Auditory Research Bulletin

Auditory Research Bulletin
Published by Advanced Bionics® Corporation, Valencia, California, USA.
© 2005. All rights reserved.
Advanced Bionics, the Advanced Bionic logo, CLARION, CII Bionic Ear,
HiResolution, HiRes, the HiRes logo, HiRes 90K, HiFocus, HiFocus Helix,
HiFocus Apex, Auria, Platinum BTE, T-Mic, FireFly, IntelliLink, SoundWave, and
Sound Bursts are trademarks or registered trademarks of Advanced Bionics,
incorporated in the United States and other countries.
Production design and editing assistance for this publication were provided by
ReMarks Editing, Bainbridge Island, Washington. E: [email protected]
Contents
Preface. . . . .v
Foreword . . . .vi
Acknowledgements. . . . .ix
Introduction. . . . .x
List of Research Summaries. . . .xii
Medical & Surgical Studies. . . .19
Objective Measures. . . .41
HiResolution Sound. . . .89
Novel Processing. . .117
Bilateral Cochlear Implants. . .141
Ear-Level System Features. . .155
Patient Assessment & Training. . .169
New Study Initiatives. . .185
Contributing Authors. . .190
Participating Research Centers. . .194
Research Staff Worldwide. . .200
Preface
This biennial report reflects the hard work, dedication and commitment of hundreds of individuals around the globe in the field of auditory science and technology. The research represented in this publication
was made possible by the efforts of dedicated professionals with the goal
of improving the lives of people with hearing impairments and deafness.
On behalf of everyone at Advanced Bionics, I extend our deepest appreciation for the work represented here by the committed men and women
of the auditory community. We are proud to be a part of this community—built on friendships, professional relationships, and mutual
respect that is never limited by geographical or disciplinary boundaries.
To every investigator, researcher, and clinician: We thank you for your immense
efforts that advance the science of cochlear implantation. Your commitment to
improving outcomes helps all patients achieve their maximum hearing potential.
Jim Miller
President, Auditory Division
Advanced Bionics Corporation
v
Foreword
COCHLEAR IMPLANTS: past accomplishments...
In 1957 the first cochlear implant operation was performed by Dr Charles Eyriès at
the University of Paris in a patient who had become totally deaf in both ears after
mastoid surgery. The implantable single channel device was constructed by Professor
André Djourno, and consisted of an induction coil capable of stimulating the cochlear
nerve with sinusoid current of any frequency. When the device was activated three
days later, the patient could discriminate some sounds and understand a few simple
words. Subsequently, Djourno developed a transtympanic needle for promontory
stimulation, a real-time speech analyzer, and a two-channel stimulator, publishing his
findings in a series of papers in the proceedings of the French Society of Biology. Sadly,
his research met an untimely end through lack of funds. A few years later, following
Djourno’s lead, research groups around the world began to build implantable devices,
using the recently available transistors to transmit radio signals through the scalp.
Cochlear implants became more widely available in the 1980s. Since then
they have provided increasing degrees of speech perception for adults with
severely impaired hearing and congenitally deaf children. The population able
to benefit from cochlear implants continues to expand. Smaller, behind-theear processors have been developed, implant reliability has improved, and
speech perception has steadily improved. Speech processing strategies with a
choice of stimulation patterns allow implants to be tailored to individual users.
Today’s relatively sophisticated cochlear prostheses have come a long way from
Djourno’s first implant. In terms of ability to communicate, they provide huge
benefits for thousands of individuals worldwide. Djourno died in 1996, well aware
of the impact that his seminal early work has had in the new science of cochlear
implantation, reflected in the body of work contained in this research bulletin.
John Graham, M.D., FRCS
Royal National Throat, Nose and Ear Hospital
London, UK
vi
....future challenges.
The phenomenal growth of scientific knowledge underlying cochlear implantation in its nearly 50-year history has tracked the explosive clinical growth of
implantation worldwide. This growth has been beyond the wildest dreams of
even the earliest pioneers in the field. While predicting the future of implantation has in the past proven unreliable, it should be readily apparent from
the pages that follow that the roadmap for the future of implantation looks
as remarkable as its history. Given that the original cochlear implants were
developed independently in three continents, it is not surprising that this
roadmap spans the globe. Signal processing strategies that enhance speech
in noise, music, tonal language perception, and binaural listening appear on
the near horizon. Streamlined, remote, and self-programming techniques
for adults as well as improved methods and technologies for programming
young children are all under active investigation. Improved imaging, both
in cadaveric temporal bones and in vivo, combined with telemetered physiologic and electric field data, provide unprecedented information about the
electro-anatomic relationship between electrode arrays and their target
neurons. The clinical benefits of such synergy should soon be substantial.
With its multiple independent current sources, exceedingly fast signal
processing speeds, flexible and investigator-friendly research interfaces,
and telemetry, the HiResolution Bionic Ear System is an exceptional
implanted hardware architecture—favorably positioned to exploit current
and future advances in technology and clinical applications. This bulletin
provides an overview of worldwide research made possible by this enabling
technology—enabling for patients because of the benefits it provides,
and enabling to the global research community by its superb capabilities.
Jay Rubinstein, M.D., Ph.D.
Director, Virginia Merrill Bloedel Hearing Research Center
Professor of Otolaryngology & Bioengineering
University of Washington, Seattle, Washington, USA
vii
Acknowledgements
Approximately 14 years have passed since the first patient was implanted with the
first generation cochlear implant manufactured by Advanced Bionics. At that time,
the CLARION® device represented a major step forward in cochlear implant
technology by implementing two completely different sound processing strategies
(compressed analog and continuous interleaved sampling) in a single processor
using transcutaneous transmission. Since that first device, significant advances
have been made in electrode technology, sound processing, device programming, and miniaturization of the internal and external system components.
Today’s HiResolution® Bionic Ear System consists of the HiRes 90K® internal
device with HiFocus® electrode, the Auria® behind-the-ear sound processor,
and the body-worn Platinum Sound Processor. The system with HiResolution Sound is programmed using the SoundWave™ Professional Suite software.
This publication is a compilation of reports on studies from around the world
with the HiResolution Bionic Ear technology. Included in this work are contributions of surgeons and clinicians who describe the everyday benefits experienced by Bionic Ear users—and from researchers who are exploring new
applications that will benefit Bionic Ear recipients today and in the future.
Advanced Bionics is indebted to these investigators for their efforts and willingness to allow their research to be summarized in this publication. (An
alphabetical index of contributing authors and a list of participating research
centers appear at the end of the bulletin.) Advanced Bionics also is grateful to
members of its Research and Development and Clinical Research departments
for their commitment to furthering cochlear implant technology through collaborative research—and for their assistance in making this publication possible.
We hope you enjoy this biennial edition of the Auditory Research Bulletin, and
we look forward to working with you, our clinical and scientific collaborators,
as we explore further the capabilities of the HiResolution Bionic Ear System.
Mary Joe Osberger, Ph.D.
Global Director, Clinical Research, Auditory Division
Advanced Bionics Corporation
ix
Introduction
The HiResolution Bionic Ear System is a unique and technologically innovative cochlear implant with built-in flexibility and upgradeability. Its flexibility arises from 16 independent circuits that allow each channel to be
powered and programmed separately with any type or combination of
waveforms, stimulation patterns (simultaneous or nonsimultaneous), and
electrode configurations. Its upgradeability lies in the untapped capability of the CII and HiRes 90K implanted electronics. Consequently,
recipients have access to HiResolution Sound today—and to new sound
processing strategies in the future simply through software upgrades.
The Auditory Research Bulletin highlights work conducted with clinical patients
and research subjects. Each section may include highlights of multicenter studies,
investigator-initiated research, or new studies that recently have been initiated.
Most of the clinical studies have been conducted with patients programmed
with HiResolution Sound using the commercial SoundWave Professional
Suite. Other studies have explored additional capabilities of the HiResolution
Bionic Ear System using special research platforms developed by the Research
and Development team at Advanced Bionics. These hardware and software
tools provide researchers a technological platform for conducting innovative research with today’s CII and HiRes 90K users. Moreover, they provide
investigators the opportunity to probe the capabilities of the system and to
develop sound processing algorithms and diagnostic tools that may benefit
HiResolution Bionic Ear users in the future. These research tools include:
• Clarion Research Interface (CRI—versions 1 and 2)
The CRI allows direct, programmable, transcutaneous communication to
the implanted electronics (without requiring a percutaneous connection)
for acute laboratory experiments. Use of the CRI requires engineering and
software programming expertise. The CRI typically is used to explore novel
stimulation patterns and sound coding strategies.
• Bionic Ear Programming System (BEPS)
BEPS is a flexible software tool for programming HiResolution and other
advanced sound processing strategies for the CII or HiRes 90K implants.
It allows the investigator to control pulse width, pulse shape, grounding
configurations, AGC options, electrode grouping, and adjustable filter
coefficients for creating customized stimulation programs.
x
Advanced Bionics® Auditory Research Bulletin 2005
• Bionic Ear Data Collection System (BEDCS)
BEDCS is a software program that controls sequenced stimulus presentation
and data acquisition for the CII and HiRes 90K implants. BEDCS was
developed originally to implement protocols for measuring neural responses.
In its current implementation, BEDCS also includes a psychophysics data
collection engine with sequencing and seeking capabilities, making it suitable
for a variety of psychophysical experiments.
• Sound Processing Algorithm Integrated Development Environment (SPAIDE)
This software was developed at the University of Antwerp as a real-time
platform for conducting advanced sound processing research with the CII
and HiRes 90K implanted electronics.
• Speech Processor Application Framework (SPAF)
This tool is designed to implement new sound processing strategies.
SPAF enables the researcher to develop new strategies on a PC that can
be compiled to run on a sound processor in a simulated mode. Once the
simulated strategy is validated, it then can be run in real time on a sound
processor.
• Electric Field Imaging and Modeling Tool (EFI-M)
The EFI-M software is designed to make highly accurate electric field
measurements and to estimate parameters for creating electrical network
models of current flow within the electrically stimulated cochlea.
Many of these research tools were used in studies that are summarized
in this publication, which mainly appear in the Objective Measures and
Novel Processing sections of this book. Some of the research interfaces also
were used during the early development of HiResolution sound processing.
xi
List of Research Summaries
Medical & Surgical Studies
19
Comparison of HiRes 90K with the HiFocus Ij versus HiFocus Helix Electrodes .......................... 20
Evaluation of Advanced Bionics Cochlear Implant Electrodes
in a Human Temporal Bone Model ......................................................................................... 22
Evaluation of the HiFocus Helix Slim Electrode .......................................................................... 24
Minimally Invasive Surgery with the HiRes 90K .......................................................................... 25
Temporal Bone Results with the HiFocus 4L Electrode ................................................................ 26
High-Resolution Micro-CT Imaging: Innovative Technique for Visualising
Intracochlear Tissues with Inserted Cochlear Implant Electrodes ............................................. 28
Multisection CT as a Valuable Tool in the Postoperative Assessment
of Cochlear Implant Patients ................................................................................................... 30
Statistical Analysis of Cochlear Implant Reliability ...................................................................... 32
Patient Performance with the Advanced Bionics HiRes 90K Device
with HiFocus Helix Electrode ................................................................................................. 34
Audiological Findings Following Cochlear Reimplantation ......................................................... 36
Magnetoencephalographic Study of Auditory Cortical Plasticity
in Postlinguistically Deafened Adults ...................................................................................... 39
Objective Measures
Worldwide Objective Measures Study: European Update ........................................................... 42
Using Objective Tests to Optimize Implant Function and Benefit ................................................ 44
Clinical Use of Objective Measures with Children and Adults Using HiRes ................................ 46
Evaluating HiResolution Sound Processing Based on Objective Measures .................................. 48
Can NRI and ESRTs Be Used to Optimize Program Parameters? .................................................. 50
Relationship Between Objective Measures and HiRes Programming Levels ................................ 52
xii
Advanced Bionics® Auditory Research Bulletin 2005
41
Objective Measures—continued
Setting Upper Programming Levels in Children ........................................................................... 54
Neural Response Imaging (NRI): Correlation with Cochlear Implant Threshold Levels................. 57
Intraoperative and Postoperative Objective Measures: Relationship
to HiResolution Program Parameters ...................................................................................... 58
Neural Response Imaging: Evolution of Thresholds and Their Relationship
to HiResolution M Levels ....................................................................................................... 60
Exploring Changes in the “1st NRI” Over Time .......................................................................... 62
Relationship Between Electrical Stapedial Thresholds and HiRes Programming Levels ................ 64
Predictive Value of Intraoperative ESRTs for Setting Postoperative HiRes M Levels ...................... 66
Electrically Evoked Brainstem and Cortical Potentials: Relationships to Fitting Parameters
and Performance .................................................................................................................... 68
The Relationships of ECAP and EABR Thresholds to Hearing and Speech Perception
in Cochlear Implant Users ...................................................................................................... 70
Use of Single and Multi-Electrode Stimulation to Compare ECAP
and Behavioral Programming Levels ....................................................................................... 72
Effects of Simultaneous Stimulation of Multiple Electrodes
on Behavioral and ECAP Thresholds ....................................................................................... 74
Banded Neural Response Imaging (NRI): Preliminary Results...................................................... 76
Investigation of the Neural Response for a Burst Stimulus ........................................................... 78
Psychophysical versus Physiologic Forward Masking in Cochlear Implants ................................. 80
Channel Interaction in Children.................................................................................................. 82
Unraveling the Electrically Evoked Compound Action Potential .................................................. 84
Relationship Between Electrical Field Models and the Cochlear Anatomy
of Clarion CII Subjects............................................................................................................ 86
—continued on next page—
List of Research Summaries
xiii
List of Research Summaries—continued from previous page
HiResolution Sound
89
Performance of Adults with HiResolution Sound Processing ....................................................... ..90
European HiResolution Multicenter Study: Sound Quality Results .............................................. ..92
The Effect of Number of Channels on HiRes Performance ........................................................... ..94
Perception of Sinusoidal Modulation with High-Frequency Pulse Trains by CII Users .................. ..96
Speech Perception at Varied Stimulation Rates with the Clarion CII Cochlear Implant ................. ..98
Multicentre Music Perception Study in France ............................................................................ 100
Speech Recognition Performance of Children with HiRes ........................................................... 102
Speech Perception Results in Children Using HiResolution and Standard Resolution Modes:
A Two-Year Follow-up Report ................................................................................................. 104
HiRes Performance in Mandarin-Speaking Children ................................................................... 106
Use of the Bark Transform to Measure Vowel Productions in a Child Transitioned
to HiRes Sound Processing ..................................................................................................... 108
HiResolution Programming and Performance in Young Children ................................................. 110
Experience with the HiResolution Bionic Ear in Children............................................................ 112
A “Benchmark” for Performance with the CII HiResolution Cochlear Implant ............................. 114
Novel Processing
Current Steering and Spectral Channels in HiResolution Bionic Ear Users ................................. 118
Discrimination of Single- and Dual-Electrode Stimuli in Clarion CII Users .................................. 120
Virtual Channels: Improvements in Frequency Resolution........................................................... 122
Preliminary Results of Pitch Strength with HiRes 120 Spectral Resolution ................................... 124
Conditioning Pulse Trains in Cochlear Implants: Effects on Loudness Growth ............................. 126
Effects of High-Rate Conditioning Stimuli on Frequency and Intensity Discrimination
in Cochlear Implant Users ...................................................................................................... 128
A Novel Speech Processing Algorithm Based on Phenomenological Models of the Cochlea ....... 130
Chronic Evaluation of a Low-Power Strategy ............................................................................... 133
xiv
Advanced Bionics® Auditory Research Bulletin 2005
117
Novel Processing—continued
SPAIDE: A Real-Time Research Platform for the Clarion CII
and HiRes 90K Cochlear Implants .......................................................................................... 134
Speech Recognition with a Cochlear Implant Using Triphasic Charge-Balanced Pulses ............... 136
Effects of Cochlear Implantation on Auditory Nerve Synapses
in Congenitally Deaf White Cats ............................................................................................ 138
Bilateral Cochlear Implants
141
Bilateral Benefit in Adult Users of the HiRes 90K Bionic Ear System ........................................... 142
Effect of Sound Processing on Bilateral Performance with the CII Cochlear Implant .................... 143
Development of a Direct-Input System to Evaluate Spatial Unmasking and Sound Localization
in Bilateral Implant Users ....................................................................................................... 144
Changes Over Time in the Benefit of Contralateral Amplification
in Unilateral Cochlear Implant Users...................................................................................... 146
Changes in Fusion and Localization Performance When Transitioning
from Monolateral to Bilateral Listening ................................................................................... 148
Bilateral Benefit in Twins Implanted with HiRes 90K Devices Before One Year of Age ................ 149
HiRes Benefit in a Bilaterally Implanted Adult ............................................................................ 150
HiRes versus Conventional Strategy Benefit in a Bilaterally Implanted Adult ............................... 152
Ear-Level System Features
155
The Usefulness of a Pinna Microphone Position for Sound Localization
in Bilateral Cochlear Implant Users ........................................................................................ 156
Comparison of Benefit for the Auria T-Mic, Auria BTE,
and Platinum Headpiece Microphones .................................................................................. 158
Performance of the Auria T-Mic and the Behind-the-Ear Microphone in Noise ............................ 160
Alternate Microphone Position for Enhanced Listening
with a Behind-the-Ear Speech Processor ................................................................................. 162
Pediatric Evaluation of the HiRes Auria Sound Processor ............................................................ 164
Evaluation of an Induction (T-Coil) Module for the Auria Ear-Level Sound Processor ................... 166
—continued on next page—
List of Research Summaries
xv
List of Research Summaries—continued from previous page
Patient Assessment & Training
169
Monitoring Expectations and Perceived Quality of Life Changes
in Adult Cochlear Implant Users............................................................................................. 170
Outcome Measures in Prelinguistically Deafened Adults with Cochlear Implants ....................... 172
Everyday Cochlear Implant Benefits in Prelinguistically Deafened Adults ................................... 173
Self-Reported Benefit in Prelinguistically Deafened Adults .......................................................... 174
Tinnitus Suppression and Cochlear Implants: Study of Patients
with Postlinguistic Hearing Loss ............................................................................................. 176
Schooling and Educational Performance in Children and Adolescents
Wearing Cochlear Implants .................................................................................................... 178
Use of the PRISE in Evaluating Preverbal Development in Infants with Cochlear Implants .......... 180
Teaching Nursery Rhymes with Music to Young Children Using Cochlear Implants..................... 182
New Study Initiatives
Multicenter Study in Asia: Direct-Connect Testing with the CII/HiRes 90K Implants .................... 186
Ongoing Cochlear Implant Studies: Beijing Tongren Hospital, China .......................................... 187
Bilateral Implant Benefit in Adults and Children ......................................................................... 188
Multicenter Study in Colombia: HiRes 90K Benefit in Children .................................................. 188
Radiographic and Behavioral Frequency-to-Place Alignment in Cochlear Implant Subjects......... 189
Development of an Inferior Colliculus Implant ........................................................................... 189
xvi
Advanced Bionics® Auditory Research Bulletin 2005
185
Medical & Surgical
Studies
The HiRes 90K is the most powerful and versatile cochlear implant
available. The package design of the HiRes 90K and its insertion
tools makes it relatively simple to implant, MRI-safe, and small
enough for the youngest of children.
With 16 independent current sources (one for each electrode
contact) under digital control, the HiRes 90K is highly upgradeable
and provides an ideal platform for clinical research. The patented
HiFocus Electrodes feature 16 platinum-iridium contacts intended
to focus stimulation toward the neural elements. Thinner models
of the HiFocus Electrode, designed for preservation of residual
hearing, are under development.
Comparison of HiRes 90K with the HiFocus Ij versus
HiFocus Helix Electrodes
Thomas Lenarz, Prof., M.D., Ph.D.
Carolin Frohne-Büchner, Ph.D.*
Andreas Büchner, Ph.D.
Rolf-Dieter Battmer, Prof., Ph.D.
Medizinische Hochschule Hannover
Hannover, Germany
* also with Advanced Bionics Corporation, Europe
In recent years, Advanced Bionics has responded to
the decreasing age of implantation with a progression of redesigns to the cochlear implant package,
culminating in the HiRes 90K device. The implant
electronics are contained within a hermetically sealed
titanium case with a removable magnet as well as a
telemetry coil that is attached and encased in silastic.
The implant maintains the full capacity of the CII
Bionic Ear® electronics, including update rates of
up to 83,000 Hz and objective telemetric measures.
Most recently, Advanced Bionics has designed a perimodiolar electrode array—the HiFocus Helix™—
to meet a continuing design goal: to position the
contacts close to the modiolus safely and consistently across surgeries. Perimodiolar placement
has been shown to provide several advantages.
These include lower programming levels, reduced
current spread, and possibly reduced channel
interaction. The electrode array contains 16 active
contacts, has a total length of 24.5 mm, and has a
distance between contacts of 0.85 mm. The array
was designed for minimum trauma with an insertion depth of at least one turn. The device comes
packaged on a preloaded insertion tube with
disposable tools (Figure 1), thereby reducing the
time required to prepare the device for insertion.
Figure 1. (A) HiRes 90K device and surgical tool set and (B) a
closeup view of the HiFocus Helix electrode array preloaded
onto the insertion tool.
20
Our temporal bone experiments showed no damage
to the basilar membrane. The majority of contacts
were positioned close to the modiolus, although a
small distance in the mid array region was observed.
In this report, 22 postlinguistically deafened adult
subjects have been implanted with the HiRes
90K, using a minimally invasive approach. Half of
patients received the standard HiFocus Ij design,
and the other half received the HiFocus Helix. The
two groups were matched demographically. The
mean age at implantation in the HiFocus Ij group
Advanced Bionics® Auditory Research Bulletin 2005
was 55 years compared to 47 years in the Helix
group. Mean duration of hearing impairment for
the implanted ear was 29 years versus 25 years,
and duration of deafness was 12 years versus 13
years for the HiFocus Ij and Helix groups, respectively. All subjects have been programmed with the
HiRes strategy and use their implants routinely.
Postoperative performance is summarized in Figure
2. Initially, the performance of the group using the
Helix electrode array was slightly superior. However,
after three months, the results for subjects with
the HiFocus Ij catch up. During the whole study
period, both groups achieve significantly better
performance compared to CII users in conventional processing modes. The better performance of
the HiRes 90K users can be explained by the more
advanced speech coding strategy. The CII users, in
the time period shown in this report, were initially
fitted in standard mode. The patients have all been
converted to HiRes and show significantly improved
performance. Since this report, results of a larger
group of Medizinische Hochschule Hannover
patients are indicating equivalent performance
between patient groups with the two designs so that
the differences observed for the patients reported
here appear to be a result of the small sample size.
“...the HiRes 90K device
is much simpler to implant...
is compatible with the
minimally invasive surgical approach
and hence is particularly well suited
for implantation in young children.”
Figure 2. Postoperative performance of both study groups.
Mean scores for the HSM sentences test in quiet and in noise at
four postoperative test intervals.
Overall, our experience has been that the HiRes
90K device is much simpler to implant than the
previous CII device, is compatible with the minimally invasive surgical approach, and hence is
particularly well suited for implantation in young
children. Intraoperatively, a clearly reduced stapedius
reflex threshold was found for the Helix electrode,
as well as significantly lower programming level
postoperatively (172 CU vs. 155 CU). The averaged
insertion depth of 397° confirmed the target insertion depth for the Helix—although not as deep
as the 413° achieved with the HiFocus Ij design.
Medical & Surgical Studies
21
Evaluation of Advanced Bionics Cochlear Implant
Electrodes in a Human Temporal Bone Model
C. Gary Wright, Ph.D.
Peter S. Roland, M.D.
University of Texas at Southwestern Medical Center, Dallas, TX, USA
“...the thin electrode arrays represent
a significant design improvement
over previous, standard designs.”
Figure 1. Cochlear dissection with segments of the osseous
lamina and basilar membrane removed to show a standard
HiFocus Helix electrode situated in scala tympani. The electrode contacts are numbered from 1-16 in black; the distance
of each contact from the modiolus is represented in millimeters, shown in red. The brownish appearance of the more
apical region of the electrode is the result of osmium staining.
22
This study was performed to assess the insertional properties of Advanced Bionics cochlear
implant electrodes in human cadaveric temporal
bones using cochlear microdissection. The
HiFocus Helix electrode array as well as two
“thin” electrode prototypes (the thin version of
the Helix design and a thin lateral design) were
evaluated during the course of this investigation.
All electrode insertions were performed via standard
cochleostomies placed in glutaraldehyde-fixed human
temporal bones. After insertion, the preparations
were osmium stained and opened in such a way that
the electrodes could be studied and photographed in
situ in scala tympani. As illustrated in Figure 1, this
method offered direct, three-dimensional viewing
of the entire length of the implanted electrode to
determine insertion depth and position with respect
to relevant intracochlear structures. The specimens
also allowed careful evaluation of any trauma that
might have occurred during the insertion process—
including spiral ligament or basilar membrane injury,
fracture of the osseous lamina, or modiolar damage.
In the six specimens implanted with standard
HiFocus Helix electrodes, no perforations of the
basilar membrane or evidence of modiolar injury
were found. However, in two preparations, the electrodes elevated the spiral ligament or the basilar
membrane at the basal cochlear turn. In one case,
the osseous lamina was fractured in the lower
base as a result of slight buckling of the electrode.
Advanced Bionics® Auditory Research Bulletin 2005
Typically, the standard HiFocus Helix electrodes
showed good modiolar proximity in the lower basal
turn, tended to deviate toward the lateral wall of the
scala tympani in the middle-to-upper basal turn, and
then again achieved closer proximity to the modiolus
in the area occupied by the apical electrode contacts.
“In regard to the design goal of
easy, atraumatic
surgical insertions,
these prototype electrodes
appear very promising.”
In contrast, no evidence of electrode-induced
trauma was seen in any of the eight temporal bones
implanted with the thin prototype electrodes, and
full insertions were easily achieved. Four thin Helix
and four thin lateral arrays were included in this
phase of the study. Cochlear dissections revealed that in
each of the specimens the electrode was situated beneath
the intact basilar membrane with no disturbance or
injury of any soft tissue structures. The safety and
ease of insertion of these electrodes was confirmed
during additional trials using preparations in which
the cochleas were opened prior to electrode insertion
to permit direct observation and video recording of
electrode behavior during the insertion process.
Our laboratory findings therefore indicate that
the thin electrode arrays represent a significant design improvement over previous, standard designs. In regard to the design goal
of easy, atraumatic surgical insertions, these
prototype electrodes appear very promising.
Medical & Surgical Studies
23
Evaluation of the HiFocus Helix Slim Electrode
J. Thomas Roland, M.D.
New York University School of Medicine, New York, NY, USA
“...the Helix Slim electrode
provides a smaller diameter
surgical alternative
to the Helix...”
Figure 1. The HiFocus Helix Slim electrode (right) has a smaller
diameter profile than the standard Helix electrode (left).
The HiFocus Helix Slim electrode is a thinner
profile modification of the precurved perimodiolar Helix electrode (Figure 1). Like the Helix,
the Slim Helix electrode array is held straight by
a stylet on the insertion tool. The Slim electrode
is preloaded on the insertion tool and is intended
to achieve optimal placement of the stylet in the
cochlea. The tool defines the depth of the stylet
inside the cochlea to 4.5 mm only—to avoid tissue
damage with an insertion that is too deep or electrode foldback with an insertion that is too shallow.
The surgeon uses an advance-off-tool technique
to achieve a final insertion depth of up to 450°.
Working with the Advanced Bionics engineering
team, we have evaluated the Helix Slim electrode
using fluoroscopy and microscopy methods. We
have assessed (1) real-time insertion dynamics and
electrode positioning in the cochlea, (2) cochleostomy size and location, and (3) the mechanical
and hydraulic forces exerted on the intracochlear
tissues. Gross histology was used to assess electrode position and any trauma imposed upon the
structures in and around the scala tympani. For
the gross histology assessment, cochleas in which
the Helix Slim electrode had been inserted were
embedded in a polymer and cured without heat
in a vacuum. The bone was cut perpendicular to
scala tympani to look especially at the electrode
tip, the cochleostomy, and the proximal basal turn.
The fluoroscopy/microscopy experiments showed
that a 1.0 mm cochleostomy is sufficient for proper
insertion. Otherwise, the electrode can be pinched
by the fins on the insertion tool. The electrode
was inserted easily with minimal outer wall force
generation. Moreover, the insertion tool was easy
to use and to reload. Gross histology revealed reliable and repeatable perimodiolar positioning
(Figure 2), and no osseous fractures were seen.
Figure 2. Perimodiolar positioning of the HiFocus Helix Slim
electrode in scala tympani.
24
In sum, the Helix Slim electrode provides a smaller
diameter surgical alternative to the Helix electrode.
Evaluations of the new electrode design are ongoing.
Advanced Bionics® Auditory Research Bulletin 2005
Minimally Invasive Surgery with the HiRes 90K
Brian P. Perry, M.D.
Kelly Hernandez, M.S.
Charles A. Syms III, M.D., M.B.A.
Susan M. King, M.D.
James E. Olsson, M.D.
Ear Medical Group, San Antonio, TX, USA
This published report retrospectively studied postsurgical safety and device outcomes for HiRes
90K implantation using a minimally invasive (MI)
surgical approach followed by early device fitting
compared to a standard surgical (SS) approach
followed by more traditional initial programming
schedules. A total of 29 patients (8 adults and 21
children) were included in the MI group and 71
patients (55 adults and 16 children) were included in
the SS group. Initial programming occurred within
one week postoperatively in the MI group and within
four-to-six weeks postoperatively in the SS group.
Surgical complications and device stability were
reviewed for both surgical approaches. The clinical
findings in both groups were comparable. For the
MI-approach group, no wound or device-related
complications were seen. Figure 1 shows an incision one day following surgery, typifying the MI
group. Patients experienced no medical difficulties while undergoing initial programming
following an abbreviated postoperative waiting
period, which ranged from one to seven days.
All 29 MI-approach patients wore their devices
continuously upon initial fitting. Analysis of the
clinical programming data indicated electrode
impedances stabilized earlier in the MI patient group
compared to the SS patient group. Psychophysical
M levels were comparable between the two groups.
“These findings indicate that
the HiRes 90K
can be safely implanted
using a minimally invasive
surgical technique...
thereby expediting the benefits delivered
to patients following surgery.
Figure 1. Incision one day following HiRes 90K implantation.
Reference
Perry BP, Hernandez K, Syms CA, King SM, Olsson JE. (2004) HiRes 90K
implantation using a minimally invasive surgical technique. In Miyamoto RT,
ed. Cochlear Implants. Proceedings of the VIII International Cochlear Implant
Conference, Indianapolis, Indiana, USA, 10-13 May, 2004. Amsterdam: Elsevier.
These findings indicate that the HiRes 90K can be
safely implanted using a minimally invasive surgical
technique. Compared to conventional (large incision) approaches, the smaller incision produces
minimal edema, speeds healing time, and allows
stable psychophysical measures—thereby expediting
the benefits delivered to patients following surgery.
Medical & Surgical Studies
25
Temporal Bone Results with the HiFocus 4L Electrode
Johan H.M. Frijns, Prof., M.D., Ph.D.
Jeroen J. Briaire, M.Sc.
Leiden University Medical Center
Leiden, The Netherlands
Andrzej Zarowski, M.D.
Advanced Bionics Corporation, Europe
Berit M. Verbist, M.D.
Leiden University Medical Center
Leiden, The Netherlands
Janusz Kuzma, M.E.
Advanced Bionics Corporation, Valencia, CA, USA
Perimodiolar electrodes are designed to place the
contacts in close proximity to the excitable nerve
fibres, to reduce stimulation levels, to produce more
selective stimulation, and to achieve better speech
perception. Good speech perception outcomes were
reported in a group of subjects implanted with the
CII Bionic Ear implant and the HiFocus electrode
with a partially inserted positioner (Frijns et al, 2002).
Partial insertion of the positioner was intended to
achieve a perimodiolar position for the basal contacts
and a lateral position for the more apical contacts—
in keeping with the findings of a computational
model study (Frijns et al, 2001) that predicted better
outcomes with such a placement. The presence of the
positioner also made a deep insertion depth possible.
Since commercial withdrawal of the positioner
in July 2002, the HiFocus I electrode has been
used alone. A recent study (Van der Beek et al,
submitted) showed that 25 subjects with the
HiFocus I electrode and partially inserted positioner obtained significantly better speech perception than a demographically identical group of 20
subjects with the HiFocus I electrode alone. After
three months of device use, the CVC monosyllabic
word recognition scores were 60% for the with-positioner group and 45% for the no-positioner group.
Figure 1. (A) HiFocus 4L electrode. The electrode contacts are
oriented medially for perimodiolar placement. The electrode
array is thin at the tip and broader at the base. (B) In surgery,
the electrode array is pushed off a stylet that does not reach the
electrode tip.
26
Consequently, a new, single-piece electrode array,
the HiFocus 4L (Figure 1A) has been designed
to achieve desired placement—similar to that
of the HiFocus I with partially inserted positioner—with minimal insertion trauma. The
HiFocus 4L design is suited to all cochlea sizes.
This report presents an analysis of placement and
insertion trauma in temporal bones, using CT scans
and microdissection. Two fresh human temporal
bones were prepared and scanned with a micro-CT
scanner and with a clinical multislice-CT scanner,
using the Leiden routine protocol. A cochleostomy was drilled (1.1 x 1.5 mm), including the
round window. The HiFocus 4L electrode prototype was inserted with a surgical tool provided by
Advanced Bionics® Auditory Research Bulletin 2005
Advanced Bionics. The cochleostomy was sealed
with soft tissue, and the electrode lead was sutured
to the bone. The bones were immediately scanned
with both micro- and multislice-CT scanners—
after which a careful dissection was performed.
The insertions were uneventful, and the surgical
tool was easy to handle with one hand. Microdissection showed full scala tympani insertions and
did not reveal any damage to the internal cochlear
structures. Only a minor lifting of the basilar
membrane without further damage was observed
in one bone at 180° insertion depth. Transillumination of the bones and multislice CT scans
(Figure 2) showed that the contacts were close to
the modiolus in the basal turn and gradually more
lateral in the more apical locations. The insertion
depth was at least one turn in both bones. These
findings were confirmed with the micro-CT scans.
The HiFocus 4L single-piece electrode tested in
this study is intended to achieve the same contact
placements as the HiFocus I with partially inserted
positioner—that is, a perimodiolar position for the
basal contacts and a lateral position for the more
apical contacts. These results have shown that the
design criteria were met, including minimal damage
to the cochlea. Further investigation of the HiFocus
4L design is planned, including histological and
clinical (speech perception) outcomes studies.
“These results have shown that
the design criteria were met, including
minimal damage to the cochlea.”
Figure 2. Multislice CT scans: basal part of the first temporal
bone specimen (left), apical portion of the first temporal bone
showing a 530° insertion depth (middle), and location of all individual contacts in the second temporal bone specimen (right).
References
Frijns JHM, Briare JJ, de Laat JA, Grote JJ. (2002) Initial evaluation of the Clarion CII
cochlear implant: speech perception and neural response imaging (NRI). Ear Hear
23 (3):184-197.
Frijns JHM, Briare JJ, Grote JJ. (2001) The importance of human cochlear anatomy
for the results with modiolus hugging multi-channel cochlear implants. Otol
Neurootol 22 (3):340-349.
Medical & Surgical Studies
27
High-Resolution Micro-CT Imaging: Innovative Technique
for Visualising Intracochlear Tissues
with Inserted Cochlear Implant Electrodes
Andrzej Zarowski, M.D.1,2
Filiep J. Vanpoucke, Dr. Ir.2*
Andrei Postnov, Ph.D.2
Nora De Clerck, Prof., Ph.D.2
Stefaan Peeters, Prof., Dr. Ir.2
Erwin Offeciers, Prof., M.D., Ph.D.1,2
1-Medish Instituut St. Augustinus, Wilrijk, Belgium
2-University of Antwerp, Antwerp, Belgium
* also with Advanced Bionics Corporation, Europe
Evaluation of the damage to the basilar membrane
and to the lateral and medial cochlear walls as a
consequence of electrode array insertion is a challenging problem. Common techniques—such as
histological slicing, grinding and polishing, or
microdissection—have some major drawbacks.
Sample preparation is time-consuming, and there is
a non-negligible risk, for example, that artefact may
be introduced by swelling of the silicone or through
cutting the metallic wires. An ideal evaluation technique would require minimal sample preparation.
Radiological techniques are therefore attractive.
However, present-day clinical Computed Tomography (CT) equipment lacks the resolution required
to visualize fine anatomical details and to assess the
position of the implanted electrode array with respect
to the endocochlear tissues. Recently a highly accurate x-ray micro-tomography scanner (manufacturer
Skyscan) has become available for in vivo scanning of
laboratory animals. Due to limitations in sample size,
the technique is only applicable to temporal bone
studies, but the resolution is excellent (up to 9 µm).
Without an inserted electrode array, the scanner
produces artefact-free images with near-histological quality. The cochlear membranes can be
readily visualized. With an inserted electrode,
the metallic parts produce image distortion. A
protocol was developed that involves scanning the
sample before and after insertion of the electrode.
Subtracting the 3-D reconstructions after alignment provides high-resolution images that enable
assessment of the position of the electrode array.
Importantly, this technique permits assessment not
only with respect to the bony tissues but also to
the soft tissues along the full length of the cochlea.
28
Advanced Bionics® Auditory Research Bulletin 2005
A study was conducted for a temporal bone
implanted with a HiFocus Helix electrode array.
“...this technique permits assessment
not only with respect to the bony tissues
but also to the soft tissues
along the full length of the cochlea.”
The array was designed for:
• atraumatic insertion
• perimodiolar positioning
• a target insertion angle of up to 460 degrees
Micro-CT evaluation of the temporal bone
showed that these design objectives were indeed
met. Figure 1 shows a Micro-CT section through
the human temporal bone prior to implantation with the Helix array. The ability to see, in fine
detail, the anatomical features of the cochlea, such
as the basilar membrane, illustrates how analysis
of insertion trauma was possible. Analysis of
the Micro-CT images indicated that there was
no trauma with the insertion of the Helix array.
To check consistency, microdissection of the inserted
bone was performed, exposing the basilar membrane.
Direct inspection under the operating microscope
confirmed that the basilar membrane was intact,
the electrode array was entirely located in the scala
tympani, the contacts were appropriately orientated
toward the modiolus with a perimodiolar location,
and an insertion depth of over one turn was achieved.
Figure 1. A Micro-CT cross section through a human temporal
bone, showing the detailed anatomical structures available with
this nearly histological resolution.
The ability to obtain the information outlined
above—with minimal preparation or, perhaps
more importantly, disturbance to the temporal
bone—gives hope that this radiographic technique may be very useful as a component in
the evaluation of future electrode array designs.
Medical & Surgical Studies
29
Multisection CT as a Valuable Tool in the
Postoperative Assessment of Cochlear Implant Patients
Berit M. Verbist, M.D.
Johan H. M. Frijns, Prof., M.D., Ph.D.
Jakob Geleijns, Ph.D.
Mark A. van Buchem, Prof., M.D., Ph.D.
Leiden University Medical Center, Leiden, The Netherlands
With improvements in electrode array technology
and programming flexibility for cochlear implant
systems, accurate documentation of the positions for individual electrode contacts within the
implanted cochlea becomes increasingly important.
While conventional radiography can resolve each
electrode contact, three-dimensional details are
not provided. Improved techniques such as phasecontrast or cone-beam radiography may provide
enhanced resolution, but are probably limited
to in vitro application for some time to come.
This paper describes work based around a Multisection
Computed
Tomography
(MSCT)
imaging scanner (Aquilion 4, Toshiba) used
to image patients following cochlear implantation. The following parameters were used:
• 4 x 0.5 mm section thickness
• 0.5 seconds rotation time
• 0.75 pitch factor, 120 kV tube voltage
• 240 mm scan field of view
Reconstruction of nominal 0.5 mm thick images
was made using a 0.3 mm reconstruction increment.
Radiation risk from this scan, including dosage to
the eye lens, is 0.8 mSv, well below annual dosage
from natural sources. Virtually isotrophic voxels
of 0.47 x 0.47 x 0.5 mm allowed reformations in
any plane with virtually no loss in resolution. Both
2-D reformations and 3-D reconstructions were
made including views in the plane of the basal turn
(thus in the plane of the electrode array)— and
orthogonal to this, providing coronal images of
the scala tympani and scala vestibuli. The latter set
of images are particularly suited to examination of
cochlear trauma following electrode array insertion.
Postoperative MSCT images were reviewed in three
cases to investigate whether clinically useful data
could be produced. Case #1 was a 2.5-year-old girl
implanted following meningitis and a diagnosis of
ossifying labyrinthitis. Prior to implantation, some
fibrous and osseous tissue was removed from the
scala tympani before the HiFocus I electrode array
could be inserted. It was only possible to implant
part of the positioner, a spacer intended to move
the electrode contacts to a perimodiolar position.
The MSCT was performed immediately following
surgery, using the same general anaesthetic as used
for the surgery. In Case #2, a progressive loss of
hearing led to deafness. Preimplant CT and MRI
imaging showed normal and patent cochleae. In
this case, the HiFocus electrode array and positioner
could be inserted smoothly. Case #3 presented
with a progressive hearing loss and long-standing
(45 years) duration of deafness. In this case, the
HiFocus I electrode was used alone (no positioner
inserted), and the array tended to be pushed back,
leading to a relatively shallow insertion depth.
For Case #1 the MSCT imaging with Multiplanar
Reconstructions (MPR) and volume-rendered
images revealed some kinking of the apical part of
the electrode array (Figure 1). Consequently, the two
most distal electrode contacts (1 and 2) were excluded
from the child’s sound processor programs. Two
years following implantation, this girl’s oral language
development is within the normal range for her age.
30
Advanced Bionics® Auditory Research Bulletin 2005
“With appropriate scanning parameters,
MSCT can image
individual electrode contacts...”
In Case #2, it was possible to obtain MPR from
the MSCT-imaging data, confirming that the more
basal electrodes were located medially—as intended
with successful placement of the positioner. For
this patient, phoneme perception on a standardized Dutch monosyllabic word test improved from
0% preimplant to 93.5% at 12 months postimplant.
Finally, for Case #3, where no positioner was used,
the entire length of electrode array was located
along the lateral wall of the scala tympani. Phoneme
scores for this patient improved from 0% preoperatively to 86% after twelve months of device use.
If MSCT is to provide high quality, clinically useful
images, two problems must be overcome—(1) image
degradation due to artefact and (2) only part of
the array appearing in any given section. We have
tackled the first problem by using high-resolution scanning and a high-resolution reconstruction filter. The HiFocus array has planar contacts
measuring 0.4 x 0.5 mm placed on a 1.1 mm pitch
along the electrode array. To distinguish individual
contacts and in-plane versus cross-plane, a resolution of at least 2.5 line pairs per mm (lp/mm) was
required. For separate visualization of neighbouring
contacts, a resolution of at least 1.2 lp/mm was
required. The scanner performance approached
the requirements for visualization of individual
contacts. As a result of blooming, the actual shape
of the contacts cannot yet be visualized accurately.
Figure 1. The postoperative MPRs of Case #1. The individual
contacts are clearly distinguishable. The tip of the electrode
was flipped over during surgery—as indicated by the numbered
electrodes.
trauma. Fine cochlear structures, such as the basilar
membrane, cannot be discerned. However, it is
possible to determine whether the electrode array
is in either the scala tympani or scala vestibuli.
With appropriate scanning parameters, MSCT
can image individual electrode contacts as well
as anatomical details of the cochlea, providing
useful clinical information. Technical advances
indicate that MSCT may be expanded to much
more routine clinical application rather than
being limited to complex or unusual cases.
The near-isotropic volumetric imaging available with MSCT was critical for the production
of high quality cross section data and, hence, the
reconstruction of arbitrary planes—particularly
useful for the analysis of Case #1. Oblique coronal
images, reconstructed parallel to the modiolus,
show promise for the assessment of cochlear
Medical & Surgical Studies
31
Statistical Analysis of Cochlear Implant Reliability
David R. Schramm, M.D.
Ottawa Hospital (Civic Campus), Ottawa, ON, Canada
Sandra S. Stinnett, Dr.P.H.
Duke University, Durham, NC, USA
Table 1. Device Failures — Pediatric
Age at
Implantation
Years with
Device
Device Type
Failure Type
4 yrs 2 mos
2.64
Nucleus 22
Integrity Test
5 yrs 1 mo
3.02
Clarion S
Integrity Test
5 yrs 10 mos
3.41
Nucleus 24
Integrity Test
4 yrs 6 mos
3.21
Clarion S &
positioner
Integrity Test
13 yrs 11 mos
3.26
Clarion
HiFocus Ij
Integrity Test
Device failures in 5 of 203 total pediatric implant surgeries.
Table 2. Device Failures — Adults
Age at
Implantation
Years with
Device
Device Type
Failure Type
57 yrs
8.07
Nucleus 22
Integrity Test
27 yrs
6.38
Nucleus 22
Integrity Test
48 yrs
3.41
Clarion S &
positioner
Integrity Test
Occasionally replacement of defective cochlear
implant devices is required, bringing the reliability
of cochlear implants into question. It is difficult to
compare data from various manufacturers because
they tend to use different methods of reporting reliability. The purpose of this study is to investigate
device reliability in our pediatric and adult populations (409 surgeries over 12 years) using the KaplanMeier product limit method of survival analysis.
Patients were “censored” from the database if they
became non-users (nine individuals), died (five
adults), were lost to follow-up (five), or if functioning
devices were removed for medical reasons (two) or at
patient request (five). In this study, cochlear implants
were defined as defective if there was objective
evidence of device failure on integrity testing. Moreover, “soft failures” were included if there was objective or subjective deterioration in speech perception
abilities and subsequent reimplantation restored
or improved speech recognition. Survival rates
for children and adults were analyzed separately.
Of 203 implant surgeries on 198 children, 5 device
failures occured. Of 206 implant surgeries on 202
adults, 3 device failures occurred. Information about
these device failures is shown in Table 1 for children
and Table 2 for adults. All patients with device failures underwent successful reimplantation with full
electrode insertion. The number of device failures was
the same for implants manufactured by Advanced
Bionics (4 of 266) and Cochlear Ltd. (4 of 143).
Device failures in 3 of 206 total adult implant surgeries.
