with moderate SNHL, non-native speakers with normal hearing,
and non-native speakers with SNHL.

The stimuli for the experiment included 75 lexically easy and 75 lexically hard
words. Ten talkers (5 men, 5 women) produced each word while speaking at a normal
conversation level. Participants were instructed to repeat the word heard after each
presentation. The ability of listeners to accommodate variations in the talkers’ voice was
assessed by comparing recognition scores for a single-talker condition to those obtained
in a multiple-talker condition.

Results revealed that participants with hearing loss, both native and non-native,
required higher threshold sound pressure levels to reach 50% correct performance levels

than their respective counterparts in the normal-hearing groups. Likewise, the hearing

14

loss groups of non-native participants required significantly higher sound-pressure levels
for 50% correct performance levels than the hearing native participants. The overall
difference in performance between the normal-hearing and hearing-loss participants was
greater than the difference between the native and nonnative participants. Furthermore,
there was a performance difference between the hearing-loss groups and the normally-
hearing groups in single vs. multiple-talker and easy vs. hard word tested conditions.
Performance of the normally-hearing groups was superior to that of the hearing-loss
groups in all tested conditions.

In the field of CIs, a few studies have addressed the impact of talker variability on
speech perception performance. A study by Chang and Fu (2006) investigated the effects
of talker variability on vowel recognition by cochlear implant (CI) users and by normal-
hearing (NH) participants listening to 4-channel acoustic CI simulations. Six
postlingually deafened CI users (4 men and 2 women; aged 47—71 years) were tested with
their clinically assigned speech processors. For the 8 NH participants (4 men and 4
women; aged 26—3 7), 3 CI processors were simulated using different combinations of
carrier type and temporal envelope cutoff frequency (noise band/160 Hz, sine wave/ 160
Hz, and sine wave/20 Hz). Vowel recognition was measured using a lZ-alternative,
closed-set identification procedure. The 12 phoneme tokens included 10 monophthongs
and 2 diphthongs, presented in a /h/vowel/d/ context (e. g., heed, hid); which were drawn
from speech samples collected by Hillenbrand, Getty, Clark, and Wheeler (1995). Test
stimuli were presented in either a single-talker context (1 talker per test block) or a multi-

talker context (4 talkers per test block), at 65 dBA.

15

Results indicated that Cl users’ vowel recognition was significantly poorer in the
multi-talker context than in the single-talker context. When noise-band carriers were used
in the simulations, NH performance was not significantly affected by talker’s variability.
However, when sine-wave carriers were used in the simulations, NH performance was
significantly affected by talker’s variability in both envelope filter conditions. The
investigators concluded that because the fundamental fiequency was not preserved by the
signal processing, both spectral and temporal cues contributed to the talker variability
effects observed with sine-wave carriers. Similarly, spectral and temporal cues may have
contributed to the talker variability effects observed with C1 participants. This area of
study needs further investigation to explore effects of talker related variability on C1
users’ speech perception performance in everyday listening conditions.

Telephone use.

Speech perception by C1 users has been extensively investigated for the special
and important case of talking on the telephone. Telephone use can be difficult for CI
users in part due to the limited frequency bandwidth of telephone signals (Liu, 2009).
Other issues include lack of access to visual cues (no lip-reading) and fewer available
contextual cues for the conversation (Anderson et al., 2006). Despite these limitations,
many CI users can initiate calls to familiar speakers, answer the telephone, and identify a
person's gender and age over the telephone. Understanding can become more difficult as
the topic and speaker become more unfamiliar (Cray et al., 2004; Faber & Grontved,
2000).

Anderson et al. (2006) evaluated the use of both landline and mobile phones in an

international sample of Cl users. Results from 196 surveys obtained in 10 different

16

countries (Austria, South Afi'ica, United Kingdom, United States of America, Argentina,
Belgium, Canada, Germany, Greece, and India) showed that there is a significant shift
from preoperative non-use of a telephone to use of a telephone postoperatively. Talking
to familiar speakers about familiar topics was found to be the easiest listening condition
on the telephone, and it was easier to recognize a voice using the landline. Many
respondents found it difficult to make a call without some assistance. Most participants
could manage to call someone in an emergency, even on a mobile phone.

For reasons noted above, talking over the telephone presents some challenges not
faced when talking to a person directly. CI users may vary in their ability to meet these

chaflenges

A survey study

Chapter 2 below reports the outcome of a survey of postlingually-deafened adult
CI users. The survey asked for self-reports regarding the cognitive effort needed to
perceive speech under various circumstances. Of particular interest were survey questions
regarding each of three factors that were reviewed in this section: (a) time since implant;

(b) background noise; and (c) talker-related variability.

An experimental study
Chapter 3 reports the outcome of an experiment that explored the cognitive effort
that must be expended by a cochlear implant user when perceiving speech in quiet and in

noise.

l7

CHAPTER 2

A Cognitive Effort Survey

Part one of the present study is a survey that was given to a broad cross-section of
postlingually-deafened adults who use C13. The individuals surveyed were native English
speakers and consistent C1 users. The survey investigated their self-perceptions regarding
the cognitive effort needed when perceiving speech under a variety of listening
conditions. This chapter presents the methods and results of the survey, and discusses the
implications of those results.

Methods

Survey instruments, or questionnaires, are used to collect data about subjects’
demographics, personal histories, knowledge, behaviors, and attitudes (Passmore,
Dobbie, Parchman , & Tysinger, 2002) . Mode of survey presentation is one of the first
decisions to be made in designing and conducting a survey. Choices are essentially
between interviewer administration (i.e. face-to face or by telephone) or self-
administration by the respondent (i.e. postal delivery to one’s home) (McColl et al.,
2001). Self-administered surveys, distributed by mail or email, can provide respondents
with privacy and anonymity (McColl et al., 2001; Passmore et al., 2002). However, these
surveys typically yield numerous unusable or incomplete responses and may require
multiple mailings to obtain adequate response rate (Passmore et al., 2002).

A self—administered survey that assessed cognitive effort in everyday life listening

situations was employed in this study. Postlingually-deafened adults who had been

18

wearing a CI for at least 6 months were asked about their listening experiences with the
device. The survey highlighted the cognitive effort required when perceiving speech.
The section below outlines the steps taken to develop, test and administer the
survey instrument. Methods followed in developing the present survey were based on
findings of other studies or opinions of experts who have found certain strategies to
increase response rates and minimize bias (Burns et al., 2008; Edwards et al., 2002;

Dillman, 1995).

The Survey Instrument

The current survey (see Appendix A) was specifically developed for this study. A
description of how the survey was designed and conducted is presented here. The
discussion includes the following topics: survey domains, question construct, response
format, appearance, consistency questions, pilot testing, and survey administration.

Survey domains.

The first step in the survey was item generation, the consideration of all potential
items or questions for inclusion in the questionnaire, with the goal of introducing
domains that help answer the research question (Kirshner & Guyatt, 1985). Three major
domains were examined (a) cognitive effort when using a CI; (b) hearing history; and (c)
demographics.

Cognitive effort.

This domain contained 16 questions that were subdivided as follows. Question, 1-
5 asked about the participant’s personal perception of cognitive effort pre-implant and at

various intervals post-implant (in a week, and in 3-6 months or longer). Question 6a-6s

19

investigated the participant’s own perception of cognitive effort in various listening
situations (in quiet vs. in low, medium and high noise levels; environmental vs. speech
background noise; familiar vs. unfamiliar conversation topic; familiar vs. unfamiliar
single talker; familiar vs. unfamiliar group of talkers; slow vs. fast talker; non-native
English talker vs. native English talker; and familiar vs. unfamiliar talker on the phone).
Questions 7- 8 investigated listening when engaged in a simultaneous task (e.g., taking
notes, or driving a car). The possible advantage of using the hearing aid at the ear
opposite to the implanted ear was explored in questions (9- 10). Questions 12 -14 were
geared toward lipreading effects on speech listening. The last questions (15-16) were
about how often participants communicated with familiar vs. unfamiliar talkers.

Hearing history.

The hearing history domain included questions 17-27. Question 17-18
investigated the etiology of hearing loss (i.e., sudden or progressive). Questions 19- 20
focused on the age at which the hearing loss was diagnosed and when the loss reached a
level of severe-to-profound in both ears. Questions 21 -23 included age at initial hearing
aid use, how often the hearing aid was used prior to getting the CI, and how often the
hearing aid was used in the ear opposite to the implanted ear. Questions 24-25
investigated the ear of implant (i.e. right vs. left) and duration of implant use. Vision
status was the focus of questions 26-27.

Demographics.

This domain included questions 28-31. It posed questions about age, whether

English was a participant’s first language, level of education and employment status.

20

Questions construct.

When constructing the survey questions, the main goal was to develop a query
that respondents would interpret similarly and respond to accurately (Dillman, 2000).
Therefore, considerable pilot work to refine wording and content was required to
generate the survey questions (Rattray & Jones, 2007).

Each question in the survey was composed of a “stem” and a “response format.”
The stem is the question that requires a response and the response format provides a
framework for subjects’ answers (Passmore et al., 2002). The wording of stems was
developed to be short - fewer than 20 words, and simple and easy to understand and
interpret (Stone, 1993). Simple vocabulary was used to make sure that it could be
understood by most people. Ambiguity was avoided (Burns et al., 2008), and the stem
wording was developed to mean the same thing to both the respondent and the inquirer
(Stone, 1993). Also, an equal number of positive and negative categories for the scalar
questions was used to convey that disagreement is an acceptable answer (Dillman, 2000).

Questions were asked one at a time (Passmore et al., 2002) and double-barreled
questions were avoided (Dillman, 2000). Double-barreled questions contain two
components that require one answer, but about which a respondent may think differently
(e.g., Do you wear a hearing aid with the CI and use it all the time?) There is evidence
that item non-responses are more frequent for these types of questions (Bowling, 1997).
Consideration was given to the order in which items and domains were presented. For
example, to engage participants and prevent boredom, demographic and/or clinical

questions were presented at the end of the survey (Rattray & Jones, 2007).

21

Response format.

Response formats used in the survey were both closed and open. An open
response format allows subjects to answer a question in free text (questions 17, 19, 20,
and 21). Closed-response formats give subjects a structured way to answer items by
requiring the subject to choose from a list of options (Passmore et al., 2002) (questions 1-
16, 22-27, and 29-30). Questions 25 and 30 were partially-closed ended and several
response categories were listed while allowing respondents to add other answers
(Dillman, 2000).

Answers were generally provided in a 5-point Likert-type scale. The traditional
Likert scale, with a statement as a stem followed by responses ranging fiom strongly
disagree to strongly agree, is commonly used in surveys (Passmore et al., 2002) . The
Likert- type format used in the cognitive effort domain measured attitude towards
cognitive effort when listening in different conditions in terms of being
effortful/effortless (strongly effortful, somewhat effortful, neither effortful or effortless,
somewhat effortless, strongly effortless and no opinion) (questions 1- 8). The likert-type
response scale (all the time, most of the time, some of the time, rarely and never) was
used when respondents were asked to rate the fi'equency with which they use a hearing
aid in conjunction with a CI (questions 9), lipreading (questions 12-13), and
communication with familiar/unfamiliar speakers (questions 15-16). In the hearing
history and demographic domains, other response formats such as open-ended answers
(questions 19-21 and 28) and yes/no answers (questions 29) were provided. See

Appendix A.

22

Survey appearance.

The questionnaire was printed as a booklet (i.e. on sheets of paper folded, and
stapled through the spine). This format is recommended by survey experts (e. g. Dillman,
2000; Sudman, & Bradburn, 1982), as it provides ease in reading and turning pages, and
reduces the risk of losing pages. Also Bourque and Fielder (1995) suggested that a
booklet presented a more professional appearance.

Creating visual navigational guides to assist respondents in adhering to the
prescribed navigational path and correctly interpreting the written information was
considered in the appearance of the survey construction. Dillman (2000) suggested a
number of aspects such as using dark print for questions and light print for answer
choices, separating optional or occasionally needed instructions fiom the question
statement by font or size variations, and using simple answer boxes for recording
answers. Survey experts also agreed that vertical answer formatting is easier than the
horizontal formatting for self-completing the survey (Bourque & Fielder, 1995; Dillman,
2000; Sudman & Bradburn, 1982).

Instructions were placed where needed and not at the beginning of the
questionnaire (Bourque & Fielder, 1995). Following Dillman’s (2000) recommendation,
text was organized so that the question, associated instructions and response categories
all appeared on a single page. He also recommended instructions regarding how the
answer questions should be provided to distinguish instructions from the question
distinctive format of printing (e. g., using upper case or italics). Instructions of skip
patterns were given to respondents to guide them to the appropriate questions. Special

attention was given to the front cover of the survey that contained, as recommended by

23

Dillman (2000) and Sudman et al. (1995), the title of the survey (which should convey its
purpose in an interesting but neutral manner), a graphic illustration, and brief
instructions.

Consistency questions.

Consistency is important to test reliability and validity (Fowler, 1995). Fowler
(1995) indicated that a way to measure the consistency of survey questions is simply to
ask the same person the same questions twice, generally with minor differences in
wording. Inconsistent answers to the same questions, are considered to be evidence of
invalid reporting. In the current survey, two consistency questions (question 4 and 60)
were presented. The first consistency question (question 4) “After using your cochlear
implant for six months or longer, did listening to speech become more or less effortful
than it was before the implant?”, represented a consistency check when compared to
question 5, “Did speech listening become more or less effortfiil when you became more
accustomed to hearing speech with your cochlear implant?”. The second consistency
question (question 60), “How effortful or effortless is listening to someone using English
as a first language” was compared to question 6q, “Did speech listening become more or
less effortful when listening to someone using your native language?”

Pilot testing.

The questionnaire was pre-tested with a group of users to ensure respondents
comprehended questions as intended in the study by carrying out cognitive interviews.
Cognitive interviewing is a technique that is used to determine whether respondents
comprehend questions as intended by the survey sponsor and whether questions can be

answered accurately (Dillman, 2000). There are two main cognitive techniques: think

24

aloud interviewing and probing. In the think-aloud approach the respondent is asked to
think-aloud while completing the questionnaire, whereas the probing method involves the
investigator asking questions or probes to elicit how the respondent went about answering
the question. In the think-aloud approach, respondents concurrently describe what they
are thinking and answer the survey questions. Probing can be used either concurrently or
retrospectively (Collins, 2003). In the current survey both methods were used to elicit
answers and feedback from respondents. These techniques were effective in identifying
whether respondents understood the questions in a consistent way. Questions reported by
respondents to be a problem were either deleted or reworded.

Survey administration.

A review by Heberlein and Baumgartner (1978) showed that the total number of
contacts (including both prenotification and follow-up) had the strongest impact on
response rates. More recent reviews (Edwards et al., 2002) and experts in the field of
survey methodology (Dillman, 2000) also agreed that it is important to make multiple
contacts and that the total design method should involve four contacts in total: the
prenotification initial mailing and up to three follow-ups.

In the present study, initial contact with respondents was via a simple,
personalized cover letter that was mailed with the questionnaire. The cover letter content
included thanking the subject for answering the questionnaire questions, explaining the
purpose of the survey, explaining why that person was chosen to complete it, and why
that person’s participation was important to the study. A stamped, self-addressed pre-paid
envelope was included with the questionnaire. Respondents were requested in the cover

letter to complete and return the survey questionnaire in the pre- paid envelope.

25

Edwards and colleagues (2002) reported that sending a personalized contact letter
with the questionnaire, sending the questionnaires by first class post, and providing a pre-
paid return envelop are associated with increased response rates. Each questionnaire was
assigned a serial number to ensure security of the contacted names and to keep track of
who returned the survey and who did not.

Two to four weeks after mailing the questionnaire, a follow-up contact was
attempted. A follow-up contact letter and a replacement questionnaire were mailed to
participants who did not respond. The follow up letter indicated that the person’s
completed questionnaire had not been received and urged the recipient to respond.
Nakash and colleagues (2006) identified, in a systematic review of 15 randomized trial
reviews in health research that reminder letters have a favorable impact on response rates.
Regarding the increase in response rate fi'om making multiple contacts, Edwards and
colleagues (2003) reported similar results in their systematic review.

The questionnaire, cover letter, and second contact letters were approved by the
Institutional Review Board (IRB) of the University of Michigan and Michigan State
University prior to mailing. Contact letters were printed on letterhead stationery of the

Department of Communicative Sciences and Disorders, Michigan State University.

26

Results and Discussion

Questionnaires were mailed to 260 prospective participants soliciting information
on cognitive effort, demographics, and hearing history. Out of the 260, 138 participants
(53%) responded to the survey. Of the 138 respondents, 88 were female (63%) and 51
were male (37%). Surveys returned by 12 of the respondents (1 male, 11 females) were
excluded from further analysis because of inconsistent responses to the “consistency
questions” (see the survey instrument section above), or because there was other clear
evidence that the user had misunderstood the survey instructions. Hence the present
report is based on an analysis of the survey answers provided by 127 participants.

There was a total of 31 questions on the survey, several with multiple parts.
Appendix A provides descriptive statistics regarding the participants’ responses to all of
these questions. The focus of this dissertation — and the focus of a set of analyses
reported in this chapter — was on responses given to a subset of questions that assessed
perceived cognitive effort in relation to three factors of primary interest: (a) time since
implant; (b) level and type of background noise; and (c) talker-related variability.
Aspects of the participants’ hearing history and demographics that may have influenced
their responses to questions regarding these three primary factors were also considered

here.

27

Participants

Survey participants were accessed from the University of Michigan CI Program
patient database. All identified participants had received their CIs and subsequent therapy
through the University of Michigan CI program. Prospective participants had received a
unilateral CI and had used the device for more than 6 months.

Demographics.

All of the participants were native speakers of English and ranged in age from 23-
90 years (mean=63). Of the 127 participants who responded to the survey, 69 (54%) were
5 60 years old, 54 (43%) were > 60 years, and 3% were age anonymous.

Education level varied widely between participants and was distributed as follows:
elementary school (2%), junior high school (2%), high school (or vocational high school)
(29%), some college (34%), or 4 years of college (13%).The highest level of education
was more than 4 years of college (16%). Only (1%) of the participants did not identify
their level of completed education.