32
Advanced Bionics® Auditory Research Bulletin 2005
“The Kaplan-Meier product limit
method of survival analysis...
is considered to be more accurate than
the life-table (actuarial) method...”
Device survival may be partially dependent
on the surgical technique used for implantation. At our center, a recessed bed is created,
and the implant is secured with non-absorbable suture(s). When possible, the electrode
lead is secured using the split incus technique.
Figure 1 shows the Kaplan-Meier survival distribution for devices in children and adults. The graph
plots the ratio of surviving devices to total number
of implants as a function of number of years postimplant. The five-year cumulative survival for adults
is 0.9899—and for children it is almost as high
(0.9628). Up to eight years postimplant, cumulative
survival is slightly greater for adults than children.
The Kaplan-Meier product limit method of survival
analysis can be used when the time at which the device
fails is known, as is the case for our database. The
method is considered to be more accurate than the
life-table (actuarial) method of survival analysis. It is
recommended that this common standard be used for
comparison of device reliability across manufacturers.
Figure 1. Cumulative survival plotted as the ratio of surviving devices to total number of implants as a function of number of years
postimplant. At five years postimplant, the cumulative survival of devices is 0.9899 in adults and 0.9628 in children.
Medical & Surgical Studies
33
Patient Performance with the Advanced Bionics
HiRes 90K Device with HiFocus Helix Electrode
Christiane Séguin, M.Sc.
David R. Schramm, M.D.
Elizabeth Fitzpatrick, M.Sc.
Shelley Armstrong, M.Cl.Sc.
Josée Chénier, M.O.A.
Ottawa Hospital (Civic Campus), Ottawa, ON, Canada
The HiFocus Helix electrode is designed to
achieve highly focused stimulation to the auditory nerve through perimodiolar placement of
the electrode. Sixteen stimulating contacts face
the modiolar wall, and the surgical insertion technique is intended to minimize risk of damage to
the basilar membrane or lateral wall of the cochlea.
A multicenter study of patients using the HiFocus
Helix electrode and HiRes 90K cochlear implant
was completed in September, 2004. Data on 32 adult
subjects were collected at our center as part of that
study. The present study assesses the safety of the
device and compares the preoperative to postimplant
speech recognition performance of our patients.
Subjects were adults (18 years of age or older)
with severe or profound sensorineural hearing
loss (mean pure tone average 100 dB or greater)
acquired postlinguistically (six years of age or
older). To qualify for the study, patients achieved
preoperative CNC word recognition scores no
better than 50% in the best aided condition.
Patient performance was measured pre- and postoperatively with open set tests including CNC
words and the HINT (Hearing in Noise Test)
sentences in quiet and in noise (+10 dB signal-tonoise ratio). All tests were presented at 70 dB SPL.
Figure 1. Individual results of CNC word recognition testing,
rank ordered according to six-month postimplant scores.
34
Results of the study have shown that the desired perimodiolar placement of the HiFocus Helix electrode
was achieved in all 32 patients, with no postoperative complications. Figures 1-3 show the individual
Advanced Bionics® Auditory Research Bulletin 2005
“Open-set speech recognition...
is significantly improved in
nearly all (postlinguistic) patients
after six months use of the
HiRes 90K cochlear implant
with the Helix electrode.”
speech recognition scores for 19 of the study patients
who have completed their six-month follow-up
evaluations. For one French-speaking patient who
was unable to complete the tests in English, scores of
0 were assigned on all speech perception measures.
For the group, mean CNC word scores were 3.5%
preoperatively compared to 50.1% at six months
postimplant. Mean scores for the HINT in quiet
improved from 15.2% to 82.9% and for the HINT
in noise from 7.1% to 64.0%. As the six-month
individual scores show, 10 of the 19 patients had
achieved CNC word scores greater than 50%.
Moreover, 14 subjects scored higher than 80% for
the HINT sentences in quiet, and 6 subjects scored
higher than 80% for the HINT sentences in noise.
Figure 2. Individual results of HINT sentences in quiet testing,
rank ordered according to six-month postimplant scores.
In summary, the HiFocus Helix electrode can
be safely implanted with successful perimodiolar
placement of the 16 stimulating contacts. Openset speech recognition in quiet and in noise is
significantly improved in nearly all (postlinguistic) patients after six months use of the HiRes
90K cochlear implant with the Helix electrode.
Figure 3. Individual results of HINT sentences in noise
(+10 dB SNR ratio) testing, rank ordered according to sixmonth postimplant scores.
Medical & Surgical Studies
35
Audiological Findings Following Cochlear Reimplantation
Patrizia Mancini, M.D.
Deborah Ballantyne, Ph.D.
Ersilia Bosco, Cl. Psych.
Chiara D’Elia, M.D.
Roberto Filipo, Prof., M.D.
University of Rome La Sapienza, Rome, Italy
“...no negative findings were encountered
for the entire study group.”
The aim of the study was to verify audiological findings following reimplantation in relation to updated
technology and new coding strategies. Included in
this study were 13 cochlear implant wearers (4 children, 9 adults). The original device types included:
Clarion, Med-El, 3M/House, and Ineraid. The
Med-El users were reimplanted with Med-El
devices, the remainder with Clarion devices. Ten of
the subjects were originally implanted in our clinic.
Three were implanted in another center. Three of the
13 subjects were reimplanted in the opposite ear.
The paediatric speech perception test battery
was based on Erber’s hierarchical model. For this
specific analysis, only open set results (recognition and comprehension) will be reported and
discussed. The use of acoustic feedback has been
assessed according to age and ability of the individual child. Testing was carried out using live
voice by the same operator; a black mesh screen
was used to cover the speaker’s face. Tests included:
• Perceptive Ability Test (Italian) (TAP)—
an Italian adaptation of the GASP (Arslan,
1997) used in children ages 4 years and older.
• Italian Common Protocol—used in this study
to evaluate children ages 5 years and older.
• Bi-Trochee-Polysyllabic word test (BTP)—
(Bosco, 1997) used in children 2-5 years of age.
The adult open-set test battery included the
Speech Perception Test in quiet—consisting
of bisyllabic phonetically balanced (PB) words
and interactive sentence lists (questions).
Stimuli were presented at 70 dB SPL via CD audio
in a silent room. Phonemic scoring was applied to
words and key-word scoring was used for sentences.
Scores were subsequently converted into percentage
values. The Speech in Noise Test consisted of bisyllabic PB words and sentences with babble as background noise. Testing was carried out at 35 dB SL
in a sound field (MCL). The full protocol involved
36
Advanced Bionics® Auditory Research Bulletin 2005
testing at various signal-to-noise ratios (from +25
to 0 dB). In the original implant group (wearing
body-worn processors), the competing noise came
from behind. For the reimplant group (wearing earlevel processors), the primary signal was presented
from the front with the competing noise presented
both ipsi- and contralaterally (90° azimuth).
The replacement of a cochlear implant has proven
to be safe and effective. Nonetheless, counseling in
such circumstances is complex in that it must take
into account possible changes in speech perception related to variables such as electrode reinsertion, number of active channels, and technological
differences between initial and reimplanted devices
such as speech processor, signal transfer hardware,
and speech processing strategy. Balkany et al (1999),
in a study on patients reimplanted with the same
multichannel device, showed that open-set word
and sentence scores for adults and children were
higher following reimplantation, but the differences were not significant. Parisier et al (2001), in
a study of 19 children reimplanted with the same
or technologically different devices (Nucleus 22
and Clarion), showed that the speech perception
abilities of all children either remained the same
or improved following reimplantation. No correlation with a possible change in speech perception
strategy or type of implant was made in either study.
In our study group, replacement of a cochlear
implant has proven to be a safe procedure. In all
cases it was successful in restoring hearing. In fact,
no negative findings were encountered for the
entire study group. A full insertion of the replacement electrode was accomplished in all but three
subjects (1 Med-El, 2 Clarion). Nevertheless,
speech perception performance was similar or even
better than that obtained with the first implant,
as shown in Figure 1. Mean hearing thresholds
before and after reimplantation were 33.7 and
27.8 dBHL, respectively (t-test value p = 0.06).
“...speech perception performance
was similar or even better
than that obtained
with the first implant...”
Figure 1. Open set speech perception scores. Points above
diagonal represent relative improvement in speech recognition
in quiet for words (circles and dotted trendline) and sentences
(blue squares and solid trendline).
—continued on next page—
Medical & Surgical Studies
37
—continued from previous page—
In the adult group, a significant improvement
(> 10%) was seen in speech perception in quiet and
in noise when reimplantation was performed with
a technology that increased the number of channels or the stimulation rate. Figure 2 displays speech
perception scores for the four adult subjects who
obtained a full insertion and were upgraded to the
HiRes strategy following implantation. Only one
subject showed no significant improvement, perhaps
as a result of obtaining only a partial insertion.
When reimplantation was performed with the same
implant model (Clarion 1.2 or Med-El Combi 40+),
no significant change was observed in speech perception performance. In three subjects, reimplantation
was performed in the opposite ear, where acoustic
deprivation was greater. Despite this, the replacement
implant’s performance was influenced by technological improvement and the coding strategy in use.
In the paediatric group, results were less homogeneous. Two reimplanted children (Advanced
Bionics) obtained an increased number of channels
and stimulation rate. One of these children showed
significant improvement after three months. The
other child showed no significant improvement in
speech perception. One child (Med-El) received
updated technology but only had a partial reinsertion. One child (Advanced Bionics) was reimplanted with same device. Neither of these two
children reached open set speech understanding
with the first implant, and no improvement on
initial findings was seen following reimplantation.
In conclusion, improved technology, especially in terms of increased number of channels and pulse rate, can often be associated with
improvement in speech perception performance.
References
Arslan E, Genovese E, Orzan E, Turrini M. Test Abilità Percettive (TAP). (1997) In
EDS. Amplifon Valutazione della percezione verbale nel bambino ipoacusico. Bari:
Ecumenica Editrice.
Balkany TJ, Hodges AV, Gomez-Marin O, Bird PA, Dolan-Ash S, Butts S, Telischi F,
Lee D. (1999) Cochlear reimplantation. Laryngoscope 109: 351-355.
Figure 2. Speech perception in noise (SNR +10) evaluated for
upgrade to HiRes (subjects 5,6, 7, and 13) with first implant
(signal front, noise from behind 180°) and second implant (signal front, noise 90° ipsi- and contralateral to the implant side).
38
Bosco E, Ballantyne D, Schlögl M. (1997) Bi-Trochee-Polisyllabic word test (BTP).
In: Allum-Mecklenburg DJ, ed. Evaluation of Auditory Responses to Speech (EARS),
Electronics Ges.m.b.H.
Parisier SC, Chute P, Popp AL, Suh GD. (2001) Outcome analysis of cochlear
implant reimplantation in children. Laryngoscope 111: 26-32.
Advanced Bionics® Auditory Research Bulletin 2005
Magnetoencephalographic Study of Auditory Cortical
Plasticity in Postlinguistically Deafened Adults
Christo Pantev, Prof., Dr. Ing.
Antoinette G. Dinnesen, Prof., M.D.
Bernhard Ross, Ph.D.*
Andreas Wollbrink, Dipl. Ing.
Arne Knief
University of Münster, Münster, Germany
*
also with The Rotman Research Institute, Toronto, ON, Canada
During the last 20 years or so, research has shown
that the functional organization of both the developing and mature auditory cortex is not statically
fixed, but changes over time in response to behaviorally relevant stimulation. In postlinguistically deafened adults who receive a cochlear implant, many
recover their hearing abilities well enough to use
the telephone. Thus, these subjects provide a model
for understanding the process of regaining auditory function after prolonged sensory deprivation.
Magnetoencephalography (MEG) provides an
objective tool for studying the plasticity of the auditory cortex in adults who have been implanted.
MEG is a neuroimaging method that (1) is noninvasive, (2) does not generate acoustic noise, (3) is
highly reproducible, and (4) has high temporal resolution, thereby allowing distinction between different
components of evoked activity. MEG cannot be
used with conventional cochlear implants because
they contain a strong permanent magnet that holds
the transmitter coil in place. However, Advanced
Bionics has developed a magnet-free device with
an RF shield that prevents interference between
the implant transmitter and the MEG scanner.
“These results indicate that the
cochlear implant provided
effective stimulation to the
auditory cortex, and that
neural plasticity remains
even in adults with
prolonged auditory deprivation.”
months after implantation, thereby demonstrating
the neural plasticity of the auditory cortex. By two
years postimplantation, the MEG-estimated source
waveforms for primary and non-primary auditory
cortices were commensurate with a group of ten
normal-hearing control subjects. Moreover, the time
course of the development of “normal” cortical function and the improvement of word recognition skills
were similar. These results indicate that the cochlear
implant provided effective stimulation to the auditory cortex, and that neural plasticity remains even
in adults with prolonged auditory deprivation.
Reference
Pantev C, Dinnesen A, Ross B, Wollbrink A, Knief A. (2005) Dynamics of auditory
plasticity after cochlear implantation: a longitudinal study. Cerebral Cortex Advance
Access. Oxford University Press. Available online at http://cercor.oxfordjournals.
org/cgi/content/abstract/bhi081v (accessed 16 June, 2005).
In this published study, MEG recordings and
speech perception were evaluated over a period
of two years after surgery in two postlinguistically deafened adults who were implanted with the
magnet-free device. MEG responses showed rapid
change in evoked brain activity over the first six
Medical & Surgical Studies
39
Objective Measures
In addition to programming the HiResolution Bionic Ear System,
the SoundWave Professional Suite software offers tools for testing
the implanted electrode array and for objectively assessing the
function of the hearing nerve.
Neural Response Imaging (NRI) measures the response of the
hearing nerve to electrical stimulation. Single-channel NRI
recordings can be made with clinical SoundWave software, whereas
research software has been designed to measure banded NRI
responses by stimulating multiple electrodes simultaneously. Banded
NRI allows rapid assessment of the entire electrode array and yields
input-output functions that more closely parallel psychophysical
loudness growth functions obtained with Speech Bursts compared to
single-channel NRI input-output functions.
Moreover, SoundWave’s unique speech burst stimuli can be used to
elicit electrical stapedius reflexes—the thresholds of which relate
closely to everyday, most comfortable listening levels.
Worldwide Objective Measures Study: European Update
Setting stimulation levels for the HiRes system
requires users to make psychophysical judgments of
comfortable loudness levels. Because it is sometimes
difficult to obtain reliable loudness judgments, clinicians have turned to objective measures of hearing
function—in particular evoked compound action
potential (ECAP) and evoked stapedius reflex
threshold (ESRT)—for additional information in
programming cochlear implants. Both ECAP and
ESRT have been shown to be related to the listening
levels used in programming implants (Firszt et
al, 2003; Novak et al, 2003; Overstreet, 2004).
The objectives of this multicentre study are:
• To explore the relationships over time between
ECAPs, ESRTs, and programming parameters
in HiRes users
• To develop guidelines for using ECAPs and
ESRTs in device fittings, thereby maximizing
user benefit.
Table 1. Stable M levels as a percent of intraoperative
tNRI, according to stimulation site.
Stim/Record
Electrodes
Stable M
% of tNRI
3/1
151%
7/5
150%
11/9
116%
15/13
89%
Mean
123%
So far, 51 subjects have been enrolled in this study.
In this report, all objective measures have been
obtained intraoperatively and are compared to
behavioural M levels obtained at first fitting and
after 3, 6, and 12 months of device use. To ensure
a sufficient duration of follow-up, 41 subjects (30
children, 11 adults) are included in the present
analysis. All subjects in the study use either a CII
Bionic Ear or HiRes 90K implant unilaterally.
The ESRT recordings were performed using Speech
Burst™ waveforms and the ECAP was measured
through Neural Response Imaging (NRI)—both
features of the SoundWave clinical programming software. NRI was measured on stimulating/
recording electrode pairs 3/1, 7/5, 11/9, and 15/13.
For this study, two NRI measures have been defined:
(1) the threshold of NRI (tNRI) in which the NRI
response waveform should have a zero-amplitude—
and (2) the “1st NRI” in which the NRI waveform is
the smallest response that can be identified visually.
Overall, the ESRT was obtained at a success
rate of 73%—in line with the other investigations (e.g. Cullington, 2003). Stimulation at the
cochlear apex was more likely to elicit a reflex
(80%) than stimulation at the base (66%). Similarly, the NRI responses were robust (with an
overall success rate of 84%) and were more often
obtained at the apex (89%) than at the base (79%).
Observations from the mean NRI and M level
data in Figure 1 (and Table 1) are as follows:
• Both the 1st NRI and tNRI values were higher
at the base compared to the apex.
• 1st NRI values were on average 30 CU higher
than the tNRI across the array.
• M levels were roughly equivalent across the
array, with a slight decrease at the base.
• M levels at three months were on average 43
CU higher than the Ms obtained at first fitting.
• Compared to tNRI, the first-fitting M levels
were higher at the apex and lower at the base.
Figure 1. M levels obtained at first fitting and at three months compared to
intraoperative tNRI and 1st NRI in children. On average the stable M levels (at three
months) were 123% of the intraoperative tNRI. (See Table 1.)
42
• Compared to tNRI, M levels at three months
were higher across the array, except at the most
basal part. On average, the M levels were 123%
higher than the tNRI.
Advanced Bionics® Auditory Research Bulletin 2005
Observations from the mean ESRT, NRI, and
stable M levels (obtained after three months
of use) in Figure 2 (and Table 2) are as follows:
Participating Centres
Ghent University, Ghent, Belgium
• M levels were roughly equivalent across the
array, with a slight decrease towards the base.
Hôpital Edouard Herriot, Lyon, France
Hôpital Robert Debré, Paris, France
Hôpital Charles Nicolle, Rouen, France
• Compared to M levels, ESRT values were
lower at the base.
• Compared to M levels, tNRI values were higher
at the base.
• On average, the stable M levels fell between the
intraoperative tNRI and ESRT levels—except
at the most basal electrodes, where the M levels
fell slightly below the tNRIs.
Perhaps the lower M levels at the base are the
result of greater energy in the speech burst stimuli
presented to the basal electrodes compared to the
energy in the long-term, high frequency speech
spectrum. The group M levels were in the normative range (100 to 300 CU) although slightly below
the average (180 CU) previously reported (Arnold
& Boyle, 2005). This finding may be explained
by the high number of children in this sample. In
the cited study, we observed that basal Ms tend to
increase over time as higher pitches become more
readily accepted. Hence, for these subjects, the basal
Ms may yet fall into line with the rest of the array.
In these data, stable M levels (obtained after more
than three months of device use) fell consistently
between intraoperative ESRT and tNRI values.
This preliminary result indicates that ESRT and
tNRI values may be considered in routine, clinical
HiRes fittings. However, more long-term data
are necessary to refine clinical recommendations,
especially in setting levels for basal electrodes.
Unfallkrankenhaus Berlin, Berlin, Germany
Desa’s Hospital, Mumbai, India
Hadassah University Hospital, Jerusalem, Israel
Chaim Shebah Medical Centre, Tel Hashomer, Israel
Schneider Children’s Medical Centre, Petah Tikva, Israel
U.O. Audiologia, Ferrara, Italy
Clinique Rachidi, Casablanca, Morocco
Erasmus MC, Rotterdam, Netherlands
Hacettepe University, Ankara, Turkey
SSK (SB) Ismir Hospital, Izmir, Turkey
Guy’s and St. Thomas’ Hospital, London, United Kingdom
Principal Investigator
Hôpital Robert Debré, Paris, France
Table 2. Stable M levels as a percent of intraoperative
ESRT and tNRI, according to stimulation site.
Stim/Record
Electrodes
Stable M
% of ESRT
Stable M
% of tNRI
3/1
55%
127%
7/5
58%
130%
11/9
67%
108%
15/13
65%
88%
Mean
60%
112%
References
Arnold L, Boyle P. (2005) “Intelligent programming” concept with SoundWave.
Poster presented at the 10th Symposium on Cochlear Implants in Children, Dallas,
Texas, 15-19 March, 2005.
Cullington H, ed. (2003) Cochlear Implants: Objective Measures. Chichester, UK:
Whurr Publishers.
Firszt JB, Runge-Samuelson CL, Reider A, Raulie J, Wackym P, Overstreet E. (2003)
Comparisons of ECAP and High Resolution programming levels in the Clarion CII
using single and multielectrode stimulation techniques. Poster presented at the
Conference on Implantable Auditory Prostheses, Pacific Grove, CA, 17-22 August,
2003.
Novak MA, Overstreet EH, Thomas JF, Rotz LA, Black JM. (2003) EABR and ECAP
thresholds and growth function slopes: correlations with HiResolution program
settings. Poster presented at the Conference on Implantable Auditory Prostheses,
Pacific Grove, CA, 17-22 August, 2003.
Overstreet EH. (2004) New objective measurement techniques and their relationship to HiResolution program settings. Poster presented at the 7th European Symposium on Paediatric Cochlear Implantation, Geneva, Switzerland, 2-5 May, 2004.
Figure 2. Intraoperative ESRT and tNRI compared to stable M levels obtained
after more than three months of use. On average the stable M levels are between the
intraoperative ESRT (60%) and tNRI (112%) values. (See Table 2.)
Objective Measures
43
Using Objective Tests to Optimize
Implant Function and Benefit
Study Sites in North America
Carle Foundation Hospital, Urbana, Illinois
Houston Ear Research Foundation, Houston, Texas
L’ Hôtel-Dieu de Québec, Québec City, Québec
Integris Health, Oklahoma City, Oklahoma
Sunnybrook & Women’s College Health Sciences Center
Toronto, Ontario
University of Massachusetts, Amherst, Massachusetts
University of Minnesota, Minneapolis, Minnesota
Neural Response Imaging (NRI) is a software
module in the SoundWave clinical programming system that allows measurement of the electrical compound action potential (ECAP) with
the implanted electrode. The ECAP reflects the
response of the auditory nerve to electrical stimulation delivered by the implant, and does not
require a behavioral response from the patient.
This multicenter clinical study is being conducted to
characterize the NRI-elicited ECAP and to evaluate
electrical stapedial reflex thresholds (ESRTs) and
electrical auditory brainstem responses (EABRs)
in a large group of Bionic Ear recipients (both
adults and children). ECAP responses, ESRTs, and
EABRs are compared to speech perception ability,
HiRes program parameters, and demographic information (1) to determine the characteristics of ECAP
responses as a function of user profile and benefit, and
(2) to determine the relationships between ECAPs,
ESRTs, EABRs, and HiRes programming levels.
The overall goal is to develop clinical guidelines for
using NRI, ESRTs, and EABRs in Bionic Ear recipients—especially for patients who cannot provide
reliable behavioral responses to electrical stimulation.
According to study protocols, objective measures
for each patient are obtained as follows:
• ECAP input-output functions are measured
using (SoundWave) NRI on each of four
programming channels.
• ESRTs are elicited using (SoundWave)
Speech Bursts delivered to four groups of four
adjacent electrodes and measured using clinical
impedance equipment.
• EABRs are elicited by using pulsatile stimuli
delivered to four electrodes along the implanted
array and measured using standard scalp EEG
electrodes.
44
Advanced Bionics® Auditory Research Bulletin 2005
ECAP responses, ESRTs, and EABRs then are
compared to each patient’s speech recognition
ability, HiRes program parameters, and demographic
information. Participating study sites can administer
all or part of the test battery depending upon their
clinical interests. Centers also have the option of
administering the NRI, ESRT, and EABR testing
intraoperatively, or for administering the test battery
at multiple sequential test sessions in order to determine any longitudinal changes that might occur.
“These preliminary data suggest that
ESRTs elicited by Speech Bursts
most closely approximate
behavioral M levels.”
Figure 1 shows average speech-burst M levels,
single-channel tNRI values, single-channel ESRTs,
and speech-burst ESRTs for 22 adults. These
preliminary data suggest that ESRTs elicited by
Speech Bursts most closely approximate behavioral M
levels. The tNRI values are slightly below M levels.
As additional data are compiled, guidelines
will be developed to assist clinicians in using
these objective measures to estimate psychophysical comfort levels, to verify behavioral
data, and to counsel patients appropriately.
Figure 1. Mean behavioral comfort (M) levels, speech-burst ESRTs, single-channel ESRT, and tNRI responses (measured on electrodes 3, 7, 11, and 15) for 22 adults—expressed in HiRes Clinical Units (CU).
Objective Measures
45
Clinical Use of Objective Measures
with Children and Adults Using HiRes
Jean Thomas, M.S.
Lee Ann Rotz, M.A.
Carle Foundation Hospital, Urbana, IL, USA
Michael A. Novak, M.D.
Carle Clinic Association, Urbana, IL, USA
Edward Overstreet, Ph.D.
Advanced Bionics Corporation, Valencia, CA, USA
The electrical compound action potential (ECAP)
and electrical auditory brainstem response (EABR)
have been utilized intra- and post-operatively
across cochlear implant devices to verify neural
responsiveness and to aid in device programming
(Brown 2003). In the Carle Expanding Children’s
Hearing Opportunities (ECHO) cochlear implant
program, we face the unique challenges of programming the devices in infants and young children. It
has been our experience that objective measures
can be helpful in setting appropriate program
levels when limited behavioral data are available.
To guide our clinical use of objective measures,
we sought to investigate the relationships between
ECAP and EABR thresholds and program levels
in HiRes users through a retrospective study. We
reviewed data from a group of 32 CII and HiRes 90K
users who had been fit with HiRes sound processing
during 2003. The demographics of this group were
analyzed by age and duration of deafness. Three
groups were identified—children, adults with shortterm deafness, and adults with long-term deafness.
Data from 19 postlinguistically deafened adults
were reviewed for a total of 20 ears (one patient had
bilateral implants). Ages at the time of implantation ranged from 17.5 to 88.4 years with a mean
age of 52.8 years. One 17.5-year-old was included
in the adult analyses because the patient’s language
level and ability allowed the patient to participate
in adult speech perception tests. Adults were separated into two groups by duration of severe-toprofound hearing loss in the implanted ear (PTA >
90 dB). Nine adults had short-term deafness (less
than 10 years) and 10 adults had long-term deafness (greater than 10 years). In addition, data from
13 pediatric patients were reviewed. Ten of the 13
children were prelinguistically deafened (onset of
deafness before age three years). Four children had
used HiRes from initial stimulation (age at implant
7-14 years; device experience of 4-12 months) and
nine children had been reprogrammed in HiRes
after using conventional strategies (age at implant
10 months to 14 years; device experience 2-9 years).
Intraoperative EABR and postoperative NRI thresholds (obtained during the first six months after initial
stimulation) were reviewed for an apical, medial,
46
Advanced Bionics® Auditory Research Bulletin 2005
“For all groups, tNRI and 1st NRI
tended to fall
at an audible level below the M level.”
and basal electrode, typically electrodes 1, 7, and 13.
Behavioral HiRes M levels were examined at initial
stimulation and at three months after initial stimulation. M levels were defined as the most comfortable loudness level used in the patient’s everyday
program. Objective-measure and M-level values
were averaged across the three electrodes. NRI and
EABR thresholds were compared to HiRes program
settings by converting all stimulation levels into
units of charge per phase (HiRes Clinical Units).
Figure 1 shows the averages and standard deviations for M levels, NRI response thresholds, and
EABR thresholds for the three demographic
groups. Notably, the pediatric users tended to have
lower M levels than the adults. Both the M levels
and NRI responses were significantly different
between the pediatric and long-term deafened
adult groups (p < .05). For all groups, tNRI and
1st NRI tended to fall at an audible level below
the M level. The EABR thresholds tended to fall
higher in the dynamic range than the NRI thresholds. However, there was significant intrasubject
variability. In some patients, individual NRI and
EABR levels occasionally exceeded the M levels.
Because ECAP and EABR thresholds occur within
an audible range, these objective tests can aid the
clinician in choosing a level at which a young child
can be conditioned to sound. The normative ranges
can also be used to verify the accuracy of behavioral
M levels. If objective measure results fall within
their expected range, and there is a large discrepancy
between behavioral M levels and the expected Mlevel range, then M levels may require reevaluation.
Figure 1. Average (with standard deviations) M levels, NRI
thresholds, and EABR thresholds for children (n = 13), STD
(short-term deafened) adults (n = 9), and LTD (long-term deafened) adults (n = 10, 11 ears). Data are from the three-month
test interval.
Reference
Brown CJ. (2003) Clinical uses of electrically evoked auditory nerve and brainstem
responses. Curr Opin Otolaryngol Head Neck Surg 11:383-387.
Objective Measures
47
Evaluating HiResolution Sound Processing
Based on Objective Measures
Jace Wolfe, Ph.D.
Heather Kasulis, Au.D.
Integris Health, Oklahoma City, OK, USA
“... the proximity between M levels
derived behaviorally and those based
on ECAP (NRI) and ESRT measures
suggests that these objective tools
can be used effectively to guide
HiRes programming.”
The HiResolution (HiRes) cochlear implant system
provides tools for objective hearing tests that can
aid device programming. However, the relationships
between objective test results, programming levels,
and implant benefit are not well defined. For objective
measures to be applied effectively, it is important to
understand how they correlate with behavioral “most
comfortable” loudness levels (M levels). Of particular clinical interest are two objective indicators of
hearing function: electrical stapedial reflex thresholds
(ESRTs) and electrical compound action potential
responses (ECAPs)—also designated tNRIs when
obtained with the Neural Response Imaging (NRI)
software module in the SoundWave programming
system. While ECAPs (or tNRIs) provide useful
clinical information, the ESRTs may be ideally
suited for predicting programming levels because
these responses are elicited with the same high-rate
stimuli that are delivered in HiRes sound processing.
In our clinic, we are studying the efficacy of using
ESRTs and ECAPs to set program levels in HiRes
users. In keeping with our study protocol, subjects
with at least three months of HiRes experience
participated in two test sessions. In the first session,
M levels, ESRTs, ECAPs, and speech perception
48
Advanced Bionics® Auditory Research Bulletin 2005
were measured. If behavioral M levels and objective
test results were disparate, the M levels were reprogrammed based on the objective values. After one to
two months, subjects returned for the second session
at which time speech perception was retested.
To date, 15 adults and 2 teenagers have been evaluated. For most patients, the HiRes programs that
were based on ECAP (tNRI) and ESRT results
required little change in M levels. As shown in
Figure 1, the mean speech perception scores were
slightly higher after the programs were adjusted,
but the differences were not significant (p < .01).
However, for a few individual patients, speech
recognition performance did improve when M levels
were adjusted according to the NRI and ESRT
values, as illustrated by patient A6 in Figure 2.
Figure 1. Mean scores for 17 subjects at the initial test session
and after reprogramming using objective measures. HINT in
noise presented at 60 dB SPL, +8 dB SNR; CNC words presented at 50, 60, and 70 dB SPL.
Overall, the proximity between M levels derived
behaviorally and those based on ECAP (NRI) and
ESRT measures suggests that these objective tools
can be used effectively to guide HiRes programming. In cases where substantial differences between
ESRTs and behavioral M levels existed, speech
recognition was improved by creating new programs
with M levels set closer to the ESRT values.
Figure 2. Initial M levels, tNRIs (ECAPs), ESRTs, and final
(adjusted) M levels for patient A6. This patient’s CNC scores (at
70 dB SPL presentation level) were 16% before and 56% after
program modifications based on objective measures.
Objective Measures
49
Can NRI and ESRTs Be Used to Optimize
Program Parameters?
Gül Caner, M.D.
Levent Olgun, M.D.
SSK Izmir Hospital, Izmir, Turkey
Laure Arnold, M.Sc.
Advanced Bionics Corporation, Europe
Programming in children can be challenging,
particularly in congenitally deaf children who
have no concept of loudness. Clinical experience
suggests that comfort levels may be set too high
when based only on behavioral loudness judgments. Therefore, fitting guidelines should include
both objective measures along with behavioral
responses, especially over time, to avoid setting
comfort levels too high. The question is whether the
programs based upon objective measures combined
with behavioral responses provide improved sound
awareness and greater implant benefit in children than programs based solely on behavioral
responses. In this study, correlations between Neural
Response Imaging (NRI) responses, electrical
stapedius reflex thresholds (ESRTs), and psychophysical measurements were evaluated in order to
develop guidelines to optimize HiRes programs.
Included in this study were 18 patients—16 children and 2 adults who were implanted with a HiRes
90K device. NRI responses were measured on four
electrodes intraoperatively, at first fitting, and after
three and six months of implant use. Subjects were
fitted using the SoundWave default parameters—
Speech Bursts and automatic calculation of thresholds (10% of most comfortable M levels). ESRTs
were determined visually during the surgery. No
ESRT measurements were made postoperatively.
ESRTs and tNRI values were compared with the
levels used to program the sound processors-—
that is, the M levels for HiRes-P (paired) and
HiRes-S (sequential) programs. The M levels used
in the analyses were those used by each patient in
their everyday HiRes programs. Changes of tNRI
values over time also were monitored. CAP and
MAIS scores were noted. All subjects reached
the three-month evaluation. Data from the sixmonth evaluation were available for 13 subjects.
NRI responses could be recorded in 91.7% (66/72)
of measurements intraoperatively. Postoperative
success rate increased to 92.3% (181/196). NRI
responses were more easily obtained at the medial
and basal electrode locations than the apical locations, the same result that we have observed during
studies in adult patients. The stapedius reflex could
be observed in 76.4% of the subjects intraoperatively.
Figure 1. Comparison of ESRTs, tNRI values, and live-speech M
levels intraoperatively and at first fitting. Stimulation was delivered to four electrodes along the array from apex (electrode 3)
to base (electrode 15).
50
NRI data obtained intraoperatively showed an
increase in tNRI toward the basal end of the
electrode array. ESRTs were higher at the basal
and apical locations than when stimulating the
middle part of the electrode array (Figure 1).
Advanced Bionics® Auditory Research Bulletin 2005
First fitting M levels were slightly higher at the
apical end of the electrode array and tNRI levels
were higher at the basal end of the electrode array.
On average, M levels were 15% higher on the
apical electrodes and 24% lower on the basal electrodes than the tNRI values. First fitting Ms were
72% of intraoperative tNRI values overall. tNRI
levels decreased after surgery, then increased postoperatively (maximum reached at six months),
and tended to stabilize afterwards. A 9% increase
was observed from first fitting to six months.
Comparison of all measurements revealed that tNRI
values increased from the apical to the basal end
of the electrode array and from first fitting to six
months. Changes in M levels over time were similar
across the array. A 25% increase was observed in
all locations. Sequential M levels were higher than
paired M levels (21% at first fitting and 19% at six
months: stable difference). First fitting M levels
were, on average, 15% higher than tNRI levels for
the apical electrodes and 24% lower for the basal
electrodes. The six-month M levels were, on average,
63% of the intraoperative ESRTs (70% for sequential
and 56% for paired). That ratio remained somewhat
stable for the medial electrodes, that is, paired M/
ESRT was 63%, and sequential M/ESRT was 80%.
“NRI is a very useful method for
programming cochlear implants
with very young children
who cannot demonstrate
a reliable judgment of loudness.”
basal). The overall ratio of six-month sequential M
levels and six-month tNRI values was 103%. CAP
scores increased rapidly after first fitting. MAIS
scores also increased rapidly after the first fitting,
but the increase slowed during subsequent months.
NRI is a very useful method for programming
cochlear implants with very young children who
cannot demonstrate a reliable judgment of loudness. Stable M levels for HiRes-S nearly equaled
the tNRI values. The ESRT values vary across the
array. With the use of an impedance bridge intraoperatively, lower values and better correlations
might be expected between ESRTs and M levels.
This study indicates that tNRI values can be used to
predict M levels across the array. Differences in M
levels and tNRI levels across the array (+15% apical
and -24% basal) at first fitting may serve as a guideline to start and then can be revised with live speech
stimulation during the fitting. First fitting M levels
were 72% of the intraoperative tNRI levels along the
electrode array, ranging from 101% to 67% (apical to
Objective Measures
51
Relationship Between Objective Measures and
HiRes Programming Levels
De-min Han, Prof. M.D., Ph.D.
Xiao-tian Zhao, M.D.
Xue-qing Chen, M.D.
Sha Liu, M.D.
Bo Liu, M.D.
Ling-yan Mo, M.D.
Yong-xin Li, M.D.
Ying Kong, M.D.
Tongren Hospital, Beijing, People’s Republic of China
“...EART using Speech Burst stimuli
provides the best estimate of
most comfortable loudness levels.”
There is a growing trend to use objective measures
to assist with cochlear implant device programming,
especially with very young children who cannot
perform psychophysical tasks traditionally used to
set loudness levels. Neural Response Imaging (NRI)
and electrically evoked auditory reflex thresholds
(EART) are two objective measures that can be
used with the HiResolution Bionic Ear System.
In order to use these objective measures effectively,
it is necessary to know their relationship to most
comfortable loudness levels (M levels) derived
from the behavioral responses of each patient. The
present study examined the relationship between
NRI, EART, and behaviorally set comfort (M)
levels in patients who received the CII implant.
The subjects consisted of 11 profoundly hearingimpaired children and adults, ranging in age from
2 to 29 years, who had between one and nine
months experience with HiRes sound processing.
The NRI software was used to measure the
compound action potential in response to electrical
stimulation (ECAP). Incremental levels of stimulation were used to establish a neural growth function. The lowest current level at which a response
was detected was the1st NRI. A regression line was
fitted to the neural growth function, and its x-intercept was taken as the threshold response (tNRI).
Figure 1. tNRI plotted as a function of live speech M levels
from 53 stimulating electrodes.
52
NRI responses were measured in 10 of the 11
subjects. Complete NRI (1st NRI and tNRI) and
behavioral M-level data were obtained from 8 of the
11 subjects. In Figure 1, tNRI results are plotted as a
function of M level for live speech stimuli. The solid
line represents the regression line fit to the data.
The slope of the regression line is 0.597, and the
correlation coefficient of tNRI to M-level is 0.676.
Advanced Bionics® Auditory Research Bulletin 2005
In Figure 2, a similar plot is presented with 1st
NRI shown as a function of M level. The solid line
represents the regression line. The slope of the line
is 0.772, and the correlation of 1st NRI to M level
is 0.738. These data indicate a closer correlation
between the M level and the 1st NRI than the tNRI,
respectively. The M levels, tNRI, and 1st NRI tend
to follow similar patterns across the electrode array.
EART testing was accomplished using Speech
Bursts, the same stimuli that are used to obtain
M levels during programming.
Speech Bursts
were delivered to four groups of four electrodes. For each group, the stimulus amplitude was increased until a reflex was observed.
Figure 2. 1st NRI plotted as a function of live speech M levels
from 53 stimulating electrodes. The solid line represents the
regression line fit to the data.
EARTs were elicited in 7 of the 11 subjects.
Responses could not be obtained in three children with cochlear malformation and in one child
because of excessive head movements. In Figure
3, EART values are plotted as a function of M
levels. The regression line fitted to the data (slope
= 0.9057) is shown and indicates a high correlation (r = 0.89) between EART values and M levels.
These results across Figures 1 to 3 indicate that
EART using Speech Burst stimuli provides the
best estimate of most comfortable loudness levels.
However, in this study, EART responses were
obtained in a fewer number of subjects than the
NRI responses. In addition to setting comfort
levels, the NRI data can be used for troubleshooting and other programming functions. Thus,
both NRI and EART are useful tools to assist
with device programming in the clinical setting.
Figure 3. EART values are plotted as a function of M levels. The
solid line represents the regression line fit to the data.
Objective Measures
53
Setting Upper Programming Levels in Children
Teresa Zwolan, Ph.D.
University of Michigan, Ann Arbor, MI, USA
Edward Overstreet, Ph.D.
Advanced Bionics Corporation, Valencia, CA, USA
“Based upon these data,
we have derived normative ranges
that clinicians can use as guidelines
in setting M levels
for children and adults.”
Setting most comfortable (M) levels in a cochlear
implant program has a significant effect on patient
outcomes. With the Advanced Bionics devices, M
levels should be set to a loudness level that is “most
comfortable” for the patient. With children, these
upper programming levels can be difficult to set as
children are often unable to provide precise feedback regarding the loudness and comfort level of
a sound. Thus, when setting M levels, the child is
watched closely for any adverse reactions to sound,
possibly indicating that M levels have been set to
a level that is too loud or uncomfortable. In most
clinics, M levels are set using either objective or
subjective measures or a combination of the two.
The implant user’s perception of sound will change
during the first few weeks of device use. Thus, most
clinics conservatively set M levels at device activation and then gradually increase M levels over
subsequent clinic visits. Children’s speech processors are often programmed with successively louder
programs, and parents are instructed to gradually
change to the louder programs over a period of
days or weeks. It is important to monitor M levels
over time because they should eventually stabilize,
making increases in M levels no longer necessary.
Unnecessary increases in M levels can have adverse
consequences, such as discomfort, decreased desire
to use the device, facial nerve stimulation, adaptation
of loudness, channel interaction, signal distortion,
voltage compliance issues, and reduced battery life.
In order to provide clinicians with guidelines for
setting M levels, we measured and compiled psychophysical responses for a large number of implant
users so that a normative M-level range could be
defined. Subjects included 33 adult and 44 pediatric patients from our clinic who used the previous
generation C-I cochlear implant. Also included were
25 adult CII users who participated in a multicenter
clinical trial evaluating the HiResolution (HiRes)
sound processing strategy. In addition, M levels at
six months for 45 adult subjects using the HiRes
54
Advanced Bionics® Auditory Research Bulletin 2005
���
Based upon these data, we have derived normative ranges that clinicians can use as guidelines in
setting M levels for children and adults. Specifically, in users of earlier generation (C-I) pulsatile strategies, approximately 80% of M levels fall
below 450 CU and above 180 CU. In CII users,
90% of M levels fall below 300 CU and above
100 CU. In addition, C-I users tended to have
higher “measured” levels and greater scatter when
compared to cohorts using the CII/HiRes 90K
technology. These differences are likely due to some
nonlinearity in the C-I current sources as compared
to the improved CII/90K current sources—and to
the much slower stimulation rates in C-I programs.
Although providing patients with louder programs
can be helpful at first, the provision of successively
louder programs may not be necessary after a child’s
program has stabilized, which usually occurs after
3-6 months of device use. Moreover, providing
successively louder programs can be detrimental
because many parents will continue to work the
��������������
���
���
���
�
���
���
��������������
Because patients used different pulse durations
in their programs, all stimulation levels were
converted to charge units (nC per phase) and then
to HiRes Clinical Units (CU)—thereby enabling
cross-program comparisons. Figure 1 shows mean
thresholds (T) and most comfortable (M) levels for
C-I users at initial stimulation as well as 12 and 24
months after initial stimulation. Figure 2 (on the
following page) shows the mean T and M levels for
adult HiRes 90K users at six months postactivation.
Mean T and M levels for all electrodes in 25 CII
users are shown in Figure 3 as a function of time.
���
���
���
���
�
���
���
��������������
90K device also were included in this report. M
levels were compiled for all available electrodes from
initial activation and from 12, 15, or 24 months
post-activation. Measurements were combined for
children and adults for the C-I data sets. Prelinguistically deafened adults as well as subjects who had
cochlear anomalies were excluded from the analysis.
���
���
���
�
Figure 1. Average T and M levels at initial stimulation (A), at 12
months (B), and at 24 months post initial stimulation (C) plotted
as a function of electrode for 53 Clarion (C-I) implant users.
—continued on next page—
Objective Measures
55
—continued from previous page—
“If a child is making adequate progress
...clinicians should refrain from
automatically increasing M levels...”
Figure 2. Average T and M levels for 45 HiRes 90K users as a
function of electrode at six months post initial stimulation.
Figure 3. Average levels for all electrodes as a function of time
for 25 CII users who participated in the HiRes clinical trial. The
first three data points show the T and M levels of these users
running a conventional pulsatile strategy. At three months, the
patients were crossed over to HiRes (initial levels shown as “initial HiRes”). These users continued to use HiRes. The last time
interval reflects T/M levels after 12 months of HiRes use (for a
total of 15 months of device use).
child up to the loudest program based upon a false
conception that “louder is better” and because their
child’s reaction to sound may be more noticeable.
Even though small increases in M levels may seem
inconsequential, clinicians should be aware that
such increases may negatively affect performance.
Clinicians should be aware of the potential pitfalls
and warning signs of oversetting stimulation levels.
If a child is making adequate progress with the
device, clinicians should refrain from automatically
increasing M levels or from making unnecessary
adjustments to the child’s program. Instead, parents
should be advised about the potential detriments of
oversetting M levels. Several patient management
steps also can be followed. Clinicians can limit the
upper “set-volume” setting to prevent parents from
overstimulating the child. If testing indicates a need
for higher M levels, clinicians can increase loudness
in very small steps and monitor progress to determine if greater increases might be beneficial. If a
child demonstrates higher than average M levels, the
clinician should consider successively softer programs
and encourage parents to gradually work their way
towards the softer program—particularly if the child
is not making adequate progress in developing speech
and language or speech perception skills. If the
child continues to demonstrate higher than average
M levels, the manufacturer should be contacted
and internal device integrity should be evaluated.
Finally, clinicians should use information derived
from objective measures such as NRI (neural
response imaging) and stapedial reflex thresholds
to assist with setting M levels. Increases in baseline objective measure thresholds over time may
indicate that M levels are set too high. Regular
evaluations of sound field thresholds, speech perception, and speech and language skills should be
conducted to monitor performance with the device
and to help determine if program parameters or
sound processing strategies should be changed.
In conclusion, continued evaluation of upper
stimulation levels in adults and children is
needed to better understand the programming variables that may affect performance.
56
Advanced Bionics® Auditory Research Bulletin 2005
Neural Response Imaging (NRI):
Correlation with Cochlear Implant Threshold Levels
Amparo Platero, M.D.1
Antonio Morant, M.D.1
Maria I. Pitarch, M.D.1
Manuel Tomás, M.D.2
Jaime Marco, M.D., Ph.D.1
Paz Martinez, M.D.1*
1-Hospital Clínico Universitario de Valencia,
Valencia, Spain
2-Hospital Son Dureta, Palma de Mallorca, Spain
* also with Advanced Bionics Corporation, Europe
Neural response imaging is incorporated into
the SoundWave programming software for the
Advanced Bionics CII and HiRes 90K cochlear
implants. It is used to obtain compound action
potentials of the auditory nerve (ECAP) evoked as
the result of the electric stimulation of the cochlea,
providing electrophysiological information about
the stimulated neural tissue. Using NRI, one of
the 16 electrodes is used to produce a stimulating
current while a different electrode is used to record
the evoked neural response. These recordings are
easy and fast to obtain, providing information
about the stimulating thresholds when programming. This is extremely important data, especially
for those patients unable to respond reliably to
stimuli for example when programming young
children. The goal of this study is to evaluate the
correlation between the NRI values and programming parameters—with the final objective being to
program the cochlear implant based on NRI data.
We have performed recordings in four electrodes
(electrodes 3, 6, 9, and 15)—evaluating the basal,
medial, and apical turns of the cochlea, respectively. These recordings were obtained during
surgery and at 1, 3, 6, and 12 months postoperatively—providing information about the neural
tissue and the number of inserted electrodes in 20
patients (12 children and 8 adults) implanted in
our hospital with Advanced Bionics CII and HiRes
90K cochlear implants. The cochlear implant fittings
have been based upon psychophysical and behavioural methods and have not used the NRI values
as a reference. We have evaluated the NRI variation
over time and its correlation with the programming M levels in relation to subject demographics.
Analyses of the data reveal great variability across
subjects and also across pairs of tested electrodes.
The time evolution trend of tNRI shows that (1)
the values tend to get lower in all studied channels,
(2) this trend is observed in both adult and paediatric
groups, and (3) the stabilization takes place after six
months. On the contrary, M levels grow, reaching
stability in the same time interval. The ratio between
both (tNRI/M) is high within the first months
with a mean (within the channels) around 85%,
falling to 65-70% between 6-9 months (Figure 1).
Figure 1. tNRI as a percent of M levels over time.
Creating programs for children can be difficult.
Therefore, it is extremely advantageous to use
the tNRI values. These results show that we have
obtained stable programs for all our patients after 912 months of CI experience. We can use our tNRI
data to predict the M levels during the early phases
of CI programming and, therefore, to obtain optimal
auditory efficiency in the early months of device activation. However, it is important that the objective
measurements (tNRI) are used together with the classical behavioural tests to create optimal CI programs.
Objective Measures
57
Intraoperative and Postoperative Objective Measures:
Relationship to HiResolution Program Parameters
William H. Shapiro, M.A.
Betsy Bromberg, M.A.
New York University School of Medicine, New York, NY, USA
“Intraoperative ESRTs and tNRI
measures appear to be useful
in guiding program settings
at initial stimulation.”
As the criteria for implant candidacy expand (especially to include very young children) and the number
of programming parameter options available to the
clinician increases, objective measures may play an
even greater role in programming patients efficiently
and accurately. Two objective measures that have
proven useful are electrical compound action potentials (ECAPs) and electrical stapedial reflex thresholds (ESRTs) (e.g., Brown, 2003; Hodges et al, 2003).
The purpose of this study was (1) to determine the
relationships among ECAPs and ESRTs obtained
during surgery and most comfortable levels (M
levels) measured at initial stimulation, and (2) to track
changes in those responses over time postoperatively.
Subjects were 17 adults who had been implanted
with the HiRes 90K device with the HiFocus electrode and two adults who had been implanted with
the HiRes 90K device with the HiFocus Helix electrode. The mean age of implantation was 59.7 years
(range: 30-75 years). All subjects had full electrode
insertions, no complications, and were full-time
users of their devices throughout the study. For all
subjects, data were collected intraoperatively, at
initial stimulation, and at the one-month postoperative interval. In addition, seven subjects provided
data at the three-month postoperative interval.
ECAPs were measured using the Neural Response
Imaging (NRI) software and standard NRI procedures. NRI responses were measured for stimulation
on electrodes 2, 8, and 14, and tNRIs were determined. The tNRI is the x-intercept of a regression
line fit to selected points of the ECAP growth
function. For measuring M levels and ESRTs,
Speech Bursts were delivered to four banded channels (electrodes 1-4, 5-8, 9-12, and 13-16). Speech
Bursts consist of white noise passed through the
same filters and envelope detectors that are used
58
Advanced Bionics® Auditory Research Bulletin 2005
when processing sound during everyday HiRes
use. The spectral and temporal characteristics of
Speech Bursts are more representative of real sounds,
unlike the slow pulse trains conventionally used for
setting threshold and comfort levels. ESRTs were
measured during postoperative sessions by delivering the electrical stimulus to the implant while
recording the stapedial reflex from the contralateral
ear with standard clinical impedance equipment.
During surgery, ESRTs were determined visually.
NRI and ESRT data were obtained successfully in
almost all cases. Figure 1 shows a comparison of
the intraoperative objective measures to psychophysical results at initial stimulation. ESRTs were
higher and more variable than either the tNRI
or the psychophysical M levels. Figure 2 shows
results over time for the seven subjects who had
reached the three-month test interval. These results
indicate that ESRTs tended to change more over
time than M levels or the tNRI. Note, however,
that the higher ESRTs intraoperatively may be a
result of using visual detection compared to the
impedance-bridge method used postoperatively.
Figure 1. Means and standard deviations for intraoperative
ESRTs and tNRIs, and initial-stimulation M levels. All levels are
in HiRes charge units (Clinical Units).
In summary, ESRTs and NRI responses can be
obtained in a short amount of time in the operating room in most patients. Intraoperative ESRTs
and tNRI measures appear to be useful in guiding
program settings at initial stimulation. As a general
guideline, initial stimulation M levels should be
below intraoperative ESRT levels and near intraoperative tNRI levels for patients using HiRes.
References
Brown CJ. (2003) Clinical uses of electrically evoked auditory nerve and brainstem
responses. Curr Opin Otolaryngol Head Neck Surg 11(5): 383-387.
Figure 2. ESRTs, tNRIs, and M levels over time for seven
patients. Levels are shown in HiRes Clinical Units relative to
M levels at initial stimulation.
Hodges AV, Butts SL, King JE. (2003) Electrically evoked stapedial reflexes: utility
in cochlear implant patients In Cullington HE, ed. Cochlear Implants: Objective
Measures, London: Whurr Publishers, pp. 81-93.
Objective Measures
59
Neural Response Imaging: Evolution of Thresholds
and Their Relationship to HiResolution M Levels
Patrizia Mancini, M.D.
Chiara D’Elia, M.D.
M. Barbara, M.D., Ph.D.
Roberto Filipo, Prof., M.D.
University of Rome La Sapienza, Rome, Italy
The aim of this study was to investigate the evolution of NRI thresholds over time in order to
confirm the utility of intraoperative NRI measurements as a guide for setting postoperative program
parameters, especially for children during their
first fitting sessions (see Frijns et al, 2002).
The study group consisted of 36 adults implanted with
a CII device, three adults implanted with a HiRes
90K, and 22 prelinguistically deafened children (19
implanted with a CII, 3 implanted with a HiRes
90K). The mean age at implant for the children was
3.5 years (range 2-10 years). All subjects were High
Resolution (HiRes) users. Determination of HiRes
most comfortable (M) levels and threshold (T) levels
was typically carried out at initial stimulation, and
subsequently at one, three, six, and 12 months after
initial stimulation. All subjects were fit using the
following program variables: 12-16 active channels,
pulse width between 11 and 32 µs, and average pulse
rate of 2,017 pulses per second per channel (pps/ch).
Table 1. Relationship between M levels and tNRI
across the electrode array at three time periods.
Adults
Children
Comparison
Apex
Mid
Base
Apex
Mid
Base
Initial M vs
Intraop tNRI
92
[53]
63
[21]
106
[44]
62
[13]
50
[18]
55
[17]
12-month M vs
Intraop tNRI
103
[51]
84
[39]
124
[48]
69
[30]
64
[15]
86
[25]
12-month M vs
12-month tNRI
110
[52]
99
[59]
143
[65]
91
[40]
96
[14]
86
[18]
The numbers represent M levels as a percentage of the tNRI
values. Corresponding standard deviations are in brackets.
60
NRI measurements were carried out during surgery
and then repeated during postoperative sessions.
For adult subjects, recordings were carried out at
initial stimulation (one month after surgery) and
then after 3 and 12 months. Behavioral thresholds
also were measured at the postoperative sessions.
In children, postoperative NRI evaluations were
carried out at initial stimulation and at the 12month follow-up appointment. For both adults and
children, three stimulation electrodes were tested
(electrodes 3, 9, and 15) corresponding to the apical,
mid, and basal areas of the cochlea, respectively.
Thresholds of the NRI response, termed tNRI, were
determined for each electrode. Where no threshold
could be recorded, the next recording electrode was
substituted. The default recording electrodes were
two contacts apical to the stimulation electrode. A
biphasic pulse with a 32 µs per phase duration was
delivered at a rate of 29 pps. Stimulus intensity never
exceeded 800 µA to avoid any possible damage of
the nerve or surrounding structures. The tNRI was
visually confirmed, and narrow stimulus amplitude
steps were used (16 µA) for greater precision. The
tNRI and HiRes M levels were converted into
Clinical Units (CU = pulse width x µA x 0.0128447)
in order to produce a common unit for determining
the relationships among behavioral measures and
tNRI over time. The goal was to predict the levels
to be used as a starting point during initial stimulation through examining the relationships among the
psychophysical and objective measures over time.
Some of the comparisons between tNRI values and
M levels are listed in Table 1.The tNRI was measurable in 95.5 % of children (21/22) and in 81.5 %
of adults (32/39). The tNRI values decreased over
time, and a statistical significance was found both
in the adult (p < .001) and pediatric (p < .02) study
Advanced Bionics® Auditory Research Bulletin 2005
“...this preliminary study demonstrates
how intraoperative NRI thresholds
may provide useful landmarks
for guiding
cochlear implant programming.”
groups (Figures 1 and 2). In contrast, M values
showed a trend of increasing values over time
both for adults (p < .01) and children (p < .001).
Neural telemetry is routinely used now for obtaining
objective threshold values, estimating the degree of
neural survival, and confirming function at the neural
interface. As can be seen from these preliminary
data, tNRI values tend to decrease over time. In fact,
the relationship between M levels and tNRI values
also tends to give lower percentages in the intraoperative measurement session, particularly for the
pediatric study group. Specifically, M levels are 56%
and 87% of tNRI for children and adults, respectively. Percentages obtained in children are lower
compared to adults both in intraoperative and postoperative sessions. These results might be influenced
both by differences in neural survival (presumed to
be higher in children) and also by the reliability of
behavioral responses in children. Furthermore, in
order to avoid excessive loudness, NRI recordings
were not carried out for all children, thereby possibly
introducing a bias in the resulting comparisons.
HiRes M levels increased and stabilized during the
first year of implant use, which also influenced the
relationships between tNRI values and M levels.
In sum, this preliminary study demonstrates
how intraoperative NRI thresholds may provide
useful landmarks for guiding cochlear implant
programming. Further research is required to
establish
dependable
relationships
between
NRI thresholds and behavioral M levels during
fitting of the speech processor in children.
Figure 1. tNRI values and M levels over time for three
electrodes in adults.
Figure 2. tNRI values and M levels over time for three
electrodes in children.
Reference
Frijns JHM, Briaire JJ, de Laat J, Grote JJ. (2002) Initial evaluation of the Clarion CII
cochlear implant: speech perception and Neural Response Imaging. Ear Hear 23:
184-197.
Objective Measures
61
Exploring Changes in the “1st NRI” Over Time
Miriam Adler, M.A.
Cahtia Adelman, M.Sc.
Haya Levi, M.A.
Hadassah University Hospital, Jerusalem, Israel
“Most subjects showed
stable or lower thresholds
between test sessions
for at least one electrode.”
Neural response imaging (NRI) is a software tool
that can be used to support the fitting of CII and
HiRes cochlear implant recipients. NRI uses telemetry with standard speech processors to elicit and
record electrically evoked compound action potentials (ECAPs). NRI can be easily administered intraand postoperatively. NRI measurements are useful in
creating programs for very young or uncooperative
patients and provide a means of monitoring the
long-term physiological responses for all patients.
The objective of this study was to explore changes
in NRI responses over time. Retrospective comparisons of NRI response thresholds across test sessions
were made using “1st NRI” values, defined as the
lowest stimulus levels that elicit a visually detectable ECAP. NRI measurements were made with
the Platinum Sound Processor using the SoundWave fitting software. Included in the study were
10 implant recipients in our clinic population who
had completed at least two evaluation sessions and
for whom responses were obtained on one to four
electrodes. The study group included six children,
three teenagers, and one adult. The etiologies were
varied, including genetic (four subjects), meningitis
(one subject), and unknown causes (five subjects).
For each recording, the 1st NRI was determined by
independent judgments of experienced observers. The
time interval between recordings ranged from three
to 13 months. In three subjects, the first recording
was obtained intraoperatively. NRI recordings from
different sessions were compared. If the two 1st NRI
values were within an increment of each other, the
responses were rated the same. If the 1st NRI values
differed by more than one increment, the responses
were rated as different. Stimuli were delivered on
electrode E3, E7, E11, and E15 so that there was
one channel stimulated for each Speech Burst group.
62
Advanced Bionics® Auditory Research Bulletin 2005
A greater number of recordings was obtained
from the apical electrodes because the recordings were more robust than those obtained
elsewhere along the electrode array. As shown
in Table 1, most thresholds at the apical electrode (E3) remained stable across test sessions.
Most subjects showed stable or lower thresholds
between test sessions for at least one electrode.
The exception was the one subject with meningitis whose values increased on all electrodes with
time. In the three subjects with intra- and postoperative recordings, the values for E3 remained
stable between test intervals of three to five months.
In one of these three patients, E7 and E15 also
remained stable. In the other two patients, the
postoperative values were lower on E7, E11, and
E15. Neither the length of the time between
sessions nor age of implantation affected the results.
It may be speculated that more robust (and more
stable) recordings were obtained from the apical
channels because subjects typically had some lowfrequency residual hearing preoperatively—and
thus greater neural reserves for responding to electrical stimulation. Such speculation is supported
by Nadol (1997), who reported that histological quantification of surviving spiral ganglion
cells in profoundly deaf subjects indicated that
neural degeneration was more severe in the basal
turn compared to the apical turn of the cochlea.
Table 1. Trends in comparison of NRI values over time.
Channels
Stim (Record)
No Change
in Values
Increase
in Values
Decrease
in Values
E3 (1)
7
2
1
E7 (5)
4
1
3
E11 (9)
2
1
4
Number of responses showing no change, increase, or
decrease in values between test sessions. More responses are
obtained at the apex (E3) where responses were more robust.
Reference
Nadol JB Jr. (1997) Patterns of neural degeneration in the human cochlea and
auditory nerve: implications for cochlear implantation. Otolaryngol Head Neck
Surg 117:220-228.
We are currently enrolled in a multicenter study
exploring tNRI responses over longer periods of
time. We anticipate that comparisons between tNRI
and 1st NRI results will yield interesting findings.
Objective Measures
63
Relationship Between Electrical Stapedial Thresholds and
HiRes Programming Levels
Lisa Buckler, M.A.
Kristen Dawson, M.A.
Midwest Ear Institute, Kansas City, MO, USA
“...speech burst-elicited ESRTs
provide an excellent guide
for predicting HiRes M levels...”
Objective measures are becoming standard tools
to assist in programming implant speech processors. One useful measure is the electrical stapedial
reflex threshold (ESRT). Previous studies have
shown a relationship between ESRTs and behavioral comfort levels for single-electrode stimulation
(e.g., Hodges et al, 2003; Stephan & Welzl-Müller,
2000). The HiRes system, however, is programmed
using Speech Bursts, which consist of white noise
passed through the same filters and envelop detectors that are active in everyday HiRes use. The
spectral and temporal characteristics of these
speech-burst stimuli are more representative of real
sounds, unlike the slow pulse trains conventionally
used for setting threshold and comfort levels (and
used in previous ESRT studies). The purpose of this
research was to examine the relationship between
ESRTs and behaviorally measured comfort levels
for electrical stimulation using the Speech Bursts
feature of the SoundWave programming software.
Subjects included 18 adults who were implanted
with the HiRes system (CII or 90K device) and who
used HiRes sound processing. ESRTs were obtained
using Speech Bursts delivered to four groups of four
adjacent electrodes. For each group of four electrodes,
the stimulus amplitude was increased until a reflex
was observed. Then, a bracketing technique was used
to find the ESRT. Each patient’s program was activated using those ESRT levels. The volume then was
adjusted until the patient reported a most comfortable level for speech. The final comfort levels, or M
levels, then were compared with the original ESRT
levels that were obtained during the same test session.
64
Advanced Bionics® Auditory Research Bulletin 2005
���
Figure 2 shows the psychophysical loudness rankings for speech burst stimuli presented at the
same levels that elicited the ESRTs. The loudness
rankings for the 18 listeners ranged between 6
(most comfortable) and 8 (loud but comfortable).
���
����������������������
Figure 1 shows the ESRTs and M levels for the
18 subjects. There are at most four data points for
each subject, corresponding to the four groups of
stimulated electrodes (some patients had stapedial
reflexes for some electrode groups but not others).
The solid line represents a perfect correlation (r =
1.0). Data above the line indicate an ESRT higher
than the M level, whereas data below the line
indicate an ESRT below the M level. The solid
line is the best-fit line to the experimental data.
The correlation coefficient r for the best-fit line
is 0.97, indicating an almost perfect relationship
between ESRTs and M levels in these subjects.
���
���
���
���
���
�
� ��
����
����
����
����
����
����
���
�������������������
Figure 1. ESRTs and M levels (CU = Clinical Units) for 18 adult
subjects . There is a strong relationship between the stimulus
levels eliciting ESRTs and the stimulus levels judged as most
comfortable.
Taken together, these data indicate that speech
burst-elicited ESRTs provide an excellent guide
for predicting HiRes M levels. Thus, these ESRT
measures may be useful in programming HiRes
for young children and other patients who
cannot provide reliable behavioral responses.
References
Hodges AV, Butts SL, King JE. (2003) Electrically evoked stapedial reflexes: utility
in cochlear implant patients. In Cullington HE, ed. Cochlear Implants: Objective
Measures. London: Whurr Publishers, 81-93.
Stephan K, Welzl-Müller K. (2003) Post-operative stapedius reflex tests with
simultaneous loudness scaling in patients supplied with cochlear implants.
Audiology 39:13-18.
Figure 2. Psychophysical loudness ranking for four bands of
Speech Bursts presented at levels corresponding to the stimulus
levels that elicited ESRTs (6 = “most comfortable” and
8 = “loud but comfortable”).
Objective Measures
65
Predictive Value of Intraoperative ESRTs
for Setting Postoperative HiRes M Levels
Rolf-Dieter Battmer, Prof., Ph.D.
Martina Brendel, M.Sc.
Andreas Büchner, Ph.D.
Carolin Frohne-Büchner, Ph.D.*
Thomas Lenarz, Prof., M.D., Ph.D.
Medizinische Hochschule Hannover, Hannover, Germany
* also with Advanced Bionics Corporation, Europe
HiResolution (HiRes) sound processing has been
clinically available since 2001. It uses 16 channels with stimulation rates up to 83,000 pulses
per second (pps). HiRes has been evaluated in
clinical studies with adult users, and the results
demonstrated improvements in speech perception
compared to standard strategies (SAS, CIS and
PPS) (Koch et al 2004; see also Büchner et al at
page 98 of this bulletin). One of the new features
of HiRes is the option to fit M levels with Speech
Bursts, consisting of white noise passed through the
same filters and envelope detectors that are used
when processing sound during everyday HiRes use.
Speech Bursts allow up to four channels to be stimulated at the same time and are more representative
of real sounds heard in everyday life compared to
the standard approach of stimulating one channel
at a time with a tone burst. Although the HiRes
software provides tools for fitting the implant with
a small number of measurements, further improvements to the fitting process are possible if objective
measures can be used for estimation of M levels.
� ��
��
��
��
� ��
��
��
� ��
� �
��
�
��
��
��
��
��
��
�
��
��
��
��
�
��
��
��� ��� ��� ��� ��� ��� ��
��
��� ��� ��� ��� ��� ��� ��
��
�
�
�
�
Figure 1. SoundWave allows ESRTs and M levels to be measured using tone bursts delivered to each channel separately
(top) or by using Speech Bursts delivered to 3-4 channels at
the same time (bottom). In this study, seven different groups of
Speech Bursts were used to elicit the ESRTs (bottom).
66
This evaluation concentrated on using electrically evoked stapedius reflex thresholds (ESRTs)
as a guide for setting M levels. In the past, it
has been shown that there is a good correlation between single-channel ESRTs and singlechannel M level measurements (Battmer et al,
1990). The objective of this restrospective study
was to record ESRTs intraoperatively using speech
burst stimulation and to investigate the predictive value for determining postoperative M levels.
To record ESRTs, stimuli were presented via SoundWave using a sequential 16-channel program with
default settings. The stimuli were seven different
speech burst groupings, shown graphically in
Figure 1. Each speech burst grouping consisted of
filtered noise presented to four channels (or three
channels if an odd number of electrodes were switched
off ). The speech burst groups started from the basal
end of the array by stimulating the first four (or
three) basal electrodes. The next speech burst group
contained four (or three) channels that had been
shifted apically by half the number of electrodes in
the band. Measurements were taken intraoperatively
in 20 adults, and the results were analyzed thereafter.
Postoperative fitting was performed using normal
clinical procedures. The M levels used in the analysis
were the means for the (four/three) channels within
each of the seven different speech burst groups. Each
group contained the same channels as those groups
shown for the ESRT measurements in Figure 1. The
clinician was unaware of the results of the intraoperative ESRT measurements to avoid any bias.
The intraoperative ESRT values were correlated
to the M levels used in each subject’s 16 channel
sequential program three months after first fitting.
On average, the speech burst M levels were at 80%
of the speech burst ESRT values. Of the 20 subjects,
Advanced Bionics® Auditory Research Bulletin 2005
two were outliers. Both of these subjects had severe
hearing impairment since early childhood. Their
data were removed from the data set. The subsequent
analysis then showed a correlation coefficient of
r = 0.53 between the M levels used in the program
at three months and the intraoperative ESRTs.
“...using speech burst stimulation
not only decreases fitting time
but also permits rapid measurement
of the ESRT profile along the array.”
The profiles of the ESRTs were compared to the
profiles of the M levels. To characterize the profiles
of the M levels versus the ESRTs, both were shifted
by their means, i.e. both averages were set to zero.
Then the differences in values between the corresponding ESRTs and M levels for each of the seven
speech burst groups were calculated. A criterion
value of 20 CU was used because our exploratory
studies showed that a difference of 20 CU in stimulation current is not noticeable in live speech for
the majority of implanted subjects. The number of
speech burst groups per subject (out of a total of 7
groups for each of the 18 subjects) were counted
where the difference between the profiles exceeded
20 CU. The resulting proportion of subjects with
the respective number of groups exceeding 20 CU
is shown in Figure 2. In general, the profiles of
the ESRTs fit well to the profiles of the M levels.
In summary, using speech burst stimulation not
only decreases fitting time but also permits rapid
measurement of the ESRT profile along the array.
The absolute ESRT values give a reasonable indicator for the postoperative M levels, but the profile
is an even more reliable predictor of M levels. For
fitting children, the clinician can start with the
ESRT profile at very low levels and can then increase
M levels based on behavioural reactions of the child.
Figure 2. Proportion of subjects (out of a total of 7 groups x 18
subjects) where the intraoperative ESRTs and postoperative M
levels differed by more than 20 CU.
References
Battmer R, Laszig R, Lehnhardt E. (1990) Electrically elicited stapedius reflex in
cochlear implant patients. Ear Hear 11(5):370-374.
Koch DB, Osberger MJ, Segel P, Kessler DK. (2004) HiResolution and conventional
sound processing in the HiResolution Bionic Ear: using appropriate outcome
measures to assess speech recognition ability. Audiol Neurootol 9:214-223.
Objective Measures
67
Electrically Evoked Brainstem and Cortical Potentials:
Relationships to Fitting Parameters and Performance
Jeanne Guiraud, M.Sc.
Lionel Collet, Prof., M.D., Ph.D.
Université Claude Bernard Lyon 1
Hôpital Edouard Herriot
Lyon, France
Eric Truy, Prof., M.D.
Hôpital Edouard Herriot, Lyon, France
Laure Arnold, M.Sc.
Patrick Boyle, M.Sc.
Advanced Bionics Corporation, Europe
The overall objective of this study was to explore
the relationships between electrically evoked
potentials recorded from the brainstem and the
cortex with stimulation parameters and behavioral
measures obtained for adult cochlear implant users.
Ten adults using either CII or HiRes 90K cochlear
implants for a minimum of one month participated in this study. The electrical auditory brainstem response (EABR, waves III and V) and the
electrical late auditory response (ELAR, N1-P2
complex) were recorded from electrodes 3 (apex),
7, 11, and 15 (base) at perceptual threshold (T)
and at 40%, 70%, and 100% of dynamic range—
between T and the upper limit of comfortable
loudness (ULCL). Each subject’s dynamic range
and most comfortable (M) levels were determined behaviorally with the same stimulation
parameters used to obtain the evoked potentials.
Phoneme intelligibility was evaluated using the
French Lafon words lists (12 lists, each list containing
17 words, and each word consisting of three
phonemes) presented at 60 dB SPL in quiet and
noise (S/N +10 dB). Analysis of variance (ANOVA)
for repeated measures and correlational analyses
were used to determine statistical significance.
EABRs could be recorded in 9 of the 10 subjects,
while ELARs could be recorded from all subjects.
The recordings in Figure 1 for one patient typify
the data collected. No differences between potentials from CII and HiRes 90K users were noticed.
68
Advanced Bionics® Auditory Research Bulletin 2005
Significant correlations in the data were the following:
• The higher the intensity of stimulation, the
shorter the latencies of wave V and complex
N1-P2 latencies, whose amplitudes also
increased with louder stimulation.
“...the nerve structures involved
below the brainstem (wave V)
are very important to the
speech perceptual abilities
of cochlear implant users.”
• The shorter the latencies of waves III and V
obtained at comfortable but loud stimulation,
the lower the perceptual threshold.
Also seen in the data was a relationship between
evoked responses and phoneme perception with
best performers having shorter wave V, interpeak III-V, and complex N1-P2 latencies.
From this study we have gained a better understanding as to the influence of stimulation level and
electrode location on EABR and ELAR responses.
We have also demonstrated that the nerve structures involved below the brainstem (wave V) are
very important to the speech perceptual abilities of
cochlear implant users. Moreover, the EABR wave
V responses may be useful clinically in programming T levels. Further investigation is needed in a
larger sample to confirm these preliminary findings.
Figure 1. EABRs (left) and ELARs (right) recorded at perceptual threshold (T) and at 40%, 70%, and 100% of the dynamic range
from electrode 3 in one subject.
Objective Measures
69
The Relationships of ECAP and EABR Thresholds
to Hearing and Speech Perception
in Cochlear Implant Users
Dietmar Basta, M.D., Ph.D.
Unfallkrankenhaus Berlin, Berlin, Germany
Andreas Dahme, Ph.D.
Hearing-Therapy-Center, Potsdam, Germany
Ingo Todt, M.D.
Arne Ernst, Prof., M.D.
Unfallkrankenhaus Berlin, Berlin, Germany
The ECAP and EABR thresholds obtained during
cochlear implant stimulation are useful for speech
processor fitting in very young or noncompliant
patients. However, the correlation between physiological data and single-electrode psychophysical
measurements is very poor in some patients for
reasons that are not clearly understood. It has been
speculated that ECAP and EABR measurements,
which reflect function at the lower levels of the
auditory pathway, depend on individual anatomical
and physiological properties and their development or rehabilitation. In contrast, psychophysical
tasks require loudness (behavioural) judgments
that involve higher levels of the auditory system.
The aim of the present study was to investigate the
correlation between objective single-electrode electrical evoked thresholds and the psychophysically
determined current levels used for device programming. A group of 22 subjects participated in this
study. ECAPs were measured intraoperatively using
hardware and software provided by Advanced Bionics
or Cochlear Ltd. EABR thresholds were obtained
one year postoperatively. Psychophysical T and M or
C levels were obtained six weeks, three months, and
one year after surgery. All subjects were capable of
providing reliable loudness judgment during device
programming. Speech perception (monosyllabic
words presented at 65 dB SPL) and pure tone tests
were performed at three months and one year postoperatively. Pure tone responses were averaged (1000,
2000, and 4000 Hz) for use in subsequent analyses.
Subjects were divided into two groups—good
performers and poor performers—based on their
speech perception performance at three months
postoperatively. Correlation coefficients were calcu-
70
Advanced Bionics® Auditory Research Bulletin 2005
lated between ECAP data and program levels
obtained six weeks and three months postoperatively. The group that attained high speech perception scores within a short period of time (“good
performers”) showed a significantly higher correlation between ECAP thresholds and program levels
at six weeks and three months than the poorer
performing group, as shown in Figure 1. Although
not displayed in the figure, the good performers also
showed significantly lower pure tone thresholds than
the poorer performers at the three-month interval.
The different correlation coefficients between good
and poor performers were not influenced by other
factors like age, duration of deafness, or hearing aid
use (tested by discriminant analysis). One year postoperatively, the correlation between EABR thresholds and psychophysical measurements did not
differ between good and poor performers. In addition, the pure tone thresholds for the two groups
were no longer different at the one-year test interval.
In summary, these results suggest a relationship between the status of lower and higher levels
of the auditory pathway and the development of
speech recognition skills with a cochlear implant.
However, continued improvement in speech
recognition will be more influenced by higher
level speech (and cognitive) processes that are
not assessed with “threshold” measurements estimated by pure tone audiometry, EABR, or ECAP.
“...these results suggest a relationship
between the status of
lower and higher levels of
the auditory pathway
and the development of
speech recognition skills with a
cochlear implant.”
Figure 1. Correlation coefficients between ECAP thresholds and
psychophysical comfort levels (tNRT-c) and thresholds (tNRT-t)
in “good” and “poor” performers. Each bar represents the mean
correlation coefficient of 242 electrodes. All differences shown
between the patient groups were significant (p < 0.05).
Objective Measures
71
Use of Single and Multi-Electrode Stimulation
to Compare ECAP and Behavioral Programming Levels
Jill B. Firszt, Ph.D.
Christina L. Runge-Samuelson, Ph.D.
P. Ashley Wackym, M.D.
Medical College of Wisconsin, Milwaukee, WI, USA
Optimal programming of the HiResolution sound
processing strategy remains under investigation.
Psychophysical measures of loudness incorporated in speech processor programs can be made
using either single or banded (e.g., four electrodes
activated simultaneously) electrode stimulation.
Because multiple electrodes are typically stimulated when a cochlear implant recipient listens
to speech, obtaining measures of loudness with
single electrodes may overestimate the current
levels needed during everyday communication.
Determining behavioral measures of loudness can
be difficult for cochlear implant patients, especially in young children as well as some adults.
The electrically evoked compound action potential (ECAP) is a direct measure of auditory nerve
Figure 1. Behavioral loudness growth functions for single electrodes 5, 6, 7, 8 and banded electrodes 5-8 for one subject.
The x-axis displays stimulus level and the y-axis shows behavioral measures of loudness where values of 1 to 3 represent
threshold to maximum comfort level.
72
activity in response to an electrical signal (Abbas
et al, 1999) and has been used to assist with optimization of speech processor settings. Single-electrode ECAP thresholds have been shown to fall
within the behavioral electrical dynamic range, that
is, between psychophysical measures of threshold
and comfortable loudness (Brown et al, 2000;
Franck & Norton, 2001; Hughes et al, 2000). The
ECAP also can be recorded to stimuli delivered
simultaneously to multiple electrode contacts. The
purpose of this study was to determine the utility
of (1) multi-electrode banded stimulation to obtain
behavioral measures used for programming and
(2) electrophysiological responses (the ECAP).
A within-subject repeated-measures design was used
to compare data collected with multi- and singleelectrode stimulation. Subjects were profoundly
hearing-impaired children and adults who had
received the Clarion CII device. For each subject,
behavioral measures of threshold (T) and comfortable loudness (M) were made on electrodes 5-8
using banded as well as single-electrode stimulation.
For the same electrode groups, ECAP input/output
functions (I/O) were obtained for single- and multielectrode stimulation within the behavioral dynamic
range. The stimulation rate was 29 Hz for both
behavioral and ECAP measures. Stimuli were 32
µs/phase biphasic pulses with a monopolar configuration. For ECAP measures, the alternating polarity
paradigm was used for stimulus artifact rejection.
Within subjects, the single-electrode behavioral
measures of loudness were similar for adults and
children. When electrodes were banded, loudness measures were present at consistently lower
levels. Figure 1 shows an example of a behavioral
loudness growth function for single electrodes
compared to banded electrodes for one adult subject.
Advanced Bionics® Auditory Research Bulletin 2005
“...because multiple electrodes
are activated in response to speech,
the use of banded
rather than single-electrode measures
may better approximate current levels
needed for everyday listening.”
Compared to single electrodes, the banded I/O
functions were present at lower levels and showed
steeper growth. Figure 2 shows an example of the
ECAP I/O functions obtained for single and
banded electrodes within the behavioral dynamic
range for the same adult subject. The slopes of
the I/O functions for the banded electrodes were
approximately four times steeper than for the
single electrodes. This result indicates a cumulative effect on the ECAP amplitude with increases
in stimulus level for the banded condition.
In conclusion, because multiple electrodes are
activated in response to speech, the use of banded
rather than single-electrode measures may better
approximate current levels needed for everyday
listening. One potential clinical advantage of using
banded stimuli is that less time will be required
to adjust speech processor settings because, typically, M levels are overestimated when based on
single-electrode measures. A second potential
benefit may be improved optimization of individual processor programs by combining behavioral and physiologic (i.e., ECAP) responses.
Figure 2. ECAP I/O functions for single electrodes 5, 6, 7, 8
and banded electrodes 5-8 for one subject. The x-axis displays
stimulus level and the y-axis shows ECAP amplitude. The
legend indicates the stimulating and recording electrode for
each test measure (for example, e5-4 = stimulating electrode 5,
recording electrode 4).
Acknowledgements
Supported by intramural funds from the Department of Otolaryngology and
Communication Sciences, Medical College of Wisconsin and Advanced Bionics
Corporation.
References
Abbas PJ, Brown CJ, Shallop JK, Firszt JB, Hughes ML, Hong SA, Staller SJ. (1999)
Summary of results using the Nucleus CI24M implant to record the electrically
evoked compound action potential (EAP). Ear Hear 20(1):45-59.
Brown CJ, Hughes ML, Luk B, Abbas PJ, Wolaver A, Gervais J. (2000) The
relationship between EAP and EABR thresholds and levels used to program the
Nucleus 24 speech processor: data from adults. Ear Hear 21(2):151-163.
Franck KH, Norton SJ. (2001) Estimation of psychophysical levels using the
electrically evoked compound action potential measured with the neural response
telemetry capabilities of Cochlear Corporation’s CI24M device. Ear Hear
22(4):289-299.
Hughes ML, Brown CJ, Abbas PJ, Wolaver AA, Gervais JP. (2000) Comparison of
EAP thresholds with MAP levels in the Nucleus 24 cochlear implant: data from
children. Ear Hear 21(2):164-174.
Objective Measures
73
Effects of Simultaneous Stimulation of Multiple Electrodes
on Behavioral and ECAP Thresholds
Christina L. Runge-Samuelson, Ph.D.
Jill B. Firszt, Ph.D.
P. Ashley Wackym, M.D.
Medical College of Wisconsin, Milwaukee, WI, USA
“...the simultaneous monopolar
stimulation of multiple electrodes
affects both behavioral and ECAP
thresholds—depending on the
number of electrodes stimulated
and the distance between them.”
Traditionally, loudness mapping for cochlear
implants is conducted by obtaining psychophysical
measures, e.g., threshold and maximum comfortable
levels, on single electrodes. Advances in cochlear
implant technology increase the possibilities for
simultaneous presentation of electrical stimulation to multiple electrodes as a viable programming
technique. However, loudness percepts for such
complex stimuli may be affected by several factors,
including the numbers of electrodes stimulated
simultaneously and the distances between stimulated electrodes. Using bipolar stimulation, several
studies have examined the effects of systematically
broadening the stimulus current field by increasing
the distance between the active and reference electrodes. In general, thresholds decrease as electrode
separation is increased. This has been demonstrated
psychophysically in macaque monkeys (Pfingst,
1989) and Nucleus 22 users (Chatterjee, 1999;
Pfingst et al, 1995), and with the EABR in Nucleus
22 and Ineraid users (Abbas & Brown, 1991).
In recent years, we have employed software from
Advanced Bionics to investigate the clinical feasi-
74
bility of using monopolar, simultaneous electrode stimulation to elicit the electrically evoked
compound action potential (ECAP) and to measure
loudness growth in Clarion CII users. Potentially,
this method may prove to be a more time-efficient
programming method compared to conventional,
single-electrode electrophysiological and psychophysical measures. However, the relationship
between single- and banded-electrode responses is
unclear. The research software allows recordings of
the ECAP to simultaneous, multi-electrode stimulation—thereby allowing comparisons between
single and multielectrode stimulation for both electrophysiologic responses and behavioral loudness
perception. The goals of this study were to examine
the effects of number and separation of electrodes
stimulated on behavioral and ECAP thresholds.
Subjects in this study included 10 adult Clarion CII
users. Stimuli were monopolar, 32 µs/phase biphasic
pulses presented at a rate of 29 Hz and a minimum
current step size of 8 µA. (The artifact rejection
method used was alternating polarity.) Loudness
growth, from threshold to maximum comfort level,
was measured on each electrode as well as on electrodes banded in various combinations. The banding
combinations included sets of two, three, and four
electrodes at varied distances (up to 3.3 mm apart),
yielding 15 test conditions in total per subject. ECAPs
for each subject were measured at stimulation levels
spanning the subject’s behavioral dynamic range
for each banding combination. ECAP threshold
was defined as 0.04 mV (noise floor ~0.02 mV).
The data showed that, for both behavioral and ECAP
measures, increases in the number of electrodes stimulated resulted in threshold decreases. For all subjects,
the threshold decreases were fit with the exponential
Advanced Bionics® Auditory Research Bulletin 2005
decay function y = y0 + ae-bx (where y is threshold, x
is the number of electrodes, b is the decay factor, and
y0 and a are constants) with correlation coefficients
≥ 0.98.
Figure 1 shows an example from one subject (S7).
Although there is some offset between the behavioral and ECAP thresholds, the exponential decay
factors for both measures were the same (b = 0.94).
In general, the results for electrode separation distance
indicate that as separation distance between banded
electrodes increased, behavioral and ECAP thresholds also increased, as exemplified in the results for a
single subject (S7) in Figure 2. It was noted, however,
that in some subjects an electrode with a relatively
lower threshold tended to decrease the banded
threshold when it was part of the grouped electrodes.
Figure 1. Behavioral (black diamond) and ECAP (gray square)
thresholds for number of electrodes banded for Subject 7.
The x-axis displays the number of electrodes used for banded
stimulation.
In conclusion, the simultaneous monopolar stimulation of multiple electrodes affects both behavioral
and ECAP thresholds—depending on the number
of electrodes stimulated and the distance between
them. These issues need to be considered when using
the banded stimulation method in a clinical setting.
Acknowledgements
Supported by intramural funds from the Department of Otolaryngology and
Communication Sciences, Medical College of Wisconsin and Advanced Bionics
Corporation.
Figure 2. Behavioral (diamonds) and ECAP (squares) thresholds
as a function of electrode separation distance for Subject 7.
References
Abbas PJ, Brown CJ. (1991) Electrically evoked auditory brainstem responses:
Growth of response with current level. Hear Res 51(1):123-137.
Chatterjee M. (1999) Effects of stimulation mode on threshold and loudness growth
in multielectrode cochlear implants. J Acoust Soc Am 105(2):850-860.
Pfingst BE. (1989) Psychophysical constraints on biophysical/neural models of
threshold. In: Miller JM, Spelman FA, eds. Cochlear Implants: Models of the
Electrically Stimulated Ear. New York, NY:Springer-Verlag, 161-183.
Pfingst BE, Miller AL, Morris DJ, Zwolan TA, Spelman FA, Clopton BM. (1995)
Effects of electrical current configuration on stimulus detection. Ann Otol Rhinol
Larngol Suppl 166:127-131.
Objective Measures
75
Banded Neural Response Imaging (NRI):
Preliminary Results
Jeanne Guiraud, M.Sc.1,2
Eric Truy, Prof., M.D.2
Laure Arnold, M.Sc.3
Patrick Boyle, M.Sc.3
Lionel Collet, Prof., M.D., Ph.D.1,2
1 Université Claude Bernard Lyon 1, Lyon, France
2 Hôpital Edouard Herriot, Lyon, France
3 Advanced Bionics Corporation, Europe
“Compared to single-channel NRI,
banded NRI may provide
a more efficient clinical tool
for setting programming levels...”
Single-channel NRI consists of measuring ECAPs
with stimulation on one electrode while recording
(typically) from an electrode located two electrodes away in the apical direction. (For example,
stimulation on E3 would be recorded on E1.)
It has been shown that thresholds for singlechannel NRI-elicited ECAPs fall within the electrical dynamic range (e.g., Novak et al, 2003).
For banded-NRI measurements, three or four
consecutive electrodes are stimulated simultaneously, and ECAPs are recorded (typically) on an
electrode located two electrodes away in the apical
direction. (For example, stimulation on E13-16,
would be recorded on E11.) Previous studies have
reported that banded ECAPs may be more related
to clinical programming levels because the same
pattern of stimulation (multielectrode Speech Bursts)
76
is delivered by the SoundWave fitting program.
Indeed, Firszt et al (2003) have shown that growth
function slopes are more similar to the loudness
growth responses with banded NRI measurements. They also showed that growth function
curves are steeper and tNRI (threshold NRI) values
lower with banded NRI in a proportional way
depending on the number of electrodes stimulated.
The objectives of this study were to investigate
the relationships between (1) single-channel and
banded NRI measures, (2) banded-NRI measurements and clinical programming parameters, and
(3) banded-ECAP responses and speech perception. We also wanted to examine the effect of
electrode band location on the NRI response.
Thus far, six adult subjects have participated in this
study. The subjects ranged in age from 31 to 66 years
(mean = 49 years), in duration of profound bilateral
hearing loss from 4 to 33 years (mean = 16 years),
and in duration of cochlear implant use from 1 to
36 months (mean = 15 months). The research platform (BEDCS) was used to deliver biphasic, 32 µs
per phase pulses at a rate of 29 Hz. The alternating
polarity approach was used for artifact reduction.
For single-channel NRI measures, the electrodes
of stimulation (and recording) were E4(2), E5(3),
E9(7), and E13(11). For banded NRI recordings, electrode bands (and recording electrode)
were E1-4(6), E5-8(3), E9-12(7), and E13-16(11).
During ECAP measures, behavioral thresholds (T)
and most comfortable loudness (M) levels were
obtained, using a seven-step loudness scale. Speech
perception scores were obtained using the French
Lafon word lists (12 lists, each containing 17 words
and each word consisting of three phonemes) at
60 dB SPL in quiet and in noise (S/N +10 dB).
Advanced Bionics® Auditory Research Bulletin 2005
Single-channel and banded NRI responses were
successfully recorded in all subjects except for one
electrode and its corresponding band in two subjects.
Our results (shown in Figure 1 a-d) replicated
those of Firszt et al (2003). Specifically, banded
tNRI values were lower than single tNRI (ratio
3.7:1) and growth curves were steeper for banded
NRI measures (ratio 4:1). We found that banded
tNRI values were lower for apical electrode bands
compared to other locations along the array. We
also found that banded tNRI values fell within the
behavioral dynamic range and that they more closely
approximated T levels (compared to single-channel
tNRI values). Correlation was found between banded
tNRI and T levels, while no correlation was found
between banded tNRI and speech perception scores.
Results from this study showed that banded NRI
can be reliably recorded. Compared to singlechannel NRI, banded NRI may provide a more
efficient clinical tool for setting programming levels
in that values are obtained in less time and appear
to more closely approximate behavioral T levels.
However, these results are only preliminary and
require further investigation in a larger study sample.
References
Firszt JB, Runge-Samuelson CL, Raulie J, et al. (2003) Comparisons of eCAP
and High Resolution programming levels in the Clarion CII using single and
multielectrode stimulation techniques. Poster presented at the Conference on
Implantable Auditory Prostheses, Pacific Grove, CA, 17-22 August, 2003.
Novak MA, Overstreet EH, Thomas JF, Rotz, LA, Black JM. (2003) EABR and ECAP
thresholds and growth function slopes: correlations with HiResolution program
settings. Poster presented at the Conference on Implantable Auditory Prostheses,
Pacific Grove, CA, 17-22 August 2003.
Figure 1. Input/output functions for banded and singlechannel NRI in one subject along the electrode array: apical
(A), mid-apical (B), mid-basal (C), and basal (D). Growth curves
are steeper and tNRI measures are lower for banded NRI.
Objective Measures
77
Investigation of the Neural Response for a Burst Stimulus
Andreas Büchner, Ph.D.
Carolin Frohne-Büchner, Ph.D.*
Lutz Gärtner, M.Sc.
Martina Brendel, M.Sc.
Timo Stöver, M.D.
Thomas Lenarz, Prof., M.D., Ph.D.
Medizinische Hochschule Hannover, Hannover, Germany
* also with Advanced Bionics Corporation, Europe
Often it has been hypothesized that high stimulation rates, above 1500 pps, introduce a stochastic
response pattern in the auditory nerve. Conceptually,
if individual fibres cannot follow a high stimulation
rate, desynchronisation of the neural response to burst
stimuli should be visible in recorded responses. Several
studies have already shown that the optimal stimulation rate is not necessarily the highest possible rate.
Recently, Büchner et al (2005) reported that
the loudness for a given stimulus level does not
increase monotonically with the stimulation rate;
instead loudness reaches a maximum at a certain
rate that corresponds roughly to the optimal rate.
Neural Response Imaging (NRI) may serve as
an objective indicator of optimal stimulation
rate. The purpose of this study was to record the
response of the auditory nerve following stimulation with a burst signal presented at varied high
rates—implemented with the Bionic Ear Data
Collection System (BEDCS) research platform.
For these experiments, the stimuli were constructed
in a manner similar to those used with the clinical
programming system (SoundWave) for singlechannel stimulation. The timing was controlled by
the pulse width as the only variable parameter. A
12-channel program was created resulting in an
interstimulus interval of 22 times the pulse width.