Employment status also varied substantially. Participants were either employed
(25%), unemployed (6%), owned their own business (4%), students (2%), retired (54%)
or involved in other. roles (8%). As in education, (1%) of the participants did not give
their employment status.

Hearing history.

Hearing loss onset and duration of deafness. The participants reported that their
hearing loss (whether sudden or progressive) was identified as severe-to-profound as
early as year 1 and as late as year 86 (mean=42 years). Duration of deafness ranged

between 0-65 years (mean=21). Of the 127 participants, 27 (21%) reported having a

28

sudden hearing loss and a 100 (79%) reported having a progressive hearing loss (i.e.
worsened over time).

Etiology of hearing loss. The most common etiologies of hearing loss reported
included heritable causes (syndromic, non-syndromic), noise exposure, meningitis,
enlarged vestibular aquaduct, ototoxic drugs, head trauma, Meniere’s disease, chronic ear
infection, viral disease, nerve deafiiess, and otosclerosis. The rest of the participants
reported deafness of unknown cause. Table 2-1 gives a frequency count of the etiologies

of hearing loss across the 127 participants.

Table 2-1. The etiologies of hearing loss across participants.

 

Etiology of HL No. of participants Average percentage
Heritable 32 26%
Noise exposure 7 6%
Meningitis 7 6%
Ototoxic drugs 1 1%
Enlarged Vestibular aquaduct 2 2%
Head Trauma 3 2%
Meniere’s disease 1 1%
Chronic ear infection 1 1%
Viral disease 1 l 9%
Nerve deafiiess 13 10%
Oto sclerosis l 1 %
Unknown 48 38%
Total 127 100%

 

29

Hearing aid use before implantation. The median age at which participants
received a hearing aid before implantation was 29 years; the average for 1- 81 years.
Most participants used their hearing aid in the implanted car all the time (61%). Other
participants used their hearing aids most of the time (8%), some of the time (3%), rarely
(5%) or never (21%). Of the participants, 2% gave no response to this question.
Participants using a hearing aid in the opposite ear used a hearing aid prior to Cls all the
time (54%), most of the time (6%), some of the time (3%), rarely (6%), or never in their
life (31%). The question was not responded to by 2% of the participants.

Hearing aid use after implantation. After implantation, only 14% of the
participants reported using a hearing aid in the opposite ear all the time, 3% most of the
time, and 2% some of the time. In comparison, the highest percentage (75%) reported
never using a hearing aid in the opposite ear post-implant. The rest of the participants
(3%) reported rarely using a hearing in the opposite ear. Only a few participants (2%) did
not respond to this question.

Ear of implantation. Slightly more participants (54%) had a right ear implantation
compared to the 44% who received an implant in the left ear. The remaining 3% gave no
response to this question. What might explain the higher percentage of implant in the
right ear is the right ear advantage (REA). For more than 50 years, the REA has been
reported in the auditory processing literature (Gadea, Espert, & Chirivella, 1997). The left
hemisphere is dominant for speech and language processing and the contralateral
auditory pathways are stronger. Therefore, when sound from the right ear is sent to the

left hemisphere (via contralateral pathways) a right ear advantage is often apparent in

30

speech, language, and dichotic presentations of language-based sounds, particularly in
younger people (Tadros et al., 2005; Henkin et al., 2008).

C1 use. All participants had been using their Cls for a period longer than 6
months. Specifically, 4% of the participants have been wearing the device for 7-12
months, 9% for 13-18 months, 5% for 19-24 months and 80% for more than 2 years.
Only 2% of the participants did not report length of time they have been using their

device.

Cognitive Effort Questions

Response rate.

The present analysis focused on 20 self-evaluation questions regarding cognitive
effort. Table 2-2 lists the response rate to each of these questions. The response rate was
90% or above for all but one question (Question 65). In that instance, there was an 88%

response rate. The average response rate for all 20 cognitive effort questions was 97%.

31

Table 2-2. Summary 3f cognitive eflort response rates (based on 127 survey participants)

 

 

Question focus # of Responses %

Before C1 124 98%
lst Week after CI 1 19 94%
3-6 Months after CI 125 98%
>6 Months after CI 124 98%
Quiet 123 97%
Low noise 125 98%
Medium noise 126 99%
High noise 125 98%
Familiar topic 124 98%
Unfamiliar topic 126 99%
Familiar talker 123 97%
Unfamiliar talker 124 98%
Familiar group 125 98%
Unfamiliar group 126 99%
Familiar talker on phone 1 14 90%
Unfamiliar talker on phone 112 88%
Slow talker 124 98%
Fast talker 125 98%
Native English talker 122 96%
Non native English talker 117 92%
Average 123 97%

 

“No Opinion” rate.

Answers to the cognitive effort questions were given on a 6-point scale. Steps 1-5
of the scale gave self-ratings of cognitive effort needed when performing a particular
listening task. Step 6 was a “no opinion” option. The “no opinion” option was available

to respondents who felt they could not rate cognitive effort for a particular situation.

32

Table 2-3 reports the frequency of “no opinion” responses given to each cognitive effort

question and related statistics. Figure 2-1 shows a plot of “no opinion” response rate.

Table 2-3. SummaQ/ of “No Opinion ” reports for each of the cognitivgfiort questions.

 

 

Question focus # of Respondents # Reporting %
no-opinion

Before CI 124 3 2%
lst week after Cls 119 8 7%
3-6 weeks after Cls 125 2 2%
> 6 months after C13 124 3 2%
Quiet 123 4 3%
Low noise 125 2 2%
Medium noise 126 l 1%
High noise 125 2 2%
Familiar topic 124 l 1%
Unfamiliar topic 126 l 1%
Familiar talker 123 4 3%
Unfamiliar talker 124 l 1%
Familiar group 125 1 1%
Unfamiliar group 126 1 1%
Familiar talker on phone 114 9 8%
Unfamiliar talker on phone 1 12 10 9%
Slow talker 124 2 2%
Fast talker 125 l 1%
Native E. talker 122 3 2%
Non native E. talker 117 12 10%
Average 123 4 3%

 

Respondents who gave “no opinion” for a question were excluded from any
statistical analysis associated with that question. Table 2-3 shows that the number of

exclusions of this kind was small, typically <5% (mean=3%). The largest “no opinion”

33

rate for any question was 10% (question 6p). The rates of 3 other questions (2, 6r, 6s)

were also relatively high (2= 7 %, 6r= 8%, 6s= 9%).

12% "
10% 4
8%“
6%"
4%"

2%“

“No-opinion” Response Rate (%)

 

12 3 4 6a6b6c6d6g6h 6i 6j6k 61 6r6s6m6n6o6p

Questions #
Figure 2-1. Percentage of “no opinion” responses given to the cognitive effort
questions. '

The rating scale: data coding.

The five response options for cognitive effort rating were as follows: strongly
effortful, somewhat effortful, neither effortful nor effortless, somewhat effortless, and
strongly effortless. For the present study, those options were assigned numbers on a 5-
point scale of increasing effort as follows:

‘61,,

“Strongly effortless” was coded as and is referred to here as “low” effort.
“Somewhat effortless” was coded as “2”.
“Neither effortful nor effortless was coded as “3” and is referred to as “mid”.

“Somewhat effortless” was coded as “4”.

“Strongly effortful” was coded as “5” and is referred to as “high” effort.

34

Data Analysis — cognitive effort questions

Respondents’ ratings regarding cognitive effort, calculated on this 5-point scale,
were the primary data for this study. Means and standard deviations for various
conditions are reported graphically (and compared statistically) throughout this chapter.
Figure 2-2 provides an example of the format for graphical presentation. Responses are
shown for two questions about perceived cognitive effort, effort “before Cl” (prior to
implant) and “1 week after Cl” (one week post-implant). The mean response “before
CI” was 4.73 (SD. = 0.78) which represents “high” effort, near the maximum possible
perceived cognitive effort on this scale. This compares to a mean rating of 2.35 (SD.

=1 .22) for cognitive effort needed just “1 week after CI”.

High -

Cognitive Effort Rating
3
D.

E‘
2

 

 

 

 

Before C 1 After a week
Time Since Implant

Figure 2-2. Cognitive effort before receipt of the cochlear implant and one week post-
implant (mean and standard deviation).

35

Repeated measures analysis of variance (RM-AN OVA) was used to analyze the
results of this study (SPSS software) (Field, 2009). Most factors in the analyses
(described below) were binary, i.e. they had only two levels. There were, however,
several factors with 3 or more levels. In such cases, there is potential for overestimation
of statistical significance due to violation of the “sphericity” assumption of RM-AN OVA.
The Greenhouse-Geisser correction takes this into account and reduces the degrees of
freedom in the statistical analysis accordingly (Field & Hole, 2003). The Greenhouse-
Geisser correction was used throughout this study, whenever three or more levels of a
factor were analyzed.

Time since implant.

As discussed in Chapter 1 (see Time Since Implant section) a number of previous
studies have provided evidence of improving speech perception for the first 1-2 years
post-implant for postlingually-deafened adults using multichannel Cls. Typically, no
further improvements or declines in speech perception are seen thereafter.

The current study participants were asked to rate their cognitive effort when
listening to speech at each of 4 points in time: before implant, one week post-implant, after
3-6 months, and after more than 6 months. Figure 3 shows the results. Perceived cognitive
effort before implant (Question 1) was reported to be extremely high (mean = 4.73, SD.
= 0.78). After a week of CI use (Question 2), the perceived effort score was substantially
reduced (mean = 2.35, SD. = 1.22). After 3-6 months post-implant (Question 3) the effort
was reduced further (mean= 1.62, SD. = 0.88). Finally, after 6 months (Question 4) the

effort rating was reduced to its lowest reported value (mean = 1.40, SD. = 0 .78).

36

High-

 

   

 

 

on
.S
’5 _
Of.
‘5
8: .
Lt.) Mld-l
o
.2
-
DD
0
U

Low

 

 

 

   

 

Before CI After 1 week 3-6 mos >6 mos

Time Since Implant

Figure 2-3. Cognitive effort and time since implant.

A statistical analysis of the results shown in Figure 2-3 found a highly significant
main effect of time since implant (F (2.43, 281.82) = 342.23, P<.0001). To examine
differences between the individual means, Bonferroni-corrected post-hoe tests (paired t-
tests) were conducted. Table 2-4 shows the results of these tests, conducted on time-
adjacent intervals (e. g., before implant vs. one week after). Comparison 1 reported in this
table (before implant vs. one week after) shows that participants experienced a significant
reduction in cognitive effort immediately post-implant. Comparison 2 shows that further
significant reduction in cognitive effort was experienced after 3-6 months. Finally, there
was no significant difference between the mean effort scores reported at 3-6 months and >

6 months. Overall, the pattern of results found here shows a marked reduction in cognitive

37

effort immediately post-implant and some further reduction in effort over the first 6 months

as the users accommodate to listening with a CI.

Table 2-4. Pairwise comparisons between adjacent time since implant means
(Bonferroni-protected t-tests).

 

 

Time since implant comparison Mean difference SE Significance
Before vs. After a week 2.41 0.14 0.0001
After a week vs. After 3-6 months 0.74 0.10 0.0001
After 3-6 months vs. After >6 months 0.19 0.09 0.173

 

These self-report results generally correspond well with previous studies of speech
perception which indicate that the greatest improvement in speech perception performance
takes place during the first 6 months post-implant (Hamzavi et al. 2003; Oh et al., 2003;
Ruffin et al., 2007; VaLimaa & Sorri, 2001). Speech perception often continues to
improve after 6 months; whereas, no significant improvement in cognitive effort was
reported here beyond 6 months. In other words, these results indicate that improvement in
cognitive effort takes place somewhat faster than it does in speech perception, and may
asymptote after 6 months of implant use.

The largest improvement seen here took place in just a week post-implant. Hence,
consistent with a number of other reports (e. g., Hamzavi et al., 2003), the participants here
reported immediate and dramatic improvements in listening post-implant. What perhaps
explains this immediate improvement in cognitive effort in postlingually-deafened adults
are the top-down listening skills in speech and language that they had developed prior to

deafiiess. Top-down listening skill involves a listener's ability to bring previous information

38

to bear on the task of understanding the "heard" language (Morley, 2001). This may
enhance the immediacy of improvement, as reported here. In summary, results indicated a
dramatic reduction in perceived cognitive effort immediately post-implant and continuing
improvements thereafter over approximately the first 6 months post-implantation.

Background noise.

Studies have agreed that speech performance deteriorates rapidly with increased
levels of background noise for CI users (See chapter 1, speech perception section).
Overall, listening to speech in competition with noise, especially noise from other talkers,
is challenging for most CI users. A CI listener cannot gain an advantage in distinguishing
among multiple talkers by perceiving their different voice pitches (Nelson et al., 2003;
Stickney, Zeng, Litovsky, & Assmann, 2004) and this inhibits their ability to understand
target speech (Nelson et al., 2003). Other contributing factors include limited frequency
resolution with an implant and possibly spectral “smearing.”

The present study assessed CI users self-perceptions of the effort required to
perceive speech under various background noise conditions. Specifically, participants
were asked to rate cognitive effort for each of four levels of background noise: quiet, low
noise, medium noise, and high noise (Questions 6a, 6b, 6c, and 6d). The results are shown

in Figure 2-4. In quiet, cognitive effort was reported to be low (mean =1.55, SD. = 1.04).

In low background noise, cognitive effort was higher than in quiet (mean = 2.13, SD.

1.13). In medium background noise, cognitive effort increased to the mid-level (mean
3.07, SD. = 1.25), and cognitive effort was given its highest rating (mean = 4.26, SD. =

1.12) in high background noise.

39

High -

Mid e

 

 

 

Cognitive Effort Rating

Low1

 

   

 

 

Quiet Low .Medium High
Background Noise

Figure 2-4. Cognitive effort ratings for listening in quiet and in three different
levels of background noise.

Overall, there was a highly significant main effect of background noise on
cognitive effort (F (2.16, 259.14) = 273.79, P<.0001). This is consistent with numerous
speech perception studies which have reported background noise to have a negative
impact on speech recognition performance in adults using Cls. The presence of
background noise reduces the intelligibility of a target talker’s speech and at the same
time increases the need for the listener to apply top-down strategies to compensate. The
present findings show that Cl users are conscious of this cognitive demand and sense that
it becomes increasingly great when the noise level increases (and the signal-to- noise
level decreases).

Table 2-5 reports post-hoc tests between all adjacent means shown in Figure 5. All

of the pairwise differences were statistically significant beyond p<0.001. The present

40

results are consistent with Fetterman and Domico’s (2002) speech perception study, in
which speech recognition scores were found to be significantly different between
different background noise levels. Clearly, the current survey participants were aware of

the perceptual challenges posed by increasing levels of background noise.

Table 2-5. Pairwise comparisons between adjacent noise level means.

 

 

Background noise comparison Mean difference Std. Error Significance
Quiet vs. Low -0.57 0.06 0.0001
Low vs. Medium -0.96 0.09 0.0001
Medium vs. High -1.21 0.10 0.0001
Noise type.

Participants were also asked to compare the cognitive effort required when
listening in background noises of different types. Specifically, they compared speech
noise to environmental noise. Numerous speech perception studies have reported that the
more similar the spectral characteristics of a masking noise are to the speech signal, the
greater is the masking effect (Hygge, Ronnberg, Larsby, & Arlinger, 1992; Larsby &
Arlinger, 1994). Participants in the current study were asked to rate cognitive effort when
background noise was environmental (e. g., car horn) and when there was speech noise
(e. g., multiple talkers in a restaurant). Results are shown in Figure 2-5. Ratings for the two
noise types were similar, although the effort rating was slightly and statistically
significantly higher for the speech noise (F (l, 124) = 8.32, P<0.0005). Apparently, the

users are aware of somewhat greater cognitive challenges when the background noise is

41

speech. However, the high effort rating for both noise types (speech: 4.0, environmental:

3.78) point up their general difficulty when dealing with background noise of any kind.

High-

 

 

Mid -

Cognitive Effort Rating

Low ~

 

 

 

 

 

Environmental Speech

Type of Noise

Figure 2-5. Cognitive effort for listening in environmental vs. speech background
noise.

Talker-related variability.

Physical differences between talkers introduce variability into the speech signal.
Variability also results from changes in the speaking rate and from the speaker’s
language history (native vs. non-native speaker). In the current survey, a participant’s
perception of the importance of talker related variability was investigated. Specifically,
participants were asked to rate cognitive effort according to (a) the “familiarity” of topic
and talker, and (b) “speaking characteristics” of the talker (fast vs. slow talker, native vs.

non-native talker).

42

Familiarity effects.

Possible effects of familiarity on cognitive effort were evaluated. Specifically,
participants were asked whether there was a perceived difference between listening in a
familiar condition vs. the corresponding unfamiliar condition. The results for a series of
comparisons are given below.

Familiarity of topic. Participants were asked to rate cognitive effort when listening
to a familiar topic (question 6g) vs. an unfamiliar topic (question 6h). Figure 2-6 shows
the results. Cognitive effort was reported to be within the low-mid range (mean = 2.60,
SD. =1.26) when listening to a familiar topic. Cognitive effort increased significantly (F
(1,123) = 135.58, P< .0001) to the high-mid range (mean response= 3.60, SD. =1 .22)

when the topic was unfamiliar.

Highz
on
.E -
*5
M
t a
g? Mid ‘
Lu
0.)
.2.
ED
0
U
Low —

 

 

 

 

 

Familiar Unfamiliar

Topic of Conversation

Figure 2-6. Cognitive effort and level of familiarity of the conversation topic.

43

A few perceptual studies (e.g., Tye-Murray & Tyler, 1988) have examined the
impact of topic familiarity on performance in individuals with hearing loss. These studies
have documented that unfamiliar content is more difficult to perceive than familiar
content. Results fiom the current study agree with those results.

Familiarity with one or more talkers. Cognitive effort was rated when listening
to a single familiar talker (Question 6i) vs. a single unfamiliar talker (Question 6j), and
when listening to a familiar group of talkers (Question 6k) vs. an unfamiliar group of
talkers (Question 61). Results are given in Figure 2-7. Cognitive effort was reported to be
at the mid range when listening to an unfamiliar talker (mean = 2.63, SD. =1 .21) and to
substantially decrease, to the low range (mean = 1.79, SD. =1.13), when listening to a
familiar talker. Similar results were reported in the case of listening to a group of talkers.
Cognitive effort was at the high-mid range when listening to an unfamiliar group (mean =
3.90, SD. =1.16) and decreased to the lower-mid range (mean = 3.28, SD. =1.28) when

the group was familiar.