Stimulus rate was varied between 900 and 3500 pps,
meaning that a pulse width between 45 and 12 µs
and an interstimulus interval between 990 and 260 µs
78
were used. The stimulation level was kept constant
in charge at a level that approximated comfortable
loudness (M level). The stimulus was presented on
channel 7, and the NRI response was measured on
channel 5. Alternating stimulus polarity was used
for artifact reduction. Measurements were taken
starting with one pulse, as with the conventional
NRI technique, and then increasing the number of
pulses to a maximum of 20. Because of loudness
summation in some cases, the number of pulses
often was limited to between three and seven pulses.
Overall, it was found that the response decreases
with increasing number of pulses in the burst.
This decrease is not monotonic but shows some
oscillating effect, e.g., an odd number of pulses
leads to a bigger response than an even number
(Figure 1). Out of the 14 subjects tested, in two
the NRI response could only be recorded for a
single pulse but not for a burst. Twelve subjects
show the alternating pattern reported in other
studies. In eight subjects this zigzag pattern disappears for higher stimulation rates. Preliminary
data for four subjects indicate that the rate range
in which the zigzag pattern disappears may correspond to the optimal stimulation rate (Figure 2).
These findings correspond to a previous study by
Rubinstein et al (1999) reporting on the recorded
neural responses to burst stimuli in animals as well as
for Ineraid subjects. They explained that at stimulation rates above 1000 pps, the individual nerve fibres
can no longer follow each electrical stimulus. Thus,
Advanced Bionics® Auditory Research Bulletin 2005
“Preliminary data for four subjects
indicate that the rate range in which
the zigzag pattern disappears
may correspond to
the optimal stimulation rate.”
for the first stimulus in a burst, all fibres are ready to
fire. For the second pulse, nearly all fibres are in the
refractory period, leading to a very small or no neural
response. For the third stimulus, nearly all fibres are
ready to fire again, leading to a response that falls
between the first two. This mechanism leads to the
zigzag pattern shown in the recorded response.
One can only speculate why for very high rates the
alternating pattern disappears. Maybe differences in
the individual properties of the fibres play a more
important role than for lower rates as the time window
between the stimuli gets smaller. Whereas at lower
rates nearly all fibres are out of the refractory period
for the third pulse, at higher rates the majority may
still not be ready to fire. Further studies are necessary
to investigate the correlation of the optimal stimulation rate and the change in the burst-NRI pattern.
Figure 1. Example of a zigzag NRI pattern evoked by stimulation with a burst signal. (Subject’s history: hearing impaired
since age 12 years; treatment at age of 31 years with ototoxic
medications; implanted at the age of 40 years with a HiRes 90K
one year prior to participation in this study.)
References
Brendel M, Büchner A, Frohne-Büchner C, Gärtner L, Lenarz T. (2005) Evaluation
der Möglichkeit zur Bestimmung der optimalen Stimulationsrate über die Lautheit
im HiRes-System. Paper presented at: 8. Jahrestagung der Deutschen Gesellschaft
für Audiologie, Göttingen, 24–26 February, 2005.
Rubinstein JT, Wilson BS, Finley CC, Abbas PJ. (1999) Pseudospontaneous activity:
stochastic independence of auditory nerve fibres with electrical stimulation. Hear
Res 127:108-118.
Figure 2. Example of a transition from the alternating to a more
complex burst-NRI pattern. (Subject’s history: hearing impaired
since the age of 17 years; onset of deafness at age 40; unknown
etiology; implanted at the age of 50 years with CII three years
prior to the experiment. In a previous crossover study, 2500 pps
was found to be the subject’s optimal stimulation rate.)
Objective Measures
79
Psychophysical versus Physiologic Forward Masking
in Cochlear Implants
Michelle L. Hughes, Ph.D.
Lisa J. Stille, M.A.
Kelly R. Barrow, M.A.
Boys Town National Research Hospital, Omaha, NE, USA
One potential problem with multichannel stimulation in cochlear implants is that multiple electrodes
stimulate overlapping populations of neurons. This
overlap (or interaction) can be measured either
physiologically or psychophysically. It is reasonable to assume that the amount of interaction
should be similar for either method of measurement. The purpose of this study was to compare
the amount of electrode interaction measured
in electrically evoked compound action potentials (ECAPs) versus psychophysical thresholds
obtained with a forward masking technique.
Thus far, 14 adult cochlear implant recipients have
participated in this study, including 8 Nucleus
24R(CS), 3 Clarion CII, and 3 HiRes 90K users.
For the psychophysical portion of the study,
unmasked probe thresholds were obtained for a
basal, middle, and apical electrode in each subject.
Testing involved a three-interval, two-alternative forced-choice (3I-2AFC) task requiring the
subject to indicate the interval in which the sound
stimulus (either 2 or 3) was heard. Next, masked
thresholds were obtained using the 3I-2AFC
task where the subject indicated which interval
(either 2 or 3) contained the “different” stimulus.
Table 1. Masker and probe electrode combinations
tested for Clarion and Nucleus devices.
Clarion CII/90K
Nucleus 24R (CS)
Probe
Electrode
Masker
Electrode
Probe
Electrode
Masker
Electrode
5
1, 3, 5, 7, 9
7
1, 4, 7, 10, 13
9
5, 7, 9, 11, 13
11
5, 8, 11, 14, 17
12
8, 10, 12, 14, 16
16
10, 13, 16, 19, 22
The same pairs were used for both the physiologic (ECAP)
and psychophysical forward masking experiments. Maskers
were spaced two electrodes apart in Clarion measurements
and three electrodes apart in Nucleus measurements—
representing roughly equivalent spacing for the two cochlear
implant designs.
80
Five masker electrodes were tested for each of the
three probe electrodes, as shown in Table 1. Masker
and probe stimuli consisted of a 1000-pps train of 25or 50-usec/phase biphasic pulses. Masker duration
was 300 ms and probe duration was 20 ms following
a 2-ms delay. Masker levels were fixed at 80% of the
dynamic range for the masker stimulus. The amount
of masking for each masker-probe pair was calculated by subtracting the unmasked (probe-alone)
threshold from the masked threshold. The amount
of masking was normalized to the masker-equalsprobe condition and then plotted as a function of
masker electrode for each probe (Figure 1, dark blue).
For the physiologic aspects of this study, ECAPs were
obtained using a traditional forward masking technique described previously (Abbas et al, 1999, 2004).
Masker and probe stimuli for ECAP measures each
consisted of a single biphasic current pulse of the
same duration used for the psychophysical portion of
the experiment (either 25 or 50 usec/phase). All stimulation was in monopolar mode. Masker and probe
levels were fixed at 80% of the dynamic range for the
ECAP stimulus. Like the psychophysical measures,
the probe was fixed while the masker varied in location along the electrode array. Peak-to-peak ECAP
amplitudes were measured and then normalized to
the amplitude of the masker-equals-probe condition, yielding the amount of masking. The amount
of masking for the ECAP was then plotted as a
function of masker electrode (Figure 1, light blue).
Results to date have shown that both physiologic and
psychophysical measures typically demonstrated the
greatest amount of masking (or interaction) when
masker and probe were on the same electrode—with
less masking at greater masker-probe separations.
Figure 1 shows ECAP (light blue) and psychophysical (dark blue) forward masking patterns for
the three probe electrodes obtained in an example
subject (a Clarion CII user). For each individual,
the normalized amount of psychophysical masking
Advanced Bionics® Auditory Research Bulletin 2005
“... ECAP measures may provide
an efficient alternative to
time-consuming behavioral measures
of electrode interactions.”
was plotted relative to the normalized amount of
ECAP masking for all probe electrodes tested. The
normalization points (masker-equals-probe conditions) were removed and correlation coefficients
were obtained, which ranged from 0.13 to 0.85
across subjects. When data from all subjects were
pooled, there was a significant correlation between
physiologic and psychophysical forward masking
(r = 0.55, p < 0.0001). These results are consistent
with those reported by Cohen et al (2003). Correlations remained significant when data were analyzed
as a function of electrode place, with the strongest
correlations occurring for basal and middle electrodes.
mance. This result was consistent across all three
speech perception measures. Currently, the data are
being examined further to investigate this finding.
To determine whether these results were related to
speech perception performance, the correlation coefficient for the ECAP-psychophysical comparison
was plotted relative to speech perception performance
for three tests: BKB-SIN (sentences in noise), CNC
words, and CNC phonemes. Data analysis revealed
an interesting trend: subjects with the best correlation between ECAP and psychophysical forward
masking had the poorest speech perception performance, and subjects with the poorest correlations
demonstrated the best speech perception perfor-
Acknowledgement
In conclusion, this study showed a significant correlation between physiologic and psychophysical
measures of channel interaction. Based on these
data, ECAP measures may provide an efficient
alternative to time-consuming behavioral measures
of electrode interactions. Further study is needed
to evaluate how these measures can be used to
better predict speech perception performance.
This study was funded by the NIH-NIDCD grant RO3 DC007017-01A1.
References
Abbas PJ, Brown CJ, Shallop JK, Firszt JB, Hughes ML, Hong SH, Staller SJ. (1999)
Summary of results using the Nucleus CI24M implant to record the electrically
evoked compound action potential. Ear and Hear 20:45-59.
Abbas PJ, Hughes ML, Brown CJ, Miller CA, South H. (2004) Channel interaction
in cochlear implant users evaluated using the electrically evoked compound action
potential. Audiol & Neurotol 9:203-213.
Cohen LT, Richardson LM, Saunders E, Cowan RSC. (2003) Spatial spread of neural
excitation in cochlear implant recipients: comparison of improved ECAP method
and psychophysical forward masking. Hear Res 179:72-87.
Figure 1. Normalized amount of masking as a function of masker electrode for psychophysical forward masking (dark blue) and
ECAP forward masking (light blue) for a Clarion CII subject (C7).
Objective Measures
81
Channel Interaction in Children
Kevin H. Franck, Ph.D.
The Children’s Hospital of Philadelphia, Philadelphia, PA, USA
Marc D. Eisen, M.D., Ph.D.
University of Pennsylvania School of Medicine, Philadelphia, PA, USA
“...channel interaction data in adult
implant users cannot be extrapolated
appropriately to children.”
Channel interaction is defined as the breakdown
in discrimination of differences between electrodes
at the level of perception. Relevant to understanding channel interaction, children have a more
consistent onset of deafness (congenital), duration of hearing loss, and likely a more uniform
etiology of deafness than adults. Because of these
inherent differences, channel interaction data in
adult implant users cannot be extrapolated appropriately to children. Therefore, the goal of our
work is to define properties of channel interaction
characteristic of the pediatric implant population.
We have studied channel interaction in children
using both electrophysiologic and psychophysical
approaches (Eisen & Franck, 2005). Our subjects
consisted of children who use the C-I device and
HiFocus electrode (with or without positioner),
and children who use the Nucleus 24 Contour and
Nucleus 24 Straight arrays. In the electrophysiologic studies, we measured the electrically evoked
compound action potential using a masker-probe
paradigm. By varying the position of the masker
electrode with a fixed probe location, the resulting
interaction functions were used to quantify electrode
interactions. The electrophysiologic results showed
that electrode interaction was affected significantly
by stimulus intensity, electrode location, and array
type. Electrode interaction increased with stimulus
intensity. Electrode interaction also increased toward
the apical end of the electrode array compared with
the basal end. A small increase in interaction was
seen in the with-positioner data, but no other differences in interaction were attributed to array type.
In the psychophysical studies, we measured
frequency and electrode discrimination using twointerval adaptive forced-choice tests with a video
game graphical user interface. The software gener-
82
Advanced Bionics® Auditory Research Bulletin 2005
ates acoustic stimuli for frequency discrimination
and electric stimuli for electrode discrimination.
With this video game platform, we were able to
determine difference limens for both acoustic
frequencies and inter-electrode distances. It was
challenging to use the adaptive forced-choice testing
with children less than five years of age, and difference limens were more variable and larger in this
young population than in older children. Acoustic
frequency discrimination was poorer for implanted
children than for normal hearing children at all
ages tested, but all children showed improvements
with increasing age. Frequency difference limens in
implanted children approached electrode spacing
and signal processing characteristics of the implant.
Electrode discrimination improved with stimulus
intensity level, and varied with electrode location.
“Electrode interaction depends on
stimulus intensity and
location along the electrode array.
Discrimination depends on the
child’s age and stimulus intensity.”
In summary, we have measured electrode interaction at the level of the cochlear nerve and
discrimination at the level of perception in pediatric congenitally deaf subjects. Electrode interaction depends on stimulus intensity and location
along the electrode array. Discrimination depends
on the child’s age and stimulus intensity. Ongoing
work aims to identify the contribution of electrode
interaction at the level of the cochlear nerve to
channel discrimination at the level of perception.
Acknowledgement
This research was supported by the Deafness Research Foundation (KHF) and the
American Academy of Otolaryngology (MDE).
Reference
Eisen MD, Franck KH. (2005) Electrode interaction in pediatric cochlear implant
subjects. J Assoc Res Otolaryngol 6(2):160-170.
Objective Measures
83
Unraveling the Electrically Evoked
Compound Action Potential
Jeroen J. Briaire, M.Sc.
Johan H.M. Frijns, Prof., M.D., Ph.D.
Leiden University Medical Center, Leiden, The Netherlands
“... typical responses just above threshold
and to higher stimulation levels
do not contribute to the N1-to-P1 difference
as much as the P0 peak...”
This work aimed at deriving a fundamental understanding of the processes that underlie the recording
of evoked compound action potentials (ECAP)
from humans. The intention was to determine the
contribution of the individual nerve fibres to the
overall signal and also to examine the extent to
which this signal provides clinically relevant information. Work was based on a detailed, two-part
computer model of the human cochlea, including
(1) a three-dimensional, volume conduction model
and (2) an active, nonlinear, auditory nerve fibre
model. The model was developed at the Leiden
University Medical Center (Frijns et al, 2000 a-b;
Briaire & Frijns, 2000). The model was extended to
better represent the human cochlea in a number of
aspects, including (1) morphology contributing to
cross-turn stimulation, such as an unmyelinated cell
body; (2) a much longer peripheral process impacting
latency; and (3) an unmyelinated presomatic region
allowing the action potential (AP) to bridge the larger
capacitor formed by the unmyelinated cell body.
Each modeled fibre represented 100 actual nerve
fibres providing spatial resolution of 100 µm. A
perimodiolar electrode array based on the HiFocus
design was used with the standard condition:
stimulating at electrode 12 (mid basal turn) and
recording from electrode 1 (the most apical contact
at around 1.5 turns of insertion). A biphasic stimulation pulse width of 37.5 µs per phase and delivered at 6 dB above threshold was used, apart from
84
a variety of current levels used to examine amplitude series of signal against stimulation level. The
currents at each modeled node of Ranvier were
composed of four elements: (1) membrane capacitance, (2) sodium ionic channels, (3) potassium
ionic channels, and (4) the leakage conductance.
Each node’s current was summed to determine a
single fibre action potential (SFAP). Finally, all
SFAPs were summed to determine the potential at a recording electrode. All ECAP responses
were sampled at a 10 µs interval. To mimic the real
recording situation, an artefact-reduction scheme
also was modeled. In order to best suit the model,
a scaled artefact-subtraction approach was adopted.
Comparisons were made between the model used
in previous studies and the one developed for this
study—a closer approximation of the human cochlea.
Between the two models, similar signal shapes and
profiles were found. The main differences observed
were an upward shift in stimulation threshold and
a change from peripheral to central excitation
at threshold stimulation for the updated model.
Using both anodic- and cathodic-leading biphasic
stimulation pulses as well as the unmyelinated and
myelinated cell bodies (UMCB, MCB), four situations were studied. In comparing cathodic- versus
anodic-leading stimulation, a difference of some
70 µs for the arrival at the central end of a modeled
fibre was found—the cathodic-leading pulse having
the shorter latency. A hypothesis based on the negative-leading phase producing stimulation was not
supported by the large time difference. Large differences in SFAP characteristics were found between
MCB and UMCB fibre morphologies. These differences included latency, relative size of positive and
negative peaks, and the presence of an initial (P0 )
peak prior to the expected N1 and P1 peaks. The P0
peak only appeared for larger stimulation current
Advanced Bionics® Auditory Research Bulletin 2005
strengths, originating from fibre populations at the
centre of the stimulation area for both UMCB and
MCB conditions. Two APs were found to propagate
in different directions: one moving antidromically
towards the organ of Corti and one orthodromically towards the central end of the axon—the relative origins, sizes, and imposed delays on each of the
APs leading to widely different signal morphologies.
The finding that a long and homogeneous nerve fibre
produced SFAPs very comparable in shape with more
fundamental predictions and actual measurements
(Schoonhoven & Stegeman, 1991) indicates that
the active nerve fibre model is functioning correctly.
Peak latencies in the full model (better fitting
previous guinea pig measurements) are too short
for the human situation. While overall increases
in latency have been produced by the morphological model changes outlined above, it appears that
the introduction of truly human kinetics into the
nerve model will be necessary to produce appropriate overall latency as well as interpeak latency.
The current emitted by the cell body dominates the
response in the UMCB condition. The AP trajectory
might closely resemble the AP plot for degenerated
nerve fibres, i.e., without peripheral processes. This
result could explain the absence of ECAP measurements in subjects with neural degeneration who still
have normal auditory responses and perform well
with a cochlear implant. With further examination, it appears that the dendrite is responsible for
generation of the P0 peak at higher current levels.
response determined by the second phase. Because
the response can be dominated by the antidromically propagating AP, the exact placement of the
recording electrode will also be important. Apical
recording will favour orthodromic APs whereas
basal recording will favour antidromic APs.
A central finding of this study—that typical
responses just above threshold and to higher stimulation levels do not contribute to the N1-to-P1
difference as much as the P0 peak—has potential
clinical application. ECAP input/output curves
tend to show a shallower slope than would be
predicted by the increased number of excited fibres.
This may extend to saturation or even a decrease in
response at higher stimulation levels. The various
response types necessarily undermine the unitary
response theory where every fibre contributes
the same amount to the whole nerve response.
References
Briare JJ, Frijns JHM. (2000a) 3D mesh generation to solve the electrical volume
conduction problem in th eimplanted inner ear. Simpra 8:57-73.
Briare JJ, Frijns JHM. (2000b) Field patterns in a 3D tapered spiral model of the
electrically stimulated cochlea. Hear Res 148(1-2):18-30.
Frijns JHM, de Snoo SL, Schoonhoven R. (2000) Improving the accuracy of the
boundary element method by the use of second-order interpolation functions. IEEE
Trans Biomed Eng 47(10):1336-1346.
Schoonhoven R, Stegeman DF. (1991) Models and analysis of compound action
potentials. Crit Rev Biomed Eng 19(1):47-111.
The anodic-leading stimulus excites the nerve fibre
in a single place, close to the electrode contact. The
cathodic-leading stimulus induces two APs just next
to the stimulation site, similar to responses seen
for monophasic stimulation with the polarity of
Objective Measures
85
Relationship Between Electrical Field Models and the
Cochlear Anatomy of Clarion CII Subjects
Filiep J. Vanpoucke, Dr. Ir.
Advanced Bionics Corporation, Europe
Johan H.M. Frijns, Prof., M.D. Ph.D.
Leiden University Medical Center, Leiden, The Netherlands
Stefaan Peeters, Prof., Dr. Ir.
University of Antwerp, Antwerp, Belgium
Figure 1. Electrical field image recorded in a subject at 40 µA
peak current and 6000 Hz sine wave stimulation. Each of the
16 curves represents the voltage measured across the electrode
array when a single contact is stimulated.
Figure 2. Electrical model for the conduction through the
cochlear tissues. Each segment corresponds to the area
between two consecutive electrode contacts. The longitudinal
resistors (rL) model current flow along or parallel to the scala
tympani. The transversal resistors (rT) model current through
bony cochlear walls.
86
We previously presented an accurate measurement
technique to quantify the intracochlear potentials
by using the standard measurement capabilities
of the CII electronic platform (Vanpouke et al,
2004 a). The combination of careful recording and
use of state-of-the-art signal processing tools results
in a high quality intracochlear potential map, also
known as electrical field imaging (EFI), illustrated in
Figure 1. The peak when measuring on a stimulating
contact is caused by the additional contribution of
the contact impedance itself, i.e., the voltage that is
needed to pass electrical current through the contact.
An initial study showed that substantial variability
in the electrical field images exists across subjects.
In order to better understand the causes for this
variability, we, with other colleagues, have proposed
a model that relates the EFI to the anatomical and
electrical properties of the cochlea (Vanpoucke et al,
2004b). We have shown that the cochlear tissues can
be represented accurately (up to 95%) by a simple
resistive ladder network consisting of longitudinal
and transversal components. The longitudinal resistors model the current component parallel to the
scala tympani, whereas the transversal resistors
represent current leakage through the bony cochlear
walls (Figure 2). The model parameters reflect the
local conductivities in the implanted cochlea. It is
therefore theoretically feasible to gain insight into
factors that influence these local conductivities, such
as tissue formation or anatomical deformations.
A first survey was conducted on a group of 25
subjects at the Leiden University Medical Center.
All subjects were wearing a Clarion CII device with
HiFocus I electrode array. Approximately half of the
subjects had a positioner implanted. On average the
“standard” electrical field model indicates a value
of 100-250 Ohm for the longitudinal resistors and
more than 10 kOhm for the transversal resistors. This
difference of almost two orders of magnitude indi-
Advanced Bionics® Auditory Research Bulletin 2005
cates that the perilymph acts as a highway shunting
off the current. Only a small percentage of the
current effectively is traversing the lateral or modiolar walls. The longitudinal resistors typically show
a slight increase in the region of the cochleostomy,
potentially indicating some tissue formation near the
cochleostomy site. The most basal transversal resistor
is on the order of 3 kOhm. The base of the cochlea
is therefore the preferred pathway for the current
to flow from the cochlea to the reference electrode.
“We have shown that
the cochlear tissues can be
represented accurately (up to 95%)
by a simple resistive ladder network
consisting of longitudinal and
transversal components.”
Approximately 40% of the subjects also showed
a dip in the transversal impedance near the end of
the basal turn. In a few subjects, the mid-cochlear
conductance path is well localized and high. We
have postulated that the cause of this dip might
be the proximity of the facial nerve canal. For one
subject, the position was confirmed via CT scan.
In the absence of a Positioner, the value of the
basal resistor was +/- halved because the base
of the cochlea is more open from an electrical
point of view. On the contrary, there is a subject
with a partially ossified scala. The region of ossification is clearly seen in the resistance values.
The EFI model for one particular subject contains
very high values for the resistivity near the base
of the cochlea. In this case the cochlea was
drilled out during surgery. The CT scan of the
subject showed that the region of increased resistivity coincided with the region of ossification.
Results for thes subjects have been instrumental
to understanding the relationship between the
cochlear anatomy and the corresponding EFI
model. Although the applicability of the technique in clinical settings still has to be proven,
in a number of cases, the EFI modeling approach
produced
reliable
and
consistent
results.
References
Vanpoucke F., Zarowski A., Casselman J., Frijns J., Peeters S. (2004a) The facial
nerve canal: an important cochlear conduction path revealed by Clarion electrical
field imaging. Otol Neurotol 25(3):282-9.
Vanpoucke F., Zarowski A., Peeters S. (2004b) Identification of the impedance
model of an implanted cochlear prosthesis from intracochlear potential
measurements. IEEE Trans Biomed Eng 51(12):2174-83.
Objective Measures
87
HiResolution Sound
The HiResolution Bionic Ear System is the only cochlear implant
capable of delivering HiResolution (HiRes) Sound to the auditory
system.
HiRes is optimized for each implant recipient using the SoundWave
Professional Suite software, which simplifies the fitting process
and reduces fitting time. HiRes is designed to offer a wide,
programmable dynamic range, preservation of spectral and temporal
details of sound, and stimulation rates of up to 83,000 pulses per
second.
Reports from around the world demonstrate that HiRes is
surpassing previous-generation technology in providing improved
language skills, speech perception, and speech intelligibility to
implant recipients of all ages. Furthermore, HiRes listeners are
reporting music appreciation benefit previously thought to be an
unrealistic expectation for many cochlear implant users.
Performance of Adults with HiResolution
Sound Processing
The clinical trial of HiResolution (HiRes) sound
processing in North America was the first investigation examining the benefits of this new sound
processing scheme (Koch et al, 2004). One of the
aims of the study was to compare performance
between HiRes and conventional strategies (CIS,
MPS, SAS). Fifty-one postlinguistically deafened
adults participated in the study at 20 sites in the
United States and Canada. After being implanted
with the CII Bionic Ear System, the subjects were
fit initially with their preferred conventional strategy
and assessed on a battery of speech perception tests
after three months of use. Patients then were fit with
HiRes and assessed again after three months of use.
Study Sites in North America
Beth Israel Medical Center, New York, New York
California Ear Institute, Palo Alto, California
Glenrose Rehabilitation Hospital, Edmonton, Alberta
House Ear Clinic, Los Angeles, California
Houston Ear Research Foundation, Houston, Texas
Jackson Ear Clinic, Jackson, Mississippi
Johns Hopkins University, Baltimore, Maryland
Medical College of Wisconsin, Milwaukee, Wisconsin
Midwest Ear Institute, Kansas City, Missouri
New York Presbyterian Hospital/Columbia University Medical Center,
New York, New York
New York University, New York, New York
Otology Group of San Antonio, San Antonio, Texas
Ottawa Hospital (Civic Campus), Ottawa, Ontario
Mean three-month scores with conventional
and HiRes sound processing are summarized in
Tables 1 and 2. The mean improvement from
conventional to HiRes processing was significant for all speech recognition tests (monosyllabic words and sentences in quiet and in noise)
(p < .01). Notably, subjects who were the lowest
performers with conventional sound processing
showed the greatest improvements with HiRes.
Spokane Ear, Nose & Throat Clinic, Spokane, Washington
Sunnybrook & Women’s College Health Sciences Centre,
Toronto, Ontario
Tampa Bay Hearing & Balance Center, Tampa, Florida
University of California, San Francisco, California
University of Iowa, Iowa City, Iowa
University of Miami, Miami, Florida
University of Minnesota, Minneapolis, Minnesota
Table 1. Monosyllabic word and sentence recognition preoperatively and
after three months of conventional strategy use followed by three months of HiRes use.
CNC Words
CID Sentences
Pre
Conventional
HiRes
Pre
Conventional
HiRes
Mean
3.4%
41.4%
50.0%
21.4%
75.6%
83.6%
Std Dev
7.1%
25.2%
25.1%
23.0%
30.6%
26.4%
n
47
50
51
48
50
51
Scores are expressed as percent correct.
Table 2. HINT sentence recognition in quiet and in noise preoperatively
and after three months of conventional strategy use followed by three months of HiRes use.
HINT in Quiet
HINT in Noise (+10 dB SNR)
Pre
Conventional
HiRes
Pre
Conventional
HiRes
Mean
11.8%
68.8%
80.0%
2.7%
47.0%
60.8%
Std Dev
15.3%
30.7%
25.2%
6.6%
32.3%
28.4%
n
51
50
51
43
49
51
Scores are expressed as percent correct.
90
Advanced Bionics® Auditory Research Bulletin 2005
Figure 1 shows the three-month conventional
sound processing results for subjects who scored
50% or lower (n = 10) on CID sentences, an easy
test of open-set sentence recognition. After using
HiRes for three months, 5 of these 10 subjects
improved their performance between 40% and
70%, and five subjects scored 70% or higher.
Figure 2 shows the results for subjects who scored
30% or less (n = 19) on HINT sentences in speechspectrum noise (+10 dB signal-to-noise ratio) with
conventional processing. After using HiRes for three
months, 11 of these 19 subjects improved their scores
on HINT sentences in noise by 20-62%, and eight
of the subjects scored 49% or higher with HiRes.
Moreover, 96% of all subjects preferred HiRes at
the end of the study, including those patients who
achieved relatively high scores with conventional
processing. These study results indicate that cochlear
implant recipients have the potential to benefit
from improvements in sound processing technology.
Reference
Koch DB, Osberger MJ, Segel P, Kesser DK. (2004) HiResolution and conventional
sound processing in the HiResolution Bionic Ear: using appropriate outcome
measures to assess speech recognition ability. Audiol Neurotol 9:214-223.
Figure 1. Individual scores on CID sentences for subjects who scored 50% or less (n = 10) after using conventional processing for
three months (left) compared with their performance after using HiRes for three months (right). Scores are rank ordered from lowest
to highest performer in each group.
Figure 2. Individual scores on HINT sentences in noise (+10 dB SNR) for subjects who scored 30% or less (n = 19) after using convent
rank ordered from lowest to highest performer in each group.
HiResolution Sound
91
European HiResolution Multicenter Study:
Sound Quality Results
Participating Study Sites
Azienda Ospedaliera di Padova, Padova, Italy
Centre Hospitalier Universitaire de Rouen, Rouen, France
Guy’s and St. Thomas’ Hospital, London, United Kingdom
Leiden University Medical Center, Leiden, The Netherlands
Medizinische Hochschule Hannover, Hannover, Germany
Ospedale Civile di Rovereto, Rovereto, Italy
Policlinico 1 Roma, Rome, Italy
Universitätskrankenhaus Eppendorf Hamburg, Hamburg, Germany
A European multicenter clinical trial was conducted
to investigate the benefit of HiResolution sound
processing compared to standard strategies (SAS,
PPS, CIS). A total of 37 postlinguistically deafened adults who had been using standard strategies participated in the study at eight sites. Each
subject listened with three HiResolution programs
in a balanced crossover design. The study explored
the use of different numbers of channels (8, 12,
16) and rates of stimulation (1500, 2000, 3000
pps/per channel) on speech perception. The study
was designed to separate the effects of channel
number and stimulation rate. Hence, the variation
of channel number was always with a 2,000 pps/per
channel stimulation rate, and the variation of rate
was always for an 8-channel program. Each program
was experienced for one month, and a counterbalanced presentation order was used. Different
study sites explored the effects of different parameters within a series of interlocking subprotocols.
As part of the overall study protocol, subjects’
perceptions of sound quality for their preferred
HiResolution program and conventional sound
processing were evaluated. Several questionnaires
were administered including a general sound quality
questionnaire, a male voice quality questionnaire,
a female voice quality questionnaire, and a music
quality assessment test. Language-appropriate sound
samples were used for the male and female voice
quality test, and three music tracks were used for
92
Advanced Bionics® Auditory Research Bulletin 2005
“...the main subjective benefits
of HiResolution
were for sound clarity
and listening in the presence of
background noise.”
the music quality assessment (mellow jazz passage,
dynamic classical passage, unaccompanied female
vocal passage). All questionnaires had a similar
response structure where sound quality was rated
on a variety of attributes (e.g., clarity, pleasantness,
overall quality, boominess, tinniness, naturalness)
using a scale of 0-10 in which a low score was negative and a high score was positive. The general sound
quality questionnaire also assessed noise interference, degree of echo, and quality of own voice.
Mean ratings are summarized in Figures 1 and 2.
On the general sound quality questionnaire, subjects
reported that the main improvements between standard strategies and HiResolution were for clarity
of sound and reduction of noise interference for
speech understanding. The ratings were significantly different (p < 0.05) for these characteristics.
(Other categories evaluated did not show significant
differences.) The benefits of HiResolution for music
listening were encouraging. The greatest differences
between HiResolution and conventional processing
were reported for the dynamic classical music track.
HiResolution received higher ratings for pleasantness,
and music was perceived as less boomy and less tinny.
Figure 1. Mean sound quality ratings for clarity and noise interference. Higher scores signify that the sound is clearer and that
there is less interference from noise when listening to speech.
The HiRes ratings were significantly higher than the ratings for
conventional strategies (p < .05).
In summary, the main subjective benefits of HiResolution were sound clarity and listening in the presence of background noise. HiResolution provided
improvements in perceived quality for certain
musical aspects of challenging recorded music tracks.
Figure 2. Mean quality ratings for dynamic classical music
passage.
HiResolution Sound
93
The Effect of Number of Channels on HiRes Performance
Johan H.M. Frijns, Prof., M.D., Ph.D.
W. Martin C. Klop, M.D.
Raymond M. Bonnet, M.D.
Jeroen J. Briaire, M.Sc.
Leiden University Medical Center, Leiden, The Netherlands
The effect of HiResolution (HiRes) stimulation
rate and number of channels on patient benefit
was evaluated by measuring speech perception in
quiet and in noise in nine CII implant users. All
subjects had at least three months (range: 3-11
months) experience using standard resolution sound
processing. Prior to the study, eight subjects used
Continuous Interleaved Sampling (CIS) with 833
pps/channel, monopolar, 75 µs/phase—and one
subject used Paired Pulsatile Stimulation (PPS)
with 1666 pps/channel, monopolar, 75 µs/phase.
This study (Frijns et al, 2003) used a blind crossover design to evaluate three HiRes strategies with
different numbers of channels (8, 12, and 16). The
pulse rate per channel was fixed (1400 pps), the
pulse width was 21 µs, and the interpulse interval
was varied (43, 21, or 0 µs) according to the number
of channels to maintain the constant pulse rate per
channel. In Leiden, the fitting approach aims to
avoid cross-turn stimulation and to maximize the
amount of high frequency information delivered to
the electrode array. To ensure that the subjects did not
know how many channels were being used during
each study phase, thresholds and most comfortable
levels were measured for all electrodes at each fitting
session. Subjects were divided into three groups, and
the fitting order for the three HiRes strategies was
balanced across the groups according to a modified
Latin-square design to avoid learning effects. Each
strategy was used as the exclusive strategy for one
month, after which speech performance was assessed
using CVC word lists presented at 65 dB SPL in
quiet and in noise (+10, +5, 0, and –5 dB SNR).
Three optimal programs for each user were
defined based on performance in noise, as follows:
• Program with highest performance at +5 dB
SNR.
• Program with optimal weighted average of
scores at +10 dB and +5 dB SNR.
• Program with optimal speech reception
threshold (SRT), defined as the SNR at which
50% of the phonemes were understood.
94
Advanced Bionics® Auditory Research Bulletin 2005
Speech perception data are shown in Figure 1. In
general, performance deteriorated with increased
noise level for all three HiRes strategies. Performance in quiet was the same for standard resolution
compared to all HiRes strategies, and no one set of
programming parameters was optimal for all users
in all noise conditions. Nonetheless, performance
with all HiRes strategies was better than with standard resolution. In the +5 dB SNR condition, word
recognition with the 8- and 12-channel HiRes
programs showed a significant improvement over
word recognition with standard resolution (p < 0.05).
“...even nonoptimized HiRes programs
provided improved benefit
over standard resolution programs.”
Reference
Frijns JHM, Klop WMC, Bonnet RM, Briaire JJ. (2003) Optimizing the number of
electrodes with high-rate stimulation of the Clarion CII cochlear implant. Acta
Otolaryngol 123:138-142.
In sum, although all subjects had very good results
with standard resolution before the study (average
CVC monosyllabic word score of 66% in quiet), they
all demonstrated improved benefit from HiRes, especially in noise. A ceiling effect was observed for tests
in quiet. The optimal number of channels was dependent on the individual user and the criterion used to
define the optimal program. Whether objective and
subjective measures of electrode selectivity can be
used to choose the optimal electrode configuration
and to reduce the workload for audiologists and users
is now being studied. However, it is important to note
that even nonoptimized HiRes programs provided
improved benefit over standard resolution programs.
Figure 1. Dutch monosyllable word scores at a +5 dB SNR for three strategies. For the majority of subjects, even the poorest performance with a HiRes program was superior to performance with the standard resolution strategy.
HiResolution Sound
95
Perception of Sinusoidal Modulation with High-Frequency
Pulse Trains by CII Users
Leonid Litvak, Ph.D.
Edward Overstreet, Ph.D.
Andrew Voelkel
Advanced Bionics Corporation, Valencia, CA, USA
“These results suggest that
CII/HiRes users
are able to derive pitch information
from the temporal modulation
of high-rate carriers.”
In contemporary cochlear implants, the acoustic
signal is transformed and delivered to the electrode contacts within the cochlea via modulated
pulse trains. Many investigations have demonstrated that a relationship exists between pulse rate
and pitch perception. For low rates, pitch increases
as the pulse rate increases up to a “pitch-knee”
frequency, beyond which pitch stays constant as
rate increases (e.g., Landsberger & McKay, 2005;
Shannon, 1983). For modulated pulse-train carriers
(< 1000 pps), temporal modulation produces pitch
percepts in the same frequency range as those for
pulse trains, typically below 300 Hz (e.g., McKay
et al, 1994; Shannon, 1992; Wilson et al, 1994).
The HiRes system can support carrier pulse rates
greater than 5,000 pps per contact, which may
allow more signal information to be encoded in
the modulation pattern. Moreover, these high pulse
rates may cause auditory nerve responses to be more
desynchronized, thereby producing responses more
similar to a normal ear than responses to low-rate
pulse trains (Rubinstein et al, 1999; Litvak et al,
2003). However, it is unclear how much information can be encoded and used by cochlear
implant listeners when high carrier rates are used.
This study evaluated pitch-knee frequencies and
their dependence upon carrier rate and stimulation
place in nine CII users. Two carrier rates were used
(3,000 pps and 10,000 pps), and pitch was evaluated by stimulating at least three contacts along the
electrode array. Subjects listened to two standard
stimuli—one in which the carrier was modulated by
100 Hz and one in which the carrier was unmodulated. Subjects were instructed that the pitch associated with the modulated standard was assigned
a pitch ranking of “0” and that the unmodulated
96
Advanced Bionics® Auditory Research Bulletin 2005
standard was assigned a pitch ranking of “100.” The
test stimuli were the carriers modulated at frequencies of 100 Hz and higher. Subjects assigned a pitch
ranking to the test stimuli between 0 and 100. The
pitch ranking data were fit with a model in order
to determine the pitch-knee frequencies for each
subject and each test condition. For this experiment, pitch-knee was defined as a frequency that
was ranked as 90% of the rank assigned to a 1000Hz modulation (which was typically near 100).
Figure 1 shows the distribution of pitch-knee
frequencies for seven of the nine subjects who could
perform the task. (Two subjects perceived the two
standards as very close or having the same pitch.)
For 27 ranking measurements, the pitch-knee
frequencies ranged from very low to over 800 Hz.
Over 50% of the trials produced pitch-knee frequencies of 300 Hz or greater and 15% indicated pitches
over 600 Hz. Pitch-knee frequency was not strongly
dependent on carrier pulse rate, but was highly
influenced by place of stimulation for some subjects.
These results suggest that CII/HiRes users are able to
derive pitch information from the temporal modulation of high-rate carriers. Moreover, some subjects in
this study were able to perceive higher pitch information than has been demonstrated previously for pulse
trains or modulated low-rate carriers (< 1000 pps).
Figure 1. Distribution of pitch-knee frequencies from a total of
27 pitch rankings.
References
Landsberger DM, McKay DM. (2005) Perceptual differences between low and high
rates of stimulation on single electrodes for cochlear implantees. J Acoust Soc Am
117(1):319-327.
Litvak LM, Smith ZM, Delgutte B, Eddington DK. (2003) Desynchronization of
electrically evoked auditory-nerve activity by high-frequency pulse trains of long
duration. J Acoust Soc Am 114(4):2066-2078.
McKay CM, McDermott HJ, Clark GM. (1994) Pitch percepts associated with
amplitude-modulated current pulse trains in cochlear implantees. J Acoust Soc Am
96(5):2664-2673.
Rubinstein JT, Wilson BS, Finley CC, Abbas PJ. (1999) Pseudospontaneous activity:
stochastic independence of auditory nerve fibers with electrical stimulation. Hear
Res 127(1-2):108-118.
Shannon RV. (1983) Multichannel electrical stimulation of the auditory nerve in
man I. Basic psychophysics. Hear Res 11(2):157-189.
Shannon RV. (1992) Temporal modulation transfer functions in patients with
cochlear implants. J Acoust Soc Am 91(4):2156-2164.
Wilson BS, Lawson DT, Zerbi M, Finley CC. (1994) Recent developments with the
CIS strategies. In Hochmair-Desoyer I & Hochmair ES, eds, Advances in Cochlear
Implants. Proceedings of the Third International Cochlear Implant Conference,
Innsbruck, Austria, April, 1993. Manz, Wein, Austria: Datenkonvertierung,
Reproduktion and Druck, pp. 103-112.
HiResolution Sound
97
Speech Perception at Varied Stimulation Rates
with the Clarion CII Cochlear Implant
Andreas Büchner, Ph.D.
Carolin Frohne-Büchner, Ph.D.*
Rolf-Dieter Battmer, Prof., Ph.D.
Thomas Lenarz, Prof., M.D., Ph.D.
Medizinische Hochschule Hannover, Hannover, Germany
* also with Advanced Bionics Corporation, Europe
“...speech perception performance
increased remarkably
when patients changed
from conventional coding strategies
to high-rate processing.”
Currently most speech processing strategies are
based on either a Continuous Interleaved Sampling
(CIS) approach or an “n-of-m” approach. One
predominant property common to these two
approaches is the sequential pattern of stimulation—meaning that only one electrode fires at any
point in time. The sequential stimulation paradigm
effectively eliminates electrical channel interaction that often occurs in simultaneous stimulation,
such as Simultaneous Analogue Sampler (SAS).
However, because all implants have a minimal pulse
width that cannot be overcome, stimulation rates
are limited, which in turn limits overall the amount
of information delivered to the auditory system.
The Clarion CII and the HiRes 90K implant
electronics are capable of delivering pulse rates
of up to 83,000 pulses per second (pps). The
increased update rate allows better resolution of
the time structure of the audio signal compared
to previous cochlear implant technologies.
The purpose of these studies was to investigate
patient performance as a function of pulse rate.
Only postliniguistically deafened adults with a
patent cochlea and no other handicaps were selected.
All subjects have been implanted with the Clarion
CII device and the HiFocus electrode with positioner and had at least three months of experience in
the standard Clarion speech coding strategies (CIS,
PPS, or SAS) prior to participating in these studies.
The results reported here are gathered from two
different high-rate studies: The first investigated
three different rates—1500, 2000, and 3000 pps—
in a randomized crossover paradigm. In all cases,
8-channel sequential high-rate programs were used.
The second study involved four different conditions: 8 channels @ 2500 pps, 8 channels @ 5000
pps, 16 channels @ 2500 pps, and 16 channels @
5000 pps—the latter using paired stimulation. In
the paired mode, two electrodes fire simultaneously,
thereby doubling the pulse rate per channel. Thirteen
patients were tested in these conditions in a crossover
98
Advanced Bionics® Auditory Research Bulletin 2005
study—six of whom also participated in the first study
previously described. Because these six patients had
already completed several high-rate conditions, the
3000 pps (instead of 2500 pps) condition was retested
once again and averaged with the previous results.
On average, speech perception performance
increased remarkably when patients changed from
conventional speech coding strategies to high-rate
processing (Figure 1). The performance increase of
just below 20% in the HSM sentence test in noise
is highly significant. Interestingly, SAS users showed
better results at very high rates (5000 pps), whereas
the group of former CIS users had their best performance between 2000 and 3000 pps (Figure 2). These
findings are statistically significant (two-tail Student’s
t-test), as indicated with asterisks in the figures.
In viewing individual performance, it becomes
apparent that stimulation rate plays an important
role—even at rates above 1500 pps/channel. Performance increases as great as 43% were observed within
the high rate (1500 to 5000 pps). However, the
fastest possible rate does not necessarily lead to the
best hearing performance for each individual subject.
Figure 1. Individual as well as averaged sentence scores for
(A) six subjects who took part in both phases of the study and
(B) seven subjects who took part in the second phase only.
The conditions shown are eight channels, sequential only. The
paired condition is not displayed. Note: * = p < 0.05 and ** =
p <0 .01.
Figure 2. Sentence scores for all subjects who (A) had CIS as their preferred strategy and (B) who had SAS as their preferred strategy
before the trial. Note: * = p < 0.05 and ** = p < 0.01.
HiResolution Sound
99
Multicentre Music Perception Study in France
Participating Centres
L’Ecole Normale Supérieure, Paris
Pelegrin Hospital, Bordeaux
Avicenne Hospital, Bobigny
Beaujon Hospital, Clichy
Gui de Chauliac Hospital, Montpellier
Principal Investigator
Purpan Hospital, Toulouse
Although significant advances have been made for
cochlear implant users, music perception remains
unsatisfactory for most implant recipients (McDermott, 2004; Leal et al, 2003; Gfeller et al, 2000). Still,
the possibility remains for improvement through
further innovations in sound processing strategies.
In this clinical study, a special HiRes sound
filter bank, called frequency alignment (FA)—
was used. The FA filter bank was designed to
match more closely the tonotopic distribution
of individual subjects. The expected outcome
of this approach is improved perception of
the harmonic relationships in musical sounds.
This study is designed as a six-month, balanced crossover, within-subject investigation. Thirty subjects
expected to enter the protocol will be randomly
separated into two groups. Subjects belonging to
Group A start with the default filter setting for
the first three months and then are given the FA
option for the following three months. Conversely,
subjects in Group B start with the FA settings and
after three months are given the default frequency
allocation. Music perception is evaluated at four
intervals: one month, three months, four months,
and six months after initial device activation.
The test battery includes pitch, melody, and timbre
comparisons, developed in collaboration with Centre
National de la Recherche Scientifique, to gather data
in an interactive software called CI-Music. For each
test, the result is a measure of “just noticeable difference” as a given parameter is varied. The pitch test is
a two-alternative “discriminate the higher pitch” task,
where the two spectral regions (f1 and f2) assessed
correspond to two groups of the four most apical
electrodes. For the melody task, random melodies
of four notes are played twice, with a fundamental
frequency difference introduced at random on one of
the notes in the repeat. Subjects are asked to identify
100
Advanced Bionics® Auditory Research Bulletin 2005
which note has changed. The same spectral regions
(f1 and f2) used in the pitch task also are assessed
for the melody task. Additional tests are administered via an audio CD, including discrimination and
identification of rhythms, identification of instruments in a closed set, and recognition of singers.
Results are shown in Figures 1 and 2 for three
subjects in Group A (default settings for three
months followed by FA settings for three months)
who were first to complete the study protocols. As
seen in Figure 1, all three subjects, after some months
of device experience, perform reasonably well on the
pitch discrimination task—with the best performance seen in Subject S2A who could discriminate
intervals of 2.3 semitones. However, none of the
subjects could perform the melody task (Figure 2).
These subjects’ inability to perform the melody
task cannot be explained by their pitch discrimination thresholds alone. The results may reflect
either unspecific music cognition impairments or
may indicate that these subjects used timbre rather
than pitch cues upon which to base their responses.
Figure 1. Results from the pitch ranking test (frequency band
corresponding to four most apical channels) for the first three
subjects in Group A to complete the study. Testing at one- and
three-month intervals were with default HiRes filter settings—
and at four- and six-month intervals with the experimental (FA)
filter settings. Performance is expressed in terms of interval
discrimination in semitones. After six months, subject S2A discriminates intervals of 2.3 semitones. Subject S3A shows great
improvement for this task when switching from HiRes to HiRes
FA at the four- and six-month test intervals.
Although it is too early to conclude that any benefit
in music perception results from filter modifications, these first patients have shown improved
pitch discrimination abilities over time. These
initial findings reinforce our interest in this study
of subjects implanted with the HiRes 90K system.
References
Gfeller K, Christ A, Knutson JF, Witt S, Murray K, Tyler RS. (2000) Musical
backgrounds, listening habits, and aesthetic enjoyment of adult cochlear implant
recipients. J Am Acad Audiol 11:390-406.
Leal MC, Shin Y, Laborde ML, Lugardon S, Andrieu S, Deguine O, Fraysse B.
(2003) Music perception in adult cochlear implant recipients. Acta Otolaryngol
123(7):826-835
McDermott HJ. (2004) Music perception with cochlear implants: a review. Trends
in Amplification, 8(2):49-82.
Figure 2. Results from the melody test (frequency band corresponding to four most apical channels) for the first three
subjects in Group A to complete the study. Despite improvements in pitch discrimination, the subjects show no significant
improvement (at four and six months) for the melody identification task with the HiRes FA filter settings.
HiResolution Sound
101
Speech Recognition Performance of Children with HiRes
Mary Joe Osberger, Ph.D.
Advanced Bionics Corporation, Valencia, CA, USA
Amy McConkey Robbins, M.S.
Communication Consulting Services, Indianapolis, IN, USA
Results of a retrospective study revealed that children implanted between 12 and 18 months of age
who used HiRes from the time of initial stimulation achieved significantly higher postimplant scores
on the IT-MAIS than their age-matched peers
who used previous generation (conventional) sound
processing schemes (SAS, CIS, MPS) (Osberger
& Koch, 2004). This study also compared performance as a function of sound processing strategy
but included older children who could be tested on
traditional measures of open-set word recognition.
Figure 1. Mean pre-and postimplant (at three months) phoneme and word scores on the MLNT for older children using
either HiRes or conventional (SAS, CIS, MPS) sound processing
(n = 9 in each group). Differences were significant for phonemes (p < 0.01) and words (p < 0.05).
Figure 2. Mean pre- and postimplant (at three months) phoneme and word scores on the PBK test for older children
using either HiRes or conventional (SAS, CIS, MPS) sound processing (n = 9 in each group). Differences were significant for
phonemes (p < 0.01) and words (p < 0.05).
102
As in the previous study, data were collected during
clinical trials sponsored by Advanced Bionics Corporation and were analyzed retrospectively. Age at
implant and each child’s preimplant PBK (Phonetically Balanced Kindergarten) (Haskins, 1949) word
score were used to match HiRes children to their
peers who used conventional processing (SAS, CIS,
MPS). There were nine children in each group. The
mean age at implant for both the HiRes and conventional group was 8 years (range = 3 to 14 years in
the HiRes group and 4 to 14 years in the conventional group). The mean preimplant PBK word score
was 6% and 7% for the HiRes and conventional
group, respectively (range = 0-12% in both groups).
Performance was evaluated on two measures of
open-set word recognition—the PBK test and the
MLNT (Multisyllabic Lexical Neighborhood Test)
(Kirk et al, 1995)—administered with recorded
materials at 70 dB SPL. Performance on both
measures was scored for percentage of phonemes
and words correctly understood by the child.
Group comparisons were made between preimplant and three-month postimplant performance.
The group results are shown in Figures 1 and 2. T-tests
revealed no significant differences in performance
between the two groups of children preimplant.
After three months of device experience, the mean
Advanced Bionics® Auditory Research Bulletin 2005
MLNT-phoneme, MLNT-word, and PBKphoneme scores of the HiRes group were significantly higher than those of the conventional group.
“The children in each group show a range
of performance; however, the scores of the
relatively poor and moderate performers
are higher in the HiRes group...”
Figures 3 and 4 show the individual scores of the children, rank ordered from lowest to highest performers.
The children in each group show a range of performance; however, the scores of the relatively poor and
moderate performers are higher in the HiRes group
than the conventional processing group. This pattern
of performance is similar to that observed in adults
after they switched to HiRes sound processing
(Koch et al, 2004) and in young children who used
HiRes from the time of initial stimulation. The
superior performance of children using HiRes also
is consistent with the findings of Bosco et al (2005).
Figure 3. Individual phoneme and word scores at three months
on the MLNT for older children using either HiRes or conventional (SAS, CIS, MPS) sound processing—rank-ordered from
lowest to highest performer within each group.
References
Bosco E, D’Agosta, Mancini P, Triascl G, D’Elia C, Filipo R (2005) Speech
perception results in children implanted with Clarion devices: HiResolution and
Standard Resolution modes. Acta Otolaryngol 125:148-158.
Haskins H. (1949) A phonetically balanced speech discrimination test for children.
Master’s thesis. Northwestern University, Evanston, IL.
Iler-Kirk K, Pisoni DB, Osberger MJ. (1995) Lexical effects on spoken word
recognition by pediatric cochlear implant users. Ear Hear 16:470-481.
Koch DB, Osberger MJ, Segel P, Kessler D. (2004) HiResolution and conventional
sound processing in the HiResolution Bionic Ear: Using appropriate outcome
measures to assess speech recognition ability. Audiol Neurotol 9:214-223.
Osberger MJ, Koch DB. (2004) Effect of sound processing on performance of
young children with cochlear implants. In: Miyamoto RT, ed. Cochlear Implants.
Proceedings of the VIII International Cochlear Implant Conference, Indianapolis,
Indiana, USA, 10-13 May, 2004. Amsterdam: Elsevier.
Figure 4. Individual phoneme and word scores at three months
on the PBK test for older children who used either HiRes or
conventional (SAS, CIS, MPS) sound processing—rank-ordered
from lowest to highest performer within each group.
HiResolution Sound
103
Speech Perception Results in Children Using
HiResolution and Standard Resolution Modes:
A Two-Year Follow-up Report
Ersilia Bosco, Cl. Psych.
Patrizia Mancini, M.D.
Luciana D’Agosta, Sp. Th.
Gabriella Traisci, Sp. Th.
Chiara D’Elia, M.D.
Roberto Filipo, Prof., M.D.
University of Rome La Sapienza, Rome, Italy
A battery of auditory tests was administered preand postimplant—at switch-on and at 3, 6, 9, 12, and
24 months. Age-apropriate tests were selected from
the battery that consisted of the following measures:
• Listening Progress Profile (LLP)
One primary indicator of cochlear implant benefit
is the ability to perceive speech. Results from many
studies have shown a wide range of speech perception abilities achieved by implanted children. In fact,
the development of speech perception by implanted
children depends on many variables, including
nonauditory as well as audiological factors (see e.g.,
Bosco et al, 2005; Geers et al, 2003; O’Donoghue
et al,1998; Waltzman et al, 1998). In a previous
investigation (Bosco et al, 2005), we examined
the effect of different speech coding strategies on
postimplant performance in children with Clarion
cochlear implants. The results showed that after 12
months of device experience, children who used
HiResolution (HiRes) sound processing demonstrated greater implant benefit than their peers
who used previous-generation sound processing
strategies. The purpose of the present investigation was to compare the performance of the
children after 24 months of implant experience.
Included in this study were 37 children. Sixteen
children implanted with the Clarion 1.2 device
using either CIS or SAS were in the standard
resolution mode (SRM) group. The other 21
children implanted with the CII using HiRes
processing were included in the HiRes mode
(HRM) group. For the postimplant analyses, each
strategy group was divided further according to
age at implant—under or over five years of age.
All children had prelinguistic onset of deafness.
104
• Test Abilità Percettive (Perceptive Ability
Test; TAP)—for ages 4 years and over; an
Italian adaptation of the Glendonald Auditory
Screening Procedure (GASP).
• Test Identificazione Parole Infantili (Children’s
Word Identification Test—TIPI)—for ages
4 years and over; an Italian adaptation of the
Northwestern University Children’s Perception
of Speech (NU-CHIPS).
• Bi-Trochee-Polysyllabic word test (BTP)—
for ages 2 to 5 years.
• Test Abilità Uditive Varese (Varese Auditory
Skills Test, TAUV)—explores all speech
perception skills.
• Protocollo Comune (Common Protocol)
For all subjects, ages one year through
adulthood.
• Speech Audiometry—phonetically balanced
bisyllabic words, in closed- and open-set
formats.
Results were pooled and analyzed across the
different tests according to the hierarchy of auditory skills described by Erber (1982): detection,
identification, recognition, and comprehension.
The results reported here are limited to those
obtained preimplant and at 12 and 24 months postimplant. Figure 1 shows preimplant performance for
SRM and HRM groups. Figures 2 and 3 show the
postimplant results for the children younger than
age 5 years and older than age 5 years, respectively.
Advanced Bionics® Auditory Research Bulletin 2005
Results shown in Figures 2 and 3 indicate that the
children using the HiRes sound strategy developed
better speech perception skills by 12 and 24 months
postimplant compared to SRM children. (Results
for detection tasks are not shown because the children in both groups performed comparably well in
this skill area.) For identification tasks, the performance of the HRM group was better and less variable than that of the SRM group. Performance for
auditory recognition and comprehension tasks was
also better for the HRM children than for SRM
children. This trend of better results was statistically significant at 24 months for children implanted
before and after age 5, both for recognition (t-test:
<5 yrs p = 0.02; >5 yrs p = 0.02) and comprehension
(t-test: <5 yrs p = 0.01; >5 yrs p = 0.05). Notably,
the children implanted after age 5 showed better
performance in the area of comprehension than
the younger group, especially if they used HiRes.
This latter finding perhaps reflects developmental
factors such as cognitive and language skills.
The
tent
that
dren
results of this investigation are consiswith the findings of Bosco et al (2005)
reported on the performance of chilwith only 12 months of device experience.
Figure 1. Results of speech perception skills obtained prior to
implantation for SRM (standard) and HRM (HiRes) groups.
Figure 2. Results of speech perception skills for SRM (standard)
and HRM (HiRes) groups of children implanted before age 5.
References
Bosco E, D’Agosta L, Mancini P, Traisci G, D’Elia C, Filipo R. (2005) Speech
perception results in children implanted with Clarion devices: Hi-Resolution and
Standard resolution mode. Acta Otolaryngol (Stockh) 125:148-158.
Erber, NP. (1982) Auditory training. Washington, DC: Alexandar Graham Bell
Association for the Deaf.
Geers A, Brenner C, Davidson L. (2003) Factors associated with development of
speech perception skills in children implanted by age five. Ear Hear 24 (1S):24S35S.
O’Donoghue GM, Nikolopoulos TP, Archbold SM, Tait M. (1998) Speech
perception in children after cochlear implantation. Am J Otol 19:762-767.
Waltzman SB, Cohen NL, Gomolin RH, Green JE, Shapiro WH, Hoffman RA,
Roland JT. (1997) Open set speech perception in congenitally deaf children using
cochlear implants. Am J Otol 18:342-349.
Figure 3. Results of speech perception skills for SRM (standard)
and HRM (HiRes) groups of children implanted after age 5.
HiResolution Sound
105
HiRes Performance in Mandarin-Speaking Children
Yong-xin Li, M.D.
De-min Han, Prof., M.D., Ph.D.
Xiao-tian Zhao, M.D.
Xue-qing Chen, M.D.
Liang Wang, M.D.
Ying Kong, M.D.
Tongren Hospital, Beijing, People’s Republic of China
“...the HiRes children
demonstrated superior performance
even though they had used
their sound processing strategy for
a much shorter period of time...”
The purpose of the present study was to assess the
development of auditory skills in children implanted
with the Bionic Ear (CII device) and to compare
their performance to children who used other
commercially available sound processing strategies.
Four profoundly hearing-impaired children, ranging
in age from 2 to 9 years, comprised the HiRes group.
Duration of HiRes experience ranged from 1.5 to
10 months—with a mean length of device use of
8 months. Seven children, ranging in age from 2
to 8 years, were included in the non-HiRes group.
The Meaningful Auditory Integration Scale (MAIS)
and the Infant-Toddler Meaningful Auditory Integration Scale (IT-MAIS) were used to evaluate the
acquisition of auditory milestones in the two groups
of children. Parents were interviewed to determine
the frequency of occurrence of “target” behaviors
in everyday situations. Based on the information
obtained from the parents, the clinician assigned
a rating, scaled according to the frequency with
which a given behavior was observed, as follows:
• 0 = never observed
• 1 = rarely observed
• 2 = occasionally observed
• 3 = frequently observed
• 4 = always observed
Scores were averaged across the MAIS/IT-MAIS
because the two measures assess the same behaviors
and differ only on the response criteria appropriate
for either older (MAIS) or younger (IT-MAIS)
106
Advanced Bionics® Auditory Research Bulletin 2005
children. Results were grouped into three categories of behavior: Vocalization (Questions 1 and 2),
Alerting to Sound (Questions 3-6), and Deriving
Meaning from Sounds (Questions 7-10)—with
scores averaged across questions within each category.
Figures 1 to 3 show the results for the three categories of behavior for the HiRes and non-HiRes
groups. The data show that the ratings for the HiRes
group were substantially higher than those of the
non-HiRes group in all three categories of behavior.
Notably, the HiRes children demonstrated superior
performance even though they had used their sound
processing strategy for a much shorter period of
time than their peers in the non-HiRes group. These
results suggest that the development of auditory
skills appears to occur earlier with HiRes processing
than with other sound coding schemes—and the
rate at which skills are acquired is more rapid.
Figure 1. Mean ratings assigned to HiRes and non-HiRes groups
on MAIS/IT-MAIS questions that assess vocalization.
Figure 2. Mean ratings assigned to HiRes and non-HiRes groups
on MAIS/IT-MAIS questions that assess alerting to sounds.
Figure 3. Mean ratings assigned to HiRes and non-HiRes
groups on MAIS/IT-MAIS questions that assess deriving meaning
from sound.
HiResolution Sound
107
Use of the Bark Transform to Measure Vowel Productions
in a Child Transitioned to HiRes Sound Processing
Ross Tonini, Au.D.
Texas Children’s Hospital, Houston, TX, USA
Charles Ballay, M.D.
Baylor College of Medicine, Houston, TX, USA
Spiros Manolidis, M.D.
Columbia University, New York, NY, USA
In accordance with critical band theory, Zwicker
(1961) proposed a method of vowel production
analysis by dividing the audible frequency range of
the human ear into 24 bands or Barks—after Heinrich Barkhausen, the creator of the loudness scale.
In analyzing vowel productions, the F1-F0 dimension
corresponds well to vowel height or openness, and
the F3-F2 dimension (in American English) distinguishes front and back vowels. Vocal tract dimensions
vary widely and dynamically between men, women,
and children. However, the Bark transform compensates for these wide variations and allows for data
normalization not possible with an acoustic Hertz
scale (Sydral & Gopal, 1986; McCaffrey & Sussman,
1994).The Bark transform is calculated as follows:
Bark (B) = 13 arctan (0.76f ) + 3.5 arctan (f/7.5)2
Conventionally, the “vowel triangle” (vowel space)
is defined by the cornerstone vowels /i/, /a/, and
/u/. Each vowel production can be described in
terms of a simple, Euclidian space by showing the
F1-F0 and F3-F2 Bark values as x and y coordinates, respectively, on a two-dimensional plot.
Subsequently, each vowel space can be calculated using Heron’s Formula (Kendig, 2000).
In this study, we hypothesized that changes in
speech production would be demonstrated as a
108
result of better auditory cues (finer temporal and
spectral information) provided by HiResolution
sound processing compared to previous technology.
Specifically, we predicted that improved productions of the cornerstone vowels could be measured
and tracked over time as our subject transitioned
from use of SAS to HiResolution sound processing.
The subject (RH) is a five-year-old female with
history of bilateral Mondini and Pendred’s syndrome.
At three years of age, she was implanted with the CII
Bionic Ear. After using SAS programs for 14 months,
she was transitioned to the HiResolution strategy.
At each of five test sessions, imitative speech samples
for the cornerstone vowels in isolation were taken.
Later, the frequency values of the fundamental
frequency and the first, second, and third harmonics
were averaged and converted to Bark scale values.
The F3-F2 and F1-F0 Bark differences (vowel triangles) were plotted, as shown in Figure 1. The areas
of the cornerstone vowel spaces were calculated and
plotted across each of the five test sessions (Figure 2).
These data reveal that prior to transition to the
HiResolution strategy, RH’s vowel space was markedly restricted. The data further suggest that RH
was not detecting the F3-F2 difference for the
vowel /u/ and the F1-F0 difference for the vowel /a/.
Advanced Bionics® Auditory Research Bulletin 2005
“For this child, the transition into
HiResolution sound processing
has led to significant normalization
of her overall vowel productions.”
After transition to HiRes, we noted a dramatic
opening of RH’s vowel space that has maintained within the normative range for children
her age with normal hearing (Lee et al, 1999).
Additionally, we have seen the rotation of RH’s
overall vowel triangle in the direction of the
vowel space typical of normal-hearing adults.
We conclude that changes in speech production
relative to speech processing strategies may be objectively measured using the Bark scale. For this child,
the transition into HiResolution sound processing
has led to significant normalization of her overall
vowel productions. In the future, spectrographic
evaluation of vowel productions may provide useful
insights into speech perception and may guide
individual programming (mapping) decisions.
Figure 1. RH’s vowel triangles for the cornerstone vowels over
time—shown as the narrow shaded triangle after 14 months
with SAS (baseline) and four subsequent test sessions with
HiRes. The dashed, gray triangle represents the mean, and the
gray ovals represent the range of vowel productions in normalhearing adults. Bark scale normative data for children are currently not available.
References
Kendig K. (2000) Is a 2000-year-old formula still keeping some secrets? Am Math
Monthly 107(5): 402-415.
Lee S, Potamianos A, Narayanan S. (1999) Acoustics of children’s speech:
Developmental changes of temporal and spectral parameters. J Acoust Soc Am
105(3): 1455-1468.
McCaffrey HA, Sussman HM. (1994) An investigation of vowel organization in
speakers with severe and profound hearing loss. JSHR 37(4): 938-951.
Syrdal AK, Gopal HS. (1986) A perceptual model of vowel recognition based on
the auditory representation of American English vowels. J Acoust Soc Am 79(4):
1086-1100.
Zwicker E. (1961) Subdivision of the audible frequency range into critical bands
(frequenzgruppen). J Acoust Soc Am 33(2): 248.
Figure 2. Changes in RH’s vowel space calculated using
Heron’s formula. The blue dotted lines represent the range, and
gray dashed line the mean, of vowel productions for five-yearold females with normal hearing. (Lee et al, 1999.)
HiResolution Sound
109
HiResolution Programming and Performance
in Young Children
Jennifer Mertes, Au.D.
Jill Chinnici, M.A.
Margaret Sampson, M.S.
Johns Hopkins University, Baltimore, MD, USA
Fitting HiResolution (HiRes) sound programs
is accomplished using the SoundWave software. SoundWave is fast and easy to use with
very young children because of its many automatic features, thereby decreasing programming time, minimizing stress on the child, and
reducing the number of appointments required.
Preliminary data from young children in our clinic
suggest that children using HiRes develop auditory
skills faster than children using conventional sound
processing. We studied two groups of children,
matched for age at implant, who were implanted at
less than 18 months of age. One group used conventional sound processing, and one group used HiRes.
Performance for the two groups was compared for
two measures—the IT-MAIS to assess auditory
skill development and the Reynell Development
Language Scales to evaluate verbal skill development.
The IT-MAIS is a 10-item parent-interview
schedule that assesses spontaneous auditory behaviors in daily living situations without visual cues.
Auditory skill development is evaluated in three
areas:(1) changing vocalizations in response to
sound, (2) alerting to sound, and (3) deriving
meaning from sound. Scores range from 0-40 and
a score of 40 indicates that a child has acquired all
of the skills evaluated by the test. The results on the
IT-MAIS, shown in Figure 1, indicate that children
using HiRes acquire auditory skills at a faster rate
than children using conventional sound processing.
Figure 1. Auditory skills as measured on the IT-MAIS. Three-month postimplant scores for 10 matched pairs of children—one group
using HiRes and the other group using conventional strategies.
110
Advanced Bionics® Auditory Research Bulletin 2005
“...children using HiRes are developing
good auditory and language skills, in
many cases at a faster rate than children
using conventional sound processing.”
The Reynell scales evaluate verbal comprehension
and expressive language skills. Figure 2 shows results
from the Reynell Developmental Language Scales
one year after implantation for 5 of the 10 matched
pairs of children. The data show that typically the
children using HiRes are developing receptive and
expressive language abilities faster than children
using conventional sound processing. Moreover,
standard scores above 82 (100 minus one standard
deviation) indicate skills commensurate with normalhearing peers. Three of the HiRes children are within
the normal score range for expressive language.
Overall, our experience has been that HiRes is
easy to program in young children and that children are developing good auditory and language
skills, in many cases at a faster rate than children using conventional sound processing.
Reference
Mertes J. (2004) HiRes and Children: Performance and Programming. Paper
presented at the VIII International Cochlear Implant Conference, Indianapolis,
Indiana, 10-13 May, 2004.
Figure 2. One-year postimplant Reynell Developmental Scales
receptive language (A) and expressive language (B) scores for
five matched pairs of children—one group using HiRes and the
other group using conventional strategies.
HiResolution Sound
111
Experience with the HiResolution Bionic Ear in Children
Carolyn J. Brown, M.S.
Craig A. Buchman, M.D.
University of North Carolina, Chapel Hill, NC, USA
“...the results of this retrospective study
show that children implanted with
the Bionic Ear system
derive excellent benefit from their devices
and HiRes sound processing.”
The purpose of this retrospective study was to
examine the benefit experienced by children who
received the HiResolution Bionic Ear. Fortynine children have been implanted with a CII (n
= 26) or HiRes 90K (n = 23) device at our center.
Twelve of the children were reimplanted with a
CII or 90K after failure of a previous device. This
summary focuses on the results obtained with children who received a Bionic Ear as their original
device and have been programmed with HiRes
(11 with the CII and 18 with the HiRes 90K).
The children ranged in age from 7 months to 11 years
at time of implant. Speech perception was evaluated
with the Early Speech Perception (ESP) Test and
the Phonetically Balanced Kindergarten (PBK) Test
(scored for both phonemes and words). The tests
were presented using live voice. Speech production
was evaluated with the Identifying Early Phonological Needs Test (EPNT). Numbers of children
with test scores are different among tests because
of age and the elapsed postimplant time interval
when measures were developmentally appropriate
for each child. Scores obtained at each child’s most
recent evaluation were included in the analyses.
Demographics for ESP Test
All
(n = 12)
Age at implant (yrs)
Duration of total use (yrs)
Duration HiRes use (yrs)
< 5 years
(n = 4)
≥ 5 years
(n = 8)
4.3
1.6
5.6
(0.7-11.4)
(0.7-3.3)
(2.0-11.4)
2.2
2.4
2.0
(0.6-3.5)
(0.6-3.5)
(0.7-3.3)
0.9
0.9
0.9
(0-1.4)
(0.6-1.4)
(0-1.4)
Figure 1. Demographics (mean and range) and results by age at
time of testing on the ESP Test.
112
Results are summarized together for the CII and
90K users because the electronics of the two systems
are identical and both systems implement HiRes
sound processing. Results for the ESP and PBK
Tests were analyzed by age at the time of testing
(younger or older than five years of age). Figure 1
shows that both younger and older children achieved
high levels of performance on the ESP Test and that
the test ceiling was reached on this measure. On the
PBK Test, the performance of the younger versus
the older children was comparable for phoneme
recognition. In contrast, the mean PBK-word
recognition score for the older children was higher
than the mean score for the younger children. This
finding can be anticipated because the vocabulary
on this test typically is difficult for younger children. In addition, some younger children’s scores
may have been poorer because of their inability to
produce some of the consonant blends required for
Advanced Bionics® Auditory Research Bulletin 2005
a correct response. Notably, the mean PBK-word
scores of the younger and older children were 41.3%
and 53.3%, respectively. This level of performance
indicates that all of the children obtain significant
speech recognition benefit from their implants.
Figure 3 summarizes the children’s speech production skills on the EPNT. This test is most appropriate for younger children. Thus, results were
analyzed for children implanted before and after
two years of age. Figure 3 shows that the children
in both groups accurately produced syllables, stress
and intonation, initial consonants, vowels and diphthongs. Production accuracy was lower for manner,
place, and voicing features as well as for final consonants. Overall, the results from the younger and
older children were similar although the children
implanted at an older age attained higher scores on
some of the items such as final consonants, diphthongs, and consonant voicing. However, two other
factors may have impacted the score comparisons.
First, the children implanted at less than two years of
age were approximately one year younger at the time
of testing. Second, two of the children implanted
at an older age had progressive hearing losses.
In summary, the results of this retrospective study show that children implanted with
the Bionic Ear system derive excellent benefit
from their devices and HiRes sound processing.
Demographics for PBK Test
Age at implant (yrs)
Duration of total use (yrs)
Duration HiRes use (yrs)
≥ 5 years
All
> 5 years
(n = 9)
(n = 3)
5.3
1.3
7.3
(1.2-14.9)
(1.2-1.4)
(2.0-14.9)
(n = 6)
2.3
3.1
1.9
(0.3-3.5)
(2.6-3.5)
(0.3-3.3)
0.8
0.8
0.8
(0-1.4)
(0-1.4)
(0-1.4)
Figure 2. Demographics (mean and range) and results by age at
time of testing on the PBK Test.
Demographics for EPNT
All
(n = 10)
Age at implant (yrs)
Mean age at test (yrs)
Duration of total use (yrs)
Duration HiRes use (yrs)
< 2 years
(n = 5)
≥ 2 years
(n = 5)
2.2
1.2
3.2
(0.7-11.4)
(0.7-3.3)
(2.0-11.4)
4.7
4.2
5.1
(0.7-11.4)
(0.7-11.4)
(0.7-11.4)
2.5
3.0
1.9
(0.6-3.5)
(0.6-3.5)
(0.7-3.3)
0.7
0.7
0.6
(0-1.4)
(0-1.4)
(0-1.0)
Figure 3. Demographics (mean and range) and results by age at time of testing on the Early Phonological Needs Test (EPNT).
HiResolution Sound
113
A “Benchmark” for Performance with the
CII HiResolution Cochlear Implant
Michael F. Dorman, Ph.D.
Anthony J. Spahr, Ph.D.
Arizona State University, Tempe, AZ, USA
“We conclude that ‘better’
implant processors
have some distance to go
to surpass the level of performance
afforded by the CII HiResolution system.”
When evaluating the performance of new processing
schemes, it is critical to understand the level of the
performance that can be achieved with previous
processing schemes. This summary describes the
performance of a patient fit with an Advanced Bionics
CII cochlear implant who achieved the highest CNC
score recorded in our laboratory—100% correct. The
performance of this patient on a battery of tests,
detailed below, sets the standard to be exceeded
if “new” signal processing strategies, e.g., current
steering or combined acoustic and electric stimulation, are to surpass “old” signal processing strategies.
The CII patient, HR4, noticed a hearing loss at age
23 and by age 34 was completely deaf—with tactile
sensations only at sound pressure levels exceeding
90 dB for standard audiometric frequencies (2508,000 Hz). He was implanted within a year of his
complete deafness. HR4’s processor was configured
with a 16-channel HiRes pulsatile program—with
a pulse duration of 11 µs per phase and a stimulation rate of 1449 pulses per second per electrode.
HR4’s speech recognition performance was
compared to the performance of six normal-hearing
undergraduate students at Arizona State University. Tests administered to all subjects included:
• CNC words (50 items)
• CUNY sentences (24 sentences and
approximately 200 words)
• HINT sentences (250 sentences and 1320
words, presented in quiet)
• Arizona Biomedical Institute (AzBio)
sentences (40 sentences and approximately 270
words, presented in quiet)
114
Advanced Bionics® Auditory Research Bulletin 2005
• Consonant identification (20 consonants in an
/I/-consonant-/I/ context)
• Vowel identification (13 computer-synthesized
vowels in a /b/-vowel-/t/ context)
• CUNY and AzBio sentences presented in fourtalker babble
All of the scores for HR4 are high (Table 1). His
scores for speech material presented in quiet,
including words, sentences, consonants, and
vowels, match or closely approximate the scores
for the normal-hearing group. His score for the
most difficult test used in quiet in standard clinical
practice, recognition of the monosyllabic CNC
words, is 100%. In contrast, some of his scores
for sentences presented in competing speechbabble noise are worse than normal. His score for
the CUNY sentences at a signal-to-noise ratio of
+10 dB is 98%; however, his scores for the AzBio
sentences at signal-to-noise ratios of +10 and +5
dB are below those of the normal-hearing subjects.
Such high scores overall are consistent with
HR4’s ability to communicate with ease in most
listening situations. He can understand conversations not directed to him and can identify speakers
by regional dialect. He can mimic voices and
accents that he has heard only after receiving the
implant. His speech recognition abilities are representative of the very best that can be achieved
with present-day cochlear implant systems. We
conclude that “better” implant processors have
some distance to go to surpass the level of performance afforded by the CII HiResolution system.
Table 1. Percent correct scores for implant patient HR4
and for six subjects with normal hearing.
Test
HR4
Normal Hearing*
CNC words
100%
98.3% ± 0.6
CUNY sententecs
100%
99.0% ± 0.5
HINT sentences
100%
100%
AzBio sentences
98%
99.9% ± 0.1
20 consonants
94%
99.3% ± 0.7
13 synthetic vowels
97%
99.3% ± 0.4
CUNY sentences +10 dB SNR
98%
97.1% ± 1.1
AzBio sentences +10 dB SNR
90%
99.7% ± 0.2
AzBio sentences +5 dB SNR
77%
99.5% ± 0.2
* Scores for the six normal-hearing listeners are given as the
mean and standard error of the mean.
HiResolution Sound
115
Novel Processing
Today’s HiResolution Sound (HiRes) consumes only a fraction
of the capability available in the CII and HiRes 90K electronic
platforms. Current steering for increasing the number of spectral
channels, neural conditioning for stochasticity, physiologically
based stimulation, and low-power sound processing are under
development.
All CII and HiRes 90K users will have access to advances in
HiResolution Sound through software updates without having to
undergo additional surgery.
Current Steering and Spectral Channels in
HiResolution Bionic Ear Users
Study Sites in North America
Boys Town National Research Hospital, Omaha, Nebraska
Denver Ear Associates, Denver, Colorado
L’ Hôtel-Dieu de Québec, Québec City, Québec
Houston Ear Research Foundation, Houston, Texas
Indiana University, Indianapolis, Indiana
Johns Hopkins University, Baltimore, Maryland
Mayo Clinic, Rochester, Minnesota
Medical College of Georgia, Augusta, Georgia
Medical College of Wisconsin, Milwaukee, Wisconsin
New York University, New York, New York
Ottawa Hospital (Civic Campus), Ottawa, Ontario
Sunnybrook & Women’s College Health Sciences Centre,
Toronto, Ontario
University of Iowa, Iowa City, Iowa
University of Minnesota, Minneapolis, Minnesota
Washington University, St. Louis, Missouri
Figure 1. Delivering current proportionally between two adjacent contacts stimulates a different neural population than
current delivered to each contact alone. In most subjects,
different pitches are perceived as the proportion of simultaneously delivered current is varied between adjacent electrode
contacts. This schematic illustrates the creation of an additional
channel when 70% of current is delivered to one electrode and
30% is delivered to the other electrode simultaneously. Theoretically, a subject will perceive a pitch that is intermediate to
the pitches heard when each contact is stimulated alone.
118
The number of spectral channels is defined as the
total number of distinct pitches a cochlear implant
user can perceive when current (stimulation) is
delivered to different locations along the cochlea. All
cochlear implant sound-processing strategies developed to date deliver a fixed number of spectral channels—limited by the number of electrode contacts
and the residual hearing ability of the user. Although
early research suggested that additional “virtual”
spectral channels could be created by stimulating two
electrodes simultaneously (Townsend et al, 1987;
Wilson et al, 1994), virtual channels have never been
implemented in a wearable sound processing strategy.
In the CII and HiRes 90K implants, the number
of sites of stimulation can be increased beyond the
number of electrode contacts. Through simultaneous
delivery of current to pairs of adjacent electrodes,
stimulation can be “steered” to sites between the
contacts by varying the proportion of current delivered to each electrode of the pair, as illustrated in the
schematic in Figure 1. Current steering is possible
because each of the 16 electrode contacts is powered
by an independently programmable current source.
Donaldson et al (2005) recently investigated the
use of current steering in a small number of CII
and HiRes 90K users. (See also Donaldson et al at
page 120 of this bulletin.) Their results showed that
subjects could hear multiple unique pitches when
current was steered between adjacent electrode
contacts. Moreover, their findings showed that a
constant level of current produced the same loudness
whether it was apportioned simultaneously between
two electrodes or delivered all to one contact. Thus,
when current steering is implemented in a sound
processing strategy, comfort (M) levels for the
virtual spectral channels can be interpolated from
the M levels of the two adjacent (electrode) channels.
This multicenter study is investigating in a large
sample of postlinguistically deafened adults the
number of spectral channels that can be resolved
with the HiRes 90K or CII device. According to
study protocol, each subject was tasked with loudness
balancing and pitch-ranking electrode pairs (apical 2-
Advanced Bionics® Auditory Research Bulletin 2005
3, medial 8-9, basal 13-14). A two-alternative forcedchoice (2AFC) paradigm is used where subjects
identify the tone higher in pitch while current is
varied proportionally between electrodes in each pair.
Figure 2 summarizes the data from 65 implanted
ears (8 subjects had bilateral implants) for the medial
pair of electrodes. The majority of patients can hear
at least one additional spectral channel between
the two electrodes. The average number of spectral
channels that could be distinguished was 4.5 for the
basal electrode pair, 7.1 for the mid-array electrode
pair, and 5.9 for the apical electrode pair. Assuming
that the numbers of spectral channels for these three
electrode pairs are representative of the entire array,
the potential number of spectral channels overall
can be calculated. For these 65 subjects, the potential
number of channels ranges from 7 to 451 (Figure 3).
These results indicate that additional spectral resolution can be created using current steering. More
spectral channels means that the HiRes system has
the potential to provide enhanced spectral resolution in addition to the enhanced amplitude and
temporal resolution that already exists in HiRes
sound processing. Increased spectral information
through current steering should lead to improved
speech recognition in noise and music perception. A sound processing strategy that incorporates
current steering is under development and undergoing initial evaluation in CII and HiRes 90K users.
“Increased spectral information
through current steering should lead to
improved speech recognition in noise
and music perception.”
Figure 2. Proportion of 65 ears that cannot discriminate
between two electrodes in the medial pair (15%), that can hear
different pitches for each electrode but perceive no intermediate pitches (9%), that can perceive one intermediate pitch
(26%), and that can perceive two or more intermediate pitches
(50%). Over 75% of these ears can hear at least one intermediate pitch as a result of current steering.
References
Donaldson GS, Kreft HA, Litvak L. (2005) Place-pitch discrimination of singleversus dual-electrode stimuli by cochlear implant users. J Acoust Soc Amer, in
press.
Townsend B, Cotter N, van Compernolle D, White RL. (1987) Pitch perception by
cochlear implant subjects. J Acoust Soc Amer 82:106-115.
Wilson BS, Lawson DT, Zerbi M, Finley CC. (1994) Recent developments with the
CIS strategies. In: Hochmair-Desoyer IJ, Hochmair ES, eds, Advances in Cochlear
Implants. Proceedings of the Third International Cochlear Implant Conference,
Innsbruck, Austria, April, 1993. Austria: Datenkonvertierung, Reproduktion and
Druck, 103-112. Manz: Wein.
Figure 3. Rank-ordered potential number of spectral channels
based on responses of 65 subjects.
Novel Processing
119
Discrimination of Single- and Dual-Electrode Stimuli
in Clarion CII Users
Gail S. Donaldson, Ph.D.
University of South Florida, Tampa, FL, USA
Heather A. Kreft, M.A.
University of Minnesota, Minneapolis, MN, USA
Leonid Litvak, Ph.D.
Advanced Bionics Corporation, Valencia, CA, USA
Cochlear implant users have difficulty performing
listening tasks that require good spectral resolution,
such as speech recognition in the presence of background noise. In some cochlear implant users, spectral resolution may be limited by poor neural survival;
however, in others, it may be limited primarily by the
finite number of intracochlear electrodes available
for stimulation. One strategy for increasing spectral
resolution in the latter group of cochlear implant
users is to utilize dual-electrode stimulation. In
this approach, weighted stimulation of two physical
electrodes, either simultaneously or with a brief
temporal separation, produces pitch percepts intermediate to those produced by each electrode alone.
Although previous studies have demonstrated the
feasibility of this approach (Townshend et al, 1987;
McDermott and McKay, 1994; Wilson et al, 1994),
it is not known just how many place-pitch steps
can be generated using dual-electrode stimulation
in cochlear implant users with present-day devices.
Figure 1. Thresholds for discriminating single-electrode versus
dual-electrode stimuli for medium loud stimuli at three locations along the electrode array.
120
The present study (Donaldson et al, 2005) measured
discrimination thresholds for single- versus dual-electrode stimuli as a means of estimating the number of
place-pitch steps that could be achieved with simultaneous, dual-electrode stimulation. Six postlinguistically deafened adults with the CII device were
tested. Pairs of adjacent electrodes at each of three
locations along the implanted array (basal, middle,
and apical) were tested using stimuli that were
balanced in loudness to a “medium loud” level. The
middle electrode pair also was tested using stimuli
balanced to a “medium soft” level. Stimuli were
200 ms pulse trains presented in monopolar mode.
Discrimination thresholds were obtained using a
two-alternative forced choice (2AFC) procedure in
which the subject heard a single-electrode stimulus
(apical electrode stimulated alone) and a dual-electrode stimulus (apical and basal electrodes stimulated simultaneously) in random order on each trial.
The subject’s task was to choose the stimulus with
the higher pitch; thus, a correct response occurred
when the dual-electrode stimulus was selected. The
proportion of current (α) directed to the more basal
electrode of the dual-electrode pair was varied to
determine the relative current weighting at discrimination threshold. Using this metric, a threshold value
of α = 0.2 would indicate that a pitch (or quality)
change was detected when 20% of the current was
directed to the more basal electrode of the pair.
Similarly, a threshold of α = 0.9 would indicate
that a pitch change was not detected until 90% of
the current was directed to the more basal electrode.
Thresholds varied considerably across subjects, and
in some cases across electrodes within subjects. As
shown in Figure 1, for the medium-loud stimuli,
one subject (D18) could not discriminate adjacent
Advanced Bionics® Auditory Research Bulletin 2005
“...the present results suggest that
dual-electrode stimulation could increase
the number of place-pitch steps
available to most CII users.”
single electrodes in the apical cochlear location
(α > 1) and two other electrode pairs could not be
tested. For the 15 remaining electrode pairs, discrimination thresholds ranged from α = 0.12 to α = 0.64,
with an average value of α = 0.36. A level effect was
observed in three subjects and in the mean data, with
medium-soft stimuli producing significantly larger
mean thresholds than medium loud stimuli (Figure 2).
The present results suggest that dual-electrode
stimulation could increase the number of placepitch steps available to most CII users. Five of six
subjects in this study were able to perceive an intermediate pitch between the pitches of adjacent electrodes in all three cochlear locations. Four subjects
had thresholds less than α = 0.3 for at least one
electrode pair, suggesting that three or more intermediate pitches may be possible in many cases.
Figure 2. Comparison of discrimination thresholds for mediumloud and medium-soft stimuli at the middle electrode location.
References
Donaldson GS, Kreft HA, Litvak L. (2005) Place-pitch discrimination of singleversus dual-electrode stimuli by cochlear implant users. J Acoust Soc Am, in press.
McDermott HJ, McKay CM. (1994) Pitch-ranking with nonsimultaneous dualelectrode electrical stimulation of the cochlea. J Acoust Soc Am 96:155-162.
Townshend B, Cotter N, van Compernolle D, White RL. (1987) Pitch perception by
cochlear implant subjects. J Acoust Soc Am 82:106-115.
Wilson BS, Lawson DT, Zerbi M, Finley CC. (1994) Recent developments with the
CIS strategies. In Hochmair-Desoyer IJ and Hochmair ES, eds. Advances in Cochlear
Implants. Proceedings of the Third International Cochlear Implant Conference,
Innsbruck, Austria. April 1993. Wein: Manz.
Novel Processing
121
Virtual Channels: Improvements in Frequency Resolution
Martina Brendel, M.Sc.
Corinna Habermann
Carolin Frohne-Büchner, Ph.D.*
Andreas Büchner, Ph.D.
Timo Stöver, M.D.
Thomas Lenarz, Prof., M.D., Ph.D.
Medizinische Hochschule Hannover, Hannover, Germany
* also with Advanced Bionics Corporation, Europe
Currently the Advanced Bionics HiRes system
stimulates the auditory nerve via 16 physical electrode contacts. The intention is to create 16 channels to compensate (to some extent) for lost function
resulting from damage to the thousands of inner hair
cells. The implant’s independent current sources make
it possible to stimulate adjacent electrode contacts
simultaneously—a new technique known as “current
steering” whereby current is injected in two adjacent
contacts simultaneously. Two independent currents
are delivered to the adjacent contacts and sum in the
cochlea to create an intermediate (“perceptual” or
“virtual”) channel, as shown in the Figure 1 schematic. Through current steering, the CI user may
perceive an increased number of pitches and hence
experience improved frequency resolution. Conceptually, such improved frequency resolution would
lead to better sound quality, particularly better speech
in noise recognition and better music perception.
The goals of this study were to investigate (1)
whether subjects were able to distinguish adjacent
electrodes and (2) how many subjects were able to
perceive a distinct pitch if the current was steered
between two adjacent contacts. The determination
of distinct pitch percepts was implemented with a
modified version of the SoundWave fitting software.
The SoundWave interface allows several options
in parameter settings, including the minimum
spacing between stimulating channels—0.5
(one intermediate channel), 0.25 (three), or
0.125 (seven). In this study a minimum channel
spacing of 0.5 was investigated. The number
of stimuli per channel was set to a value of 1.
Measurements started with each subject tasked to
pitch rank stimuli from channels spaced two electrodes apart. For each correct response, the distance
between channels was reduced to 1 and further to
0.5 (the medial virtual channel). Each combination
was tested eight times. With six correct answers,
the combination was assumed to be discriminable.
The results were reported in table format. This
table details the 16 physical channels, the discriminable channels, and the resolution achieved.
Figure 1. Principle of a conventional physical channel (A) and a
virtual channel realised by current steering (B).
122
The study included 46 postlinguistically deafened
adults with a mean age of 48 years (range 30-77 years)
and a mean duration of deafness of 6.7 years (range
0-36 years). All participants used HiRes processing.
Advanced Bionics® Auditory Research Bulletin 2005
“...more than 80% of the study group
could perceive and discriminate
virtual channels...”
The measurements took place between the initial
switch-on (four to six weeks postoperatively) and 3.5
years post switch-on. In a subgroup, the measurement
was repeated during the first fitting phase and at one
and three months thereafter to evaluate whether a
possible learning effect was reflected in the data.
Our results showed that more than 80% of the study
group could perceive and discriminate virtual channels—with approximately 20% distinguishing eight
intermediate channels, which is in accordance with
half of the physical channels. All participants show
areas with poorer or better resolution along the
array—with the majority showing poorer resolution
in the basal part of the electrode array compared
to the mid-region or apex (Figure 2). Improved
resolution was observed between initial and three
months post-initial activation, without any specific
training for the tasks. Figure 3 shows the data
for one of the subjects typifying these findings.
Our findings clearly reflect the potential for advanced
speech coding strategies leading to improved spectral resolution. The large majority of these study
participants perceived virtual channels despite a
Figure 2. The number (percent) of participants able to distinguish the channel at a resolution of 0.5 for each electrode
contact.
lack of any specific virtual-channel perception
training or everyday experience with virtual channel
processing. Since implant experience seems to have
an impact on discrimination abilities, one may
assume that virtual channels are probably advantageous early on—even though users cannot initially
perceive pitch differences, but may learn over time.
Figure 3. Pitch resolution ability over time for one patient. Resolution of 0.5 = can hear one medial virtual channel, resolution of
1 = can pitch rank adjacent electrodes, resolution of 2 = can pitch rank channels spaced two electrodes apart, and resolution of
> 2 = cannot pitch rank channels spaced two electrodes apart.
Novel Processing
123
Preliminary Results of Pitch Strength with
HiRes 120 Spectral Resolution
Anthony J. Spahr, Ph.D.
Arizona State University, Tempe, AZ, USA
Gulam Emadi, Ph.D.
Advanced Bionics Corporation, Valencia, CA, USA
“...current steering is a viable alternative
to the more traditional variants of
Continuous Interleaved Sampling...”
Cochlear implant patients commonly achieve
moderate-to-high levels of open-set speech understanding, but even patients achieving the highest
levels of speech understanding often report suboptimal quality of voices, environmental sounds, and
music. The perception of details in these sounds
may be limited by several factors, including a
lack of neural survival, the number of electrodes
available in modern devices (12–22), or even the
method used to deliver electrical stimulation.
For a patient with a full complement of active
physical electrodes, HiRes 120 creates currentsteered channels by pairing adjacent electrodes (12, 2-3, 3-4, and so forth). Each of the 15 electrode
pairs can be used to “steer” current to either physical
electrode in the pair or to any of seven intermediate
locations, resulting in a combined total of 120 real
and “virtual” electrode locations that can be incorporated into the strategy. Conceptually, this greater
number of effective electrode locations (as compared
to the 16 physical contacts alone) produces a more
accurate representation of frequency information than the standard HiRes strategy—resulting
in perceived improvements in sound quality.