44

 

 

 

 

 

 

 

 

 

 

Single talker Group of talkers
High .
DD 1
.E
ii
1': .
é Mid '1
Q)
4)
.Z "
'E
0.0
o d
U Low
Familiar Unfamiliar ‘ Familiar Unfamiliar
Talkers

Figure 2-7. Cognitive effort results when the talkers were known (familiar) or
unfamiliar.

Analysis of the results shown in Figure 2-7 revealed significant main effects of
talker familiarity (familiar vs. unfamiliar) (F (1,120) = 94.35; P<.0001), and number of
talkers (single talker vs. group of talkers) (F (l, 124) =54.68, P<.0001) on cognitive
effort. These results indicate that postlingually deafened CI users exert more cognitive
effort when listening to an unfamiliar talker (or talkers) than when listening to a familiar
talker, and when listening to a group of talkers than when listening to a single talker.
Present findings agree with results from previous studies (Bradlow, & Pisoni ,1999; Kirk,
Pisoni, & Miyamoto, 1997; Takayanagi, & Moshfegh, 2002) that report word recognition
performance to be consistently lower in multiple-talker conditions than in single-talker
conditions, in adults with mild-to-moderate hearing loss. Also, word recognition scores
were reported to be superior in familiar (single /group) conditions than in unfamiliar

(single/group) conditions.

45

Talking on the telephone. Familiarity regarding a talker on the telephone was also
rated. Results are given in Figure 2-8. When talking on the phone to an unfamiliar talker
(question 65), cognitive effort was reported to be in the high-mid range (mean = 3.77,
SD. =1.29). In comparison, when the telephone talker was familiar (question 6r),

cognitive effort reduced to the low-mid range (mean = 2.74, SD. =1.49).

Cognitive Effort Rating
:2:
G.

 

 

 

 

 

Familiar Unfamiliar
Talker on the phone

Figure 2-8. Cognitive effort results regarding familiarity of talker on the phone.

Statistical analysis of the results in Figure 2-8 indicated a significant main effect
of familiarity of talker on the phone (F (1,110) = 751.36; P<.0001) on cognitive effort.
Cognitive effort tends to increase on the telephone, particularly when listening to an
unfamiliar talker. These results correspond with results from other CI studies (Cray et al.,
2004; Faber & Grontved, 2000) that found understanding of conversation on the
telephone to be more difficult as the topic and speaker become more unfamiliar. A lack of

access to visual cues, as well as the phone’s limited frequency bandwidth, causes

46

understanding speech to be more difficult on the telephone in general. These factors
appear to exaggerate the difficulty.

Speaking characteristics of a talker.

Cognitive effort was also evaluated in the current study according to the following
characteristics of a talker: rate of speech (slow vs. fast talker) and native language (native
vs. non-native English speaker). Results of these comparisons are presented in Figure 2-
9. Cognitive effort was reported to be low (mean = 1.85, SD. =1.13) when listening to a
slow talker and considerably increased when listening to a fast talker (mean = 3.57, SD.
=1.23). Also, cognitive effort was indicated to be in the low-mid range when listening to
a native English speaker (mean = 2.16, SD. =1 .25) and to increase substantially (mean =
3.79, SD. =1 . 14) when listening to a non-native English speaker. Both of these
differences were statistically significant: speaker rate (F (1,123) = 951.78; P<.0001), and

native language (F (1,113) = 188.12; P<.0001).

 

 

 

 

 

Speaking Rate Native English Speaker

at) High .

.S

’5

M q

‘5

8': Mld , _
m l
o

.2

'E

8"

U LOW.

 

 

 

 

 

 

   

Slow Fast Yes
Talker Characteristics

Figure 2-9. Graph of cognitive effort results according to the talker characteristics.

47

These findings are consistent with results from prior studies (e. g., Bradlaw &
Pisoni; 1999; Kirk et al., 1997) that reported word recognition scores to be lower for
speech produced at a fast speaking rate compared to a slower or normal speaking rate for
adults with hearing loss. The present results indicate that Cl individuals are consciously
aware that they must expend extra effort when listening to fast speakers or non-native

speakers.

Additional Factor Questions

There is a good deal of variability among CI users regarding their ability to
perceive speech post-implant. All of the following factors have been reported to account
for some portion of that variability: (a) gender, (b) age, (c) cause of hearing loss, (d)
duration of implant use, (e) ear of implant, and (f) using hearing aid in the opposite ear
post-implant (Gifford, Dorman, McKams, & Spahr, 2007; Green et al., 2007; Hamzavi et
al., 2003; Henkin et a1, 2008, Morrell et al., 1995; Tyler et al., 2002; Van Dijk et al.,
1999). The present study asked whether any of these factors might also explain some of
the variability among survey participants regarding their cognitive effort reports. In a
series of analyses reported below, these factors were first tested for possible main effects,
and second tested for possible interactions with other cognitive effort factors.

For each of these factors, participants were divided into subgroups based on their
responses to survey questions 9, 18, 24, 25, 26, 28. Table 2-6 shows the participant

counts in each subgroup.

48

Table 2-6. “Other Factors ” subgroups sizes.

 

Variables Subgroup l Subgroup 2
Gender Males Females
(n=51) (n=87)
Age _<_ 60 years > 60 years
(n=54) (n=69)
Hearing loss onset type Sudden Progressive
(n=27) (n=100)
Uses a hearing aid in the opposite ear Yes No
(n=28) (n=96)
Duration of implant use S 2 year > 2 years
(n=22) (n=102)
Ear of implant Right Left
(n=67) (n=55)

 

Data Analysis- additional factors questions

Main effects tests.

The purpose of the investigation here was to look for trends in the survey data that

may be investigated further in future research. With that in mind, any statistical test with

a significance level of p S 0.25 was identified here as providing evidence of a significant

trend in the data.

“Best” Conditions.

Survey responses to the following questions were averaged together to get a

combined test-estimate of a participant’s cognitive effort rating when listening in

situations that might be expected to be the “best” and hence least challenging:

49

Question 4: After using your cochlear implant for six months or longer, did
listening to speech become more or less effortfiil than it was before the implant?
Question 6a: How effortful or effortless is listening in quiet settings?

Question 6g: How effortful or effortless is listening to a familiar conversation
topic?

Question 6m: How effortful or effortless is listening to someone who speaks
slower than normal?

Question 60: How effortfiil or effortless is listening to someone using your native
language?

Question 6r: How effortful or effortless is listening to someone familiar on the
phone?

Table 2-7 gives the rating results for each subgroup regarding listening under these “best”

conditions. The rightmost column highlights instances of significant trends.

50

Table 2-7. Mean cognitive eflort rating when listening under the “Best Conditions

 

Variable Subgroup Subgroup Difference T-value P-value Signific-
l 2 ant trend
Gender Males Females
2.20 1.89 0.31 2 0.05 *
Age S 60 yrs > 60 yrs
1.88 2.10 -0.22 -1.39 0.17 *
Hearing Sudden Progressive
loss onset 2.22 1.94 0.28 1.51 0.13 *
Uses a Yes No
hearing aid 1.95 2 -0.04 -0.28 0.81
in the
opposite
ear
Duration of S 2 yrs > 2 yrs
implant 1.90 2.02 -0.13 -0.63 0.53
Ear of Right Left
implant 1.99 2.05 -0.06 -0.39 0.70

 

Significant trends in the data were found for three factors when listening under

these best conditions. Those factors were: gender, with female participants reporting

lower effort than male participants; age, with younger participants reporting lower effort

than older participants; and hearing loss onset type, with progressive onset participants

reporting reduced effort compared to sudden onset participants. Although significant,

these effects were also small, resulting in mean rating differences of 0.31 or less. Hence it

can be concluded that each of these factors had a small but consistent impact on the

participant ratings for listening in everyday situations that presented minimum

challenges.

51

“Worst” Conditions.

Responses to the following questions were averaged to obtain an effort rating
when listening under the “worst” conditions:

Question 2: During the first week after getting your cochlear implant, did

listening to speech become more or less effortful than it was before the implant?

Question 6d: How effortful or effortless is listening in high level noise settings?

Question 6h: How effortful or effortless is listening when unfamiliar with the

conversation topic?

Question 6j: How effortful or effortless is listening in a one-to-one conversation

with an unfamiliar speaker?

Question 6n: How effortful or effortless is listening to someone who speaks faster

than normal?

Question 6p: How effortful or effortless is listening to someone with a foreign

accent?

Question 6s: How effortful or effortless is listening to someone unfamiliar on the

phone?

Table 2-8 gives the rating results for listening under these “worst” conditions.

52

Table 2-8. Mean cognitive eflort rating when listening under the “Worst Conditions

 

Variable Subgroup Subgroup Difference T-value P-value Signific-
l 2 ant trend
Gender Males Females
3.50 3.39 0.11 0.71 0.48
Age S 60 yrs 60 yrs
3.38 3.48 -0.09 -O.63 0.53
Hearing Sudden Progressive
loss onset 3.54 3.40 0.14 0.81 0.42
Uses a Yes No
hearing aid 3.10 3.52 -O.42 -2.46 0.02 *
in the
opposite
ear
Duration of S 2 yrs > 2 yrs
implant use 3.25 3.48 -O.23 -l .21 0.23 *
Ear of Right Left
implant 3.49 3.40 0.09 0.64 0.53

 

Trends were found for two factors when rating listening under the worst

conditions. The first factor was use of a hearing aid in the opposite ear post-implant.

Those who used a hearing aid had effort ratings significantly lower than those who did

not. This was the largest effect found here (mean difference= - 0.42). It was also

statistically significant at the 0.02 level. Several interaction analyses reported below also

indicate that listening with a hearing aid in the opposite ear can be especially helpful

under challenging listening conditions. These results agree with findings from other

studies that have documented gains from binaural hearing. Using a CI combined with a

hearing aid in the opposite ear can lead to improved sound localization, improved speech

53

discrimination in noise, and improved overall sound quality (Dunn, Tyler, & Witt, 2005;
Tange, Grolman, & Dreschler, 2009).

The second factor that had a significant trend here was duration of implant use,
with those who had worn an implant for S 2 years providing a slightly lower rating (-
0.23) when compared to those who had worn the implant for more than 2 years. This

trend was not confirmed in any of the subsequent interaction analyses.

Interactions between the Additional Factors and the Primary Factors

This chapter began with detailed analysis and discussion of three factors that were
of primary interest in this study: time since implant, background noise level (and type),
and talker-related variability. This section reports on a series of interaction analyses that
were conducted to ask whether any of the additional factors listed in Table 2-6 may have
interacted with any of the primary factors, and hence may offer added perspective on
their interpretation. Two-way repeated-measures ANOVAs were conducted for all
combinations of the additional factors and the primary factors. A significant trend was
defined statistically as any instance in which the F -test of the interaction had an
associated significance level of 0.25 or less. A total of 10 significant trends were found.

They are listed in Table 2-9 and examined individually below.

54

Table 2-9. Interactions for which there was statistical evidence of a significant trend.
The significance level of the interaction F -test is reported here.

 

 

Time Back- Telephone Native
since ground faririiiéirity partner 1:13;: C(i'lf Lang-
CI norse famrllarlty uage
Age 0.03 0.09
Gender 0.12 0.07
0.13 0.03
Hearing loss
onset
Uses a
hearmg aid '“ 0.13 0.013 0.07 0.04
the opposrte
ear.

 

 

 

 

 

 

 

Reduced effort in a favorable listening condition.

Four significant interaction trends were found that showed the following pattern:
Lower cognitive effort reported by one subgroup compared to another when perceiving
speech in a relatively favorable listening condition, but little or no difference between the
groups when the listening became more challenging. The four interactions that took this
form were: gender and speakers’ native language; age and telephone talker familiarity;
hearing loss onset type and speakers’ native language; and hearing loss onset type and rate
of speech. These interactions are pictured in Figures 2-10 through 2-13.

Paired t-tests confirmed significant differences in cognitive effort between
subgroups as follows: females had lower cognitive effort than males when listening to a
native talker (t(120)= 2.516, p<.013) ; younger CI users had lower cognitive effort than

older users when listening to a familiar talker on the telephone (t(109)= -2.819, p<.006) ;

55

hearing loss when listening to a native speaker (t(122)= 1.97, p<.051) as well as when
listening to speech produced at a slow rate (t(120)= 1.922, p<.057). All four of these trends
were consistent with results from main effects tests that indicated generally better
performance by the corresponding subgroups when listening under “best” listening

conditions (see main effects section above).

 

 

5,-

44
t:
8
2
E 3“
is?
2
'O
3 2
(U --4
.§ ‘
66’

1a

Native Non-Native
Talker Characteristics
Gender _ Female ...... Male

Figure 2-10. Interaction between CI users’ gender and talkers’ native language (native vs.
non-native speaker of English).

56

3"

Estimated Marginal Mean

 

l-l

 

 

 

 

I l
Familiar Unfamiliar

Telephone Partner Familiarity

 

 

Age >60 yrs <=60 yrs

Figure 2-1 1. Interaction between C1 users’ age and telephone partners’ familiarity.

41-

Estimated Marginal Mean

 

 

V V

Native Non Native
Talker Characteristics
Onset type 0f hearing 1055 Progressive -....-“ SUdden

 

Figure 2-12. Interaction between hearing loss onset type and talkers’ native language
(native vs. non-native speaker of English).

57

Estimated Marginal Mean

 

 

 

I

 

l
Slow Fast
Talker’s Rate of Speech
Onset type of hearing loss Progressive ”u" Sudden

Figure 2-13. Interaction between hearing loss onset type and speaking rate.

Reduced effort in a challenging listening condition.

There were five interactions in which a subgroup showed a reduced-effort
advantage for listening in a challenging condition, but little or no advantage for the
corresponding favorable listening condition. All but one of these interactions was
consistent with results obtained earlier in the main effects analyses of cognitive effort under
“worst” listening conditions.

The unexpected interaction was between gender and background noise level. It is
shown in Figure 2-14. The specifics of the interaction were that female CI users had lower
cognitive effort ratings than males when listening to speech in background noise, but only

at higher noise levels. The basis for this gender difference is unclear.

58

4“

3“

Estimated Marginal Mean

 

 

I I

I I
Quiet Low Medium High
Background Noise
Gender — Female """' Male

Figure 2-14. Interaction between C1 users’ gender and background noise level.

All of the remaining interactions in this category involved use of a hearing aid in
the opposite ear, and all four were anticipated by the earlier finding that there was a
significant main effect trend for this factor when respondents rated listening effort under
the “worst” conditions. Here, using (or not using) a hearing aid in the opposite ear was
found to interact with: (a) talker familiarity; (b) talker familiarity on the telephone; (c) rate
of speech; and (d) background noise level, with those who used an opposite ear hearing aid
reporting lower effort than those who did not under the most challenging of these
conditions. The details of these four interactions are presented in Figures 2-15 through 2-

18.

59

5:4

3...

/

Estimated Marginal Mean

2:4 /

 

 

 

l -+
V——_- I l o A-‘ T I u
Familiar Unfamiliar

 

Familiarity of Talker
HAin the opposite ear — No --- Yes

Figure 2-15. Interaction between using a hearing aid in the opposite ear and talkers’
familiarity.

/

Estimated Marginal Mean

 

I

I
Fast

Slow
Talker’s Rate of Speech
HA in the opposite ear — No --- Yes

Figure 2-16. Interaction between using a hearing aid in the opposite ear and speaking
rate.

60

Estimated Marginal Mean
DJ

 

 

 

Familiar Unfamiliar
Telephone Partner Familiarity
HA in the opposite ear — No —--- Yes
Figure 2-17. Interaction between using a hearing aid in the opposite ear and telephone
partners’ familiarity.

 

 

 

5 1
4<~
t:
8
2
35 3
.510 :4!
E
2
“o
8
63
e 24
L3 .
/
1 .
Qtiiet de Medium High
Background Noise
HA in the opposite ear — No --.-. Yes

Figure 2-18. Interaction between using a hearing aid in the opposite ear and background
noise level.

61

In summary, the present study found repeated evidence that use of a hearing aid in
the opposite ear can reduce the cognitive effort needed to perceive speech under
challenging listening conditions. At the same time, there was evidence that the presence of

a hearing aid afforded no advantage when listening under more favorable conditions.

Age with time since implant - a cross-over interaction.

A final significant trend found here involved a crossover interaction between age

and time since implant. It is shown in Figure 2-19.

 

 

 

545

l

4.
C
8
E
'5
.5

an .

i6 3";

2 l

“U :

2 l

<3 3

E :

2;; 24
in

14

Before CI After 1 week 3-6 mos >6 mos
Time Since Implant
Age — >60 yrs "—‘r’ <=60 yrs

Figure 2-19. Interaction between CI users’ age and time since implant.

62

equal after 3 months, and the younger group crossed-over to have lower ratings than the
older group after 6 months. A possible explanation for this pattern may be that the older CI
users exceeded the younger users in the area of listening skills at the outset, most likely
because they had had a longer time to develop those skills. Better listening skills led to
reduced cognitive effort in the older age group before and immediately post-implant. The
younger users, on the other hand, were apparently more adept at developing new listening
skills post-implant, and their development continued over at least the first 6 months, whereas

skill development was largely concluded for the older group by the end of 3 months.

Conclusion

This survey found repeated evidence that post-lingually deafened CI users are
aware that they must expend special cognitive effort to perceive speech. This is
especially so when background noise of any kind is present, when the topic of
conversation or the talker (or talkers) is unfamiliar, when the talker is a non-native

speaker, and when the talker speaks rapidly.