We have examined the effect of current steering on
pitch strength or salience using a logarithmically
spaced tone sweep of frequencies between 800 Hz
and 1600 Hz. Patients were asked to continually
evaluate the pitch strength of the tone sweep using
a subjective rating scale of 1 (lowest pitch strength)
to 9 (highest pitch strength). Each trial consisted
HiRes 120, the most recent real-time sound coding of four tone sweeps, each 15 seconds in duraenhancement from Advanced Bionics, is the first tion, with sweeps 1 and 3 increasing in frequency
attempt to improve sound quality through the use of and sweeps 2 and 4 decreasing in frequency.
current steering. Current steering refers to the simul- Tone sweeps were presented via direct input to
taneous stimulation of two closely-spaced electrodes the Platinum Series Processor at 60 dB SPL. All
in order to position stimulation with greater resolu- patients completed a practice session in order to
tion than is achievable with single electrodes alone. familiarize themselves with the task. Data were
Previous research has shown that current steering collected using five trials each for standard HiRes
can improve performance on pitch ranking proce- and HiRes 120. The condition order was randomdures (Townsend et al, 1987; Wilson et al, 1994). ized and patients were blinded to the test condition.
124
Advanced Bionics® Auditory Research Bulletin 2005
Although the stimulus consisted of only a pure
tone at any given instant in time, patients reported
that, in both strategies, the pitch strength of the
tone varied significantly with frequency. Patients
attributed low pitch-strength ratings to perceptions of multiple tones or a loss of tonal quality
(commonly reported as buzzes or squeaks). A
paired t-test revealed a significant (p < .05) effect
of strategy on pitch strength for eight of the 10
subjects. Of these eight cases, five patients exhibited greater pitch strength ratings with the HiRes
120 enhancement, and three patients exhibited
greater ratings with the standard HiRes strategy.
Figure 1 displays results from two individual subjects
whose pitch strength ratings indicated a clear preference for one strategy over the other. At the conclusion of testing, both patients were able to identify
their preferred strategy in a back-to-back comparison.
The outcomes of this study demonstrate that current
steering is a viable alternative to the more traditional variants of Continuous Interleaved Sampling
(CIS) strategies used in modern cochlear implant
systems. We are currently conducting a long-term
study to examine the effect of current steering on
speech understanding, melody identification, sound
quality, pitch strength, and frequency discrimination.
Figure 1. Pitch-strength ratings as a function of test frequency
for two subjects who indicated a strategy preference.
Patient A preferred HiRes 120. Patient B preferred standard
HiRes.
References
Townsend B, Cotter N, van Compernolle D, White RL. (1987) Pitch perception by
cochlear implant subjects. J Acoust Soc Am 82:106-115.
Wilson BS, Lawson DT, Zerbi M, Finley CC. (1994) Recent developments with the
CIS strategies. In Hochmair-Desoyer I, Hochmair E, eds, Advances in Cochlear
Implants. Proceedings of the Third International Cochlear Implant Conference,
Innsbruck, Austria, April, 1993. Manz, Wein, Austria: Datenkonvertierung,
Reproduktion and Druck, 103-112.
Novel Processing
125
Conditioning Pulse Trains in Cochlear Implants:
Effects on Loudness Growth
Robert S. Hong, M.D.
University of Iowa, Iowa City, IA, USA
Jay T. Rubinstein, M.D., Ph.D.
University of Washington, Seattle, WA, USA
High-rate conditioning pulse trains (5,000 pulses
per second) have been shown to increase the
dynamic range of cochlear implant recipients,
with the largest increase for each subject having
a mean value of 6.7 dB (Hong et al, 2003). These
conditioning pulse trains are theorized to enhance
the encoding of acoustic information by inducing
stochastic resonance in the auditory nerve (Rubinstein and Hong, 2003). However, it is unclear if
the increases in dynamic range seen experimentally
will be clinically beneficial, because the loudness
growth seen with conditioning may be relatively
shallow only at the extremes of the dynamic range.
This study characterized the effects of conditioning
stimuli on loudness growth functions in seven
postlinguistically deafened adults (ages 37-89)
who use the CII cochlear implant. The loudness
growth functions for each subject were measured
with sinusoidal stimuli, both with and without
the presence of a conditioning pulse train (5,000
pulses per second). The conditioner was presented
at the level that resulted in the largest increase
in dynamic range for each patient. Loudness was
evaluated using a subjective rating scale, with
0 = threshold and 100 = most comfortable loudness.
In general, the shapes of the loudness growth functions with conditioner demonstrated a shallower
growth of loudness across the dynamic range than
without conditioner. Figure 1 shows typical loudness growth functions from one subject. In order
to quantify the differences in loudness growth with
and without conditioner, the two functions for each
subject were fit with exponential, power, and cumulative Gaussian functions. Table 1 lists the results
from each curve fit for the seven subjects. The data
first were fit by an exponential function of the form:
loudness = K0 + K1e α (stimulus level). The data then were
fit by a power function of the form: loudness = K₀
Figure 1. Loudness growth functions for a typical subject (S7). Filled circles represent the loudness judgments with conditioner and
crosses represent loudness growth without conditioner. Data were fit with exponential, power, and cumulative Gaussian functions.
Solid lines show curve fits with conditioner. Dashed lines show curve fits without conditioner.
126
Advanced Bionics® Auditory Research Bulletin 2005
(stimulus level) β. The exponents of the exponential (α) and power (β) functions reflect the overall
rate of loudness growth, with a shallower loudness growth denoted by a smaller exponent. For
the Gaussian fit, the relative spread is a measure
of the loudness growth of the middle of the function, with the relative spread equal to the standard
deviation divided by the stimulus level that corresponds to 50% of total loudness. A larger relative spread is equivalent to a shallower loudness
growth across the middle of the loudness function.
The loudness growth was shallower with the addition of conditioner in all cases for all fitting models
(except for exponential fits in two subjects). Furthermore, shallower loudness growth was not confined
to the extremes of the dynamic range, but also was
present in the middle of the function, as evidenced
by the larger values for relative spread with conditioner. The data were fit well with exponential (R
= 0.888 to 0.968), power (R = 0.887 to 0.996), and
cumulative Gaussian (R = 0.928 to 0.996) functions.
These results suggest that signal processing strategies
that incorporate conditioning may be able to take
advantage of the enhanced dynamic range provided
by a conditioner. By requiring less compression of the
input signal, conditioned strategies may lead to less
distortion perceived by cochlear implant patients,
resulting in improved speech and music perception.
“...signal processing strategies
that incorporate conditioning
may be able to take advantage of the
enhanced dynamic range provided...
resulting in
improved speech and music perception.”
Table 1. Exponential, power, and cumulative
Gaussian fits of loudness growth data.
Increase α - Exponential
β - Power
Cond.
RS - Gaussian
S
DR (dB)
None
Cond.
None
None
Cond.
1
2.9
0.141
0.155
5.120
2
3.8
0.247
0.181
3.625 0.192
0.257
11.512 5.165 0.093
0.180
3
4.4
0.063
0.066
4.006
3.062 0.236
0.301
4
7.3
0.068
0.050
4.970
2.472 0.197
0.339
5
8.5
0.074
0.048
4.802
2.379 0.197
0.343
6
9.1
0.039
0.034
2.948
2.046 0.288
0.351
7
11.5
0.076
0.059
6.114
2.316 0.157
0.335
For each subject (S), the increase in dynamic range (DR)
with conditioner (Cond.) compared to without conditioner
(None). The exponent (α) of the exponential fit, the exponent
(β) of the power fit, and the relative spread (RS) of the Gaussian fit are given for the loudness growth functions. The boldfaced values represent shallower loudness growth.
References
Hong RS, Rubinstein JT. (in press) Conditioning pulse trains in cochlear implants:
effects on loudness growth. Otol Neurotol.
Hong RS, Rubinstein JT, Wehner D, Horn D. (2003) Dynamic range enhancement
for cochlear implants. Otol Neurotol 24:590-595.
Rubinstein JT, Hong RS. (2003) Signal coding in cochlear implants: exploiting
stochastic effects of electrical stimulation. Ann Otol Rhinol Laryngol 112:14-19.
Novel Processing
127
Effects of High-Rate Conditioning Stimuli on Frequency
and Intensity Discrimination in Cochlear Implant Users
Ted A. Meyer, M.D., Ph.D.
Medical University of South Carolina, Charleston, SC, USA
Jay T. Rubinstein, M.D., Ph.D.
University of Washington, Seattle, WA, USA
Robert S. Hong, M.D.
Haiming Chen, M.S.
University of Iowa, Iowa City, IA, USA
Cochlear implant users are limited by small dynamic
ranges and suboptimal frequency, intensity, and
temporal discrimination capabilities. Future gains
in performance may depend upon developing
processing strategies that better approximate the
neural responses of the normal auditory system.
Previous studies have demonstrated that the addition of a high-rate conditioning pulse train produces
pseudo-random activity in a computational model
of deafened auditory neurons and in animal studies
(Litvak et al. 2001, Rubinstein et al. 1999). This
pseudo-random neural activity, resulting from desynchronization of refractory periods by the high-rate
conditioner, was hypothesized to serve as a source
of noise in the auditory pathway that might be used
to enhance signal encoding (stochastic resonance).
Figure 1. Averaged frequency discrimination for 13 listeners
with and without conditioner. Error bars represent +/- S.E.M.
Sigmoidal curves were fit to the data. These listeners showed
an average improvement in frequency discrimination of 1.5 Hz
when the conditioner was added.
128
We are investigating the clinical effect of a high-rate
conditioner in cochlear implant users by evaluating
frequency and intensity discrimination. Discrimination was measured both with and without the
conditioning pulse train. Thirteen postlinguistically
deafened adult cochlear implant users participated
in this study. All subjects were implanted with the
Clarion CII device at the University of Iowa Hospitals and Clinics. All electrodes were fully inserted
into the scala tympani. Subjects had from 4 to 26
months experience with their implants. Stimuli
were delivered to the implanted electrode array
using the Clarion Research Interface for secondgeneration Clarion products (CRI-2). The software for the CRI-2 was custom designed using
MATLAB and Texas Instruments Code Composer.
Electrical stimuli were presented to the most
apical electrode pair (E1 and E2) in bipolar mode.
For the frequency discrimination experiment, the
test stimuli were two sinusoidal bursts of the same
frequency and one sinusoidal burst of a higher
frequency presented with anodal phase first to E1.
The sinusoids were presented at the listener’s most
comfortable level, and the interstimulus interval
(ISI) was one second. The conditioning stimulus
was a continuous 5,000 pps biphasic unmodulated
pulse train with a pulse phase duration of 50 µs.
The cathodal phase was delivered first to E1. The
presentation level of the conditioner was chosen to
give the listener the largest possible gain in dynamic
range on that electrode pair. A three-interval, threealternative forced-choice procedure with two references of the same frequency (202 Hz) and one test
stimulus higher in frequency than the references
(202 + df Hz) was used to determine frequency
discrimination. The listeners were asked to identify
the stimulus with the highest pitch. The difference limen (df/f ) at which each subject’s ability
reached P(C) = 65% (d’ = 1.0 for a three-alternative
forced-choice task) was chosen as the measure of
discrimination. Frequency discrimination improved
with the addition of the conditioner for 10 of the
13 listeners (1-11 Hz, 3-64%). When averaged
across the 13 subjects, frequency discrimination
improved by 13% (1.5 Hz), as shown in Figure 1.
Advanced Bionics® Auditory Research Bulletin 2005
For intensity discrimination, three 500-msec
202-Hz sinusoids were presented using an ISI of
500 msec. Stimuli were presented at levels spanning
10-90% of each listener’s dynamic range. The two
reference stimuli had identical intensities (I), and
the third was greater in intensity (I + dI). Intensity
discrimination was measured using a three-interval
three-alternative forced choice, two-down one-up
adaptive staircase paradigm. The conditioner was
identical to that used for the frequency discrimination experiment. The listeners were asked to identify the interval containing the loudest stimulus.
All listeners were able to make fine intensity judgments in the subthreshold region, which reflected the
expanded dynamic range created by the conditioner.
Three subjects showed improvement, five had no
difference, and four showed some decrement in intensity discrimination with the conditioner compared
to without the conditioner (Figure 2). Word (CNC)
recognition was correlated with frequency discrimination (r = 0.55; p < .05). No significant relation was
found between intensity discrimination and word
(CNC) or sentence (HINT) recognition scores.
These experiments demonstrate that frequency
and intensity discrimination on a single electrode
pair can be improved by adding a conditioning
stimulus. Coupled with an average increase in
dynamic range of 7 dB (Hong et al, 2003), these
data show a substantial benefit from re-introducing stochastic resonance to the auditory nerve
in cochlear implant users. The large variability
in benefit seen in cochlear implant users may be
related to the number of surviving neurons, the
functional capacity of the surviving neurons, and
other factors such as cognitive abilities or length
of deafness. Even if stochastic resonance causes the
remaining neurons to function more like normal
auditory neurons, there will continue to be a large
amount of variability in this population. However,
if pulse conditioning helps poorer performers to a
greater extent than better performers, intersubject
variability may be reduced to some extent. Further
research will explore other potential benefits of
stochastic resonance for cochlear implant recipients.
“These experiments demonstrate that
frequency and intensity discrimination
on a single electrode pair
can be improved
by adding a conditioning stimulus.”
Figure 2. Weber functions for intensity discrimination. The
Weber fraction (dI/I) corresponding to P(C) = 71% is plotted on
the y-axis, and the presentation intensity in dB is plotted on the
x-axis. Data were averaged across subjects with better intensity
discrimination (n = 6) and poorer intensity discrimination (n =
6). Results show that all subjects can detect changes in intensity at levels below their thresholds without the conditioner.
For the better performers (lower curves), the conditioner might
actually be impeding their intensity discrimination, while for
the poorer performers (higher curves), the conditioner does not
affect intensity discrimination.
Acknowledgements
This work was supported by NIH/NIDCD program project DC00242, Neural
Prosthesis Contract DC92107, Advanced Bionics Corporation, and Texas
Instruments.
The authors thank Dan Wehner for designing the custom interface used for patient
testing and Gang Chen for subsequent modifications to the interface.
References
Hong RS, Rubinstein JT, Wehner D, Horn D. (2003) Dynamic range enhancement
for cochlear implants. Otol Neurotol 24:590-595.
Litvak L, Delgutte B, Eddington D. (2001) Auditory nerve fiber responses to electric
stimulation: Modulated and unmodulated pulse trains. J Acoust Soc Am 110:368379.
Rubinstein JT, Wilson BS, Finley CC, Abbas PJ. (1999) Pseudospontaneous activity:
stochastic independence of auditory nerve fibers with electrical stimulation. Hear
Res 127:108-118.
Novel Processing
129
A Novel Speech Processing Algorithm
Based on Phenomenological Models of the Cochlea
Gene Fridman, M.S.
University of California, Los Angeles, CA, USA
Tracey Kruger, M.S.
Leonid Litvak, Ph.D.
Advanced Bionics Corporation, Valencia, CA, USA
The unifying feature of current commercial
cochlear implant algorithms is to separate the
input signal into frequency components so that
spectral information is delivered to each electrode
according to its spatial location, thus taking advantage of the tonotopic organization of the cochlea
(Greenwood, 1990). In the proposed new algorithm, we add several details of signal transduction from the basilar membrane to the ganglion
cells, summarized in the block diagram in Figure 1.
The consequent outcome of this new stimulation
algorithm is that we introduce temporal aspects
of acoustic information to the stimulation pattern.
Figure 1. The FMS algorithm models the average firing behavior of the spiral ganglion cells in response to basilar membrane vibration. A stimulation pulse is presented at integer multiples of the period from the output of the bandpass filters in first-in first-out
order from the pulse generation queue. The number of integer multiples depends upon the amplitude of the filter output. The outer
hair cell model emulates lateral suppression by computing the output of each filter as the difference of the given channel with the
negative-weighted sum of the neighboring channels.
130
Advanced Bionics® Auditory Research Bulletin 2005
The temporal information is provided to the auditory nerve by maintaining the stimulation phase
coherence with the incoming acoustic signal. Thus
the firing rate on any given channel is modulated
in response to temporal frequency changes of the
incoming signal. We refer to our entire algorithm
as frequency modulated stimulation (FMS) to
convey the fact that we are encoding the acoustic
information by instantaneous frequency or phase.
In a healthy cochlea, the spiral ganglion cells located
at the apical end of the basilar membrane are known
to respond in phase with the vibration period of the
basilar membrane. At the higher frequency locations along the membrane (basal locations), the
neural responses are also phase-locked, but occur at
integer multiples of the basilar membrane vibration
periods. We emulate this phase-locking behavior in
our algorithm and refer to this aspect of the strategy
as the phase-locking model (PLM). This stimulation
pattern is demonstrated in Figure 2 in response to a
500 Hz pure tone. Since PLM tries to emulate phase
locking, the biphasic pulses are introduced only
once per period on each electrode. Consequently,
the number of pulses is fewer on all electrodes, in
contrast to CIS. The use of fewer pulses also reduces
the overall power consumed by the hardware.
Responses of the basilar membrane are somewhat
nonlinear. One consequence of this nonlinearity is
that one tone can suppress the response to another
tone that is close in frequency ( Javel et al, 1978).
This suppression mechanism is similar in effect to
the lateral inhibition behavior observed in other
sensory modalities (e.g., the on-center/off-surround
cells in the retina). In the FMS algorithm, the
basilar membrane nonlinearity (which is thought to
originate at outer hair cells) is modeled by a lateral
suppression network. The lateral suppression network
is designed to emphasize spectral stimulation peaks.
We refer to the lateral suppression network in
our strategy as the outer hair cell model (OHM).
“The consequent outcome of this
new stimulation algorithm
is that we introduce
temporal aspects of acoustic information
to the stimulation pattern.”
Figure 2. Comparison of 8-channel CIS and 8-channel FMS
waveforms in response to a pure tone at 500 Hz (black pulses).
Electrode 0 indicates the lowest frequency electrode, while
electrode 7 indicates the highest frequency electrode. Gray
pulses indicate the change in the stimulation pattern when the
pure tone input frequency is increased to 600 Hz.
Finally, the firing rates of auditory nerve fibers are not
constant, but increase with stimulus amplitude. The
largest rates occur near onsets of sounds and can be
as high as 1000 spikes/sec (Chimento & Schreiner,
1991; Smith 1977). For low spontaneous rate fibers,
the sustained rates can vary from 0 spikes per second
in quiet to 300 spikes per second for loud sounds
(Kiang et al, 1965). In the FMS strategy, the rate of
stimulation also increases with the incoming signal
intensity, with the loudest sounds evoking the fastest
pulse rates. Because in a healthy ear the dependence
between firing rate and sound intensity is thought to
be mediated by the inner hair cells, we refer to this part
of the strategy as the inner hair cell model (IHM).
—continued on next page—
Novel Processing
131
—continued from previous page—
“Of note is the perceived improvement
in sound quality
from frequency-modulated stimulation
in the presence of background noise
and multitalker groups.”
Figure 3. Individual and mean scores for medial consonant
identification in quiet.
We have fit several cochlear implant users with the
FMS strategy. Subjects were required to participate
in three consecutive study sessions conducted at oneweek intervals. During the first session, each subject
was fit with either his or her control program (CIS
or SAS) or the FMS program in randomized order.
Threshold levels, comfort levels, and input dynamic
range were optimized for the patient, and the
programs were downloaded to an ear-level research
processor. Following the initial session, each subject
utilized the research processor and program exclusively for one week, at which time they returned to
Advanced Bionics for speech perception evaluation
and completion of a questionnaire. Each subject then
was fit with the alternate sound coding strategy for
an additional one-week period. The final session at
Advanced Bionics consisted of performance testing
and evaluation of the second program. At the end
of the third session, subjects were returned to their
original processor and sound processing strategy.
Individual and mean percent correct scores on
consonant recognition are shown in Figure 3. These
results indicate that subjects were able to identify
medial consonants with the same accuracy for the
FMS strategy as for the control strategy in a quiet
environment. However, data shown in Figure 4
reveal a slight improvement in medial consonant
identification in the presence of background noise
(+10 dB SNR) for FMS in some subjects. This
improvement is reflected primarily in a subject’s
ability to identify place of articulation in contrast to
manner or voicing. Of note is the perceived improvement in sound quality from frequency-modulated
stimulation in the presence of background noise
and multitalker groups. This finding is consistent
with a subject’s improved ability to identify medial
consonant tests in the presence of noise with FMS.
References
Chimento TC, Schreiner CE. (1991) Adaptation and recovery from adaptation in
single fiber responses of the cat auditory nerve. J Acoust Soc Am 90:263-273.
Figure 4. Individual and mean scores for medial consonant
identification in noise (+10 dB SNR).
Greenwood DD. (1990) A cochlear frequency-position function for several
species—29 years later. J Acoust Soc Am 87:2592-2605.
Javel E, Geisler CD, Ravindran A. (1978) Two-tone suppression in auditory nerve of
the cat: rate-intensity and temporal analyses. J Acoust Soc Am 63(4):1093-1104.
Kiang NYS, Watanabe T, Thomas EC, Clark LF. (1965) Discharge Patterns of Single
Fibers in the Cat’s Auditory Nerve. Cambridge, MA: The MIT Press.
Smith RL. (1977) Short-term adaptation in single auditory nerve fibers: some
poststimulatory effects. J Neurophysiol 40:1098-1112.
132
Advanced Bionics® Auditory Research Bulletin 2005
Chronic Evaluation of a Low-Power Strategy
Sheena McLaren, B.Sc.
Sonelle McDonald, B.Sc.
Terry Nunn, M.Sc.
Filiep J. Vanpoucke, Dr. Ir.*
Stefaan Peeters, Prof., Dr. Ir.
Guy’s and St. Thomas’ Hospital, London, UK
* also with Advanced Bionics Corporation, Europe
The conventional clinical stimulation mode of a
cochlear implant consists of amplitude-modulated,
monopolar, biphasic pulses that are spatially distributed along the electrode array. Charge-balancing is
essential to safe stimulation. However, considerable
freedom exists in shaping pulses—such as triphasic or
asymmetric—to achieve charge-balanced stimulation.
three subjects, no major differences in perceived
sound quality were found between the strategies.
Conversely, the other three subjects showed deteriorated speech perception scores and reported
degradations in overall sound quality as well.
This study evaluated speech perception and sound
quality for a strategy designed to reduce power
consumption. With the flexibility of the CII
electronics and the Bionic Ear Data Collection
System (BEDCS) research platform, a stimulation strategy was designed with pulses that feature
a +/- 100 µs interphase gap, as illustrated in Figure
1. Such pulses have the potential to increase
the efficiency of the neural response—thereby
reducing power consumption and yielding longer
battery life. However, these unconventional pulse
shapes may increase channel interaction and may
compromise patients’ speech perception abilities.
In this study, a group of six subjects participated
in a one-month chronic trial with a low-power
research strategy. Prior to participating in this study,
all patients had at least three months of experience
with a fast-rate HiResolution program. For each
subject, power consumption with the two strategies
was determined by comparing fitting levels. Speech
understanding was evaluated before switchover
and after one month of use with the research
strategy. Speech perception was measured objectively with the BKB sentences presented at 70 dBA
in quiet and in noise. Perceived sound quality was
assessed subjectively with a patient questionnaire.
Results of the study indicated that the lowpower strategy reduced power consumption of the
implant stimulator by a factor of +/- 10 (Figure
2). Speech perception scores and sound quality
ratings were more variable. Three subjects showed a
significant increase in speech understanding, especially in noisy conditions. Interestingly, for these
University of Antwerp, Antwerp, Belgium
These preliminary data show that pulse shaping
can reduce power reduction, indicative of increased
neural efficiency. However, the clinical results
achieved are highly variable. The factors underlying variability in speech perception and sound
quality are not well understood. One observation
is that the latter group had worn the HiResolution strategy for a longer time. Further analysis is
needed to understand how fitting methodology,
stimulation rate, channel ordering, and interphase
gap duration factor into patient performance.
Figure 1. Pulse shape used to increase neural efficiency.
Figure 2. Charge (nC) across the electrode array for HiRes and
the experimental(BifGap3) strategies.
Novel Processing
133
SPAIDE: A Real-Time Research Platform for the Clarion
CII and HiRes 90K Cochlear Implants
Luc Van Immerseel, Dr. Ir.1,2
Stefaan Peeters, Prof., Dr. Ir.1,2
Philippe Dykmans, M.Eng.1
Filiep J. Vanpoucke, Dr. Ir.1,2
Peter Bracke, M.Eng.1
1-Advanced Bionics Corporation, Europe
2-University of Antwerp, Antwerp, Belgium
SPAIDE (Sound Processing Algorithm Integrated
Development Environment) is a platform for
advanced research of sound processing and electrical
stimulation parameters with the Clarion CII and
HiRes 90K cochlear implants. The platform hides
the complexity of implant hardware and communication protocols. Also, it supports streaming off-line
processed data as well as real-time processing of live
input on a PC (file, sound card) while streaming
the processing results to the cochlear implant. The
hardware setup consists of (1) a PC, (2) a programming interface (PI) with USB support (currently
a SBC67 DSP board from Innovative Integration), (3) a portable speech processor (PSP), and
(4) a CII or HiRes 90K cochlear implant device.
SPAIDE implements both sound processing as well
as stimulation strategies. A sound processing strategy
is defined by the number of audio channels and the
different processing steps in each of these channels
(including preemphasis filtering, band-pass filtering,
and so forth). A stimulation strategy specifies the
number of stimulation channels, their stimulation
sequence, and their temporal and spatial definitions.
A stimulation channel is defined as a set of electrode
contacts that simultaneously carry the same electrical
stimulus waveform—though not necessarily with the
same amplitude and sign. The temporal definition
of a channel describes the electrical stimulus waveform. The spatial definition of a channel specifies
the weights with which the waveform is multiplied
for the different electrode contacts in the channel.
Processing functions are implemented in so called
“feature blocks”—each of which implements a specific
processing step (e.g., a filter bank). The sequence of
the different feature blocks and their parameters are
configured in a topology file. Fitting specifies the
patient-dependent parameters—such as connection between audio and stimulation channels and
current mapping in each of the stimulation channels.
SPAIDE includes a simulation mode that allows
verification of the whole configuration without the
need for hardware or data streaming to the implant.
Figure 1 shows four possible configurations (a-d) of
SPAIDE:
a.Typical for research applications where real-time
processing is feasible.
b. Input consists of audio-channel data that
was generated and saved to file earlier (e.g.,
by SPAIDE or by another application like
MATLAB); useful when the audio-processing
complexity is too large for implementation in a
real-time processing system on a PC.
c. Streaming application where the preprocessed
data contains the amplitude values that must
be transmitted to the implant; the preferred
configuration when the experimenter wants
maximum control over the stimulated
currents as is often the case in psychophysics
experiments.
d. Pure sound processing application where no
data is streamed to the implant.
134
Advanced Bionics® Auditory Research Bulletin 2005
“SPAIDE implements
both sound processing as well as
stimulation strategies.”
SPAIDE uses a frame-based paradigm to process its
input data. The input is first chopped in frames of
typically 50-100 ms, which are processed as specified in the topology and fitting. The results are transmitted over the USB link to the PI that handles the
data stream to the PSP and the implant in a timely
manner. It implements a buffering and synchronization mechanism that manages jitter in the processing
delay on a PC running Windows–at an overall latency
of 300-400 ms. Stimulation timing is controlled by
the PSP with microsecond accuracy, and the PSP
manages the forward telemetry to the implant.
The platform can be used in a variety of research
investigations into sound processing, complex
(simultaneous) stimulation strategies, and psychophysics. The platform does not offer a ready-made
solution for all possible research demands, but the
software is both expandable and, to a great extent,
reusable in other applications. Extending the possibilities of SPAIDE with new processing functions
(feature blocks) is one way to adapt the platform,
but necessitates C/C++ programming knowledge.
Another way to expand SPAIDE functionality is
to use it with an existing or new Windows application that can handle Win32 dynamic link library
(DLL). This latter use of SPAIDE, of course, also
demands programming skills, but such applications
can be written in the preferred language (such as
C/C++, C#, Visual Basic, Delphi, MATLAB, etc.)
Improvements to both the hardware and software
that are under development include the following:
• New CPI that supports USB such that no
special DSP board is needed.
• USB uplink will be implemented to support
microphone input from the sound processor.
• Reduced latencies from input to output for
better synchronization between visual (lipreading) cues and auditory perception with
microphone inputs.
• Support for binaural experiments.
Figure 1. Typical configurations of SPAIDE (FB = feature block).
Novel Processing
135
Speech Recognition with a Cochlear Implant
Using Triphasic Charge-Balanced Pulses
Raymond M. Bonnet, M.D.
Johan H.M. Frijns, Prof., M.D., Ph.D.
Leiden University Medical Center, Leiden,The Netherlands
Stefaan Peeters, Prof., Dr. Ir.
University of Antwerp, Antwerp, Belgium
Jeroen J. Briaire, M.Sc.
Leiden University Medical Center, Leiden, The Netherlands
“In comparing performance
across strategies,
statistically significant improvements
were found for the triphasic strategies.
The most common way of delivering stimulation
current in a cochlear implant system is through
the use of charge-balanced biphasic pulses—
typically using monopolar electrode coupling. To
reduce channel interaction in the popular Continuous Interleaved Sampler (CIS) strategy, current is
never delivered to two electrodes at the same time.
Cathodic monopolar stimulation provides a highly
efficient means of delivering electrical stimulation—
but with a risk that the stimulus may destroy cochlear
tissue as a result of unrecovered electrical charge. The
phase reversal of a biphasic pulse can prevent the
generation and propagation of an action potential
(AP) whereas a monophasic pulse of the same current
strength and phase duration would elicit a response.
Another advantage of monophasic pulses is the
ability to double the stimulation rate for a given phase
duration because the second phase may be left out.
In this study, triphasic pulses were used—each pulse
having a 33 ms active cathodic phase sandwiched
between two passive anodic phases. Each anodic
136
phase was twice the duration and, hence, one-quarter
of the amplitude of the active phase. Overlapping
the passive phases (but not active phases) allowed
charge to be recovered by as many as four channels
simultaneously. To reduce channel interaction, the
stimulation order was staggered: electrodes 1, 8, 3,
10, 5, 12, 7 and so forth. This stimulation paradigm
resulted in a pulse rate of 1,900 pulses per second
per channel (pps/ch) for a 16-channel program.
The standard clinical strategy was the HiRes™
strategy—a fast CIS-like strategy having either 12
or 16 channels active, depending on the subject. The
standard clinical strategy (labeled HR-FF) used a
biphasic stimulation pulse at 21 ms/phase, resulting
in a 1,400 pps/ch stimulation rate. When implementing strategies, the same electrodes were activated in the experimental programs as were included
in each subject’s clinical program. Two experimental
triphasic strategies were implemented: (1) TPHWR with half wave rectification and (2) TP-NoR
with polarity reversals at the filter outputs, thereby
allowing polarity reversals of the triphasic waveforms.
The effectiveness of charge balancing in these strategies was studied through a 24-hour laboratory soak
test in artificial perilymph. Residual voltages from
the triphasic strategies were found to be lower than
those for the clinical HiRes strategy. The subjects’
clinical programs as well as their triphasic experimental programs were implemented with a laboratory-based system—the Clarion Research Interface
(CRI-2) and SPAIDE research software (Advanced
Bionics NV, Antwerp). (See also Van Immerseel et
al at page 134 of this bulletin.) Emulation of each
subject’s clinical strategy with the CRI-2/SPAIDE
system was referred to as the HR-CRI strategy.
Patient performance was evaluated with CVC words,
scored by phonemes correct, for speech presented in
quiet and noise (+5 and 0 dB signal-to-noise ratios).
The noise was steady-state with a speech-shaped
spectrum. All testing was acute due to the non-porta-
Advanced Bionics® Auditory Research Bulletin 2005
bility of the research system hardware. No familiarization period was allowed with the new strategies.
All CD-based test materials were presented via a
direct electrical connection rather than free field—
as a function of the research system configuration.
Seven HiRes users with between 12 and 22
months of implant experience were included in
this study. The results of testing are summarized in
Figure 1. While a small reduction in performance
was found for the HR-CRI strategy compared
to the free-field tested clinical processor strategy
(HR-FF), the differences (for 0 and +5 dB SNR
test conditions) were not significant. Subjectively,
the participants reported that the HR-CRI strategy
sounded almost identical to the HR-FF clinical
strategy—the strategy with which they were familiar.
In comparing performance across strategies, statistically significant improvements were found for
the triphasic strategies. In the +5 SNR condition, the mean TP-HWR phoneme score was
74% compared to 67% (p<0.05) with HiResCRI. In the more difficult noise condition (0 dB
SNR), both the TP-NoR and TP-HWR strategies showed statistically significant improvements
over HR-CRI. An SRT evaluation could not be
completed since several subjects scored well above
50% correct, even in the 0 dB SNR test condition.
In comparing power requirements across strategies, average threshold (T) levels for TP-NoR
and TP-HWR were 85% and 90% of the HiRes
clinical strategy, although there was considerable
intersubject variability. The most comfortable (M)
levels were more consistently reduced to 41% and
50% of the HiRes levels for TP-NoR and TPHWR strategies, respectively. This amounted to
average power levels at the electrode of only 36%
and 30% of those required for HiRes for the TPNoR and TP-HWR, respectively. Since the power
Figure 1. Mean scores for phoneme recognition in quiet and
in noise (0 and +5 dB signal-to-noise ratios) with four sound
processing strategies: HiRes free field (FF), HiRes implemented
with the Clinical Research Interface (CRI), Triphasic with
reverse polarity (TP-NoR), and Triphasic with half-wave rectification (TP-HWR).
to the electrodes amounts to only a small proportion of the whole system power, this advantage does
not hold a considerable impact at the processor
battery—at least for present-day system designs. In
the future, however, such power reductions may be
critical to the viability of fully implantable systems.
In this study, the improved patient performance
in noise with the triphasic strategies may be
attributed, at least in part, to the higher stimulation rates offered by these strategies. However, the
TP-HWR strategy, which was preferred overall,
halved the effective stimulation rate to only 950
pps/channel for a 16-channel program—making
it the most power efficient strategy, requiring only
around one-third of the HiRes strategy’s power.
Finally, the rectification involved in the TP-HWR
may also have produced a more controlled stimulation pattern, avoiding phase reversals in the
triphasic pulses and unpredictable phase relations
in sequential stimulation periods. Future chronic
studies should incorporate an active (rather than
passive) charge-balancing mechanism to further
control this factor—even for the TP-HWR strategy.
Novel Processing
137
Effects of Cochlear Implantation on Auditory Nerve
Synapses in Congenitally Deaf White Cats
Erika A. Kretzmer, B.S.
Karen L. Montey, B.S.
Tan Pongstaporn, B.S.
David K. Ryugo, Ph.D.
Johns Hopkins University, Baltimore, MD, USA
Our long-term goal is to understand how hearing
and deafness affect the development of structure and function in the central auditory system.
Congenital deafness has been shown to cause
neuronal death, induce abnormal circuits, and create
synaptic abnormalities. Such pathologic changes
may underlie the failure of some cochlear implant
recipients to benefit from their devices. Also, a body
of evidence exists revealing the importance of age
in predicting the benefits of cochlear implantation—that is, younger recipients benefit more than
older recipients. Current experiments are aimed
at determining which, if any, deafness-induced
changes can be prevented or reversed by cochlear
implantation. Our studies focus on the brain locus
where auditory nerve fibers make functional contact
with brain cells. This region is important because it
forms the gateway to the central auditory pathways.
Any disruption in signaling at this site will influence processing throughout the auditory system.
To date we have implanted each of five congenitally deaf white kittens (aged 3 or 6 months) with
a unilateral, six-channel cochlear implant. These
implants have shorter and thinner electrode array
systems than those designed for human cochleas.
Each kitten has received electrical stimulation for
35 hours per week over a period ranging from 9 to
24 weeks with Advanced Bionics CII sound processors delivering HiRes stimulation rates (38615155 pps per channel at 21.6 µs pulse widths).
Electrode selections and M-level settings were
based upon behavioral cues, ECAPs, and EABRs.
After several weeks of experience, kittens could be
called to eat in response to a learned sound association, indicating successful implant function.
At the end of the experiment, we examined a
prominent terminal of the auditory nerve—the
endbulb of Held and its synapse—using light and
electron microscopy. The endbulb is a large axosomatic synaptic ending that is formed by auditory
nerve fibers in vertebrate animals studied (Figure 1).
Its size and widespread phylogenetic distribution
suggest a fundamentally important role for sound
processing. Specifically, the endbulb has been implicated in maintaining the precise timing between
neural activity and acoustic events—a function that
mediates sound localization and the processing of
speech cues such as prosody, stop gaps, and glides.
Figure 1. Photomicrograph of stained endbulb of Held. These
large endings embrace the cell body of a single neuron (stained
purple) in the cochlear nucleus. Their large size and many
synaptic contacts ensure that there is no loss of signal fidelity
across the synapse.
138
The endbulb has been described in deaf as well
as hearing cats. It is highly branched in normal
hearing animals, but loses many of its branches as
a consequence of congenital deafness. In the elec-
Advanced Bionics® Auditory Research Bulletin 2005
“We have observed that
cochlear implant stimulation
to the auditory system
largely prevents or reverses
the synaptic abnormalities
induced by congenital deafness.”
tron microscope, only part of the structure can be
observed at any one time because of high magnification levels and very thin tissue sections. The endbulb
appears as a large profile adjacent to the cell body
of a spherical, bushy cell (Figure 2). In hearing
cats, the profile is filled with large, round synaptic
vesicles. The apposed membrane is often interrupted
by intercellular cisternae. The membrane apposition can be marked by a thickening on the bushy
cell side and a concavity facing the endbulb. This
thickening houses the transmitter receptors and
is called the postsynaptic density (PSD). Typically,
the endbulbs in deaf animals exhibit fewer synaptic
vesicles, an absence of intercellular cisternae, and
PSDs that are larger and less curved in appearance.
Using electron microscopy, we have examined
endbulbs in three implanted deaf cats. On the
side ipsilateral to the implant, we have observed
(1) a return of the punctate, concave PSD, (2)
the presence of intercellular cisternae, and (3) a
normal density of synaptic vesicles. Thus, HiRes
stimulation “rescued” many (but not all) of the
synapses of auditory nerve fibers. Statistical
analysis of synapses confirmed no significant
differences between hearing and implanted cats.
We have observed that cochlear implant stimulation to the auditory system largely prevents or
reverses the synaptic abnormalities induced by
congenital deafness. The mechanism by which
electrical stimulation serves to preserve the central
auditory system remains to be determined, but
synaptic restoration at this juncture must have a
positive effect on sound processing through the
implant. We have proposed further studies to
determine the potential role that “age at implantation” plays in habilitation of deaf individuals.
Figure 2. Electron micrographs of auditory nerve terminals
(endbulbs) in the cochlear nucleus. Note that terminals from
an implanted cat appear identical to those from a normal-hearing cat—in stark contrast to those of an untreated deaf cat with
abnormal synapses. (EB = endbulb; SBC = spherical
bushy cell.)
Acknowledgements
Supported by NIH grants RO1 DC00232 and F31 DC005864, the Emma Leipmann
Endowment Fund, and Advanced Bionics Corporation.
Novel Processing
139
Bilateral
Cochlear Implants
The degree to which cochlear implant patients can benefit from
binaural implantation is related to sound processing dynamics.
The HiResolution Bionic Ear and HiResolution (HiRes) Sound
offer particular advantages to patients with two devices because
the system is designed to preserve the intensity, spectral, and
timing cues important for binaural hearing. IntelliLink™ locks
the customized HiRes programs to the internal device so that an
implant cannot be activated with the incorrect sound processor.
Preliminary results show that users of two Bionic Ears demonstrate
unprecedented benefit when using HiRes compared to conventional
sound processing.
Notably, binaural implantation has necessitated the development
of new tools for the assessment of binaural benefit, such as hearing
in noise and sound localization. These tools can be implemented
through direct input to the Auria sound processor, thereby
precluding the need for a sound room and a multiple loudspeaker
array in the test set-up.
Bilateral Benefit in Adult Users of the HiRes 90K
Bionic Ear System
Investigators at the University of Iowa have reported
that bilaterally implanted adults demonstrated
significant improvement in speech-in-noise scores
with HiRes sound processing after only one month
of use compared to their scores after long-term (1824 months) experience with CIS (Dunn et al, 2005).
Study Sites in North America
Carle Clinic Association, Urbana, IL
Columbia University Medical Center, New York, NY
Dallas Otolaryngology, Dallas, TX
House Ear Institute/House Ear Clinic, Los Angeles, CA
Let Them Hear Foundation, Palo Alto, CA
Massachusetts Eye & Ear Infirmary, Boston, MA
Mayo Clinic, Scottsdale, AZ
Medical College of Wisconsin, Milwaukee, WI
Midwest Ear Institute, Kansas City, MO
New York University, New York, NY
Spectrum Health, Grand Rapids, MI
Tampa Bay Hearing & Balance Center, Tampa, FL
University Hospitals of Cleveland, Cleveland, OH
University of Iowa, Iowa City, IA
University of Massachusetts, Amherst, MA
University of Miami, Miami, FL
University of Minnesota, Minneapolis, MN
University of Pennsylvania, Philadelphia, PA
University of Washington, Seattle, WA
Vanderbilt University, Nashville, TN
The purpose of this multicenter study is to replicate
these findings in a larger group of newly implanted,
postlinguistically deafened adults who receive two
HiRes 90K implants during the same surgery. A
prospective counterbalanced between- and withinsubjects design is used to evaluate bilateral listening
benefits and to compare the benefits of sound
processing mode (CIS vs. HiRes). The study is a
six-month crossover design (three months with
each processing mode) with an additional onemonth period in which subjects reevaluate the two
processing modes (two weeks with each mode) and
indicate a preference. Subjects then are evaluated at
one and four months after using their preferred mode.
142
A unique aspect of the study is use of a direct-connect
system for postimplant testing designed by Soli and
colleagues (Soli et al, 2005) to eliminate the need for
a sound booth or a speaker array, thereby allowing
speech recognition and localization tests to be
administered quickly and easily. The direct-connect
system is based upon a family of head-related transfer
functions (HRTFs) measured with KEMAR at
source locations corresponding to loudspeaker positions appropriate for unilateral or bilateral testing.
Left-ear and right-ear HRTFs appropriate to the
selected source location are applied to the selected
signal and presented via direct connection to the
auxiliary input of the Auria sound processor at a
specified level. Other unique aspects of the study
are the use of adaptive procedures to assess speech
perception in noise and localization, and testing
sentence-in-noise performance at lower presentation levels. These procedures will allow fine differentiation between bilateral and unilateral performance,
and between HiRes and CIS sound processing.
Preliminary study results indicate that bilateral
implantation is advantageous and that sound
processing mode may have a significant effect on
bilateral benefit, consistent with the Iowa results.
(See also Peters and Lake at page 150 of this
bulletin.) These early conclusions await verification from a larger group of study participants.
References
Dunn C, Tyler R, Witt S, Gantz B. (2005) The effect of increasing the number of
channels and rate on bilateral cochlear implant performance. Paper presented at
the American Otological Society, Boca Raton, FL, 12–14 May, 2005.
Soli SD, Chan JCY, Vermiglio AJ, Freed DJ, Kessler DK. (2005) Assessment of spatial
unmasking and sound localization in bilateral implants using direct input. Poster
presentation at the Conference on Implantable Auditory Prostheses, Pacific Grove,
CA, 30 July–4 August, 2005.
Advanced Bionics® Auditory Research Bulletin 2005
Effect of Sound Processing on Bilateral Performance with
the CII Cochlear Implant
Camille C. Dunn, Ph.D.
Richard S. Tyler, Ph.D.
Shelley A. Witt, M.A.
Bruce J. Gantz, M.D.
University of Iowa, Iowa City, IA, USA
Results obtained from unilateral CII/HiRes 90K
implant recipients who were converted from
conventional strategies to HiResolution (HiRes)
sound processing revealed that most subjects
performed better when using the greater number of
channels and faster stimulation rate implemented
in HiRes (Koch et al, 2004). Based upon these
results, the University of Iowa recently conducted
a study to determine if increased stimulation rate
and a greater number of channels would provide
enhanced benefit to patients who had been simultaneously implanted with bilateral CII implants.
This report summarizes the HiRes results in adult
bilateral patients (Dunn et al, 2005). Subjects were
seven individuals (two men and five women) who
received bilateral CII devices during a single operation. Subjects ranged in age from 28 to 69 years
and all wore the CII BTE sound processor. All
subjects had worn the conventional CIS strategy
for at least 18 months. Before conversion to HiRes,
speech perception was measured in multi-talker
speech babble noise using recorded City University
of New York (CUNY) sentences. The sentences
were presented in the most difficult listening situation with the noise from the front (0° azimuth).
Signal-to-noise ratios (S/N) were set for each individual to avoid ceiling and floor effects and were
held constant for each subject thereafter throughout
the study. In addition, localization ability was
evaluated using everyday sounds presented from
one of eight loudspeakers that formed a 108° arc
centered in front of the listener. The subject was
told to identify the speaker from which the sound
originated, but not to identify the sound itself.
“...listeners were able to tolerate
a more difficult signal-to-noise ratio
with HiRes
than with conventional CIS.”
sion to HiRes. Then they were directed to alternate daily between the HiRes Paired and HiRes
Sequential strategies for one month. After one
month, speech perception and localization were
assessed again using the same methods. Subjects
then chose to continue to wear both HiRes strategies, or to wear only the preferred HiRes strategy.
Subjects were reevaluated 3-6 months later.
Both the HiRes Sequential and Paired strategies
resulted in dramatic improvements in speech perception in noise after one month of use. These results
indicate that listeners were able to tolerate a more
difficult signal-to-noise ratio with HiRes than with
conventional CIS. Moreover, the increases in scores
were much greater for these bilateral subjects than
for unilateral implant users who were converted from
conventional sound processing to HiRes. In contrast,
localization abilities did not improve significantly.
After three months of HiRes use, speech perception improvements stayed consistent. Further work
is needed to determine the independent beneficial
effects of stimulation rate and number of channels.
References
Dunn CC, Tyler RS, Witt SA, Gantz BJ. (2005). Effect of increasing the number of
channels and rate on bilateral cochlear implant performance. Paper presented at
the American Otological Society, Boca Raton, FL, 12-14 May, 2005.
Koch DB, Osberger MJ, Segel P, Kessler DK. (2004) HiResolution and conventional
sound processing in the HiResolution bionic ear: using appropriate outcome
measures to assess speech recognition ability. Audiol Neurotol 9:214-223.