63

CHAPTER 3

An Experimental Study of Cognitive Effort

Part two of this dissertation was an experimental study of the cognitive effort that
must be expended by a cochlear implant user when perceiving speech. Audiometric tests
of speech intelligibility are commonly used to assess the listening abilities of individuals
with hearing loss and more recently of individuals with CIs. Measures of cognitive effort
have been found to provide a useful complement to these audiometric tests, particularly
when the goal is to compare groups and/or listening conditions that may be subtly
different (Feurestein, 1992; Hick & Tharpe, 2002; McFadden, & Pittman, 2008; Rakerd,
Seitz, & Whearty, 1996; Rossiter et al., 2006). The present study made an assessment of
the cognitive effort needed to perceive speech when listening in quiet and under various
levels of competition from background noise.

Effects of noise on speech perception by C1 users, both children and adults, have
been investigated in a number of previous studies (e. g., Bento et al., 2005; Dowell,
Dettman, Blarney, Barker, & Clark, 2002; Most & Peled, 2007; Wolfe & Kasulis, 2008).
A regular finding of these studies is that background noise presents particular challenges
for CI users, far more so than for normally-hearing listeners (see Chapter 1, Level of
Background Noise for a review). The present experimental study asked a complementary
question: How do the challenges of listening in noise affect a CI user’s ability to multi-
task speech perception with other activities? As an example, consider someone who must
take notes in a classroom setting while listening to a lecture. Then imagine that noise

arises in the background, due to events going on either inside or outside of the lecture

64

THE!“

 

hall. Additional demands on any listener’s cognitive resources are likely to result
(Paccia-Fuller, 2003a). However, given the special difficulties that Cl users can have
with noise, it is quite possible that these added demands will be significantly greater for
them than for normally-hearing listeners in this same situation. That possibility was

directly tested in this study.

Dual-task Testing

One method for measuring the cognitive resources that someone must commit to
complete any task is to have that task be carried out simultaneous with a second task.
There is good evidence that an individual has limited cognitive resources to allocate to
these dual-tasks (e. g., Kahneman, 1973; Baddeley, 1986; Carpenter, Miyaki, & Just,
1994). Consequently, when one of them demands increased resources, fewer are left for
the other task. In studies like the present one, the cognitive demands of one task are of
primary interest (the demands of speech perception are of primary interest here), and
performance of the second task then becomes a source of information about those
demands. Specifically, reduced performance on the second task provides evidence of
increased cognitive resource commitments to the primary task.

Rakerd, Seitz, and Whearty (1996) used this paradigm to compare the cognitive
effort needed by hearing impaired listeners to that of normally-hearing listeners when
both groups were attending carefully to spoken messages. Their primary task was a
speech comprehension task in which subjects were required to listen for understanding to
minute-long spoken passages. The secondary task was a memory task in which subjects

learned and later recalled strings of digits. The index of cognitive effort for the speech

65

THE,"

 

task was the number of digits that were forgotten. Results showed that the hearing
impaired listeners allocated significantly more cognitive resources to speech
comprehension than did a comparison group of normally-hearing listeners. Test
protocols similar to those used by Rakerd et al. (1996) were employed in the present
study of postlingually-deafened adults who use Cls.

Here, a group of CI participants and a comparison group of normally-hearing
participants were engaged in a dual-task activity that required that they: (a) listen
carefully to sentence-long speech samples, and (b) simultaneously perform a visual
reaction time task that indexed the difficulty of speech listening. The speech listening
task was performed in quiet and in several different levels of competing background

noise.

Methods

Participants
There were two groups of participants, one who were cochlear implant users

(group CI) and a second group who had normal hearing (group NH).

The cochlear implant users group.
Twelve postlingually-deafened adults comprised the CI group. All participants
were recruited from the University of Michigan CI Program. There were 8 females and 4

males. Their ages ranged between 23—82 years old (mean=51.8) and they were all native

66

TH?

 

speakers of American English. Table 3-1 provides information about the etiology of

hearing loss for each CI participant, and about other aspects of their hearing history.

Table 3-1. Demographics and hearing history of the individuals in the CI group.

 

Participant Etiology Age Age when Age when Type HL
HL HL became
identified profound
l Hereditary 47 5 42 Progressive
2 Hereditary 40 2.5 2.5 Sudden
3 Ushers syndrome 50 6 45 Progressive
4 Unknown 5 1 5 20 Progressive
5 Meningitis 3 1 5 12 Progressive
6 Unknown 55 22 53 Progressive
7 Toxic drugs 82 20 20 Progressive
8 Otosclerosis 72 30 40 Progressive
9 Noise 5 l 39 44 Progressive
1 0 Hereditary 23 5 1 6 Progressive
l 1 Hereditary 59 30 Unknown Progressive
1 2 Unknown 56 57 1 7 Sudden

 

Inclusion criteria.

To be eligible for participation in this study, a CI user had to be a postlingually-
deafened adult who used a unilateral multichannel C1 with full electrode insertion, and
who had been using the device for 12 months or longer. It has been reported that the most
dynamic enhancements of speech perception usually takes place in the first 12 months
after fitting of a CI (Hamzavi, Baumgartner, Pok, Franz, & Gstoettner, 2003). A similar
conclusion was reached in the survey study reported above (see Chapter 2).

Another inclusion criterion for this study was that the CI user must have scored

over 80% correct on the open-set sentence recognition “Hearing in Noise Test” (N ilsson,

67

Soli, & Sullivan, 1994). This threshold was selected because it has been reported that
late-deafened adults with a unilateral CI can often achieve such a score on sentence
recognition tests (Mok, Grayden, Dowel, & Lawrence, 2006).

The last condition for participating was that the duration of deafiiess prior to
receiving a CI was 60% or less of the participant’s life. Adults implanted at a younger
age and with smaller percentages of life in deafness have been found to achieve the
highest levels of short-term postoperative speech recognition success (Blarney et al.,
1996; Green, Julyan, Hastings, & Ramsden, 2005; Oh et al., 2003).

The normally-hearing (NH) group.

The NH group consisted of 18 individuals. Participants were 21 to 59 years old
(mean = 36.16). There were 8 females and 10 males. They were either members of the
Michigan State University community or were acquaintances of the CI participants. All
NH participants passed a hearing screening test, with criteria as follows: binaural hearing
thresholds of 25 dB HL or better at 1000, 2000, and 4000 Hz.

Recruitment procedures.

An information letter was mailed to possible participants explaining the study’s
purpose, the confidentiality with which the study would be conducted, the absence of any
known risk factors, and the financial compensation for participants ($35). Interested
individuals from the NH and Cl groups then contacted the study investigator to schedule
an appointment to be tested. Participants were given the choice to be tested in one of two
locations, either at the University of Michigan Cochlear Implant Program site or at

Michigan State University’s Department of Communicative Sciences and Disorders.

68

The Dual-task

The participants performed two tasks during the experiment, a speech perception
task that is described in the next section, and a visual reaction time task that is described
in the immediately following section. As noted above, the cognitive demands of
perceiving speech were of primary interest in this study, and those demands were
assessed by means of their impact on a participant’s visual reaction time when the two
tasks were in dual-task competition.

The speech perception task.

Stimuli.

The stimuli for the speech perception task were sentences drawn from the BKB
speech recognition test set (Bench, Kowal, & Bamford, 1979). These stimuli were
presented to a listener in quiet, and in different levels of background noise. In all cases,
the background noise type was multi-talker “speech babble” (14 talkers). Multi-talker
babble was selected because of its effectiveness as a masker of speech (Carhart, Tillman,
& Greetis, 1969; Lewis, Benignus, Muller, Mallott, & Barton,, 1988).

The fiill BKB speech test set is made up of 21 lists, each having 16 sentences and
50 keywords (3 or 4 per sentence). The sentences were compiled originally from the
utterances of children with hearing loss and contain straightforward vocabulary and
syntax. Although designed for testing children, these sentences have also proved usefirl
and appropriate in testing adults (Cainer, James, & Rajan, 2008; Holden, Skinner,
Holden, & Binzer, 1995).

For the current study, a subset of the BKB test sentences was recorded by an adult

make talker with a General American dialect. The sentences were spoken in a quiet

69

environment, and digitized at the time of original recording (PCM 16-bit recording,
sample rate = 22,050 Hz). Digital audio editing software (GoldWave V5.2) was then
used to equalize the levels of the individual sentences, and to organize them into stimulus
lists with a separation of approximately two seconds between the succeeding sentences in
a list. An individual BKB sentence could appear in one list only. Five different stimulus
test lists of 32 sentences each were created for the experiment. For each listener, these
lists were randomly assigned to the single-task and dual-task test conditions as needed
(see below), with the constraint that a list was never repeated. A practice list of 16
sentences (different from any of the test lists) was also created (See Appendix B for a
complete listing of test lists and practice list).

Procedure.

Participants were tested individually, while sitting quietly in a sound-attenuated
booth. Stimuli (speech + noise) were presented from a loudspeaker directly in front of
and one meter away from the participant. The task was to attend carefully to each
sentence that was presented and to repeat the sentence back immediately upon its
completion. A directional microphone and a digital audio recorder (Zoom H4 Digital
Recorder) captured the listener’s responses. Those responses were later scored for
accuracy of reporting of the BKB key words. For each condition, the participant
completed 32 trials (one test list).

The CI users were tested while wearing their Cls. One user who also wore a
hearing aid in the opposite ear was asked to take the aid off because the current study

aimed at assessing performance of C1 users using a unilateral cochlear implant. The

70

binaural benefit and additive speech information obtained from using a hearing aid in

conjunction with a CI was not investigated.

Visual reaction time task.

For the visual reaction time (RT) task, a visual stimulus (a graphic image “X” that
filled approximately two thirds of the display) was presented at random intervals on a
computer screen. The participant was instructed to monitor continuously for this symbol
and to press a response button immediately whenever it appeared. The computer screen
was in clear view, directly in front of the participant and one meter away.

The RT task was performed singly to obtain a baseline estimate of a participant’s
reaction time, and also in dual-task competition with the speech task. In both instances,
there were a total of 24 test trials. The reaction time for a trial was calculated from the
moment of stimulus onset to the moment of the participant’s response (a button press). If
there was no response within 1500 milliseconds, a trial was scored as a “miss” and the
trial was dropped. The maximum number of misses made in any condition by any
listener was 3. The average number of misses was 1. The reaction time for a trial was
rejected as “impossibly fast” if it was less than 150 milliseconds. Rejections of this kind
were rare (28 total for the two participant groups combined). The reaction time for a

condition was calculated as the average of the RT’s obtained on all of the valid trials.

Visual stimulus presentation.
As noted, there were 24 RT test trials per condition. For the dual-task test

conditions, the visual stimuli were presented in coordination with the presentation of the

71

BKB sentences. A maximum of one stimulus presentation could occur during any one
sentence, and it did so on the following schedule. The visual symbol was either not
presented at all (8 trials, 25% of all sentences) or it was presented with equal probability
(8 trials each) during the first third, middle third, or final third of a sentence. For the
single-task reaction time test the visual stimuli were presented on the same time schedule
as for a dual-task test. (They were in fact coordinated with one of the speech test lists,

with the sound turned off.)

The testing procedures

All testing, both single-task and dual-task was completed during a single test visit
that lasted approximately 90 minutes. The sequence of events for the visit is described in
this section.

Informed Consent.

A participant consent form was reviewed and signed at the start of the visit. Any
questions regarding the consent were fully addressed prior to signing. The consent form
was approved by the Institutional Review Boards of the University of Michigan and
Michigan State University.

Hearing Screening or survey completion.

Normally hearing participants were given a hearing screening (details provided
above) at the conclusion of the signing of the consent form and prior to their participation

in the main experiment.

72

Cochlear implant participants completed a survey regarding their self-perceptions
of the cognitive effort needed to perceive speech. This was the same survey that was
reported on in Chapter 2 above.

F amiliarization with the Stimuli.

Next, participants were familiarized with the audio and visual stimuli that were to
be used in the experiment. Sixteen BKB sentences, four at each of four different signal-
to-noise ratios covering the range of those to be heard during testing, were presented as
examples. Also, four visual stimuli were presented.

Practice trials.

The participant next completed a series of eight single-task practice trials (four
speech trials and four visual RT trials) and eight dual-task practice trials. Any questions

about the procedures were fully addressed at this time.

Tests
Single-task tests.
The single-task tests were completed next. First, a participant completed one 32-

trial speech test. Then, the participant completed a 24-tria1 visual reaction time test.

Dual-task tests.
Favorable listening conditions.
Next, dual-task testing was conducted for all participants (both NH and CI) under

the two most favorable listening conditions: ( 1) speech presented in quiet, and (2) speech

73

presented at a signal-to-noise ratio (S/N) of +15 dB. All of the participants from both
groups were able to successfully complete dual-task testing under these two most
favorable listening condition tests.

The test in quiet was conducted first. Then, after a break (approximately 5
minutes), the test at +15 dB was conducted. With one exception (a dual-task test of
several CI participants conducted at S/N=+10 dB) (see below), dual-task testing was
sequenced to progress from the least challenging condition (speech in quiet) through
progressively more challenging conditions (decreasing S/N) to accommodate the needs of
CI participants in particular. NH participants were tested in this same sequence for
comparison.

Dual-task test at SW = +5 dB.

All of the CI participants and 14 of the NH participants were next tested at the
more challenging S/N of +5 dB. Every NH participant was able to perform the dual tasks
at this S/N. However, this condition proved more challenging for most of the Cl
participants. A number of them had difficulty accurately identifying the speech, and in
one instance a participant asked that the test be terminated.

Additional dual-task tests.

CI participants testing at S/N = +10 dB. Following the completion of the S/N =
+5 dB test, several of the CI participants were given a dual-task test in which the noise
level was reduced by 5 dB (i.e, S/N was set at +10 dB). This was done to reduce the
challenge of speech listening somewhat, possibly to a point where accurate speech
perception could be maintained under dual-task conditions by C1 users. Six of the twelve

Cl participants were tested in this way.

74

NH participants testing at S/N =0 dB. Following their completion of the S/N = +5
dB test, all 14 of the NH participants were tested at a more challenging S/N level to
assess the possible consequences for their dual-task performance. Specifically, they were

tested at S/N = 0 dB.

75

Results and Discussion

As noted above, participants in the present study were engaged in two tasks, a
speech perception task that was of primary interest here, and a visual reaction time task
that provided information about the cognitive demands of speech perception. For the
speech task, a participant attended to and repeated back sentence-long speech stimuli
(Bench, Kowal, & Bamford, 1979). For the visual reaction time task, the participant
responded as quickly as possible (with a button press) to a visual stimulus that randomly
appeared on a computer screen. These tasks were performed both separately (single-task
testing) and simultaneously (dual-task testing). The speech task was performed in quiet

and when listening against several different levels of background noise.

Raw Scores and Normalized Scores

Data of two types are presented below. The first type are raw scores. For the
speech task, the raw scores for a participant were the percentages of keywords correctly
identified in the single-task test and in the dual-task tests conducted at various signal-to-
noise ratios. For the reaction time task, the raw scores were the participant’s mean
reaction times for the single-task test and for the dual-task task tests.

The second type of data considered here are normalized scores. Normalized
scores were calculated for the various dual-task test conditions based on a participant’s
performance on the corresponding single-task test. For the speech task, the normalized
score was a proportion calculated by dividing the dual-task percent-correct score for a
condition by the single-task percent-correct score. For the reaction time task, the

normalized score was a difference, calculated by subtracting the single-task reaction time

76

from the corresponding dual-task reaction times. Raw scores and normalized scores were
both statistically compared across conditions by analysis of variance and, as called for, by

t-ICSIS.

Favorable Listening Conditions

This section presents the raw score and normalized score results obtained for all
tests (both single-task and dual-task) that were conducted under favorable listening
conditions. Every participant was tested under these conditions. Results are presented

first for the NH group and then for the CI group.

Normally-hearing listener group.

Speech perception test results.

Raw scores. Figure 3-la shows the NH listeners’ raw-score speech perception
test results (means and standard errors) for three tests. Test one was the single-task
speech perception test that was conducted in quiet. This test is identified in Figure 3-1a
and in a series of subsequent figures as STzSp. Tests two and three were dual-task tests
conducted under favorable listening conditions. For one of these tests; there was no
background noise; i.e., the speech was presented in quiet (condition DT:Q). For the other
test the S/N was set at +15 dB (DT:+15).

Consistent with the description of these as favorable listening conditions, the NH
listeners had mean scores of greater than 99% correct for all three of these tests.
Accordingly, an AN OVA found no significant difference among these conditions (F (1,

17) =1; p < 0.33).

77

Normal Hearing

 

 

 

‘00 ' § § g
E E 90 4
-= 0
§ 0
,g— e", 80 -
70 r .
ST:Sp DT:Q DT: +15

Test Condition

Figure 3-1a. NH listeners’ (n=18) raw score results for the speech test (mean +/-l SE)
performed as a single-task (ST:Sp) and in dual-task in competition with speech in quiet
(DT:Q) and speech presented at +15 dB S/N (DT:+15).

Normalized scores. For the dual-task conditions, a data transformation was
carried out to derive normalized speech perception test scores fiom the raw scores.
Specifically, each dual-task raw score was divided by the raw score for the single-task
speech test (ST:Sp) to calculate a proportion score. A proportion score less than 1.0
would indicate that some speech information was lost to a participant in a dual-task test
relative to the single-task baseline. This could be due to the added background noise
present in most of the dual-task test conditions (but not in condition DT:Q) or possibly
due to the added cognitive complexity of the dual-task test situation itself.

Figure 3-lb shows the NH listeners’ normalized scores for the two favorable
listening dual task tests. The means for both tests were very close to 1.0 (DT:Q mean =
1.00; DT:+15 mean = 0.98). Also, the means were not significantly different fi'om one

another (t( l 7) =1.0; p<0.33), indicating that for NH participants, speech perception at

78

S/N=+15 dB was approximately equivalent to speech perception in quiet. This finding
was expected, based on a number of prior studies of speech perception by normally-

hearing adults (e. g. Crum & Matkin, 1976; Tillman, Carhart, & Olsen, 1970).

Normal Hearing

 

Alz-
Q)

.g

wiéllt

8m
FE
3,10- F c

c:
m-E

30.9-

9.

D...

V0.8 . .

 

 

DT:Q DT:+15

Test Condition

Figure 3-lb. NH listeners’ normalized score results for the speech test.

Reaction time test results.

Raw scores. The visual reaction time task was the secondary task in the
experiment. The instruction for this task was to respond as quickly as possible (with a
button press) to a randomly appearing visual probe stimulus. Participants were also
instructed to pay equal attention to this task and the speech task in instances where the
two tasks were in dual-task competition.

Figure 3-2a shows the NH participants’ mean reaction times (+/- 1 standard error)
for three test conditions: (1) single-task visual reaction time testing (labeled STzRT); (2)

dual-task testing coincident with speech perception when the speech was presented in

79

quiet (DT:Q); and (3) dual-task testing when speech was presented at S/N=+15 dB

 

 

 

(DT:15dB).
Normal Hearing
, ‘ 400 -
is
E
q) 350 - L I
E
i: /r T
G
.g 300 -
U
8
a:
250 i r r
STIRT DT:Q DT: +15
Test Condition

Figure 3-2a. NH listeners’ (n=18) raw score results for the visual reaction time test
(mean +/-l SE) performed as a single-task (ST:Sp) and in dual-task in competition with
speech in quiet (DT:Q) and speech presented at +15 dB S/N (DT:+15)

A statistical analysis of the raw scores shown in Figure 3-2a found a significant
difference among the three conditions overall (F (1.31, 22.29) = 5.458, p< 0.009). It can
be seen that the NH participants reacted somewhat faster in the single-task test than in
either of the dual-task tests. Post-hoe tests, presented in Table 3-2 below, indicated that
there was a significant difference between the single-task score and each of the dual-task
scores, presumably because the dual-task imposed greater cognitive demands. There was
no significant difference between the two dual-task scores, indicating that the NH
participants found listening to speech at S/N=+15 dB to be no more cognitively effortfill

than listening in quiet.

80

Table 3-2. Visual reaction time score for the NH group. Post-hoc comparisons.

 

Test condition Mean Difference Std. Error t Significance
ST: Q- DT: Q -327 12.25 -2.670 0.02
ST: Q - DT: +15 -26.2 12.23 -2. 145 0.05
DT: Q-DT: +15 6.5 5.50 1.180 0.25

 

Normalized scores. The raw RT scores for the two dual-task tests were converted
to normalized difference scores (DT raw score - STzRT raw score) to show any reaction
time slowing that may have resulted from added cognitive load in a dual-task test. Figure
3-2b shows the results. There was some modest slowing relative to baseline for both dual-
task tests. Specifically, the mean normalized score for listening in quiet was +32.7 msec,
and for listening at S/N=+15 dB it was +26.2 msec. The latter RT score may have been
slightly smaller because the DT: 15 test was performed after the DT:Q test by all
participants. Seemingly, they were more practiced. The statistical difference between the
two means was not statistically different (p>0.25)

Overall, these normalized scores and the raw scores discussed above indicate that:
(a) NH listeners experienced at most a minimal cognitive load when perceiving speech
under these favorable listening conditions; and (b) for them, the load was no greater when

listening at S/N=+15 dB than when listening in quiet.

81

Normal Hearing

 

 

 

125 -
i?
a) .E
E 7’ 85 .
20; a
e 8
g3 g 45 -
r 4
5 r
DT:Q DT:+15
TestCondition

Figure 3-2b. NH listeners’ raw score results for the visual reaction time test.

Cochlear implant listeners.

Favorable listening condition results for the CI users are presented in this
section. Their speech perception testing results are considered first, then the visual
reaction time results assessing the cognitive effort required for speech listening are
presented.

Speech perception test results.

Raw scores. Figure 3-3a shows the CI participants’ raw speech perception scores.
They scored better than 80% correct for all three tests, and better than 90% correct for the
DT:Q test. This is consistent with reports that Cl users can achieve high levels of speech

perception accuracy when listening at favorable signal-to-noise ratios.

82

Cochlear Implant

100'

Speech Test
(% Correct)

80-

 

 

70 I I I
ST:Sp DT:Q DT: +15

Test Condition

Figure 3-3a. CI users’ (n=12) raw score results for the speech test.

A statistical analysis of the results shown in Figure 3-3a found a marginally
significant overall difference among the means (F (1.53, 16.85) = 4.18, p<0.04). Post-
hoc testing indicated that there was a significant pairwise difference between the means
for the two dual-task test conditions (t(11) = 3.20, p<0.008). Specifically, the scores for
the Cl users S/N=+15 dB condition were significantly lower than their scores for the
speech in quiet condition. Apparently, for CI users, even the low level of background
noise present at S/N=+15 was sufficient to produce a measurable reduction in speech
perception performance in the experiment.

Normalized scores. The normalized score results for the CI participants are
shown in Figure 3-3b. An unexpected finding regarding these speech perception tests
was that, as a group, the CI participants scored somewhat higher for the two dual-task
tests than for the single-task test. Accordingly, the normalized ratios shown in Figure 3-
3b are greater than 1.0. The most likely explanation for this finding is that dual-task
testing came later in the test session than single-task testing, which meant that the CI

users’ speech perception performance could have been aided by familiarity and practice

83

effects. Certainly the normalized scores show that the Cl users were able to maintain
good accuracy in their speech perception, even when dealing with the complexities of

dual-task testing.

 

 

Cochlear Implant
U 1.2 -
.E
.. “g 1.1-
a a
r- , -
5 a: 10
8 8 0.9-
m‘t'E
8
e 0.8 "'
E:
0.7 r
DT:Q DT:+15

Test Condition

Figure 3-3b. CI users’ normalized score results for the speech test.

There was a statistically significant difference between the normalized scores for
listening in quiet (DT:Q) and listening at S/N=+15 dB (DT:+15), with performance lower
for the +15 dB condition. This again indicates that unlike the NH listeners, CI users
experienced some measureable difficulty in performing the speech task in background
noise, even at this low noise level.

Comparisons to the NH listeners. Finally, Table 3-3 directly compares the CI
listeners’ mean scores with the NH listeners’ mean scores for all of the speech perception
tests described in this section. While the scores for the CI group tended to be lower than
those for the normally hearing group, and were, in fact, statistically significantly lower in

two instances, it is clear that that the CI users were able to maintain substantial accuracy

84

in their speech perception for all tests conducted here under favorable listening
conditions. The principal question addressed in this dissertation is how much effort the
CI users had to expend to achieve this perceptual success. The extent of that effort is
indexed by their performance on the visual reaction time task, as reported in the next

section.

Table 3-3. Comparisons of the CI and NH listener group means on several measures of
speech perception performance. (Significance indicators: * p<0.05, **p<0.01,
***p<0.001).

Ccsircigiteibn Grill; G210}: Difference t-test adj df p-value Signif.
p p (Cl-NH)

 

Raw Scores

(% Correct)

ST:Sp 88.9 98.9 -10.0 -2.624 13 0.021 "‘
DT:Q 94.2 98.9 -4.7 -2.062 18 0.054

DT:+15 84.2 98.8 -14.6 -4.l35 14 0.001 ***
Normalized

Scores

(ratio)

DT:Q 1.07 1.00 0.07 1.999 11 0.071

DT:+15 0.97 0.99 -0.02 -0.638 11 0.537

 

Visual reaction time results.
Raw scores. Figure 3-4a presents the visual reaction time raw scores for the CI
user group. There was a highly significant overall difference among the means (F (1.61,

17.67) =36.68, P<0.0001).

85

 

Cochlear Implant

 

 

 

450 -
8‘ I J
g 400 4 I
E
i: 350 .
C
'2 300
‘5 -l
8
a:

250 i I

5“" DT:Q or: +15
Test Condition

Figure 3-4a. CI users’ raw score results for the visual reaction time test.

Post-hoe testing (see Table 3-4) found that there was a significant difference
between the single-task mean and each of the dual-task means (p<0.001 in both cases).
Specifically, the CI users were significantly slowed on average relative to the single-task
baseline when reacting under both dual-task test conditions. The extent of this slowing is

shown in the normalized test scores presented in Figure 3-4b and discussed below.

Table 3-4. Visual reaction time scores for the CI group. Post-hoc comparisons.

 

Test condition Mean Difference Std. Error t Significance
ST: Q- DT: Q -102.8 11.98 -8.58 0.001
ST: Q - DT: +15 -1072 17.28 -6. 19 0.001
DT: Q — DT: +15 -44 12.60 -035 0.731

 

86

Normalized Scores. Figure 3-4b presents the Cl users’ normalized scores for the
dual-task tests conducted with speech in quiet and at S/N = +15 dB. For both of these
conditions there was notable slowing of performance relative to the single-task baseline.
Specifically, the CI users were slowed on average by 102.8 msec when listening in quiet,
and by 107.2 msec when listening at S/N=+15 dB. By contrast, the NH control group
has normalized mean scores of just 32.7 msec and 27.2 msec for these two conditions.

The clear implication of these results is that the cognitive effort required to
achieve accurate speech perception was substantial for the CI users, so substantial that it

appreciably slowed their ability to perform the competing visual reaction time task.

Cochlear lrnplant

 

 

 

125 a l
I
“g? f I
I: 4g 85‘
= m
8 r—
8 m
g g 45 “
5 r
DT:Q DT:+15

Test Condition
Figure 3-4b. Cl users’ normalized score results for the visual reaction time test.

Comparisons to the NH listeners. Table 3-5 directly compares the CI listeners’
mean scores with the NH listeners’ mean scores for all of the reaction time tests
described in this section. There were several notable findings. The first is that the two
groups had nearly identical mean raw scores for the single-task reaction time test

(difference=0.8 ms). Hence they were well matched at baseline. By contrast, the two

87

groups had significantly different raw scores for both dual-task tests. The CI group was
slower than the NH group by 70.8 msec on average when listening to speech in quiet, and
by 81.8 msec when listening to speech at S/N = +15 dB. This provides strong evidence
that the cognitive effort required to perceive speech was substantially greater for the CI
group. Finally, the statistical differences between the two groups were especially strong
with respect to their normalized scores (p<0.001 in both cases).

Taken together, the results of this experiment provide strong evidence that Cl
users must make a significantly greater than normal commitment of cognitive effort in
order to perceive speech, even when listening under acoustical conditions that are

favorable.

Table 3-5. Comparisons of the CI and NH listener group means on several measures of
visual reaction time performance. (Significance indicators: *p<0.05, **p<0.01,
***p<0.001).

Score/

Condition CI NH Difference

Group Group (CI-NH) t-test adj df p-value Slgmf-

Raw Scores (msec)

ST:Sp 312.11 311.31 0.8 0.041 23 0.967

DT:Q 414.86 344.03 70.8 2.961 19 0.008 **
DT:+15 419.31 337.53 81.8 2.627 16 0.018 *
Normalized Scores

(msec)

DT:Q 102.75 32.72 70.03 4.087 23 <0.001 ***
DT:+15 107.20 26.22 80.98 3.821 21 <0.001 ***

 

88

More Challenging Listening: Dual-Task Test at S/N = +5 dB

This section presents the results of the dual-task test that was conducted at
S/N=+5 dB. This listening condition was substantially more challenging than the two
previous conditions (listening in quiet, listening at SW = +15). Speech perception scores
for a number of the CI participants fell significantly in this condition. The test was
terminated at the request of one Cl participant. It was carried through to completion by
the other 11 CI participants, and their results are presented below. Also presented are
results for a comparison group of n=l4 NH listeners who were tested at S/N=+5 dB. All

of the NH participants were able to complete this test.

Normally-hearing and CI listeners.

Speech perception testing.

Raw scores. Figure 3-5a gives the speech test raw scores for participants in the
NH group and for participants in the CI group. There is a plot of each individual
participant’s score, and also a plot of the mean score for each group.

The NH listeners were able to maintain a high level of speech perception accuracy
when listening at S/N=+5 dB, with all but one of them scoring 98% correct or above. By
contrast, the CI users differed widely in their speech perception accuracy. One scored at
a level matching that of the CI participants, but for the majority of the C1 users listening
at S/N=+5 dB posed considerable difficulty. The mean score for the CI group was 38%
correct, which was significantly below the NH participants’ mean (t (23) =7.73;

p<0.0001).

89

100 - E]

00
O
l

O\
O
1

A
O
1

Speech Test
(% Correct)

N
O
l

 

O 1 .
NH Cl

 

Individual participant 1:] Group mean
Figure 3-5a. Speech test raw scores for the dual-task test at S/N=+5 dB. Scores are

plotted separately for each participant in the NH group (n=14) and in the CI group
(n=1 1). The mean score for each group is also presented.

Normalized scores. Figure 3-5b shows the normalized scores for the S/N = +5dB
dual-task test. The mean for the NH group was very close to 1.0 (mean = 0.98) compared
to the CI group mean of just 0.43. Generally, all participants in the NH group performed

far better than the CI participants. Only two of the CI users (18%) had normalized scores

in the range of the scores for NH participants.

90

.o "‘ t"
00 O N
1 L J

Speech Test
(Proportion, RE Baseline)
O
O5

 

 

0.0 I .
NH CI

= Individual participant B Group mean

Figure 3-5b. Speech test normalized scores for participants in both groups when tested at
+5dB S/N.

A t-test on the results presented in Figure 3-5b indicated that there was a highly
significant difference in performance between the groups (t (24) =6.83; p<0.0001). It is
clear that background noise at this level, specifically +5dB S/N, posed great difficulty for
C 1 users’ speech perception. The next section shows the consequences of this difficult
listening for cognitive effort and, in turn, for performance on the visual reaction time
task.

Visual reaction time testing.

Raw scores. Raw score results for the visual reaction time test at SW = +5 dB are
presented in Figure 3-6a. It can be seen that the NH listeners were faster on average
(mean= 363.5 msec) in their visual reaction times than the CI users (mean= 445.9 msec).
But it can also be seen that there was some variability in the reaction times for both

groups, and considerable overlap in the individual scores. Statistical analysis of the raw

91

scores indicated a statistically significant difference between the two groups at

S/N=+5dB (r (23) =-2.442; p<.035).

 

 

700 -
600 -
<1)
E 8500 -
a Cl
”“400 d
3‘: 1:1
300 -
200 i .
NH Cl
lndividualparticipant C] Group mean

Figure 3-6a. Visual reaction time raw scores in both groups when tested at +5dB S/N.

Normalized scores. Normalized scores, presented in Figure 3-6b below, show a
substantially longer visual reaction time (relative to the single-task baseline) for the CI
group (mean=136.8 msec) than for the NH group (mean= 49.83 msec). This difference
was statistically significantly (t (23) =-3.506, p<.002). The slowed responses of the CI
group in this condition provide evidence of increased cognitive resource commitments to

the primary task when listening to speech in this unfavorable acoustical condition.

92

300 -

 

A 250-
0)
ET"? 200-
i8
1— ..
gm 150 [:1
35100- ' ~
:38 50- [j
E _
0‘
-50 .

 

NH Cl

r Individual participant 13 Group mean
Figure 3-6b. Visual reaction time difference scores in both groups in +5dB S/N.

Additional Dual-task Tests

Cochlear implant listeners: testing at S/N = +10 dB.

After completing testing at +5dB S/N, several of the CI participants (n=6) were
tested at +10 dB S/N. It was clear that understanding speech at +5dB S/N was difficult
for CI users as a group, and almost impossible for several of the individual participants.
The goal here was to investigate whether Cl users might be able to perceive speech with
greater accuracy as a group when listening at +10 dB S/N. Also, the cognitive effort
needed to perceive speech at this S/N was assessed.

Speech perception test results.

Raw scores and normalized scores. Figure 3-7a shows the speech perception raw
scores for this test. Figure 3-7b shows the normalized scores. The raw score mean for the
C1 group was 64% correct, considerably higher than the mean for CI users who listened

at +5dB S/N (3 8% correct). This shows that under dual-task test conditions, CI users

93

can perceive speech with reasonable accuracy when the speech is presented in low-to-

mid-level background noise.

100‘

80‘

Speech Test
(% Correct)
0"
O
l

A
O
1

N
O
1

 

 

O

61
’~ Individual participant 13 Group mean

Figure 3-7a. Speech test raw scores for the CI group listening at +10 dB S/N.

The normalized scores mean was much closer to 1.0 for this condition

(mean=0.74) than it had been at +5dB S/N (mean=0.43).

0.6 t
0.4 1

Speech Test
(Proportion, RE Baseline)

0.2 -

 

 

0 1
CI

"1 Individual participant B Group mean
Figure 3-7b. Speech test proportion scores for the CI group listening at +10 dB S/N.

94

IP"

Visual reaction time test results.

Raw scores and normalized scores. Figure 3-83 shows the CI users visual
reaction time raw scores. The raw score mean found here was 451.7 msec, which is very
close to the mean of 445.9 msec reported earlier for the CI users who were tested at
S/N=+5 dB. However, when the raw scores shown in figure 3.8a were converted to
normalized scores based on the baseline performance of the individual subjects who were
tested here, a difference between the present condition and testing at S/N=+5 dB became

evident (see Figure 3.8b).

Reaction Time
(msec)

 

 

200 1
CI

Individual participant ‘3 Group score

Figure 3-8a. Visual reaction time raw scores in the CI group at +10 dB S/N.

The normalized mean score for S/N=+10 dB was 83.3 msec, well below the mean
of 136.8 msec for testing at S/N=+ 5dB. This clearly indicates that Cl users were more
capable of multi-tasking at 10 dB S/N and, correspondingly, that the cognitive effort

needed to perceive speech was measurably reduced here.

95

300 '1

 

 

,1 250 -
Q)
«Ea-752004
1:18 150-
0: i
= 100-
83' D
Ma 50- ..
-50

CI
a. lndividualparticipant [:1 Group mean

Figure 3-8b.Visua1 reaction time proportion score in the CI groups at +10 dB S/N.

Normally hearing listeners: testing at S/N = 0 dB.

As noted above, NH listeners were found to perceive speech with high accuracy
and minimal cognitive effort when tested at S/N = +5 dB. The level of difficulty of
speech listening, therefore, was increased for this new test of NH participants.
Specifically, the S/N was set at 0 dB. Eighteen NH listeners (14 of whom had been
previously tested at S/N = +5 dB) were tested here.

Speech perception test results.

Raw scores and normalized scores. The NH participants’ speech perception raw
scores were more variable for this test than seen previously, and also lower on average
(see Figure 3.9a below). The mean scores was 81% correct, which was a 15% drop
compared to performance at +5dB S/N. A similar pattern was seen for the normalized

scores (Figure 3-9b), which fell to a mean of 0.82 here, compared to 0.98 for S/N=+5 dB.