Subjects were reevaluated with the same speech
and localization tests immediately after conver-
Bilateral Cochlear Implants
143
Development of a Direct-Input System to Evaluate
Spatial Unmasking and Sound Localization
in Bilateral Implant Users
Sigfrid D. Soli, Ph.D.
Jenny C.Y. Chan, M.A.
Andrew J. Vermiglio, M.A.
Daniel J. Freed, M.S.
House Ear Institute, Los Angeles, CA, USA
Dorcas Kessler, M.A.
Advanced Bionics Corporation, Valencia, CA, USA
This project developed and evaluated a prototype
instrument to assess the binaural abilities of bilateral
cochlear implant (CI) wearers using signals delivered
directly to the auxiliary inputs of the sound processors. Sets of calibrated head-related transfer functions (HRTFs) from sources at various azimuths
in the horizontal plane were measured through
the T-Mic®, the built-in BTE microphone, and
the PSP headpiece microphone for the Bionic Ear
system. The Hearing in Noise Test (HINT) speech
and noise materials were processed with HRTFs for
0º, 90º, and 270º azimuths in assessments of spatial
unmasking (Figure 1). An impulse noise from
the Source Azimuth Identification in Noise Test
(SAINT) was processed with HRTFs in the horizontal plane for 12 azimuths from 90º to 270º in 15º
increments to assess sound localization (Figure 2).
Figure 1. HRTFs describe the direction-dependent acoustic
filtering a free field sound undergoes as a result of the head,
torso, and pinna. Each HRTF represents the difference between
the sound at a source originating from a particular free field
location and the sound that arrives at the ear canal, T-Mic,
BTE microphone, or PSP microphone located on a KEMAR
mannequin.
The adaptive rule for measuring reception thresholds
for sentences (RTSs) with the HINT was modified
for use with CI subjects who do not achieve 100%
intelligibility in quiet. The scoring rules for SAINT
also were modified for use with CI subjects who
are unable to identify all of the source azimuths.
Both the HRTF simulations and the modified testing procedures were validated first in 17
acoustic hearing (AH) subjects and subsequently
with five bilaterally implanted CI subjects. Signals
were delivered via sound field speakers (SF) and
144
Advanced Bionics® Auditory Research Bulletin 2005
via direct connect (DC) input. For AH subjects,
headphones were used to deliver DC signals. For
CI subjects, DC signals were delivered to the auxiliary inputs of their processors. In sum, the HRTF
simulations for HINT and SAINT produced
nearly identical SF and DC results for both the AH
and the CI subjects. These results were obtained
using the modified testing procedures for both
HINT and SAINT. DC test scores were often
slightly better than SF scores for the CI subjects.
The bilateral CI subjects differed widely with respect
to their spatial unmasking and sound localization
abilities and, in most cases, their abilities did not
fall within the range of performance for subjects
with acoustic hearing. However, in all cases, it was
possible to make accurate direct connect (DC)
assessments of these abilities for comparison
with those of normal-hearing individuals. Results
suggest that well controlled measures of CI binaural
hearing abilities may be readily obtained in clinical settings using direct connect inputs. The DC
system currently is being used in a multicenter
study of bilateral benefit in adult users of the HiRes
90K system. (See also page 142 of this bulletin.)
“Results suggest that well controlled
measures of CI binaural hearing abilities
may be readily obtained
in clinical settings
using direct connect tests.”
Figure 2. Sound sources are simulated in the horizontal
plane using HRTFs measured for 12 sources separated by
15° increments.
Acknowledgement
This project was supported by a Small Business Innovation Research grant to House
Ear Institute and Advanced Bionics Corporation from NIH-NIDCD.
Bilateral Cochlear Implants
145
Changes Over Time in the Benefit of Contralateral
Amplification in Unilateral Cochlear Implant Users
Michal Luntz, M.D.
Talma Shpak, M.A.
Hadas Weiss, M.Sc.
Bnai Zion Medical Center
Israel Institute of Technology, Haifa, Israel
The objective of this study was to evaluate changes
in the binaural-bimodal auditory abilities over time
in unilaterally implanted patients using a hearing
aid (HA) in the contralateral ear. Sentence recognition in background noise was tested in 12 patients
(described in Table 1) under three listening conditions: (1) multichannel cochlear implant (CI) alone,
(2) HA alone, and (3) CI + HA. The presentation
level was 55 dB HL at a signal-to-noise ratio of
+10 dB. Subjects were tested in two sessions—the
first session after 1-6 months of concomitant use
of both devices and the second session 6-8 months
later. (At the time of testing, all subjects had
1-11 months of use with the cochlear implant alone.)
Table 1. Patient Demographics
Preoperative PTA
CI Ear
Age
S Onset (Yrs)
no
Months of Use
HA Ear
Test 1
Test 2
no
CI
CI
HA
HA
CI
HA
CI
HA
1
Post
60
103
52
103
52
6
2
12
8
2
Post
48
107
63
93
43
1
1
9
9
3
Post
50
>120
58
83
35
6
6
12
12
4
Pre
16
120
48
105
35
11
2
17
8
5
Pre
13
108
72
98
55
6
1
12
7
6
Pre
15
112
48
112
48
2
1
8
7
7
Pre
12
117
60
113
55
3
3
12
12
8
Pre
7
100
53
95
41
6
6
12
12
9
Pre
12
123
60
115
47
6
6
12
12
10
Pre
32
>120
118
>120
100
6
6
12
12
11
Pre
26
108
30
112
48
6
6
12
12
12
Pre
11
118
57
90
38
6
6
12
12
Individual characteristics of subjects including onset of deafness (pre- or postlinguistic), age at implantation, and duration
of use with implant alone and with contralateral hearing aid
prior to each test session.
146
Figure 1 shows the individual scores for the 12
subjects in the first test session. Performance with CI
+ HA did not differ significantly from the results with
CI alone (Wilcoxon signed-rank test). However, in
the second test session (Figure 2), significant differences were found between the scores with CI + HA
and the scores with CI alone (p < 0.05, Wilcoxon
signed-rank test with Bonferroni correction).
Comparison between the group results for the first
and second test sessions showed significant improvement in both the CI-alone and CI + HA conditions
(p < 0.05, Wilcoxon signed-rank test). Relative to the
first phase of the study, the performance improvements in the second phase with CI + HA were
significantly greater than the improvements with
CI alone (p < 0.075 with Bonferroni correction).
Considerable variation in the individual data is
observed between test sessions. In the first session,
the scores in the CI-alone condition ranged from 0
to 90% (mean = 34.9%) and in the CI + HA-condition from 0 to 100% (mean = 41.1%). In the first
session, five patients could not recognize sentences
in noise in CI-alone condition, and only four
patients showed some measurable improvement
with contralateral amplification added. In the second
session, when all subjects could recognize sentences
in noise with the CI alone, seven patients showed
further improvement when contralateral amplification was added. The CI-alone scores ranged from
10% to 99% (mean = 60.6%) and the CI + HA
scores ranged from 52% to 100% (mean = 75.5%).
From these results we conclude that, for unilateral
cochlear implant recipients, the benefit of hearing
aid use in the nonimplanted ear with residual
hearing improves over time—at least during the
first year after implantation. The results of concomi-
Advanced Bionics® Auditory Research Bulletin 2005
“...the benefit of hearing aid use
in the nonimplanted ear...
improves over time—
at least during the
first year after implantation.
tant cochlear implant and hearing aid in the “ideal”
candidates selected for the present study are comparable with those of similar studies of unilaterally
implanted patients (Ching et al, 2004; Ching et al,
2001; Tyler, 2002) and also with previous studies
of bilaterally implanted patients (Gantz et al, 2002;
Müller et al, 2002). The latter, of course, represent a
different group of candidates, namely those without
effective residual hearing who are able to receive
binaural benefits only via bilateral cochlear implants.
Unilaterally implanted patients, who initially show
significant enhancements of speech perception with a
contralateral hearing aid, might later experience deterioration of residual hearing on the nonimplanted side,
with consequent deterioration of speech perception
scores in the CI + HA condition. For such patients,
contralateral implantation should be considered.
Figure 1. First test session: sentence recognition in noise after
1 to 6 months of experience with wearing both devices (preceded by 1-11 months with implant alone). No significant
difference for CI-alone (mean = 34.9%) compared to CI + HA
(mean = 41.1%).
References
Ching TYC, Incerti P, Hill M. (2004) Binaural benefits for adults who use hearing
aids and cochlear implants in opposite ears. Ear Hear 25:9-21.
Ching TYC, Psarros C, Hill M, Dillon H, Incerti P. (2001) Should children who use
cochlear implants wear hearing aids in the opposite ear? Ear Hear 22:365-380.
Gantz BJ, Tyler RS, Rubinstein JT, Wolaver A, Lowder M, Abbas P, Brown C, Hughes
M, Preece JP. (2002) Binaural cochlear implants placed during the same operation.
Otol Neurotol 23:169-180.
Müller J, Schön F, Helms J. (2002) Speech understanding in quiet and in noise in
bilateral users of the MED-EL COMBI 40/40+ cochlear implant system. Ear Hear
23:198-206.
Tyler RS, Parkinson AJ, Wilson BS, Witt S, Preece JP, Noble W. (2002) Patients
utilizing a hearing aid and a cochlear implant: speech perception and localization.
Ear Hear 23:98-105.
Figure 2. Second test session: sentence recognition in noise 6-8
months following the first test session (i.e., 7 to 12 months of
experience with wearing both devices). Significant difference
(p < 0.05, Wilcoxon signed-rank test) for CI-alone (mean =
60.6%) compared to CI + HA (mean = 75.5%).
Bilateral Cochlear Implants
147
Changes in Fusion and Localization Performance When
Transitioning from Monolateral to Bilateral Listening
Donald K. Eddington, Ph.D.
Massachusetts Institute of Technology, Cambridge, MA
Massachusetts Eye and Ear Infirmary and
Harvard Medical School, Boston, MA, USA
Becky Poon, BS
Massachusetts Institute of Technology, Cambridge, MA, USA
Victor Noel, BS
Massachusetts Eye and Ear Infirmary, Boston, MA, USA
Because normal hearing is binaural, not surprisingly the current trend is toward bilateral cochlear
implantation. The potential benefits patients might
derive from implantation of the second ear include
better sound-source localization and improved
speech reception in adverse listening conditions.
Our research is aimed at understanding the degree
to which a subject’s capability to integrate bilateral
stimuli is influenced by their listening experience.
We hypothesize that the listening strategy used by
an individual who has been using monolateral stimulation for many months will be different than that
of a subject with several months of bilateral experience. This difference is important for two reasons.
First, if experience plays a substantial role in functional ability, the current practice of comparing the
performance of monolateral and bilateral listening
conditions in long-term users of bilateral stimulation probably puts the monolateral listening condition at a disadvantage. Second, if fundamental
characteristics of sensations elicited by bilateral
stimulation depend on bilateral listening experience, it is possible that monolateral implantation
of a very young child constrains his or her brain’s
ability to develop the machinery to take advantage
of bilateral stimulation when introduced later in life.
Five adults, implanted with CII or 90K implants, are
participating in our research studies. All subjects had
near normal hearing at least through age 16. Each
subject received the first device and used it for at least
six months before the second ear was implanted. A
battery of psychophysical, localization, and speech
reception tests is administered to each subject before
they begin wearing the second sound processor. This
protocol makes it possible to test monolateral and
bilateral performance while subjects are still using
a monolaterally-developed listening strategy. Once
subjects begin using two implants, their performance
is tracked as they develop a bilateral-listening strategy.
Measures of fusion (when stimulating single interaural electrode pairs) and of localization (using
asynchronous sound processors) are obtained.
The results show changes with bilateral experience that suggest: (1) basic changes in the brain
leading to fundamental changes in the perception
of bilateral stimuli, and (2) monolateral localization performance measured in subjects using a
listening strategy developed during monolateral
listening can be substantially better than monolateral performance measured in the same subject
using a bilaterally-developed listening strategy.
Acknowledgement
Supported by the NIH-NIDCD, the Keck Foundation, and Advanced Bionics
Corporation.
148
Advanced Bionics® Auditory Research Bulletin 2005
Bilateral Benefit in Twins Implanted with
HiRes 90K Devices Before One Year of Age
Michael A. Novak, M.D. (Project Director)
Carle Clinic Association, Urbana, IL, USA
“...The four-month IT-MAIS scores have
demonstrated significant gains
in auditory skill development
for both boys.”
Adults and children who have received cochlear
implants worldwide have shown substantial benefit
from implantation in only one ear. However, the
recent trend toward bilateral cochlear implantation has demonstrated that bilateral implants have
the potential to provide binaural benefits similar
to those experienced by normal hearing listeners.
Study results in adults have indicated better sound
localization, improved speech understanding in
noise, and enhanced quality of hearing with two
implants versus one implant (e.g., Gantz et al,
2002; Litovsky et al, 2004; Schleich et al, 2004).
In young children, the plasticity of the auditory system provides an even more compelling
reason for consideration of bilateral implantation.
The aim of this study is to assess the benefit of
two implants on the development of communication skills in twin children implanted before one
year of age. The children will be followed for a
minimum of three years and assessed on a battery
of age-appropriate auditory, speech, and language
measures. Their performance will be compared to
a cohort of infants implanted unilaterally before
12 months of age (n = 7) and a group implanted
unilaterally between 12 and 18 months of age
(n = 11) at Carle Foundation Hospital. Current
data on these children support early implantation,
demonstrating that implantation by 18 months of
age can result in normal, age-appropriate speech
and spoken language skills (Hammes et al, 2002).
The identical twin boys in this study were identified at Carle Hospital with severe-to-profound
hearing loss and were fit with hearing aids at two
months of age. They were enrolled in an intensive
auditory-oral parent-infant program. Both boys
demonstrated negligible benefit from hearing
aids. Consequently, bilateral cochlear implants
were recommended. The boys received their bilateral HiRes 90K devices at 10 months of age.
In the future, postoperative evaluations will occur at
six-month intervals according to the typical protocol
for the Expanding Children’s Hearing Opportunities
(ECHO) cochlear implant program at Carle Foundation Hospital. Age-appropriate auditory, speech,
and language tests will be administered at each
assessment interval. The development of auditory,
speech, and language skills will be compared between
the twins as well as with the group of seven infants
implanted unilaterally at one year of age or younger
at Carle Hospital. As of July 2005, the boys are at the
four-month postoperative interval. The four-month
IT-MAIS scores have demonstrated significant
gains in auditory skill development for both boys.
References
Gantz BJ, Tyler RS, Rubinstein JT, Wolaver A, Lowder M, Abbas P, Brown C, Hughes
M, Preece JP. (2002) Binaural cochlear implants placed during the same operation.
Otol Neurotol 23:169-180.
Hammes D, Novak MA, Rotz LA, Willis M, Edmondson D, Thomas J. (2002) Early
identification and cochlear implantation: critical factors for spoken language
development. Ann Otol Rhinol Laryngol 111(Suppl 189):74-78.
Litovsky R, Parkinson A, Arcaroli J, et al. (2004) Bilateral cochlear implants in adults
and children. Arch Otolaryngol Head Neck Surg 130: 648-655.
Schleich P, Nopp P, D’Haese P. (2004) Head shadow, squelch, and summation
effects in bilateral users of the MED-EL Combi 40/40+ cochlear implant. Ear Hear
25:197-204.
Bilateral Cochlear Implants
149
HiRes Benefit in a Bilaterally Implanted Adult
B. Robert Peters, M.D.
Jennifer Lake, M.S.
Dallas Otolarygology, Dallas, TX, USA
Dallas Otolarygology is participating in a multicenter
study investigating the benefits of simultaneous
bilateral implantation of the HiRes 90K system.
(See also page 142 of this bulletin.). A prospective, counterbalanced between- and within-subjects
design is used to evaluate bilateral listening benefits
and to compare the benefits of sound-processing
mode (CIS vs. HiRes). The study is a six-month
crossover design (three months with each processing
mode) with an additional one-month period in
which subjects re-evaluate the two processing
modes (two weeks with each mode) and indicate a
preference. Subjects then are evaluated at one and
four months after using their preferred mode. The
test battery assesses word and sentence recognition
in quiet and in noise, as well as localization ability.
Figure 1. Bilateral localization accuracy for HiRes and CIS,
and for a group of normal-hearing listeners. The subject indicates the source location of sounds originating from headrelated-transfer-function representations of 12 speakers placed
in the horizontal plane and spanning 180 degrees from ear
to ear. The data are scored for accuracy within 2, 3, 4, 6, and
12 equal sectors of the 180 degrees. Good spatial resolution
(> 50% accuracy) was reached for 2, 3 and 4 sectors (> 45
degrees) in HiRes and for 2 and 3 sectors in CIS. The subject
scored at or below chance with one implant alone using either
HiRes or CIS.
The first subject implanted at our center has reached
the six-month test interval. This patient was fit
first with HiRes for three months, then with CIS
for three months. Figure 1 illustrates localization
ability with two implants. The localization task uses
an adaptive paradigm to avoid floor effects and to
allow quantification of localization skills over time
(Soli et al, 2005). This subject’s localization abilities with either implant alone were at chance, so he
experienced better localization with two implants
with both HiRes and CIS. Notably, his localization ability was better with HiRes than with CIS.
Figure 2 shows results for CNC words and HINT
sentences in quiet for a 60 dB SPL presentation level. Notably, when this subject was crossed
over to CIS, there was a dramatic decrease in his
performance. His CNC word scores dropped from
76% to 36%, and his HINT in noise score (60 dB
SPL, SNR = +8 dB) went from 79% to 26% immediately after the conversion. Although the CIS
scores improved after three months (six months
total implant use), they did not reach the scores
attained after the initial three months of HiRes use.
Figure 2. CNC-word and HINT in quiet scores (60 dB SPL)
after three months of HiRes use, immediately after crossing to
CIS, and after three months of CIS use. Although the CIS scores
improve with experience, they do not reach the scores attained
after the initial three months of HiRes experience.
150
Figure 3 shows the CNC word scores for each
implant alone and for both implants together.
Overall, the HiRes scores are better than the CIS
scores for all listening conditions. This patient
Advanced Bionics® Auditory Research Bulletin 2005
exhibits clear bilateral benefit with HiRes,
evidenced by higher scores when listening with
both implants compared to scores with the right or
left implant alone. Bilateral benefit is not evident
when using the CIS mode, even though he has
had longer overall use of the implants at the time
of testing (six months total). The same pattern
of results is seen for the HINT in quiet results.
Figure 4 shows the adaptive HINT in noise scores
for speech and noise coming from the front (noise
level fixed at 52 dB SPL). Performance is expressed
as the signal-to-noise ratio (SNR) that yields a
sentence recognition score of 50%. Overall, a lower
SNR is required when the patient uses HiRes than
when he uses CIS. Moreover, this subject experiences
binaural summation when using HiRes—that is, he
requires a lower SNR when using both implants
together than when using either implant alone. In
contrast, he does not experience binaural summation when using CIS. Because binaural summation—which is a central auditory phenomenon—is
experienced only when using HiRes, HiRes must
be providing essential binaural cues that are not
provided by CIS. These results, along with the data
in Figure 3, suggest that this subject should be able
to understand speech much easier in noise with
bilateral HiRes sound processing than with bilateral CIS processing in everyday listening situations.
These results demonstrate that this patient derives
the benefits of bilateral implantation and that sound
processing mode may have a significant effect on
bilateral benefit, consistent with results reported by
the University of Iowa (Dunn et al, 2005). Nonetheless, these data represent only one subject and
should be considered preliminary, pending findings from a larger group of study participants.
References
Dunn C, Tyler R, Witt S, Gantz B. (2005) The effect of increasing the number of
channels and rate on bilateral cochlear implant performance. Paper presented at
the American Otological Society, Boca Raton, Florida, 12-14 May, 2005.
Soli S, Chan J, Vermiglio A, Freed D, Kessler DK. (2005) Assessment of spatial
unmasking and sound localization in bilateral implants using direct input. Poster
presented at the Conference on Implantable Auditory Prostheses, Pacific Grove, CA,
30 July–4 August, 2005.
“...this patient derives the benefits of
bilateral implantation...
sound processing mode may have a
significant effect on bilateral benefit...”
Figure 3. CNC word scores (60 dB SPL) for each implant alone,
and for both implants used together. HiRes scores are higher
overall and the subject experiences greater speech-perception
benefit when using both implants at the same time with HiRes,
but not with CIS.
Figure 4. Adaptive HINT in noise results for each implant
alone, and for both implants used together (speech and noise
from the front). A lower SNR indicates better speech recognition in noise. The subject experiences a binaural summation
effect for HiRes but not for CIS. When using CIS, the subject
was unable to complete the task for the left implant alone
because it was too difficult.
Bilateral Cochlear Implants
151
HiRes versus Conventional Strategy Benefit
in a Bilaterally Implanted Adult
Lisa Buckler, M.A.
Kristen Dawson, M.A.
Charles Luetje, M.D.
Midwest Ear Institute, Kansas City, MO, USA
“This subject’s data demonstrate clearly the
benefits of bilateral implantation...”
Figure 1. CNC word scores for each implant alone and for both
implants used together for the first four study phases (CIS 1 =
after three months of CIS use, HiRes 1 = after three months of
HiRes use, CIS 2 = after two-week repetition phase, HiRes 2 =
after two-week repetition phase).
152
Midwest Ear Institute is participating in a multicenter study investigating the benefits of simultaneous
bilateral implantation of the HiRes 90K system.
(See also page 142 in this bulletin). A prospective,
counterbalanced between- and within-subjects
design is used to evaluate bilateral listening benefits
and to compare the benefits of sound processing
mode (CIS vs. HiRes). The study is a six-month
crossover design (three months with each processing
mode) with an additional one-month period in
which subjects re-evaluate the two processing modes
(two weeks with each mode) and indicate a preference. Subjects then are evaluated at one and four
months after using their preferred mode. The test
battery evaluates word and sentence recognition
in quiet and in noise, as well as localization ability.
We were pleased to have implanted the first patient
in this multicenter study. This 74-year-old man
has reached the seven-month test interval. He
was fit first with CIS for three months, then with
HiRes for three months, and has completed the
two-week repetition phase for both strategies.
Figures 1 and 2 summarize CNC word scores and
HINT in quiet scores (60 dB SPL) across the
four test intervals. These results indicate that the
patient shows improved performance over time
and that HiRes scores are higher than CIS for
both the primary and repetition phases of the study.
Advanced Bionics® Auditory Research Bulletin 2005
Notably, the patient experiences simple bilateral
benefit (hearing with two implants better than
either alone) for HiRes but not for CIS. After seven
total months of implant use, the patient preferred
HiRes to CIS (strength of preference = 8, where
1 = weak preference and 10 = strong preference).
“...HiRes sound processing
provides enhanced bilateral benefit
compared to CIS...”
This subject’s data demonstrate clearly the benefits
of bilateral implantation and indicate that HiRes
sound processing provides enhanced bilateral
benefit compared to CIS, consistent with the data
reported by Dunn and colleagues at the University
of Iowa (Dunn et al, 2005). These results remain to
be verified in the larger group of study participants.
References
Dunn C, Tyler R, Witt S, Gantz B. (2005) The effect of increasing the number of
channels and rate on bilateral cochlear implant performance. Paper presented at
the American Otological Society, Boca Raton, FL, 12-14 May, 2005.
Figure 2. HINT in quiet scores for each implant alone and for
both implants used together for the first four study phases (CIS 1
= after three months of CIS use, HiRes 1 = after three months of
HiRes use, CIS 2 = after two-week repetition phase, HiRes 2 =
after two-week repetition phase).
Bilateral Cochlear Implants
153
Ear-Level
System Features
The ear-level sound processors (Auria, CII BTE, and Platinum
BTE) have a variety of features designed to improve connectivity to
audio devices such as MP3 and CD players, FM systems, and other
assistive technologies that make listening easier.
The T-Mic, which places the microphone in the external ear canal,
provides direct access to cellular telephones and consumer audio
headphones without requiring a connecting cable. The T-Mic’s
location within the concha also allows listeners to take advantage
of the sound filtering effects of the pinna for better hearing in noise
and improved sound localization.
The T-Coil is a new earhook option that provides easy access to
assistive listening technology such as inductive loop systems and
hearing aid compatible telephones.
The Usefulness of a Pinna Microphone Position for
Sound Localization in Bilateral Cochlear Implant Users
Andreas Büchner, Ph.D.
Carolin Frohne-Büchner, Ph.D.*
Lutz Gärtner, M.Sc.
Anke Lesinski-Schiedat, M.D.
Rolf-Dieter Battmer, Prof., Ph.D.
Thomas Lenarz, Prof., M.D., Ph.D.
Medizinische Hochschule Hannover, Hannover, Germany
* also with Advanced Bionics Corporation, Europe
Figure 1. Auria sound processor with T-Mic earhook accessory.
Figure 2. Sound localization test setup with speaker 12 directly
in front (0° azimuth) of the subject and speaker 6 directly
behind (180° azimuth) the subject.
156
Cochlear implant systems use different microphone
positions and characteristics, none of which provide
sound detection as normal hearing listeners experience it. In the Auria T-Mic design, the microphone
is located in the pinna, close to the outer ear canal
(Figure 1). This design offers the advantages of
some favourable outer ear properties, which are
usually only accessible to normal hearing people.
A similar approach for positioning the microphone
has been suggested by Weber et al (1998) utilizing a
special headpiece for magnetless cochlear implants.
Moser et al (2002) collected similar data further
supporting the value of microphone location in the
pinna. Earlier evaluations in a unilaterally implanted
group have shown subjective preference for the TMic over the built-in microphone of the ear-level
processor (Frohne-Büchner et al, 2002). Conceptually, bilateral users may benefit more than unilateral
users from in-the-pinna microphone positioning,
as evidenced by better sound localization abilities.
In this study, sound localization was tested in a
group of four bilaterally implanted users with
ear-level sound processors in two conditions: (1)
with the built-in sound processor microphone
and (2) with the T-Mic. Three of the four subjects
normally use their ear-level devices without the
T-Mic. The fourth subject routinely uses a bodyworn processor. Sound localization was tested in
a 360° circle of 12 loudspeakers at a radius of one
meter from the subject. (Figure 2). The stimuli
were presented at 57 dB SPL five times from each
speaker (60 stimuli altogether) in random order. The
stimulus used was the short sentence “What’s your
name?” (in German). A sentence was used rather
than a short sound stimulus in order to compensate for possible AGC adjustments at stimulus
onset. The subject’s task was to identify the correct
speaker from which each stimulus emanated.
Advanced Bionics® Auditory Research Bulletin 2005
“...these data indicate that the
in-the-pinna T-Mic design
offers patients
improved directional hearing
compared to conventional
microphone placements.”
Sound localization errors for the four subjects in
the two listening conditions are shown in Figure
3. Three of the subjects made fewer errors in
the T-Mic condition. Further study of the data
shows improved localization function (fewer
errors made) with the T-Mic for sound sources
in front of and behind the subject (Figure 4).
In interpreting the data it is important to note that
no subject used the T-mic in everyday life. Three
subjects normally used the built-in BTE microphone (indicated by # symbols in Figure 3), which
artifically overestimates performance in this condition for comparison purposes. Only one subject
normally wore a body-worn processor, which means
he had equal experience in both conditions tested.
In summary, these data indicate that the in-thepinna T-Mic design offers patients improved directional hearing compared to conventional microphone
placements. In this study, three of the patients show
improved sound localization abilities despite their lack
of previous experience or training with the T-Mic.
Figure 3. Individual and group means for sound localization
errors in the front-facing BTE microphone condition compared
to the in-the-pinna (T-Mic) condition. Values are expressed as
the RMS degree of error.
References:
Frohne-Büchner C, Büchner A, Gärtner L, Boyle P, Fernald G, Battmer R, Lenarz
Th. (2002) First results with the Clarion in-the-ear microphone connected to the
behind-the-ear sound processor. Paper presented at the 6th European Symposium
on Paediatric Cochlear Implantation, Las Palmas, Spain, February, 2002.
Moser L, Müller J, Helms J, Hellmuth A, Nopp P. (2002) The effect of ear-canal
microphones on speech reception in users of the MED-EL COMBI 40+ cochlear
implant. Paper presented at the 7th International Cochlear Implant Conference,
Manchester, England, September, 2002.
Weber BP, Neuburger J, Lenarz T. (1998) Development and clinical testing of a
non-magnetic cochlear implant. Preliminary experimental studies and surgical
concept. Results in the first 10 patients. Laryngorhinootologie 77(7):376-381.
Figure 4. Mean group errors for each of the twelve sound
sources. Note the improved front and back localization function in the T-Mic for stimuli emanating directly in front of and
behind the subject (sources 5, 6, 7, and 12).
Ear-Level System Features
157
Comparison of Benefit for the Auria T-Mic, Auria BTE,
and Platinum Headpiece Microphones
Sigfrid D. Soli, Ph.D.
Jenny C.Y. Chan, M.A.
Andrew J. Vermiglio, M.A.
Daniel J. Freed, M.S.
House Ear Institute, Los Angeles, CA, USA
“These data suggest that
the Auria T-Mic can provide
improved speech understanding in noise,
especially when speech and noise
are spatially separated.”
The Auria T-Mic is an in-the-ear microphone
designed to provide easy access to cellular telephones, consumer audio electronics, and assistive
listening technology. In addition, because of the
T-Mic’s location within the concha, listeners may
be able to take advantage of the sound filtering
effects of the outer ear. Those effects can provide
better speech understanding in noise in some situations as well as improved localization of sound.
This study compared speech understanding in noise
and sound (speech) localization with the Auria TMic, the Auria built-in (BTE) microphone, and the
Platinum Headpiece (PHP) microphone. First, headrelated transfer functions (HRTFs) were measured
for the ear canal and for sound processed through
the PHP microphone, BTE microphone, and the TMic. A HRTF is the direction-dependent acoustic
filtering of a free-field sound as a result of individual
head, torso, and pinna effects. Each HRTF represents the difference between the sound at a source
originating from a particular free-field location and
the sound that arrives at the ear canal or at each of
the three microphones on a KEMAR mannequin.
Sentence reception thresholds (sSRTs) in noise and
localization ability were evaluated in 12 normalhearing listeners in free field and under headphones
where sound direction and microphone position
were simulated using HRTFs. Using the HINTadaptive test, sSRTs for speech coming from the
front were measured in the presence of noise
originating from the front and from the right side
(90-degree azimuth). Typically, separating speech
from noise results in a lower sSRT, termed “spatial
unmasking.” Spatial unmasking, in turn, helps
listeners to hear in unfavorable listening situations.
158
Advanced Bionics® Auditory Research Bulletin 2005
The mean sSRTs for noise in front and noise at the
right are shown in Figure 1. There is a large spatial
unmasking effect in sound field and for the ear
canal HRTF. That is, when the noise is separated
from the speech, the sSRT is about 7 dB lower than
when the speech and noise both come from the
front. Notably, the spatial unmasking for the T-Mic
is similar to the free-field and ear canal results. In
contrast, there is less spatial unmasking (3-5 dB)
for the PHP and BTE microphones. Therefore, the
T-Mic location allowed these subjects to take better
advantage of the filtering effects of the outer ear.
Localization ability also was better with the T-Mic
than with the PHP and BTE microphones. Figure 2
shows the accuracy with which the 12 normal-hearing
listeners could localize sound originating from 12
speakers placed in the horizontal plane and spanning
180 degrees from ear to ear. The data were scored for
accuracy within 2, 3, 4, 6, and 12 equal sectors of the
180 degrees. The ear canal and T-Mic HRTF results
again matched the sound-field results, while the
BTE and PHP HRTFs gave poorer results. Good
spatial resolution (> 75%) was reached for 6 sectors
(≥ 30 degrees) for the sound field, the ear-canal
HRTF, and the T-Mic HRTF. Greater than 45degree accuracy (4 sectors) was seen for all HRTFs.
These data suggest that the Auria T-Mic can provide
improved speech understanding in noise, especially
when speech and noise are spatially separated. In
addition, the T-Mic may provide better localization of sound than the BTE or PHP microphones.
Figure 1. Spatial masking and unmasking for 12 normal-hearing subjects represented by mean HINT sSRTs (noise front and
noise right) for sound-field listening compared to ear-canal,
T-Mic, PHP, and BTE HRTFs under headphones.
Figure 2. Mean localization accuracy for 12 normal-hearing
subjects scored by number of sectors for sound-field listening
and for ear-canal, PHP, BTE, and T-Mic HRTFs under headphones.
Ear-Level System Features
159
Performance of the Auria T-Mic and the
Behind-the-Ear Microphone in Noise
Elrietha Olivier, B.SpA, M.Th
Royal National Throat, Nose and Ear Hospital, United Kingdom
The Auria T-Mic is an accessory earhook that
attaches to the Auria (as well as CII or Platinum)
ear-level sound processor. The omnidirectional TMic is located on a flexible arm, which allows positioning inside the concha. By contrast, the built-in
behind-the-ear (BTE) microphone is a forward
facing, omnidirectional microphone housed at the
base of the earhook and positioned above the pinna.
Conceptually, the within-the-concha positioning
of the T-Mic should benefit patients when using
the telephone and headphones—and should also
improve the clarity of speech in noisy situations
compared to conventional microphone placement.
A field trial was undertaken to investigate whether
the predicted speech improvement in noise with the
T-Mic could be demonstrated. The investigation included 15 subjects (5 females and 10 males)
with a mean age of 52.1 years (range 30–77 years).
All subjects had used their BTE processors for
at least three months and scored 50% or better
on BKB sentences at their most recent postoperative assessments. All subjects had listening
experience with both the T-Mic and the BTE
microphone before onset of the investigation.
Speech recognition in quiet and in the presence
of multitalker babble was measured with digitally
recorded UCL/CUNY sentences. Performance
was scored as the number of key words correctly
repeated. Speech perception data were collected for
all 15 subjects in three randomized listening conditions—alternating between the different microphone technologies. The listening conditions were:
• Spatially coincident speech and noise, presented
at 0° in front of the subject.
• Spatially separated speech and noise, where
speech was presented from the front of
the listener at 0° and noise at 90° on the
implant side.
• Speech perception in quiet, measured at 0°.
Tests were carried out at a level of 65 dB HL,
one meter from the loudspeakers. The signalto-noise ratio ranged from +10 dB to +15 dB—
adjusted
according
to
individual
performance to minimise floor and ceiling effects.
As shown in Figure 1, the results showed an improvement in speech perception in noise for the T-Mic over
the BTE microphone when speech and noise were
spatially separated. There was no significant improvement in scores when the speech and noise were delivered from the front of the T-Mic at 0° compared to
the control condition (BTE microphone) or in quiet.
The effect of noise for each of the microphone technologies is shown in Figure 2. (Note that this is the
same data as in Figure 1, but with each technology
shown as its own comparison.) For the T-Mic, performance dropped by an average of 38.8% when noise
was introduced at 90° on the implant side, whereas
with the BTE microphone, performance dropped
by an average of 48.5% in the same condition.
Figure 1. Mean performance enhancement of 13.7% with the
T-Mic compared to the BTE microphone when speech is presented at 0° and noise at 90° on the implant side.
160
A two-way analysis of variance for microphone type
(T-Mic and BTE) and noise orientation (0° and
90°) showed a significant main effect of microphone
Advanced Bionics® Auditory Research Bulletin 2005
(F = 9.91; p = .007), a significant main effect of
noise (F = 21.84, p <.001), and a significant interaction (F = 6.77, p = .021). These results reflect what
is seen in the graphs. In general, the T-Mic scores
are higher than the BTE microphone scores, indicating that the additional pinna cues are useful in
many conditions. Of particular interest is the interaction between microphone and noise, showing
that the scores in the presence of noise at 90° for
the T-Mic are superior to those for the BTE microphone. These results indicate that improved listening
performance can be obtained with the Auria T-Mic,
particularly in noisy environments. In spatially separated speech and noise, there is a significant advantage in using the T-Mic over the BTE microphone.
Compared to the control condition, there is no
significant improvement in scores when the speech
and noise are delivered from the front or in quiet.
Previously, cochlear implant recipients using conventional microphone technologies have been unable to
benefit from pinna effects, i.e., enhancement of the
high frequencies in a signal when competing noise
comes from spatially separated angles. The T-Mic,
located inside the concha, provides the advantages
of pinna placement, which are known to contribute
to sound localisation and speech recognition,
particularly in noisy situations. These data indicate
superior performance of the T-Mic and highlight
the need for improved microphone technology in
sound processors for all cochlear implant recipients.
“These data indicate
superior performance of the T-Mic
and highlight the need
for improved microphone technology
in sound processors
for all cochlear implant recipients.”
Figure 2. Effects of noise using the T-Mic versus the BTE
microphone.
Additional Reading
Advanced Bionics Corporation. (2004) Hearing with Two Ears: Technical Advances
for Bilateral Cochlear Implantation. Advanced Bionics white paper, February 2004.
Chung K, Zeng F, Waltzman S. (2004) Using hearing aid directional microphones
and noise reduction algorithms to enhance cochlear implant performance.
Acoustics Research Letters Online. Available online at www.ucihs.uci.edu/hesp/
publications/Chung_ARLO_2004.pdf (accessed 11 July, 2005).
Dhar S, Humes LE, Calandruccio L, Barlow NN, Hiskind N. (2004) Predictability of
speech-in-noise performance from real ear measures of directional hearing aids. Ear
Hear 25(2):147-158.
Dillon H. (2001) Hearing Aids. Sydney, New South Wales: Boomerang Press.
Acknowledgement
The RNTNE gratefully acknowledges Debi Vickers from Advanced Bionics UK for
her help in analysing the data.
Dorman MF, Spahr A. (2003) A preliminary comparison of performance between
patients fit with the CII Bionic Ear® and patients fit with the Nucleus 3G System.
Paper presented at the 2003 Conference on Implantable Auditory Prostheses,
Pacific Grove, CA, 17-22 August, 2003.
Nucleus Report. (2004) Outcomes using bilateral implants in adults. Nucleus Ltd.®
July/August.
Schleich P, Nopp P, D’Haese P. (2004) Head shadow, squelch, and summation
effects in bilateral users of the Med-El Combi 40/40+ Cochlear Implant. Ear Hear
25(3):197–204.
Uhler K, Bates J, Miller C, Segel P, Wei, J, Polite C. Use of an alternate microphone
technology for enhancing BTE sound processor performance in children and
adults. Poster presented at the Ninth Symposium on Cochlear Implants in Children,
Washington, DC, 24-26 April, 2003.
Ear-Level System Features
161
Alternate Microphone Position for Enhanced Listening
with a Behind-the-Ear Speech Processor
Kristin Uhler, M.A.
University of Colorado Health Sciences Center, Aurora, CO, USA
Colleen Polite, Au.D.
University of California, San Francisco, CA, USA
“...the Auria T-Mic can provide
improved speech understanding in noise
as well as enhanced listening
when using telephones and headphones.”
This project evaluated the qualitative benefits of
a prototype version of the T-Mic—the in-theear microphone that now comes standard with
the Auria processor and as an optional accessory for the CII and Platinum BTE® (behindthe-ear) processors (Uhler et al, 2003).
Thirteen adult subjects used the prototype T-Mic with
their BTE processors for one month. They then rated
the T-Mic and built-in BTE microphone for quality
and clarity of speech in everyday situations, including:
• Listening in quiet
• Listening in noise and in group situations
• Speaking on the phone in quiet
• Speaking on the phone in noise
• Using a cell phone
• Listening to music in free field
• Listening to music with headphones
The rating scale ranged from 1 (not clear) to 10 (very
clear). Figure 1 shows the ratings for the two microphone types. Clear differences in ratings between
the T-Mic and BTE microphone were seen for
Figure 1. Mean ratings for adults using the T-Mic and BTE microphone in various listening conditions.
162
Advanced Bionics® Auditory Research Bulletin 2005
listening in noise, talking on the phone, and listening
to music with headphones. Little difference was
found in the ratings for listening to speech in quiet.
speakers. Anecdotal comments from parents and
teachers corroborated the enhanced benefit of the
T-Mic compared to the standard BTE microphone.
T-Mic and BTE microphone benefits also were
compared in five children ages 6 to 12 years. A
parent-interview schedule similar to the MAIS
(Robbins et al,1997) was developed for parents to
rate the communication skills of their children on
a scale from 0 to 4, in which 0 = never, 1 = rarely,
2 = occasionally, 3 = frequently, and 4 = always.
These results indicate that the Auria T-Mic can
provide improved speech understanding in noise
as well as enhanced listening when using telephones and headphones. These early results have
been verified by subsequent studies using the
commercial T-Mic. (See also Soli et al at page
158 and Olivier at page 160 of this publication.)
Parents rated the following communication abilities:
• Responds to voices in quiet
This project was completed when Kristin Uhler was at the Hough Ear Institute,
Oklahoma City, Oklahoma.
• Recognizes environmental sounds
References
• Responds to voices in noise
• Talks on the phone with familiar speakers
Robbins AM, Svirsky M, Osberger MJ, Pisoni DB. (1997) Beyond the audiogram:
The role of functional assessments. In Bess F, ed. Children with Hearing
Impairment. Nashville, TN: Vanderbilt-Bill Wilkerson Center Press, 105-126.
• Recognizes familiar speakers
Figure 2 shows the parent ratings for the T-Mic
and BTE microphone conditions after one month
of T-Mic use. The T-Mic showed improved skills in
noise and when talking on the phone with familiar
Uhler K, Bates J, Miller C, Segel P, Wei, J, Polite C. Use of an alternate microphone
technology for enhancing BTE sound processor performance in children and
adults. Poster presented at the Ninth Symposium on Cochlear Implants in Children,
Washington, DC, 24-26 April, 2003.
�
�
�
�
�
�
�
������
�
������
�
��������������
�������
������
������������������
���������
�����������������
Figure 2. Parents’ mean ratings of children’s listening skills with the T-Mic and BTE microphone in various listening conditions.
Ear-Level System Features
163
Pediatric Evaluation of the HiRes Auria Sound Processor
Nathalie Noël-Petroff, M.D.
Vincent Couloigner, M.D.
Sylvette Wiener-Vacher, M.D.
Martine François, M.D.
Thierry Van den Abbeele, Prof., M.D.
Hôpital Robert Debré, Paris, France
Laure Arnold, M.Sc.
Patrick Boyle, M.Sc.
Advanced Bionics Corporation, Europe
The Auria behind-the-ear (BTE) processor was
designed to address some of the wearability issues
identified in the CII BTE, especially for children.
The internal electronics of the Auria are the same
as in the CII BTE. However, the Auria features
improved styling, more user-friendly controls,
an external battery pack, extended battery life,
and a smaller headpiece. In addition, an earhook
equipped with an LED, called the FireFly®,
is available to verify proper system function.
This study evaluated the clinical application, in terms
of fitting and everyday use, of the Auria in children.
Three children, ages 8, 9, and 10 years, who had been
wearing their implants for at least eight months,
participated in the evaluation. All three used the CII
BTE. Two children used SAS and one child used
MPS. All were enrolled in oral educational programs
and had the ability to undergo open-set speech testing.
The study required three test visits. Because the Auria
is programmed with HiRes using the SoundWave
software, it was important first to switch the children from their standard sound processing strategy,
programmed with SCLIN 2000, to a HiRes strategy
programmed with SoundWave. Therefore, during
the first visit, each child was tested with their CII
BTE and conventional sound processing strategy
(SAS or MPS) with a test battery that included
free field audiometry, the Phonetically Balanced
Kindergarten (PBK) Test in quiet, the Common
Phrases Test, and the Meaningful Use of Speech
Scale questionnaire (MUSS). These test results
served as baseline performance. Then the children
164
were switched over to a HiRes program on their CII
BTE. The fitting was completed using the SoundWave automatic mode (measurement of comfort
levels only) using 16 channels. Manual adjustments
were performed as required in live speech mode for
Input Dynamic Range (IDR), threshold levels (Ts),
and high frequency most comfortable levels (Ms).
After three months of HiRes use on the CII BTE, the
children returned for the second visit. The same test
battery as used in the first test session was administered again. The subjects then were fit with the Auria
using the same HiRes program. The program levels
were altered only if initial experience suggested that a
change was necessary. Initial feedback was recorded.
After one month of experience with the Auria (four
months with HiRes), the children returned for the
third test session. The same test battery was administered. In addition, the subjects’ parents were asked
to complete a questionnaire comparing use of the
CII BTE and the Auria. The questionnaire was
divided into four sections: (1) comparison of the CII
BTE to the Auria, (2) use of the Auria, (3) satisfaction with the Auria, and (4) general comments.
It queried issues such as comfort, ease of handling,
confirmation of system function, and battery life. At
the final test session, the families decided whether
to keep the Auria or to go back to the CII BTE.
The conversion to HiRes sound processing was
straightforward for all three children using the default
parameters from SoundWave. The new programs
were immediately accepted, although it was necessary to increase M levels globally shortly after fitting.
Advanced Bionics® Auditory Research Bulletin 2005
“...conversion to HiRes and to the Auria
were accomplished without difficulty
in these pediatric subjects.”
Also, at the second test session, the Auria immediately was accepted by all subjects. The MUSS scores
were similar between the CII BTE and the Auria.
All children and families chose to keep the Auria and
no major issues were observed with the processors.
One particular benefit seen in everyday life was
the much extended battery life available with the
rechargeable batteries of the Auria compared to
the CII BTE. Battery life was sufficient for a full
school day. Moreover, the external battery pack
gives the opportunity to have an even longer duration of use. The parent questionnaire gave very
satisfying results (Figure 1). In two cases, better
speech understanding was reported. The reasons
for preferring the Auria to the CII BTE were the
longer battery life (two subjects) and better device
retention (one subject). All three children were able
to change their batteries themselves. In response to
the question—If one thing could be changed on the
Auria, what would you change?—the three answers
given were “the size,” “the weight,” and “nothing.”
Figure 1. Parent responses to questions related to everyday
use of the Auria. Parents rated each question on a scale from
0 (worst) to 10 (best).
In summary, conversion to HiRes and to the Auria
were accomplished without difficulty in these pediatric subjects. The Auria provided a longer battery
life, easier to manipulate controls, the convenience
of the FireFly earhook, and improved aesthetics.
The fact that all three children decided to keep
the Auria indicated that the Auria is a more pediatric-friendly BTE processor than the CII BTE.
Ear-Level System Features
165
Evaluation of an Induction (T-Coil) Module for the
Auria Ear-Level Sound Processor
Carolin Frohne-Büchner, Ph.D.*
Andreas Büchner, Ph.D.
Anke Lesinski-Schiedat, M.D.