96

1001
“A 80- D
‘8
88° 40‘
mv
201
0 m i

 

 

NH

Individual participant Cl Group score

Figure 3-9a. Speech test raw scores for the NH group at 0 dB S/N.

 

 

1.2 -
E 1 :
Q)
fig 53 0.8 - U
l— a
-= . 0.6 - ,
3 8
a) E 0.4 "‘
802
‘5‘: ' i
O 1
NH
-., Individual participant 1:] Group mean

Figure 3-9b. Speech test proportion score for the NH groups at 0 dB S/N.

Visual reaction time test results.

Raw scores and normalized scores. The raw score (Figure 3-10a) and normalized
score (Figure 3-10b) visual reaction time results again provide evidence regarding

cognitive effort. The mean raw score for this condition was 392.2 msec. The mean

97

normalized score was 80.9 msec. Both of these means are about 30 msec longer than the
corresponding means seen for testing at +5 dB S/N. This slowing in visual reaction time
performance shows that multi-tasking at 0 dB S/N was measurably more difficult for the
NH group. Clearly, 0 dB S/N posed some challenges and, therefore, the visual reaction

times tended to be slower at 0 dB than at any other condition tested previously.

 

 

700 -
ii) 600 -
5 ..
S if. 500 -
g E 400 - [j
‘3‘ 300 4 _

200 r

NH
Individual participant B Group mean

Figure 3-10a. Visual reaction time raw scores for the NH group tested at 0 dB S/N.

98

300 -
250 -
200 '1
150 -‘
100 -

Reaction Time
(msec, RE Baseline)

 

 

I
NH
., Individual participant [:1 Group mean
Figure 3-10b. Visual reaction time difference score for the NH group at 0 dB S/N.

The CI Group and the NH Group: A Similar Pattern of Results

This section on Additional Dual Task Tests reported on two tests. The first was a
test of CI users conducted at S/N = +10 dB. The second was a test of NH listeners
conducted at S/N = 0 dB. The results of those tests proved to be similar in a number of

ways. To highlight those similarities the results are re-plotted here side-by-side.

Speech perception test results.

The speech perception raw scores are shown in Figure 3-1 la. With the exception
of one outlying CI score, the raw scores for the two groups fell in a similar range. The
mean scores were 81% for the NH group and 64% for the CI group. The mean
computed for the CI group with the outlying score dropped was 75%, which was very

close to the NH group mean.

99

100 ..

 

 

80 4 D
8 ii .0- D
i— .
a: 8
§ U
Q °\° 40 .
m v

20 -

0 , r

NH Cl
4 lndividualparticipant [:1 Group mean

Figure 3-11a. Speech test raw scores for the NH group tested at 0 dB S/N compared to
scores for the CI group tested at +10 dB S/N.

The normalized speech perception scores shown Figure 3-11b were, again with

one exception (the CI outlier), very comparable. The mean scores for the NH and CI

groups were 0.82 and 0.74, respectively.

fl
-— is.)
L I

 

 

11‘
g
’8
E m 0.8
"= I: 0.6 a
8 g’
8' ‘E 0.4 -
8
e 0.2 -
P:
0 r 1
NH CI
. lndividualparticipant Cl Group mean

Figure 3-11b. Speech test normalized scores for the NH group tested at 0 dB S/N
compared to scores for the CI group tested at +10 dB S/N.

100

Statistical comparisons of the speech perception scores are shown in Table 3-6.

There was a significant difference between the raw scores, but no significant difference

between the normalized scores.

Table 3-6. Comparing speech test scores at: 0 dB (NH) and 10 dB (C1).

 

. . Mean Std. Error
Tested Conditions t- value df Significance . Difference
Difference
DT: 0 dB vs. DT: 10 dB 2.38 22 0.03 .18 .07
DT: 0 dB vs. DT: 10 dB
(proportional: RE 0.95 22 0.40 .08 .08

baseline)

Reaction time test results.

Visual reaction time raw scores from the two groups are shown in Figure 3-12a.

The CI group mean (451.7 msec) was about 60 msec higher than the NH mean (392.2

msec), but the ranges of scores for the two groups were very similar. The normalized

scores, shown in Figure 3-12b, were even more alike. They matched closely in both

mean value (CI mean = 83.3 msec, NH mean = 80.9 msec) and range. There was no

significant difference between the group mean results for either the raw reaction times or

the normalized times (see Table 3-7).

101

700 n

600 .1

500 .

400

1

Cl

Reaction Time
(msec)

300 ..

 

 

200 r 1
NH Cl

Individual participant B Group mean

Figure 3-12a. Visual reaction time test raw scores for the NH group tested at 0 dB S/N
compared to scores for the Cl group tested at +10 dB S/N.

 

 

300 -
250 -
73
o L5 200 -
E ii
[— m 150 "'
C
2 if 100
u .
3 8 U U
E 50 -
0 4
—50 . ,
NH Cl
Individual participant C] Group mean

Figure 3-12b. Visual reaction time test normalized scores for the NH group tested at 0 dB
S/N compared to scores for the CI group tested at +10 dB SM.

102

Table 3-7. Comparing R Tscores at: 0 dB (NH) and 10 dB (CI).

 

- - Si nifican Mean Std. Error
Tested Conditions t-value df g _ Difference
ce Difference
DT: 0 dB vs. DT: 10 dB -1.33 22 -19 '59-48 2°
DT: 0 (13 VS. DT: 10 dB _.07 22 .95 2.33 .95

(msec., RE baseline)

 

Conclusion

Results from comparison analysis show that listening to speech at 0 dB S/N for
NH listeners presented similar challenges as listening to speech at +10 dB S/N for the CI
users. These results are important for appreciating the challenges that even small amounts

of background noise may pose for CI users, in everyday life situations.

103

THF’“

CHAPTER 4

Final Thoughts

This study had two parts, a survey and an experiment. This section will discuss
clinical implications, limitations, and suggestions for further research regarding each of

these. A general summary of the study outcome will also be provided.

Clinical Implications

One potential clinical benefit of the present study is that it provides results that
may be useful in counseling individuals with C15 and their spouses and families. When
individuals with Cls perform well in speech perception tests that are conducted in a quiet
and controlled clinical setting it may be assumed by people surrounding them that they
will perform similarly well in real life listening situations. This assumption is a
misunderstanding. Listening to speech in quiet is different from listening in noise, and
listening in open situations where there may be distractions is different from listening in a
controlled environment. It was found here that greater than normal cognitive effort is
often exerted by a C I user when perceiving speech in everyday situations, especially if
competing noise is present.

Results from the current study indicate that Cl users are capable of scoring high in
speech recognition tests when listening in low-to-mid level background noise. However,
when compared to normally-hearing listeners, more cognitive effort was found to be
exerted and more time was found to be needed by C1 users to achieve these results.
Sharing of findings of this kind with the families of individuals with Cls may help them

to develop better understanding of what listening with a CI is like.

104

A specific finding of the experimental study was that the cognitive demands of
listening in moderate background noise could be roughly equated for CI users and NH
listeners if the S/N was worsened by 10 dB for the latter group. This suggests the
possibility of developing a clinical demonstration in which background noise level is
manipulated for normally-hearing listeners to properly model the cognitive challenges
faced by a CI user.

Another clinical implication of this study is that the survey results point up the
potential importance of using a hearing aid in the opposite ear. Those who used a hearing
aid often reported exerting less cognitive effort than those who did not use one, especially
when listening in situations that posed special challenges. These results concur with
findings from other studies that have documented a binaural hearing advantage. Using a
CI combined with a hearing aid in the opposite ear can lead to improved sound
localization, improved speech discrimination in noise, and improved overall sound
quality (e.g., Tange, Grolman, & Dreschler, 2009).

Finally, clinicians should be aware of the importance of top-down processing
skills for speech perception by all CI users, and perhaps especially by older users. Top-
down perception refers to perception that relies on prior knowledge and inference fi'om
context. An individual’s ability to top-down process may affect both the period of
acclimation to a new CI and also the long-term benefit that may be derived from
sustained C1 use. The present survey results suggest that top-down processing skills are
generally strong among older CI users, very possibly stronger than those of younger
users. This agrees with other CI studies that have found that older age is not a factor that

interferes with post-implant results.

105

Study Limitations

All studies have limitations. The present study has certain limitations which need
to be taken into account when considering its contributions. Addressing some of these
limitations can be seen as fruitful avenues for fiiture research. This section will discuss
separately the limitations of the survey and the experiment.

Survey.

The first limitation in the survey was the juxtaposition of some of its questions.
For example, three questions regarding listening in various levels of background noise
were presented in succession (Question 6a-6c). This was done to make the survey as
straightforward as possible for participants. It is possible that ordering the questions in
this way may have also introduced some element of bias to the responses.

Another limitation in the survey is that it was retrospective. The biggest problem
in a retrospective study is that some of the information may be hard to get because it
relies on participants recalling things that may have happened years ago. The survey
participants had worn their Cl’s an average of 2 years, and they were asked to rate
experiences as for back in time as pre-implant. Memory is a selective thing, and it can
introduce biases.

A third limitation of the survey is that data were collected here from a single
cochlear implant program. The participants’ responses could conceivably have been
affected in some way by their specific experiences as patients of the University of
Michigan Cochlear Implant Program. In a future study, collecting data from individuals
associated with multiple clinics could provide answers from a more heterogeneous group

and more representatively sample the experiences of CI users.

106

The last survey limitation concerns financial incentives. No incentives were
offered to the participants in this study. The participant response rate was 138/ 260
(53%). Although this rate was substantial, offering incentives could possibly have
increased it.

Experiment.

The first limitation in the experiment was the age range of the participants. The
majority of the participants were older adults (mean age = 51.8 years). Results might
have been in some way biased because of this age effect. Cognitive effort hypotheses
postulate age differences in the demands of memory processing to be one factor
underlying age-related deficits (Benjamin, 2005; Macht & Buschke, 1983). Memory
processing in particular may demand more cognitive'effort from older adults (Craik &
Simon, 1980; Naveh—Benjamin, 2005).

The second limitation was related to the CI devices worn by the participants. All
participants used a multichannel implant; however their CI’s differed in their processing
strategies and manufacturers. The specifics of the devices themselves were not compared
in this study. Possibly, differences in the processing strategies or the manufacturers

could have accounted for some portion of the variability among CI users seen here.

Suggestions for Further Research

Measures of cognitive effort have previously been found to provide a useful
complement to audiometric tests when assessing listeners with hearing loss in the mild to
severe range. The present study demonstrates that studies of cognitive effort may also be

informative regarding individuals with profound hearing loss who are CI users. Various

107

investigations could be extended from the present study. Future studies could address
three recommended topics: the cognitive effort requirements of bilateral CI users, the
cognitive effort requirements of children using CIs, and the benefits of using a hearing

aid in the non-implanted ear.

Summary

A survey of was conducted to answer the following question: Do postlingually
deafened CI users commonly report having to exert cognitive effort in order to perceive
speech when listening in everyday situations? It was found that they do. They
specifically reported having to exert special effort whenever background noise is present,
when listening to an unfamiliar topic of conversation or an unfamiliar talker (or talkers),
and when listening to a non-native speaker or a rapidly speaking talker. Cognitive effort
was generally reported to increase when listening in situations where any of these
challenges became extreme (e.g., when the background noise level was high). Finally,
participants reported that the cognitive effort needed to perceive speech declined rapidly
over the first few days and weeks post-implant, and declined more gradually for several
months thereafter.

Comparisons of the responses provided by different subgroups (e. g., males vs.
females, younger vs. older CI users) provided evidence of the following significant
trends. When listening to speech in good conditions (e. g., low noise, single familiar
talker) females reported exerting less effort than males, younger participants reported
exerting less effort than older participants, and progressive onset participants reported

exerting less effort than sudden onset participants. For listening under the most

108

 

challenging conditions, CI users who regularly wear a hearing aid in the opposite ear
reported exerting significantly less effort than those who do not use a hearing aid.

An experiment was conducted to determine whether postlingually-deafened adults
using a CI showed evidence of expending extra cognitive effort compared to that of
normally-hearing (NH) listeners when perceiving speech in quiet and in various levels of
background noise. Findings indicated that under favorable conditions (listening in quiet,
listening at a signal-to-noise ratio of +15 dB), CI users were able to maintain substantial
accuracy in their speech perception performance, nearly as high as that of NH listeners.
However, the cognitive effort required to perceive speech was substantially greater for
the CI group than the NH group. As the background noise level increased, CI users made
an increasing number of speech perception errors and, at the same time, showed evidence
of a far greater than normal commitment of cognitive effort to speech processing.

Finally, it was found that a group of NH listeners tested at S/N = 0 dB and a group of C1
users tested at S/N = +10 dB had similar overall results regarding both speech perception
and cognitive effort. Hence, the two groups could be roughly equated by a 10 dB
difference in S/N.

Findings from the present study indicate that analyzing cognitive effort is both
feasible and useful when investigating CI users. Results fiom this type of measurement
can give a clearer idea about how CI users may fiinction in everyday multi-tasking

listening situations, rather than just in the testing booth.

109

THF’“

APPENDIX A

Survey

Copy of the survey. Descriptive statistics regarding the participants’ responses to each
question are provided. Abbreviations: N= number of participants; NR= number who gave
no response; NO= number who chose “no opinion” option; Mn=mean; Sd= standard
deviation.

110

Listening Effort in Late-Deafened Cochlear Implanted Adults:

A Research Study

Listening Effort

Cochlear

[111 plan ts

 

Please return your completed questionnaire in the enclosed en velops to:

Rana Alkhamra
Oyer Speech-Language-Hearing Clinic
101 Wilson Road
Michigan State University
Eat Lansing Michigan, MI 48824-1220
We would like to survey you regarding how effortful (hard) / effortless (easy) you
find listening to be in different situations. Please, answer the following questions:

111

Listening Effort in Late-Deafened Cochlear Implanted Adults :
A_ Research Study

We would like to survey you regarding how hard/ effortful you find listening
to be in different situations. Therefore, please answer the following questions:

START HERE:

Listening effort using a cochlear implant

1. BEFORE getting your cochlear implant, was listening to speech effortful or

 

 

effortless?
C Strongly effortful
E Somewhat effortful N=124
:‘Neither effortful nor effortless 1133i?
[:1 Somewhat effortless Mn=l .26
D Strongly effortless Sd=0.78
:3 No opinion

 

 

 

2. During the first week AFTER getting your cochlear implant, did listening to
speech become more or less effortful than it was before the implant?

7: Strongly more effortful

 

:3 Somewhat more effortfiil 31:59
CNeither more nor less effortful NO=8

Cl Somewhat less effortful Mn=3.67
D Strongly less effortful Sd=1'20

 

 

 

D No opinion

3. During the first three to six months AFTER getting your cochlear implant, did
listening to speech become more or less effortful than it was before the implant?

 

 

 

U Strongly more effortful N=125
D Somewhat more effortful NR=O
DNeither more nor less effortfiJl :10?

m n—4.4O
L4 Somewhat less effortful Sd=0.85
I] Strongly less effortful E

 

E No opinion

112

4. After using your cochlear implant for six months or longer, did listening to
speech become more or less effortful than it was before the implant?

C Strongly more effortful
D Somewhat more effortful

UNeither more nor less effortful

'3 Somewhat less effortful
E Strongly less effortful C]
I No opinion

 

N=124
NR=O
NO=3
Mn=4.62
Sd=0.77

 

 

 

5. Did speech listening become more or less effortful when you became more
accustomed to hearing speech with your cochlear implant?

C Strongly more effortfiil
E Somewhat more effortful

ENeither more nor less effortful

D Somewhat less effortful
D Strongly less effortfulC]
C No opinion

 

N=125
NR=0
NO=2
Mn=4.49
Sd=0.92

 

 

 

6. Using your cochlear implant, how effortful or effortless do you describe listening
in the following situations: Note that numbers I through 6 correspond to the

following

a.

b.

C.

l= Strongly effortful
2= Somewhat effortful
3= Neither effortful nor effortless

(Please, mark ONE boxfor each statement)

1
In quiet settings [3
N=123
In settings with a E
low level of noise N=125
In settings with a D
medium level of noise N=126

113

2
D

NR= O

U
Nr==0

D
NR=O

4= Somewhat effortless
5= Strongly effortless
6= No opinion

3 4 5 6
U D [:1 C]

NO= 4 Mn=4.44 Sd= l .04

C] D D 3
NO=2 Mn=3.87 Sd=l.l3

D D E] D
NO=1 Mn=2.94 Sd=l.25

 

'lr

Continued/ Question # 6

1= Strongly effortful
2= Somewhat effortful
3= Neither effortful nor effortless

1

d. In settings with a CI
high level of norse N=125

e. When background noise is C
environmental sounds (e. g.,
car horn, N=126
printer, telephone ringing)

f. When background noise is D
speech sounds (e.g., in a N=125
restaurant,

in family gatherings)

g. When familiar with the D

conversation topic
N=124
h. When unfamiliar with the C

conversation topic
N= I 26

i. When in a one-to-one conversation I
with a familiar speaker

N=123
j. When in a one-to-one conversation ‘3
With an unfamiliar speaker N=124

114

(Numbers 1 through 6 correspond to the following)

4= Somewhat effortless
5= Strongly effortless
6= No opinion

[1

NR=O

L]

NR=0

E
NR=]

Ci

NR=2

:1

NR=0

m
L_J

NR=O

Ci

NR=2

Ci

NO=2

NO=l

f—“l

L_J

NO=l

C]

NO= l

'r—‘i
L4

NO=I

NO=4

[I]
NO=1

4
m
L_i
Mn=! .74
F7
Mn=2.21
Mn=2.0
B
Mn=3.4l
S
Mn=2.40
C]
Mn=4.21
C}
Mn=3.38

5 6
I] E]
Sd=l.l3
D S
sa=1.15
E- D
Sd=l.01
D E
Sd=1.26
D C]
Sd=l.22
is r?
Sd=l.l3
D D
Sd=].21

 

Continued! Question # 6

l= Strongly effortful
2= Somewhat effortful
3= Neither effortful nor effortless

k. When in a group conversation
with familiar speakers

1. When in a group conversation
with unfamiliar speakers

m. When listening to someone
who speaks slower than
normal

n. When listening to someone
who speaks faster than normal

0. When listening to someone using
English as a first language

p. When listening to someone who
speaks with a foreign accent

N=126

N=124

[J

N=125

E1

N=122

N=117

115

4= Somewhat effortless

5= Strongly effortless

6= No opinion

NR=O

NR=]

[]

NR=O

NO=1

1:1

NO=1

F

“—1

NO=2

D

NO= l

D

NO=10

Mn=2.71

C]

Mn=2. 16

1'41
L-J

Mn=4.15

D

1:1

[:1

(Numbers I through 6 correspond to the following)

1—1
L_..i

Sd=1.28

D
Sd=] .32

D

Sd=1.14

S

Mn=2.44 Sd=1.23

13

Mn=3.83

1:1

Mn=2.22

C1
Sd= l .25

E1

Sd=l.14

Continued/ Question # 6 (Numbers 1 through 6 correspond to the following)

l= Strongly effortful
2= Somewhat effortful

3= Neither effortful nor effortless 6= No opinion

1 2

q. When listening to someone using : C1

your native language N=1 l7 NR=2

r. When listening to someone familiar D D

on the telephone
N=l 14 NR=4

r—w

3. When listening to someone unfamiliar B J

on the telephone
N=l 12 NR=5

3

.

l
._a

NO=8

D

NO=9

[—1
L_J

NO=10

W

4= Somewhat effortless
5= Strongly effortless

5 6

[:1

C5

Mn=3.82 Sd=1.32

1:1

Mn=3.27

m
L__;

Mn=2.24

1:]

C1

1:]
Sd= 1 .49
[J
Sd= 1 .29

7. How effortful or effortless is identifying a topic in a conversation when you are
listening and working on something else that demands your attention (e.g.,

taking notes in a meeting, driving a car)?

I Strongly effortful

:1 Somewhat effortful

UNeither effortful nor effortless
[3 Somewhat effortless

:1 Strongly effortless

E No opinion

 

 

N=122
NR=]
NO=4
Mn=2.09
Sd=] .08

 

 

8. How often do you ask speakers to rephrase or repeat their original utterance
when you are listening and working on something else that demands your
attention (e.g., taking notes in a meeting, driving a car)?

:1 All the time

[:1 Most of the time
D Some of the time
D Rarely

:1 Never

116

 

 

N=127
NR=O
NO=O
Mn=2.83
Sd=0.70

 

 

 

9. Do you wear a hearing aid in the ear opposite to the implanted ear? -> (If never,
go to question 11)

 

 

 

 

[1 All the time N=124

. NR=2
[:1 Most of the time NO=1
D Some ofthe time Mn=4.29
DRarely Sd=l.45
:1 Never

10. Do you find that using a hearing aid in conjunction with your cochlear implant
makes it more difficult or more easy to understand speech?

 

 

 

C] Strongly more easy N=30

[:1 Somewhat more easy NR=13
ClNeither more easy nor more difficult NO=84
[3 Somewhat more difficult Mn=2'80
C1 Strongly more difficult Sd=] '49

 

C- No opinion

11. Does lipreading become especially important when you are trying to understand
speech in QUIET or in NOISY listening situations?

 

U Strongly important in quiet A consistency
D Somewhat important in quiet ”3.5“?“- The

. . , , , , majority gave 2
UNelther important in qulet nor 1n nmse different answers.
[:1 Somewhat important in noise Information is in

questions are 12&13

 

 

[:1 Strongly important in noise
:No opinion

 

12. Do you use lipreading to enhance speech understanding in QUIET listening

 

 

 

 

situations?
SAll the time N=‘26
C1 Most of the time 3:2)
1:1 Some of the time Mn=2.98
E] Rarely Sd=l . 15
U Never

117

 

13. Do you use lipreading to enhance speech understanding in NOISY listening

 

 

 

 

situations?
:All the time ”‘1”
:JM t f th t' “=0
OS 0 6 line NO=O
U Some of the time Mn=2.27
{3 Rarely Sd= ] '32
:Never

14. Do you find that lipreading increases or decreases listening effort?

 

 

 

D Strongly increases N=112
1:] Somewhat increases NR=!
UNeither increases nor decreases N0=14
[:1 Somewhat decreases 23:12,?
C] Strongly decreases

 

:- No opinion

15. Please indicate how often, on average, you spend most of your time
communicating with family and friends

 

f7 All the time

:7 Most of the time N=127

C Some of the time NR=0

D Rarely NO=0

UN Mn= l .91
ever Sd=0.8l

 

 

 

16. Please indicate how often, on average, you spend most of your time
communicating with unfamiliar people (e.g., building genitor, grocery worker,
restaurant waiter)?

 

 

 

EAll the time 2:216
r1 ' - -

in Most of the time NO=0

:1 Some of the time Mn=2.7O
C1 Rarely Sd=0.77

 

: Never

118

TP‘

Hearing History

17. What is the cause of your hearing loss (if known)? ......................

18. Was your hearing loss

C1 Sudden
DBecame worse with time

19. At what age was your hearing loss identified?

A Right ear
B Left ear

 

 

20. At what age did your hearing loss become severe-to-profound?

 

A Rightear
BLefiear

 

21. At what age did you first use a hearing aid?

 

A Right ear
B Lefi ear

 

22. Before getting your cochlear implant, how often did you use a hearing aid in the

IMPLANT ED EAR?

1:] All the time

C1 Most of the time
C1 Some of the time
D Rarely

CINever

119

 

 

 

 

 

 

 

 

I‘m

 

 

 

 

 

 

 

 

N=127
NR=O
NO=0
Mn=] .79
Sd=0.4l
A B
N=1 l8 N=l 18
NR=7 NR=
NO=2 NO=2
Mn=25.69 =26.“
Sd=19.48 “:19“
A B
N=l l6 N=1 13
NR=9 NR=10
NO=2 NO=4
Mn=40.31 Mn=40.27
Sd=21.03 Sd=l9.53
A B
N=104 N=93
NR=5 =7
NO=18 NO=27
Mn=31.54 Mn=30.01
Sd=l8.95 sd=19.10

 

 

 

 

 

 

N=125
NR=2
NO=0
=2.12
Sd=] .66

 

 

 

 

23. Before getting your cochlear implant, how often did you use a hearing aid in the

ear OPPOSITE to the implanted ear?

3 All the time

C1 Most of the time
13 Some of the time
D Rarely

1: Never

24. Which car is your cochlear implant in?

C1 Right ear
C1 Left ear

25. For how long have you had your cochlear implant?

”S 3- 6 months
3 7-12 months
5 13-18 months

'_1

1.1 19-24 months

13 If more than 2 years, please specify ...................

26. Is your vision

Normal —i (If normal, go to question 29)
UCorrected with lasses

27. If your vision is corrected, do you have more difficulty in

T; Seeing objects/people who are near you
L3 Seeing objects/people who are at a distance
C1 Seeing objects/people who are near or distant from you

120

 

 

N=125
NR=2
NO=0
Mn=2.52
Sd=1.82

 

 

 

N=122
NR=2
NO=3
Mn=1.45
Sd=0.50

 

 

 

N= 1 24
NR=2
NO=1
Mn=4.67
Sd=0.76

 

 

 

N=123
NR=3
NO=1
Mn=] .79
Sd=0.4l

 

 

 

N=78
NR=49
NO=0
Mn=] .96
Sd=0.70

 

 

 

 

 

 

 

Demographics

 

28. What is your age?

29. Is English your first language?

 

N=121
NR=6
Mn=64
Sd= l 6.27

 

 

13 Yes,
13No, please specify your first language ....................

 

N=125
NR=2
Mn=1.03
Sd= 10.96

 

30. What is the highest level of education that you have completed?

B Elementary school
CUunior high school
13 High school / vocational high school

 

 

 

 

13 Some college N=125

U4 years college NR=2

13 More than 4 years of college Mn=4-09
Sd=10.75

31. Please check the best description of your employment status.

C1 Employed

C1 Unemployed N=125

C1 Own your own business NR=2

C1 Student Mn=3.70

C1Retired Sd=l.86

UOther, please specify ............................

 

 

121

 

 

 

 

APPENDIX B

BKB Test Lists:

BKB Test: Five different stimulus test lists of 32 sentences each are listed below. An
individual BKB sentence could appear in one list only. These lists were randomly
assigned to the single-task and dual-task test conditions as needed.

122

List_A

Sentence List 1

The clown had a funny face.
The car engine’s running.
She cut with her knife.
Children like strawberries.
The house had nine rooms.
They’re buying some bread.

The green tomatoes are small.
He played with his train.

The postman shut the gate.
They’re looking at the clock.

The bag bumps on the ground.

The boy did a handstand.
A cat sits on the bed.
The lorry carried fruit.
The rain came down.
The ice cream was pink.

Sentence List 2

The Iadder’s near the door.
They had a lovely day.

The ball went into the goal.
The old gloves are dirty.
He cut his finger.

The thin dog was hungry.
The boy knew the game.
Snow falls at Christmas.
She’s taking her coat.

The police chased the car.
A mouse ran down the hole.
The lady’s making a toy.
Some sticks were under the
tree.

The little baby sleeps.
They’re watching the train.
The school finished early.

BKB Test Lists

List_B
Sentence List 3

The glass bowl broke.

The dog played with a stick.
The kettle’s quite hot.

The farmer keeps a bull.
They say some silly things.
The lady wore a coat.

The children are walking
home.

He needed his holiday.
The milk came in a bottle.
The man cleaned his shoes.
They ate the lemon jelly.
The boy’s running away.
Father looked at the book.
She drinks from her cup.
The room’s getting cold.

A girl kicked the table.

Sentence List 4

The wife helped her husband.
The machine was quite noisy.

The old man worries.

A boy ran down the path.
The house had a nice garden.
She spoke to her son.
They’re crossing the street.
Lemons grow on trees.

He found his brother.

Some animals sleep on straw.

The jam jar was full.
They’re kneeling down.

The girl lost her doll.

The cook’s making a cake.
The child grabs the toy.
The mud stuck on his shoe.

123

List_C

Sentence List 5

The bath towel was wet.

The matches lie on the shelf.
They’re running past the house.
The train had a bad crash.

The kitchen sink‘s empty.

A boy fell from the window.

She used her spoon. [
The park’s near the road.

The cook cut some onions.
The dog made an angry noise.
He’s washing his face. .

 

Somebody took the money. 1
The light went out. I
They wanted some potatoes. I‘
The naughty girl’s shouting.

The cold milk’s in ajug.

Sentence List 6

The paint dripped on the
ground.

The mother stirs the tea.
They laughed at his story.
Men wear long trousers.
The small boy was asleep.
The lady goes to the shop.
The sun melted the snow.
The father’s coming home.
She had her pocket money.
The lorry drove up the road.
He’s bringing his raincoat.
A sharp knife’s dangerous.

They took some food.

The clever girls are reading.
The broom stood in the comer.
The woman tidied her house.

List_D

Sentence List 7

The children dropped the bag.

The dog came back.

The floor looked clean.

She found her purse.

The fruit lies on the ground.
Mother fetches a saucepan.
They washed in cold water.
The young people are
dancing.

The bus went early.

They had two empty bottles.

A ball’s bouncing along.
The father forgot the bread.
The girl has a picture book.
The orange was quite sweet.
He’s holding his nose.

The new road’s on the map.

Sentence List 8

The boy forgot his book.

A friend came for lunch.
The match boxes are empty.
He climbed his ladder.

The family bought a house.
The jug stood on the shelf.
The ball broke the window.
They’re shopping for cheese.
The pond water’s dirty.
They heard a funny noise.
Police are clearing the road.

The bus stopped suddenly.

She writes to her brother.
The footballer lost a boot.
The three girls are listening.
The coat lies on a chair.

List_E
Sentence List 9
The book tells a story.
The young boy lefi home.
They’re climbing the tree.

She stood near her window.

The table has three legs.
A letter fell on the mat.
The five men are working.

The shoes were very dirty.
He listens to his father.
They went on holiday.

Baby broke his mug.
The lady packed her bag.
The dinner plate’s hot.
The train’s moving fast.

The child drank some milk.

The car hit a wall.

Sentence List 10

A tea towel’s by the sink.
The cleaner used a broom.
She looked in her mirror.
The good boy’s helping.
They followed the path.

Trial List
Sentence List 19
The family like fish
Sugar’s very sweet
The baby lay on a rug
The washing machine broke.
They’re clearing the table.
The cleaner swept the floor.
The grocer cells butter.
The bath water was warm.

He’s reaching for his spoon.
She hurt her hand.

The milkman drives a small
van.
The boy slipped on the stairs.

They’re staying for supper.
The girl held a mirror.

The cup stood on a saucer.
The cows went to market.

 

The kitchen clock was wrong.
Someone’s crossing the road.
The postman brings a letter.
The dog jumped on the chair.
They’re cycling along.

He broke his leg.

The milk was by the front
doon

The shirts hang in the
cupboard.

The ground was too hard.
The buckets hold water.

The chicken laid some eggs.

124

 

REFERENCES

Anderson, 1., Baumgartner, W.D., Boheim, K., Nahler, A., Arnoldner, C., & D'Haese, P.
(2006). Telephone use: what benefit do cochlear implant users receive? International
Journal of Audiology, 45(8), 446-53.

Allum, D. (1996). Basics of cochlear implant systems. In Allum D. (Ed.), Cochlear
implant rehabilitation in child and adults. San Diego, California: Singular Publishing
Group, Inc.

Baddeley, A. (1986). Working memory. Oxford: Clarendon.

Beijen, J .W, Mylanus, E.A., Leeuw. A.R., & Snik, A.F, (2008). Should a hearing aid in
the contralateral ear be recommended for children with a unilateral cochlear implant? The
Annals of Otolology Rhinology and Laryngology.l 17(6), 397-403.

Bench, 1., Kowal, A., & Bamford, J. (1979). The BKB (Bamford-Kowal-Bench) sentence
lists for partially-hearing children. British Journal of Audiology, 13(3), 108-12.

 

Benjamin, M. (2005). Divided attention in younger and older adults: Effects of strategy
and relatedness on memory performance and secondary task costs. Journal of
experimental psychology learning memory and cognition, 31, 520.

Bento, R. F ., De Brito Neto, R. V., Castilho, A. M., Schmidt Goffi Gomez, M.V, Giorgi
Sant'anna, S. B., Guedes, M. C., et al. (2005). Abstract Psychoacoustic dynamic range
and cochlear implant speech-perception performance in Nucleus 22 users. Cochlear
Implants International, 6(Suppl.l), 31-34.

Blarney, P., Amdt, P., Bergeron F, Bredberg, G., Brimacombe, 1., Facer, G., et a1. (1996).
Factors affecting auditory performance of postlinguistically deaf adults using cochlear
implants. Audiology and Neurootology, 1(5), 293-306.

Bourque, L.B., & Fielder, ER (1995). How to conduct self-administered and mail
surveys. Thousand Oaks, CA: Sage.

Bowling, A. (1997). Research methods in health. Open University Press, Buckingham.
Bradlow, A.R., & Pisoni, D.B.(1999).Recognition of spoken words by native and non-
native listeners: Talker—,listener-,and item- related factors. Journal of the Acoustical
Society of America, 106, 2074—2085.

Burns, K., BScPharm, M.D., Kho , M.E., Meade , M.O., Adhikari, N., et al. (2008). A

guide for the design and conduct of self-administered surveys of clinicians. Canadian
Medical Association Journal, 179(3), 245-253.

125

Cainer, K. E., James, C., & Rajan, R. (2008). Learning speech-in-noise discrimination in
adult humans. Hearing Research, 238, 155—164.

Candidacy criteria. Retrieved February 12, 2010, from
http://www.advancedbionics.com/For_Professionals/Audiology_Support/Candidacy_Crit
eria.cfm?langid=1

Carhart, R., Tillman, T. W., and Greetis, E. S. (1969). Perceptual masking in multiple
sound backgrounds. Journal of the Acoustical Society of America, 45, 694—703.

Carpenter, P.A., Miyaki, A., & Just, MA. (1994). Working memory constraints in
comprehension: evidence from individual differences, aphasia, and aging. In:
M.Gernsbacher (ed.). Handbook of psycholinguistics. San Diego, CA: Academic express.

Chang, Y.P., Fu, Q]. (2006). Effects of talker variability on vowel recognition in
cochlear implants. Journal of Speech Language and Hearing Research, 49(6), 1331-41.

Choi, S. (2005). The effect of compression on speech perception as reflected by attention
and intelligibility measures (Doctoral dissertation). Retrieved from Dissertations and
Theses database. The University of Nebraska — Lincoln.

Cohen NL. (2004). Cochlear implant candidacy and surgical considerations. Audiology
and Neurootology, 9(4), 197-202.

Collins, D. (2003). Pretesting survey instruments: An overview of cognitive methods.
Quality of Life Research, 12, 229—238.

Copeland, B.J., Pillsbury, H.C. 3rd. (2004). Cochlear implantation for the treatment of
deafness. Annual Review of Medicine, 55,157-67.

Craik, F. I. M., & Simon, E. (1980). The Roles of attention and depth of processing. In L.
W. Poon, J. L. F ozard, L. Cermak, D. Arenberg, & L. W. Thompson (Eds), New
Directions in memory and aging: Proceedings of the George T alland memorial
conference. Erlbaum, Hillsdale, NJ.

Cray, J .W., Allen, R.L., Stuart, A., Hudson, 8., Layman, E., Givens, GD. (2004). An
investigation of telephone use among cochlear implant recipients. American Journal of
Audiology, 13(2), 200-12.

Crum, M., & Matkin, N. (1976). Room acoustics: The forgotten variable? Language,
Speech, and Hearing Services in the Schools, 7, 106-110.

Cullington, H.E, & Zeng, PG. (2008). Speech recognition with varying numbers and

types of competing talkers by normal-hearing, cochlear-implant, and implant simulation
subjects. Journal of Acoustic Society of America, 123 (1), 450-61.

126

Dillman, D. (2000). Mail and internet surveys. USA: John Wiley & Sons, Inc.

Dowell, R. C, Dettman, S. J ., Blarney, P. J ., Barker, E. J ., & Clark, G. M. (2002). Speech
perception in children using cochlear implants: prediction of long-term outcomes.
Cochlear Implants International, 3(1), 1-18.

Dowell, R.C., Hollow, R., Winton, E. (2004). Outcomes for cochlear implant users with
significant residual hearing: implications for selection criteria in children.
Archeology and Otolaryngol of Head and Neck Surgery, 130(5), 575-8.

Downs, D. W. (1982). Effects of hearing aid use on speech discrimination and listening
effort. Journal of Speech and Hearing Disorders, 47, 189-193.

Dunn, C. C., Tyler, R. S., & Witt, S. A. (2005). Benefit of wearing a hearing aid on the
unimplanted ear in adult users of a cochlear implant. Journal of Speech, Language, and
Hearing Research, 48, 668—680.

Edwards, P., Roberts, 1., Clarke, M., DiGuiseppi, C., Pratap, S., et al. (2002). Increasing
response rates to postal questionnaires: systematic review. BMJ, 324, 1183-1185.
Estabrooks, W. (1998). Cochlear implants for kids. Washinston, DC: Alexander Graham
Bell Association for the Deaf.

Faber, C., & Grontved, A. (2000). Cochlear implantation and change in quality of life.
Acta Otolatyngology, suppl 543, 151-3.

Fallon, J .B., Irvine, D.R.F., & Shepherd, R.K. (2008). Cochlear implants and brain
plasticity. Hearing Research, 238, 110—117.

Fetterman, B.L., & Domico, E.H..(2002). Speech recognition in background noise of
cochlear implant patients. Otolaiyngology of Head and Neck Surgery, 126(3), 257-63.

Field, A. P., & Hole, G. (2003). How to design and report experiments (pp. 183-191).
Thousand Oaks, CA: Sage Publications Ltd.

Field, S. (2009). Discovering Statistics Using SPSS (Introducing Statistical Methods) 3rd
ed. Thousand Oaks, CA:: Sage Publications

Fowler, F. J. (1995). Improving survey questions: design and evaluation. Thousand Oaks
: Sage Publications, p.147.

Fu, Q.J., & Nogaki, G. (2005). Noise susceptibility of cochlear implant users: the role of

spectral resolution and smearing. Journal of the Association of Research in
Otolaryngology, 6(1), 19-27.

127

Friesen, L.M., Shannon, R.V., Baskent, D., & Wang, X. (2001). Speech recognition in
noise as a function of the number of spectral channels: comparison of acoustic hearing
and cochlear implants. Journal of Acoustic Society of America, 110, 1 150—1163.

F euerstein, J. F. (1992). Monaural versus binaural hearing: Ease of listening, word
recognition, and attentional effort. Ear and Hearing 13(2), 80—86.

Gadea, M., Espert, R., & Chirivella, J. (1997). Dichotic listening: elimination of the right
ear advantage under a dual task procedure. Applied Neuropsychology, 4(3), 171-175.

Gantz, B.J., Woodworth, G.G., Knutson, J .F ., Abbas, P.J., & Tyler, RS. (1993).
Multivariate predictors of audiological success with multichannel cochlear implants.
Annals of Otologv, Rhinology and Laryngologv, I 02, 909-916.

Geers, A.E., Brenner, C., & Davidson, L. (2003). Factors associated with the
development of speech perception skills in children implanted by age five. Ear and
Hearing, 24(1, suppl), 248-358.

Gifford, R.H., Dorman, M.F., McKams, S.A., & Spahr, A.J.(2007). Combined electric
and contralateral acoustic hearing: word and sentence recognition with bimodal hearing.
Utilizing a hearing aid and a cochlear implant: speech perception and localization.
Journal of Speech Language and Hearing Research, 50(4), 835-43.

Gifford, R., Shallop, J ., & Peterson, A. (2008). Speech recognition materials and ceiling
effects: considerations for cochlear implant programs. Audiology and Neuro-otology, 13,
193—205.

Gfeller, K., Turner, C. W., Woodworth, G., Mehr, M., Feam, R., et al. (2002).
Recognition of familiar melodies by adult cochlear implant recipients and normal-hearing
adults. Cochlear Implants International, 3(1), 31—55.

Green, K.M., Bhatt, Y.M., Mawman, D.J., O'Driscoll, M.P., Saeed ,S.R., Ramsden, R.T.,
Green, M.W.(2007). Predictors of audiological outcome following cochlear implantation
in adults. Cochlear Implants International, 8(1), 1-11.

Green, K. M., Julyan, P. J ., Hastings, D. L., & Ramsden, R. T. (2005). Auditory cortical
activation and speech perception in cochlear implant users: effects of implant experience
and duration of deafness. Hearing Research, 205(1-2), 184-192.

Hamzavi, J ., Baumgartner, W. D., Pok, S. M., Franz, P., & Gstoettner, W. (2003).
Variables affecting speech perception in postlingually deaf adults following cochlear
implantation. Acta Otolaryngology, 123(4), 493-498.

128

 

Heberlein, T.A., & Baumgartner, R. (1978). Factors affecting response rates to mailed
questionnaires: a quantitative analysis of the published literature. American Sociological
Review, 43, 447-462.

Henkin, Y., Taitelbaum-Swead, R., Hildesheimer, M., Migirov, L., Kronenberg, J ., &
Kishon-Rabin, L (2008). Is There a Tight Cochlear Implant Advantage? Otology &
Neurotology, 2, 489-494.

Henry, B.A., & Turner, CW. (2003). The resolution of complex spectral patterns by
cochlear implant and normal-hearing listeners. Journal of Acoustics Society of America,
113, 2861—2873.

Hicks, C. B., & Tharpe, A. M. (2002). Listening effort and fatigue in school-age children
with and without hearing loss. Journal of Speech, Language, and Hearing Research, 45,
573-5 84.

Holden, L. K., Skinner, M. W., Holden, T. A., & Binzer, S. M. (1995). Comparison of the
multipeak and spectral peak speech coding strategies of the Nucleus 22—channe1 cochlear
implant system. American Journal of Audiology, 4, 49-54.

Hygge, S., Ronnberg, J ., Larsby, B., & Arlinger, S. (1992). Normal hearing and hearing
impaired subjects ability to just follow conversation in competing speech, reversed
speech, and noise backgrounds, Journal of Speech and Hearing Research, 35, 208—215.

Kahneman, D. (1973). Attention and Effort. Englewood Cliffs, NJ: Prentice- Hall.

Kirk, K.I., Pisoni, D.B., & Miyamoto, RC. (1997). Effects of stimulus variability on
speech perception in listeners with hearing impairment. Journal of Speech Language and
Hearing Research, 40(6), 1395-405.

Kirshner, B., & Guyatt, GA. (1985). Methodological framework for assessing health
indices. Journal of Chronic Disorders, 38, 27-36.

Larsby, B., & Arlinger, S. (1994). Speech recognition and just-follow conversation tasks
for normal hearing and hearing impaired listeners with different maskers, Audiology, 33,
165—176.

Lassaletta, L, Castro, A, Bastarrica, M, Perez-Mora, R, Herran, B, Sanz, L, de Sarria, MJ,
& Gavilan, J. (2008). Musical perception and enjoyment in post-lingual patients with
cochlear implants. Acta Otorrinolaringoloy 59(5), 228-34.

Lewis, H. D., Benignus, V. A., Muller, K. E., Mallott, C. M., and Barton, C. N. (1988).
Babble and random-noise masking of speech in high and low context cue conditions.
Journal of Speech and Hearing Research, 31, 108—114.

129

 

Liu, C. (2009). Effect of bandwidth extension to telephone speech recognition in cochlear
implant users. Journal of Acoustical Society of America, 125(2), 77—83.

Macht, M., & Buschke, H. (1983). Age Differences in Cognitive Effort in Recall.
Journal of Gerontology, 38(6), 695-700.

Martin, F.N., & Clark, J .G. (2003). Introduction to Audiology. USA: Pearson Education,
Inc.

McColl, E., Jacoby, A., Thomas, L., Soutter, J ., Bamford, C., et al. (2001). Design and
use of questionnaires: a review of best practice applicable to surveys of health service
staff and patients. Health Technology Assessment, 5(31), 1- 256.

McFadden, B., & Pittman, A. (2008). Effect of minimal hearing loss on children’s ability
to multitask in quiet and in noise. Language, Speech, and Hearing Services in Schools,
39, 342—351.

Moore, J ., & Teagle, H. (2002). An introduction to cochlear implant technology,
activation, and programming. Language, Speech, and Hearing Services in Schools, 33
(3), 153-61.

Moore, BC. (2003). Coding of sounds in the auditory system and its relevance to signal
processing and coding in cochlear implants. Otolology and Neurotology, 24, 243—254.

Moller, AR. (2006). Hearing, Second Edition: Anatomy, physiology, and disorders of the
auditory system (2nd Ed.). New York: Academic Press.

Mok, M., Grayden, D., Dowel, R. C., & Lawrence, D. (2006). Speech perception for
adults who use hearing aids in conjunction with cochlear implants in opposite ears.
Journal of Speech, Language, and Hearing Research, 49, 338—351.

Morrell, C.H, Gordon-Salant, S., Brant, L.J., Metter, B.J., Klein, L.L., & Fozard, J .L.
(1995). Gender differences in a longitudinal study of age-associated hearing loss
Journal of Acoustical Society of America, 97(2), 1196-1205.

Morley, J. (2001). Aural comprehension instruction: principles and practices. In
Marianne Celce-Murcia (Ed.), Teaching English as a second or foreign language (3rd
ed.). Boston: Heinle & Heinle

Most, T., & Peled, M. (2007). Perception of suprasegmental features of speech by

children with cochlear implants and children with hearing aids. Journal of Deaf Studies
and Deaf Education, 12(3), 350-361.

130

 

Miiller-Deiler, J ., Schmidt, B.J., & Rudert, H. (1995). Effects of noise on speech
discrimination in cochlear implant patients. Annals of Otolology, Rhinology and
Laryngology, 166, 303—306.

Nakash, R.A., Hutton, J .L., Jorstad-Stein, E.C., Gates, 8., & Lamb, SE. (2006).
Maximising response to postal questionnaires: a systematic review of randomised trials in
health research. BMC Medical Research Methodology, 6, 5.

Naveh-Benjamin, M. (2005). Divided attention in younger and older adults: effects of
strategy and relatedness on memory performance and secondary task costs. Journal of
Experimental Psychology Learning Memory and Cognition, 31, 520.

Nelson, P.B., J in, S., Carney, A.E., & Nelson, DA. (2003). Understanding speech in
modulated interference: cochlear implant users and normal-hearing listeners. Journal of
Acoustic Society of America, 113, 961—968.

Nilsson, M., Soli, S. D., & Sullivan, J. A. (1994).Development of the Hearing in Noise
Test for the measurement of speech reception thresholds in quiet and in noise. Journal of
Acoustical Society of America, 95(2), 1085-99.

Niparko, J .K. (2009). Cochlear implants: principles and practices (2nd Ed.). USA:
Lippincott Williams & Wilkins.

Oh, S. H., Kim, C. S., Kang, E. J., Lee, D. S., Lee, H. J., Chang, S. 0., et a1. (2003).
Speech perception after cochlear implantation over a 4-year time period. Acta
Otolaryngology, 123(2), 148-53.

Owens, E., & Kessler, D. (1989). Cochlear implant systems. In Owens E. and Kessler D.
(Eds), Cochlear implants in young deaf children. Boston, College- Hill Press.

Passmore, C., Dobbie, A.E., Parchman, M., & Tysinger, J. (2002). . Guidelines for
constructing a survey. Family Medicine, 34, 281-286.

Pisoni, D. B. (1993). Long-terrn memory in speech perception: Some new findings on
talker variability, speaking rate, and perceptual learning. Speech Communication, 4, 75-
95.

Pisoni, D. B., & Luce, P. A. (1986). Speech perception: Research, theory and the
principal issues. Pattern Recognition by Humans and Machines. Speech Perception, 1, l—
50.

Rakerd, B., Seitz, P. F., & Whearty, M. (1996). Assessing the cognitive demands of
speech listening for people with hearing losses. Ear and Hearing, 1 7(2), 97-106.

Rattray, J ., Jones, M. C. (2007). Essential elements of questionnaire design and
development. Journal of Clinical Nursing, 16(2), 234-243.

131

 

Ruffin, C.V., Tyler, R.S., Witt, S.A., Dunn, C.C., Gantz, B.J., & Rubinstein, J.T. (2007).
Long-terrn performance of Clarion 1.0 cochlear implant users. Laryngoscope, 117(7),
1183-90.

Rauschecker, J .P., & Shannon, R.V. (2002). Sending sound to the brain. Science, 8,
295(5557), 1025-9.

Riss, D., Arnoldner, C., Baumgartner, W.D., Kaider, A., & Hamzavi, J .S. (2008). A new
fine structure speech coding strategy: speech perception at a reduced number of channels.
Otology and Neurotolology, 29(6), 784-8.

Rossiter, 8., Stevens, C., & Walker, G. (2006). Tinnitus and its effect on working
memory and attention. Journal of Speech, Language, and Hearing Research, 49, 150—
160.

Rubinstein, J .T., Parkinson, W.S., Tyler, R.S., & Gantz, 8.1. (1999). Residual speech
recognition and cochlear implant performance: effect of implantation criteria. American
Journal of Otolaryngology, 20, 445—452.

Ruffin, C.V, Tyler, R.S, Witt, S.A., Dunn, C.C, Gantz, B.J, & Rubinstein, J .T. (2007). Long-
terrn performance of Clarion 1.0 cochlear implant users. Laryngoscope, 11 7(7), 1183-90.

Sharma, A., Dorman, M.F., & Spahr, AJ. (2002). A sensitive period for the development
of the central auditory system in children with cochlear implants: implications for age of
implantation.Ear Hear. 2002 Dec;23(6):532-9.

Stickney, G.S., Zeng, F., Litovsky, R., & Assmann, P.(2004). Cochlear implant speech
recognition with speech maskers. Journal of Acoustic Society of America, 1 I 6, 1081—
1091

Stone, DH. (1993). Design a questionnaire. BMJ, 307, 1264-1266.

Sudman, S., & Bradburn, N. (1982). Asking questions: a practical guide to questionnaire
design. San Francisco, CA: Jossey-Bass.

Summerfield, A.Q., & Marshall, DH. (1995). Preoperative predictors of outcomes from
cochlear implantation in adults: performance and quality of life. Annals of Otology,
Rhinology and Laryngology Supplementary, 166, 105—108.

Tadros, S.F., Frisina, S.T., Mapes, F., Kim, SH, Frisina, D.R., et al. (2005). Loss of
Peripheral Right-Ear Advantage in Age-Related Hearing Loss. Audiology and
Neurotology, 10(1), 44-52

132

 

Taina, T., Valimaa, T.T., & Sorri, M.J. (2000). Speech perception and functional benefit
after cochlear implantation: a multicentre survey in Finland. Scandanavian Audiology,
30,112— 118.

Takayanagi, S., Dirks, D.D., & Moshfegh,A.(2002).Lexical and talker effects on word
recognition among native and non-native listeners with normal and impaired hearing.
Journal of Speech,Language,and Hearing Research, 45, 585—597.

Tange, R.A., Grolman, W., & Dreschler, WA. (2009). What to do with the other car after
cochlear implantation. Cochlear Implants International, 10 (1), 19-24.

Tillman, T., Carriart, R., & Olsen, W. (1970). Hearing aid efficiency in a competing
speech situation. Journal of Speech and Hearing Research, 13, 789-811.

Turner, C.W., Gantz B.J., Vidal, C., Behrens, A., & Henry, BA. (2004). Speech
recognition in noise for cochlear implant listeners: benefits of residual acoustic hearing.
Journal of Acoustic Society of America, 115(4), 1729-35.

Turner, C.W., Reiss, L.A., & Gantz, 8.1. (2008). Combined acoustic and electric hearing:
preserving residual acoustic hearing. Hearing Research, 242(1-2), 164-71.

Tye-Murray, N. (2004). Foundations of aural rehabilitation: children, adults and their
family members (2nd Ed.). New York: Dalmar Learning

Tye-Murray, N. & Tyler, R.(1988). A critique of continuous discourse tracking as a test
procedure. Journal of Speech and Hearing Disorders, 53, 226-231.

Tyler, R.S., Parkinson, A.J., Wilson, B.S., Witt, S., Preece, J .P., & Noble, W. (2002).
Patients utilizing a hearing aid and a cochlear implant: speech perception and '
localization. Ear and Hearing, 23(2), 98-105.

Valimaa, T.T., & Sorri, M.J. (2000). Speech perception and functional benefit t after
cochlear implantation: a multicentre survey in Finland. Scandanavian Audiology, 30,112—
118.

Van Dijk, J .E., Van Olphen, A.E., Langereis, M.C., Mens, L.H., &Brokx, J .P.,
Smoorenburg, GP. (1999). Predictors of cochlear implant performance. Audiology, 38,
109—116.

Von Bekesy, G. (1989). Experiments in Hearing. USA: Acoustical Society of America

Wilson, B.S., & Dorman, M.F. (2008). Cochlear implants: a remarkable past and a
brilliant future. Hearing Research, 242(1-2), 3-21.

133

 

Wilson, B.S., Lawson, D.T., Muller, J .M., Tyler, R.S., & Kiefer, J. (2003). Cochlear
implants: some likely next steps. Annual Review of Biomedical Engineering, 5, 207-49.

Wolfe, J ., & Kasulis, H. (2008). Relationships among objective measures and speech
perception in adult users of the HiResolution Bionic Ear. Cochlear Implants
International, 9(2), 70-81.

Zhao, E, Stephens, S.D., Sim, S.W., and Meredith, R. (1997). The use of qualitative
questionnaires in patients having and being considered for cochlear implants. Clinical
Otolaryngology and Allied Sciences, 22(3), 254-9.

Zwolan, TA. (2001). Cochlear implants. In Katz J ., Burkard R., Medwetsky L., (Ed.).
Handbook of Clinical Audiology (pp. 740-57). Lippincott Williams & Wilkins. 46.

134

   

111111111111 111111111111111111111s
3 1293113063 6405