Rolf-Dieter Battmer, Prof., Ph.D.
Thomas Lenarz, Prof., M.D., Ph.D.
Medizinische Hochschule Hannover, Hannover, Germany
* also with Advanced Bionics Corporation, Europe
Electromagnetic receivers (telecoils or “T coils”)
have long been applied in hearing aids to allow
the reception of speech and other audio signals
through inductive coupling. Compared to newer
assistive listening technologies, telecoil systems
offer the least expensive means of telephone
and large-room hearing access. Moreover, these
systems already are installed in many public places.
The Auria T-Coil is an inductive receiver module
attached to the ear-level sound processor via the
Direct Connect earhook (Figure 1). The module
includes a hearing instrument telecoil with
preamplification. The T-Coil can be rotated by
the user 90 degrees (plus or minus) from the
initial position, allowing maximum induction
of the incoming signal from a loop transmitter.
In this study, the T-Coil was tested in a group of
11 adults, ages 23 to 54 years. The subjects had
experience (from one to eight months) with
their Auria processors before participating in this
study. Nine of the 11 subjects had used hearing
aid telecoils before receiving cochlear implants.
For this study, a questionnaire was designed to
evaluate users’ subjective ratings of sound quality,
physical comfort, and listening ease with three auxiliary listening devices: (1) the Auria T-Mic, (2) the
Auria T-Coil, and (3) a standard, directly coupled
telephone adaptor. The questionnaire applied a
10-point rating scale (0 = very poor, 5 = neutral,
and 10 = very good). Subjective ratings compared
listening experiences with telephones at home and
at work, in quiet and noisy conditions. To evaluate
objectively speech perception abilities over the
telephone, the HSM sentence test (in quiet) was
administered (Hochmair-Desoyer et al, 1998).
The recorded test stimuli were delivered through a
CD player coupled to the tester’s telephone, which
was located in a room remote from the subject.
Shown in Figure 2 are individual and mean scores
for the 10 patients able to perform the HSM
sentences test. The data show similar withinsubject scores as well as group means across the
three audio inputs evaluated. Ceiling effects were
seen in four of the subjects who scored 90% or
better with all three devices. These data indicate
that speech understanding over the telephone
with the T-Coil is similar to that achieved with
the standard telephone adaptor and somewhat
improved compared to the T-Mic condition.
Figure 1. (A) The HiRes Auria sound processor with Direct
Connect earhook. (Note: T-Coil is not depicted.) (B) Cutaway
drawings: cross section details of the Auria T-Coil induction
module.
166
Subjective comparisons of sound quality for the
Auria T-Coil and T-Mic devices, in quiet and
noisy environments, are shown in Figure 3. In
quiet listening conditions, six subjects judged the
T-Coil to sound better in quality than the T-Mic,
two subjects rated the T-Coil poorer in sound
quality, and two subjects found the two inputs to
be equivalent. In noisy conditions, a strong advantage was seen for the T-Coil—with eight of nine
Advanced Bionics® Auditory Research Bulletin 2005
subjects tested judging it to provide better sound
quality over the T-Mic. During the study period,
four subjects evaluated the T-Coil with loop systems
at home, in church, or in public buildings. Two (of
the four) subjects rated the sound quality as “good”
(7), and the other two rated it as “very good” (10).
It should be noted that while testing was conducted
in quiet, the magnetic noise in the test room may
have disadvantaged the Auria T-Coil by introducing
an unfavorable signal-to-noise condition. Also, the
program settings were altered for use with the T-Coil
to mitigate reports of noise, whereas program settings
were not altered for use with the T-Mic or standard
telephone adapter. (Our clinical experience indicates
that reducing IDR may optimize use of the T-Coil.)
As shown in these data, the advantages of the TCoil are more readily seen in poor acoustic signalto-noise conditions. The questionnaire responses
regarding sound quality are encouraging—especially
because telephones in “real life” home and office
environments were used for the subjective measures.
Reference
Hochmair-Desoyer I, Schultz E, Moser L, Schmidt M. (1998) The HSM Sentence
Test as a tool for evaluating the speech understanding in noise of cochlear implant
users. Am J Otol 18:83.
“In noisy conditions, a strong advantage
was seen for the T-Coil—with eight of
nine subjects tested judging it to provide
better sound quality over the T-Mic.”
Figure 2. Individual and mean scores for the HSM sentences
in quiet using different auxiliary inputs over the telephone—
including the Auria T-Mic, a standard telephone adapter, and
the Auria T-Coil. (Note: Subject 6 was not able to perform this
test measure.)
Figure 3. Subjective ratings comparing sound quality over the telephone with the Auria T-Coil and the T-Mic in quiet and noisy
conditions. (Note: Subject 3 did not complete the questionnaire due to lack of experience; subject 11 had no opportunity for telephone use in noisy conditions.)
Ear-Level System Features
167
Patient Assessment
& Training
Early on, Advanced Bionics recognized the need for documenting
implant benefit in very young children and consequently supported
the development of the IT-MAIS. As technology has advanced even
further, patients now are reaching ceiling performance on traditional
hearing assessments. Thus, a requirement has arisen for new and
more challenging measures of implant benefit.
This section explores the development and clinical application of
assessment tools for cochlear implant recipients of all ages and
highlights the importance of auditory training as a key component
in the rehabilitation of cochlear implant recipients.
Monitoring Expectations and Perceived Quality of Life
Changes in Adult Cochlear Implant Users
Regina Presley, Au.D.
Greater Baltimore Medical Center, Baltimore, MD, USA
“...when counseled appropriately,
patients are able to form
realistic expectations
and to perceive
improved quality of life
within a short period
of cochlear implant use.”
Research and clinical experience have shown that
cochlear implant recipients demonstrate a wide
range of clinical and everyday quality of life benefits. Arguably, the variable that most influences
perceived benefit is patient expectations—and
the degree to which individual expectations are
satisfied. Consequently, a patient’s perceptions of
benefit, including quality of life enhancements,
depend largely on the effectiveness of pre- and
postoperative counseling—with establishing “realistic” patient expectations as the primary objective.
For this study, several closed- and open-ended
questionnaires were developed to monitor patient
expectations of cochlear implant benefits and
perceived changes in quality of life over time. The
questionnaires address patient expectations in
the following areas: sound awareness, lipreading,
communication, telephone use, television viewing,
background noise, music appreciation, work and
education, social life, and quality of life changes.
Each questionnaire has been completed in less
than 15 minutes and has proven simple to score.
According to the study plan (in progress), individual responses will be obtained before implant
surgery and at 3-, 6-, and 12-month intervals
postimplant. Thus far, responses from four patients
through the six-month postimplant interval
have been obtained. These patients range in age
from 44 to 74 years, with a mean age of 58 years.
The data collected to date show that for all four
patients, preimplant expectations have been met and
their perceived quality of life has been improved.
The mean expectation score increased from 58.5%
preimplant to 70% postimplant (Figure 1), and
the mean quality of life score increased from
33% preimplant to 69% postimplant (Figure 2).
170
Advanced Bionics® Auditory Research Bulletin 2005
Further study of the questionnaire responses revealed
that the highest levels of satisfaction with the cochlear
implant, as well as for everyday life in general, were in
the following categories: hearing sounds from behind,
social participation, social interactions, and workplace
interactions. These results suggest that when counseled appropriately, patients are able to form realistic expectations and to perceive improved quality
of life within a short period of cochlear implant use.
The questionnaires designed for this study have been
useful in monitoring the effectiveness of our team’s
counseling techniques and our patients’ expectations
over time. In the future, responses collected at the 12month postimplant interval will indicate the longer
effects our counseling has had on patient expectations. Tracking patient expectations and perceived
quality of life may prove useful in demonstrating
cochlear implant benefits to insurance companies,
hospital administrators, and financial supporters.
Figure 1. Individual scores for the Expectations Questionnaire
at preimplant and six months postimplant. Scores shown were
derived by converting the rating scale values into a percent of
the highest possible scores.
Figure 2. Individual scores for the Quality of Life Questionnaire at preimplant and six months postimplant. Scores shown
were derived by converting the rating scale values into a percent of the highest possible scores.
Patient Assessment & Training
171
Outcome Measures in Prelinguistically Deafened Adults
with Cochlear Implants
Paula B. Marcinkevich, Au.D.
Nancy S. Catterall, M.Sp.A.
Thomas O. Willcox, Jr., M.D.
Thomas Jefferson University, Philadelphia, PA, USA
“...open-set speech perception tests
do not always convey the benefits derived
by prelinguistic patients.
Controversy has long existed regarding cochlear
implant benefits in prelinguistically deafened adults,
for whom past studies have shown no (or extremely
limited) open-set speech recognition However, such
objective clinical measures often do not reflect the
everyday life benefits many patients experience.
We reviewed the speech perception scores and
quality of life ratings for our clinic population of
12 prelinguistic adults who ranged in age from 35
to 64 years. Ten of the patients were identified with
hearing loss before six years of age. All of the patients
use their implants daily, for at least 10 hours a day.
Figure 1. Individual scores for HINT sentences in quiet preoperatively compared to six months postoperatively.
Pre- and postoperatively, the patients were administered open-set speech perception tests—CNC
words and the HINT sentences in quiet. After
at least six months of implant use, a questionnaire was given to measure subjectively these
patients’ perceived changes in their quality of life.
Individual scores for eight patients ranged from
slight to marked improvement on the HINT
sentences test (Figure 1). Eleven patients responded
that since implantation their quality of life had
improved, as seen in Figure 2. Overall, our clinical
results agree with the findings of others: open-set
speech perception tests do not always convey the
benefits derived by prelinguistic patients. We recommend closed-set speech perception measures (such as
the WIPI and Minimal Pairs Test) as well as subjective quality of life assessments be included in standard clinical protocols for this patient population.
Figure 2. Distribution of scaled responses for the questionnaire
item regarding quality of life improvement. Scale: 0 = strongly
disagree, 1 = disagree, 2 = somewhat disagree, 3 = somewhat
agree, 4 = agree, 5 = strongly agree. (Note: The one response
scaled as ”somewhat disagree” was for a patient who answered
“no change” to the questionnaire item.)
172
Advanced Bionics® Auditory Research Bulletin 2005
Everyday Cochlear Implant Benefits in
Prelinguistically Deafened Adults
Dawn Burton Koch, Ph.D.
Mary Joe Osberger, Ph.D.
Advanced Bionics Corporation, Valencia, CA, USA
Irena Vujanovic, M.A.
Belgrade University, Belgrade, Serbia
Speech perception abilities vary widely in prelinguistically deafened adult cochlear implant users,
and clinical test results are much poorer in these
individuals than in their postlinguistically deafened counterparts. However, clinical experience
indicates that the everyday implant benefit experienced by prelinguistically deafened adults may not
be reflected accurately in audiological test results.
A questionnaire was mailed to 66 prelinguistically
deafened adults who had preimplant and onemonth, three-month, and six-month postimplant
speech perception data obtained during a clinical
trial of the CI-HiFocus I cochlear implant. The
questionnaire was designed to acquire information
about (1) implant and hearing aid use, (2) educational, family, and occupational demographics, and
(3) subjective ratings of quality of life and communication skills. Thirty-five adults (53%) returned the
questionnaire. The mean age was 33 years (range
18-60 years). There were 20 women and 15 men.
The data showed that all subjects consistently use the
auditory input provided by the implant or implant
“...prelinguistic hearing loss
does not preclude adults from experiencing
significant cochlear implant benefits.”
plus hearing aid, most were educated in “oral” environments, most completed college, and most were
employed in situations requiring spoken communication skills. Despite these demographics, the six-month
speech perception scores spanned a wide range (CNC
words: mean = 12.6%, range 0-52%; CID sentences:
mean = 29%, range 0-97%; HINT sentences:
mean = 25%, range 0-91%). The 35 respondents
were classified as “poor” or “moderate-good” users
based upon their six-month CID sentence scores.
Data analysis indicated no group effects on subjective ratings of quality of life and communication
skills (Figure 1). In other words, the respondents’
assessments of everyday benefit from the implant
were independent of their speech perception abilities. Typically, differences in communication skills
between poor and moderate-good users were
evident in listening situations lacking supplemental
cues (for example, talking on the phone with unfamiliar speakers). These data indicate that prelinguistic hearing loss does not preclude adults from
experiencing significant cochlear implant benefits.
Figure 1. Summary of self-reported implant benefit from 35 patients with prelinguistic deafness showing absence of group effects
on any item. Scale: 0 = strongly disagree, 1 = disagree, 2 = somewhat disagree, 3 = somewhat agree, 4 = agree, 5 = strongly agree.
Patient Assessment & Training
173
Self-Reported Benefit in Prelinguistically
Deafened Adults
Jennifer Mertes, Au.D.
Johns Hopkins University, Baltimore, MD, USA
“...prelinguistically deafened adults...
experience significant benefit
after cochlear implantation,
independent of performance on
traditional clinical measures.”
Clinical experience suggests that the cochlear
implant benefits received by prelinguistically deafened adults may not be assessed fully by standard
audiological tests. A questionnaire developed by
Koch et al (2004) was mailed to 113 prelinguistically deafened adults who had been implanted at
the Listening Center at Johns Hopkins University.
The questionnaire was designed to acquire information about (1) implant and hearing aid use,
(2) educational, family, and occupational demographics, and (3) subjective ratings of quality
of life and communication skills. Questionnaire data then were compared to clinical speech
perception results at one year postimplantation.
Sixty-nine individuals returned the questionnaire
(61% return rate). Of those 69 respondents, 40 used
the Clarion implant, 27 used the Nucleus implant,
and 2 used the Med-El device. The mean age was
41 years (range 20-73 years). The sample included
38 women and 31 men. (Ten patients did not have
speech perception scores at one year postimplant and
therefore were not included in the data analyses.) The
demographic data from the questionnaire indicated
that all patients use their cochlear implants consistently on a daily basis. Ten individuals use a hearing
aid in the nonimplanted ear. Most people were well
educated (primarily in oral environments), and most
had occupations that required communication skills.
Mean speech perception scores over time
for the prelinguistic group are shown in
Figure 1. Also shown for comparison are
one year mean scores for our postlinguistically deafened patients. These mean data
show that although prelinguistically deafened
patients score poorly in comparison to their
postlinguistic counterparts, they do, however,
demonstrate an increase in benefit over time.
Figure 1. Mean speech perception scores over time for 59 prelinguistically deafened survey respondents. For comparison,
mean scores at one year for 162 postlinguistically deafened
adults are shown to the far right of the figure.
174
A wide variability among the survey participants was observed in the individual speech
perception scores obtained at one year postimplantation (Figure 2). Scores ranged from
Advanced Bionics® Auditory Research Bulletin 2005
0-99% on the sentence tests (HINT mean
= 26.1%, CID sentence mean = 29.5%) and
0-82% on word recognition (mean = 16.5%).
Based on the one-year individual results, the prelinguistic survey respondents were divided into two
groups. The “poor” users were those who scored
0% on one or more test measures. The “moderateto-good” users were those who scored above 0% on
all tests. Quality of life and communication skills
ratings were compared between the poor and the
moderate-to-good user groups. Mean ratings for
quality-of-life items ranged between 2.7 (somewhat agree) and 4.1 (agree). There were no large
differences in quality-of-life assessment between
the two groups (Figure 3). Any differences in
communication skills between poor and moderateto-good users were on items that described situations requiring hearing ability (for example,
talking on the phone with unfamiliar speakers).
These results are similar to those obtained by
Koch and colleagues (2004) in a study of prelinguistically deafened adults implanted with the
CI-HiFocus I cochlear implant. Although prelinguistically deafened adults show significantly poorer
speech perception abilities compared to postlinguistically deafened adults, they do experience significant
benefit after cochlear implantation, independent
of performance on traditional clinical measures.
Figure 2. Individual rank-ordered HINT in quiet scores at one
year postimplant for 59 survey respondents. (Mean = 26.1%.)
Reference
Koch DB, King CD, Vujanovic I, Osberger MJ. (2004) Everyday cochlear implant
benefit in prelingually deafened adults. Paper presented at the American Auditory
Society, Scottsdale, AZ, 7-9 March, 2004.
Figure 3. Summary of self-reported implant benefit for patients with prelinguistic deafness (Scale: 0 = strongly disagree;
1 = disagree; 2 = somewhat disagree; 3 = somewhat agree; 4 = agree; 5 = strongly agree.)
Patient Assessment & Training
175
Tinnitus Suppression and Cochlear Implants:
Study of Patients with Postlinguistic Hearing Loss
Hassan Tavakoli, M.Sc.
Susan Abdi, M.Sc.
Amir Aalam Hospital
Tehran University of Medical Sciences, Tehran, Iran
Tinnitus results in variable degrees of disability in the
affected patient population (Bauer, 2004; Ahmad &
Seidman, 2004; Waddel & Canter, 2004.). Cochlear
implants have revolutionized the treatment of
patients with profound hearing loss, providing electrical stimulation to the cochlea based on acoustic
stimuli. Various authors have reported different
results of tinnitus suppression following cochlear
implantation (House & Urban, 1973; Ito & Sakakihara, 1994; Tyler, 1995). Our primary hypothesis
was that both types of cochlear implants currently
available in our centre would prove to be effective
in suppressing tinnitus. We wished to ascertain
whether any patients suffered any long-term exacerbation in their tinnitus subsequent to implantation.
In a longitudinal study, 57 patients (30 females
and 27 males) ranging in age from 6 to 51 years
Table 1. Degree of tinnitis severity before and after
implantation in Clarion and Nucleus patients.
Clarion
n=42
Nucleus
n=15
Severity
Before
Severe
34 (81%)
Moderate
8 (19%)
Mild
6 (14%)
None
36 (86%)
Severe
15 (100%)
1 (7%)
Moderate
11 (73%)
Mild
3 (20%)
None
Numbers of patients according to level of tinnitis.
176
After
(mean 23.5; SD 12.7) were included in this study.
All patients presented with severe tinnitus and
postlinguistic onset of deafness with a mean duration of 7.65 years (SD 6.69 years). All patients
received either Clarion or Nucleus multichannel
cochlear implants at the Amir-Aalam Hospital,
Cochlear Implantation Center, Tehran University
of Medical Sciences between the years of 1991
to 2004. Before undergoing cochlear implantation, each patient’s tinnitus was categorized
based on its severity, using a quantitative scale.
Prospectively, data was collected for all patients—
including information pertaining to the presence
and severity of tinnitus. Subsequent to successful
implant activation, each patient’s tinnitus was
evaluated and categorized using the same scale
that was used preoperatively. These data provided
the primary outcome measure of the study: the
degree of tinnitus suppression afforded by cochlear
implants. The degree of postimplantation tinnitus
was also analyzed with regard to the type of
implant used as well as the gender of each patient.
Table 1 summarizes patients’ rating of tinnitis preand postoperatively. Before surgery, 49 subjects
(85.9%) had marked tinnitus, described as severe.
Tinnitus was not a clinical factor in choosing the
ear to be implanted. The difference in severity or
duration of tinnitus between the implanted and
nonimplanted ears was not significant (p = 0.593).
Subsequent to implantation, 36 subjects (63.1%)
had a complete suppression of their tinnitus. Twenty
subjects (35%) had some suppression of tinnitus,
and only one subject (1.7%) noted no change in
tinnitus severity. Thus, statistical analysis revealed
Advanced Bionics® Auditory Research Bulletin 2005
“Our data support the hypothesis that
a significant reduction in tinnitus level
occurs in subjects
receiving a cochlear implant.”
a significant reduction in tinnitus intensity in
patients using cochlear implants—with 56 of 57
subjects (98.2%) experiencing a reduction in tinnitus
intensity (p = 0.0001). No subject suffered an exacerbation of tinnitus subsequent to implantation.
ihara (1994) reported tinnitus to be suppressed or
abolished in 77% of their patients. In our study, the
Clarion multichannel implants showed significantly
greater suppression than Nucleus implants. We could
not find any previous study to confirm this finding.
The median reduction of severity of tinnitus
for implanted and nonimplanted ears was
“moderate” and “mild” respectively, a difference
that was not statistically significant (p = 0.327).
Median reduction of tinnitus duration was small
for both implanted and nonimplanted ears.
Another interesting point is lack of significant difference in tinnitus suppression between
the implanted and nonimplanted ears. This
may be attributed directly to electrical stimulation providing useful auditory input but also to
improved quality of life with a consequential
reduction in levels of stress for the implant users.
Severity of tinnitus was reduced in both implanted
and nonimplanted ears after the implantation (p
= 0.004). Thus, the implant would appear to offer
both ipsilateral and contralateral tinnitus suppression. As shown in the table, the Clarion implants
showed a significantly greater degree of tinnitus
suppression than Nucleus implants (p < 0.05).
Our data support the hypothesis that a significant
reduction in tinnitus level occurs in subjects receiving
a cochlear implant. Cochlear implantation has not
only been a way out of silence for our recipients but
also a way of leading a much more normal life with
disappearance or marked reduction of many problems associated with deafness. Our data indicate
that 98.2% of patients experienced alleviation of
tinnitus—a therapeutic efficacy that is greater than
the 70-74% of patients reporting improvement in
other studies (Khoursandi, 1999; Brackmann, 1981;
McKerrow (1991); Khalessi, 1996; Ito, 1997). Ruckenstein et al (2001) reported a significant reduction
in tinnitus levels in patients with cochlear implants.
Prior to this, Souliere et al (1992) reported an overall
improvement in 74% of patients, and Ito and Sakak-
References
Ahmad N, Seidman M. (2004) Tinnitus in the older adult: epidemiology,
pathophysiology and treatment options. Drugs Aging 21(5):297-305.
Brackmann DE. (1981) Reduction of tinnitus in cochlear-implant patients.
J Laryngol Otol (Suppl 4):163–165.
Bauer CA. (2004) Mechanisms of tinnitus generation. Curr Opin Otolaryngol Head
Neck Surg 12(5):413-417.
House WF, Urban J. (1973) Long term results of electrode implantation and
electrical stimulation of the cochlea in man. Ann Otol Rhinol Laryngol 82: 504517.
Ito J. (1997) Tinnitus suppression in cochlear implant patients. Otolaryngol Head
Neck Surg 117(6):701-703.
Ito J, Sakakihara J. (1994) Tinnitus suppression by electrical stimulation of the
cochlear wall and by cochlear implantation. Laryngoscope 104:752–754.
Khalessi MR, Khorsandi MT, Abdi S. (1996) Cochlear implantation in Iran. Acta
Medica Iranica 34:73-76.
Khorsandi MY, Borghei H, Abdi S. (1999) Suppression of tinnitus in patients
undergoing cochlear implantation. Acta Medica Iranica 37(2):86-88
McKerrow WS, Schreiner CE, Snyder RL, Merzenich MM, Toner JG. (1991) Tinnitus
suppression by cochlear implants. Ann Otol Rhinol Laryngol 100:552–588.
Ruckenstein MJ, Hedgepeth C, Rafter KO, Montes ML, Bigelow DC. (2001) Tinnitus
suppression in patients with cochlear implants. Otol Neurotol 22(2):200-204.
Souliere CRJ, Kileny PR, Zwolan TA, et al. (1992) Tinnitus suppression following
cochlear implantation: a multifactorial investigation. Arch Otolaryngol Head Neck
Surg 118:1291–1297.
Tyler RS. (1995) Tinnitus in the profoundly hearing-impaired and the effects of
cochlear implants. Ann Otol Rhinol Laryngol Suppl 165:25.
Waddell A, Canter R. (2004) Tinnitus. Am Fam Physician 69(3):591-592.
Patient Assessment & Training
177
Schooling and Educational Performance in
Children and Adolescents Wearing Cochlear Implants
Ersilia Bosco, Cl.Psych.
Patrizia Mancini, M.D.
Luciana D’Agosta, Sp.Th.
Deborah Ballantyne, Ph.D.
Chiara D’Elia M.D.
Roberto Filipo, Prof., M.D.
University of Rome La Sapienza, Rome, Italy
Technological progress in the field of cochlear
implants has led us to evaluate more subtle yet
complex aspects of hearing function (Robbins,
2000). This approach provides us with detailed
information on specific aspects of speech perception
but can overlook the ultimate aim of the clinician—
improving the quality of life of both adults (Faber
and Grontved, 2000; Lillemor et al, 2004) and children (Beadle et al, 2000; Chmiel et al, 2000) with
cochlear implants. The cochlear implant center at
University “La Sapienza” of Rome has paid particular
attention not only to the family system (Bosco and
Filipo, 2003; Bosco et al, 1996) but also to various
aspects of schooling. The aim of the present study
was to assess the impact of implant use on schooling.
The study group consisted of 50 profoundly deaf
children implanted in our centre (48 prelinguistic,
1 perilinguistic, and 1 postlinguistic). Forty-eight
children (3 bilateral) were implanted with Clarion
devices and used either HiRes (n=21), CIS (n=15),
or SAS (n=12) sound processing. The remaining two
children were implanted with Med-El devices and
used the CIS strategy. Mean duration of implant use
at the time of this study was 3.5 years (range = 8
months to 10.5 years). A structured interview was
conducted with the parents to obtain the educational methods and the classroom settings of their
child, the support facilities present in the child’s
classroom, the method followed by speech therapists,
and the average amount of therapy in terms of hours
per week. Teachers completed a questionnaire and
rated: (1) interaction with peers (the tendency of the
child to cooperate or to be isolated and to be assertive or passive), and (2) interaction with adults (the
tendency of the child to be dependent on others or to
be self governing, and to establish relationships or be
distant from others). School report cards were used to
178
evaluate linguistic (oral and written history), logical
and mathematic, and expressive (art, theatre, music)
skills. Performance was classified as inadequate,
sufficient, quite good, good, and excellent. Nonverbal
IQ was assessed with the Raven Coloured Matrices
(1984) and Goodenough-Harris Drawing Test.
Analyses of the educational characteristics revealed
that all children attended mainstream classes in state
schools—with the exception of one Chinese child
who attended a special school for the deaf. Twentysix percent (13/50) of the children attended nursery
school; 44% (22/50) primary school; 16% (8/50)
secondary school; 8% (4/50) high school; and 6%
(3/50) university. Sixty-four percent (32/50) of the
children were following the same program as the rest
of the class, 22% (11/50), an easier version, and 14%
(7/50) a personalized version. Two university students
attended special courses for the deaf organized by the
Faculty of Engineering. After the initial two years,
they will be mainstreamed. There was no gap between
chronological age and class (grade) level attended
for 62% (31/50) of the children, whereas a gap of
only one year was seen for 28% (14/50), and a gap
of two years for 10% (5/50). Of these five, two were
university students who had decided to begin their
studies after an interruption of school for one year.
Academic performance was examined for those children in primary to high school (34/50; 68%). There
were no insufficient performers in linguistic skills.
Greater linguistic competence tended to be associated with more hours of rehabilitation. For example,
the children who were classified as good/excellent
performers in primary school received an average
of 2.4 hours per week of rehabilitation (sd=0.5),
whereas those children classified as demonstrating
sufficient linguistic skills received an average of 1.8
hours per week. In the area of logical and mathematical skills, the overall results were positive with
22 (58.8 %) classified as good performers and 14
(41 %) classified as sufficient performers. Again, the
best academic skills were associated with a greater
Advanced Bionics® Auditory Research Bulletin 2005
number of hours of rehabilitation. In fact, in primary
school, the good/excellent group received an average
of 2.4 hours of rehabilitation per week (sd=0.6),
whereas those with sufficient performance received
only 1.3 hours/week (sd=0.8). In the area of expressive skills, the results also were positive with 77%
(26/34) of the children classified between good and
excellent in an area consisting of music and drama as
well as drawing and manual activities. No association
was found between non verbal IQ and school reports.
Figure 1. Quality of social interaction with peers.
Performance of the three university students
(majoring in computer science, engineering, and
economics) were equally encouraging—receiving
on average marks of 24.3 (out of 30 possible).
Quality of social interactions is summarized in
Figures 1 and 2. The results showed that there was
normal development over the years for the social skills
determined by the teacher questionnaire. By primary
school, 70% of the children were confident and active
in their communications with peers—a trend that
continued to university. In their interactions with
adults, the children were proactive in establishing
relationships by the time they reached high school.
In conclusion, children wearing cochlear implants
showed a very positive trend in performance and
learning skills. The delay between chronological
age and class (grade) level appeared to be influenced by the interaction of many factors (history
of deafness, use or non-use of hearing aids, type
and amount of rehabilitation, IQ, type of school,
etc). As could be expected, a greater number of
hours of rehabilitation tended to coincide with
better performance for linguistic and logical skills.
Hence, it is concluded that prelinguistically deafened children wearing cochlear implants exhibit
little or no academic delay in comparison to their
normal-hearing peers and are able to establish an
adequate rapport with both peers and adults. Future
studies will compare speech perception performance
and linguistic skills in these implanted children.
Figure 2. Quality of social interaction with adults.
References
Beadle EAR, Shores A, Wood EJ. (2000) Parental perception of the impact upon the
family of cochlear implantation in children. Ann Otol Rhinol Laryngol 185 Suppl:
111-114.
Bosco E, Filipo R. (2003) Un servizio di consulenza psicopedagogica per le
famiglie dei bambini e degli adolescenti sordi impiantati in L’impianto cocleare
nelle sordità gravi e profonde. I Care 180-186.
Bosco E, Argirò MT, Ballantyne D. (1996) Rehabilitation procedures adapted
to adult and child cochlear implant users. In: Allum D, ed. Cochlear Implant
Rehabilitation in Children and Adults. London: Whurr Publishers, 31-52.
Chmiel R, Sutton L, Jenkins H. (2000) Quality of life in children with cochlear
implants. Ann Otol Rhinol Laryngol 185 Suppl 103-105.
Faber CE, Grontved AM. (2000) Cochlear implantation and change in quality of
life. Acta Otolaryngol 543 Suppl 151-153.
Lillemor R, Halleberg M, Ringdhal A. (2004) Living with cochlear implants:
Experiences of 17 adults patients in Sweden. Int J Audiol 43:115-121.
Robbins AM. Rehabilitation after cochlear implantation. (2004) In: Niparko JK, ed.
Cochlear Implants: Principles and Practices. Philadelphia: Lippincott Williams &
Wilkins, 323-367.
Patient Assessment & Training
179
Use of the PRISE in Evaluating Preverbal Development in
Infants with Cochlear Implants
Liat Kishon-Rabin, Ph.D.
Riki Taitelbaum-Swead, M.A.*
Ruth Ezrati-Vinacour, Ph.D.
Minka Hildesheimer, Ph.D.*
Tel Aviv University, Tel Aviv, Israel
* also Chaim Sheba Medical Center, Tel Hashomer, Israel
“...the data in the present study
support the feasibility of the PRISE
as a measure for assessing
preverbal vocalizations in infants.”
Preverbal vocalizations are considered to be effective predictors of later articulation and language
abilities. Any cessation of these sounds due to
hearing loss may impede formation of spoken
language. It is expected, therefore, that early assessment and intervention of infants with hearing loss
would result in increased preverbal vocalizations
and subsequent improved speech production skills.
Thus, preverbal vocalizations may provide an additional measure of cochlear implant device effectiveness in young hearing-impaired (HI) infants.
The PRISE (Production Infant Scale Evaluation)
was designed to query parents about their infants’
vocal behavior in everyday situations. It includes 11
questions that reflect developmental milestones in
preverbal vocalizations. The first questions relate to
vocal stages determined by anatomical and physiological constraints (such as phonation, cooing,
and expansion). The remaining questions reflect
the effect of auditory perception on vocalization
(such as canonical versus variegated babbling). The
last question asks whether the infant uses a permanent sequence of sounds in relation to a certain
object—thus referring to the one-word stage of
development. No questions relate to development beyond this preverbal stage, thereby avoiding
confounding factors associated with knowledge
of the language. The probes are asked in a specific,
180
hierarchical order so that scoring is cumulative
over time, with increased age. The examiner scores
the frequency of occurrence of each target behavior
according to well defined criteria. The advantage of
this questionnaire is that it assesses the behavior
of the infant in situations that are not limited to
the testing booth in a fast and reliable way and
without requiring the cooperation of the infant.
The main goals of the present study were to assess
the feasibility of the PRISE as a measure of
preverbal vocalizations in normal-hearing infants
and to compare their growth function to that of
infants before and after cochlear implantation. Two
groups of subjects participated in this study. The
first group included 163 infants between 0.5 and 20
months of age (mean age = 9 months) with normal
hearing and normal development (NH), as reported
by the parents. The second group included 18 infants
with severe-to–profound congenital hearing loss
who were implanted with a multichannel cochlear
implant (Clarion or Nucleus). The age of implantation ranged from 11 to 29 months (mean age of
implantation = 19 months) and length of implant use
ranged from 2 to 14 months (mean length of use = 7.5
months). Etiologies included CMV, genetic factors,
or unknown causes. Infants in the study group were
included if they had not reached the first-word stage
in their preimplant development and did not have
any known problems other than their hearing loss.
Parents were instructed to observe their infants’
vocal and auditory behaviors for a week prior
to interview. Parents in the study group were
administered the PRISE before and after
implantation. The clinical staff conducted the
test using a protocol that contains numerous
written probes to which clinicians must adhere.
Preverbal vocalization scores of individual infants
with normal hearing and the exponential growth
Advanced Bionics® Auditory Research Bulletin 2005
function that best fits the data (r=0.95, p<0.01)
are shown in Figure 1. The data show that 90% of
the normal-hearing PRISE performance can be
explained by age—reflecting a regular sequence of
development from birth to the emergence of words.
Also shown in Figure 1 are data for the HI infants
pre- and postimplant as a function of chronological
age. The data show that before implantation, the
infants’ PRISE scores do not exceed 45% (regardless
of age), which based on the normative data, reflects
vowel-like productions. However, after implantation,
large improvements in vocalization occur. Specifically, the infants with a short period of implant use
achieve PRISE values that correspond to consonantvowel productions (canonical stage) and some even
reach the first-word stage. Note that PRISE scores
for the HI infants six months of age or younger
prior to implant are within the normative range of
vocal behavior—thus in keeping with published
studies and reflecting the validity of the test.
Figure 1. PRISE scores of NH (open circles) and HI infants
preimplant (grey filled diamonds) and postimplant (black filled
diamonds) as a function of chronological age. Also shown is the
best-fitting exponential function (r= 0.95, p <0.01) to the NH
data (solid line) and the +/- 2 SD (dotted lines).
Figure 2 shows that PRISE scores for the study
group are in keeping with normal development
of pre-first-word vocalizations when plotted as
a function of duration of implant use. The performance of cochlear implant users—equivalent
or better in comparison to younger children
with normal hearing—may be related in part
to their older chronological age and thus their
greater anatomical and physiological maturity.
In summary, data in the present study support the
feasibility of the PRISE as a measure for assessing
preverbal vocalizations in infants. The data reported
here emphasize the importance of auditory functioning for the development of vocalizations
relevant to subsequent speech development. It also
suggests that the PRISE can assist in evaluating the
effectiveness of the habilitation process in general
and the efficacy of the implant device in particular.
Figure 2. PRISE scores of NH (open circles) and HI infants with
CI (black filled diamonds) as a function of duration of implant
use. Also shown is the best-fitting exponential function
(r = 0.95, p<0.01) to the NH data (solid line) and the +/- 2 SD
(dotted lines).
Patient Assessment & Training
181
Teaching Nursery Rhymes with Music to Young Children
Using Cochlear Implants
Susan Abdi, M.Sc.
Hassan Tavakoli, M.Sc.
Amir Aalam Hospital
Tehran University of Medical Sciences, Tehran, Iran
“...children with prelinguistic
onset of deafness...require a
thorough habilitation program
that ideally incorporates
all auditory aspects of life...”
Nursery rhymes are an indispensable part of the
development and education of children with normal
hearing—fulfilling a crucial role in their language
and speech development. The same applies to children with prelinguistic onset of deafness. These children require a thorough habilitation program that
ideally incorporates all the auditory aspects of life—
including rhythm, tonality and melody. We report
here on our experience teaching nursery rhymes with
music as part of a habilitation program for young,
prelinguistic children who use cochlear implants.
Twenty-one children, ranging in age from 2.5 to 6
years, participated in this training program. All children were implanted at the Hearing Research Centre
of the Amir Aalam Hospital. The children were
exposed to familiar nursery rhymes accompanied by
the musical melodies. The children were taught how
to play simple instruments (such as a xylophone) or
more sophisticated instruments (such as a sitar, a
traditional string instrument). The time the children
spent in music training prior to assessment ranged
from 5 to 41 months. We evaluated the children’s abilities to sing common nursery rhymes and to perform
melodies on a simple instrument. We subjectively
rated each vocal and instrumental performance on a
scale of 0 (“no ability”) to 10 (“perfect and flawless”).
182
Advanced Bionics® Auditory Research Bulletin 2005
“...young children using cochlear implants
can successfully learn nursery rhymes.”
The rating scores for vocal and instrumental
performances were totaled and averaged across
the 21 subjects. Table 1 delineates the mean scores
for each variable that we subjectively assessed to
determine overall performance. The mean score
achieved by the children for combined (vocalinstrumental) performances was 6.25, with a
median score of 8. If the better score (either vocal
or instrumental) for each child was used, the group
mean increased to 8.14 and the median to 10.
We conclude that young children using cochlear
implants can successfully learn nursery rhymes
and that learning is accelerated when the nursery
rhymes are combined with simple melodies. We
have observed that music training improves overall
vocal quality and speech fluency in these children.
Table 1. Observer Ratings of Children’s
Musical Performance and Learning
Variable
Mean
Std. Dev.
Enthusiasm
8.26
2.1
Overall Progress
8.65
1.7
Rhyme Concept
9.34
0.8
Melody Concept
8.04
2.1
Orff: Number of Lessons Learned
4.34
3.3
Other: Number of Lessons
Learned
2.04
2.1
Mistake made and Understood
1.91
2.1
Mistake made and not understood
1.30
1.4
Observers’ scoring for the different categories in which
learning and performance were defined.
Additional Reading
Abdi S, Khalessi MH, Khorsandi M, Gholami B. (2001) Introducing music
as a means of habilitation for children with cochlear implants. Int J Pediatr
Otorhinolaryngol 59(2):105-113.
Hodges AV, Dolan Ash M, Balkany TJ, Schloffman JJ, Butts SL. (1999) Speech
perception results in children with cochlear implants: contributing factors.
Otolaryngol Head Neck Surg 121(1): 31–34.
Leal MC, Shin YJ, Laborde ML, Calmels MN, Verges S, Lugardon S, Andrieu
S, Deguine O, Fraysse B. (2003) Music perception in adult cochlear implant
recipients. Acta Otolaryngol 123(7):826-835.
Koelsch S, Wittfoth M, Wolf A, Muller J, Hahne A. (2004) Music perception
in cochlear implant users: an event-related potential study. Clin Neurophysiol
115(4):966-972.
Stordahl J. (2002) Song recognition and appraisal: a comparison of children who
use cochlear implants and normally hearing children. J Music Ther 39(1):2-19.
Patient Assessment & Training
183
New Study Initiatives
Numerous studies have been initiated around the world that explore
the benefits of the HiResolution Bionic Ear System for speech
understanding and music appreciation. In addition, new implant
systems are being explored to help individuals who cannot benefit
from a conventional cochlear implant. Following are summaries of
some of these new investigations.
Multicenter Study in Asia: Direct-Connect Testing
with the CII/HiRes 90K Implants
PARTICIPATING INVESTIGATORS
Prof. De-min Han, Dr. Sha Liu
Tongren Hospital
Beijing, China
Dr. Xin Xi
General Hospital of Chinese People’s Liberation Army
Beijing, China
Prof. Sung-Kyun Moon
Ajou University School of Medicine
Suwon, Korea
Dr. Hong-Joon Park, Dr. Seung-Chul Lee,
Dr. Young-Myoung Chun, Hanah Lee,
Jee-Yeon Lee
Soree Ear Clinic, Soree Hearing Center
Seoul, Korea
Prof. T. Kubo, Dr. T. Sasaki, T. Iwaki
Osaka University Graduate School of Medicine
Osaka, Japan
M. Shiroma
International University of Health and Welfare
Tochigi, Japan
186
A direct-connect system has been developed by Soli
and colleagues (Soli et al, 2005) and is designed to
eliminate the need for a sound booth or a speaker
array, thereby allowing speech recognition and
localization tests to be administered quickly and
easily. The direct-connect system is based upon a
family of head-related transfer functions (HRTFs)
measured with KEMAR at source locations corresponding to loudspeaker positions appropriate for
unilateral or bilateral testing. Left-ear and right-ear
HRTFs appropriate to the selected source location are applied to the selected signal and presented
via direct connection to the auxiliary input of the
CII and Auria sound processors at specified levels.
Chinese, Japanese, and Korean HINT sentence
lists have been developed. This multicenter study
in Asia will evaluate methods for scoring sentence
intelligibility for use in cross-language comparisons of performance for cochlear implant recipients.
Reference
Soli S, Chan J, Vermiglio A, Freed D, Kessler DK. (2005) Assessment of spatial
unmasking and sound localization in bilateral implants using direct input. Poster
presentation at the Conference on Implantable Auditory Prostheses, Pacific Grove,
CA, 30 July–4 August, 2005.
Advanced Bionics® Auditory Research Bulletin 2005
Ongoing Cochlear Implant Studies:
Beijing Tongren Hospital, China
Beijing Tongren Hospital in China has over 100
years of history and specialization in Otorhinolaryngology. In 1996, the hospital established its cochlear implant program—among the
first in China. In 1997, Professor De-min Han
performed the country’s first pediatric surgery
with a multichannel device. The program now
has more than 600 cochlear implant recipients.
The cochlear implant team is composed of 12
experienced clinicians—including surgeons, audiologists, speech-language therapists, psychologists, and other professionals. The center has
conducted more than 80 cochlear implant
surgeries for patients with cochlear anomalies.
With the development of cochlear implant technology in China, a growing need arose to regulate
procedures including preimplant evaluations, surgery,
and postimplant programming and rehabilitation.
Tongren initiated the China Cochlear Implant Guidelines in cooperation with other hospitals in China.
Tongren is actively involved in cochlear implant
research including studies in objective measures—
EABR, NRI, and Banded NRI. Based on the
unique characteristics of Mandarin Chinese, the
center is developing speech and language test
materials for performance evaluation. The center
is also developing audiovisual rehabilitation tools
for cochlear implant recipients and their families.
New Study Initiatives
187
Bilateral Implant Benefit in Adults and Children
PARTICIPATING INVESTIGATORS
Vicente Rodríguez, M.D.
Fanny Munevar (Audiologist)
San Ignacio Hospital, Bogota, Columbia
This study will evaluate bilateral HiRes 90K benefit
in subjects implanted with two devices simultaneously. Adults and children will be evaluated
after implantation using age-appropriate speech
materials. Tests will include vowel and consonant discrimination, closed-set word identification (numbers, colors, animals, days of the week,
items of clothing), sentence recognition in quiet
and noise, Ling sounds, the Early Speech Perception (ESP) test, and tests of speechreading ability.
A Spanish version of the HINT will be used in
conjunction with the direct-connect system developed by Soli and colleagues (Soli et al, 2005).
Reference
Soli S, Chan J, Vermiglio A, Freed D, Kessler DK. (2005) Assessment of spatial
unmasking and sound localization in bilateral implants using direct input. Poster
presentation at the Conference on Implantable Auditory Prostheses, Pacific Grove,
CA, 30 July–4 August, 2005.
Multicenter Study in Colombia: HiRes 90K Benefit
in Children
PARTICIPATING INVESTIGATORS
Jose Antonio Rivas, M.D.
Adriana Rivas, M.D. (Audiologist)
Rivas Clinic, Bogota
Jorge E. Almario, M.D.
Jose Alberto Prieto, M.D
Maria Piedad Nuñez (Audiologist)
San Rafael Clinic, Bogota
Vicente Rodríguez, M.D.
Fanny Munevar (Audiologist)
San Ignacio Hospital, Bogota
188
This multicenter study will evaluate and compare
the benefits of the HiRes and conventional sound
processing in young children. Two groups of children (matched by age at implant, communication
mode, and educational philosophy) will be implanted
unilaterally with the HiResolution Bionic Ear.
Group 1 will be fit with conventional strategies (CIS
or MPS) and Group 2 will be fit with HiRes sound
processing. Auditory-skill development, speech
perception, and speech production will be evaluated
at 3, 6, 12, 18, and 24 months after implantation.
During the first month of programming, both groups
will have the option of using one of two strategies.
Group 1 will try CIS and MPS, and Group 2 will try
HiRes-P and HiRes-S. Each strategy will be used
for a period of 15 days to determine which strategy
offers better performance. When the best strategy
for each patient is determined, it will be used for a
period of two years. A between-subjects, repeatedmeasures analysis will compare benefit between
conventional and HiRes sound processing over time.
Advanced Bionics® Auditory Research Bulletin 2005
Radiographic and Behavioral Frequency-to-Place
Alignment in Cochlear Implant Subjects
PARTICIPATING INVESTIGATORS
Peter S. Roland, M.D.
C. Gary Wright, Ph.D.
Pamela Kruger, Au.D.
University of Texas Southwestern Medical Center
Dallas, TX, USA
This research study will examine the benefits of
frequency-adjusted programs that are created based
upon patients’ conventional program parameters,
music perception ability, and radiographic data
showing electrode position within the cochlea. Using
a within-subjects repeated-measures design, benefits
from the frequency-adjusted programs will be
compared to benefits from standard HiRes programs.
Development of an Inferior Colliculus Implant
PARTICIPATING INVESTIGATORS
Douglas C. Fitzpatrick, Ph.D.
Paul B. Manis, Ph.D.
Charles C. Finley, Ph.D.
Craig A. Buchman, M.D.
Harold C. Pillsbury III, M.D.
University of North Carolina, Chapel Hill, NC, USA
Gulam Emadi, Ph.D.
Michael Faltys, B.S.
Advanced Bionics Corporation, Valencia, CA, USA
This project will evaluate the feasibility of an
inferior colliculus (IC) implant. Using technology provided by Neuronexus Technologies and
Advanced Bionics, physiologic and psychoacoustic
studies of IC stimulation in an animal model will
determine basic parameters necessary for implementing IC stimulation in humans. An IC implant
may be indicated for hearing impaired individuals
who cannot derive therapeutic benefit from electrical stimulation within the cochlea. This population includes individuals with severe degeneration
of the spiral ganglion and patients who have had
their auditory nerve transected for tumor removal
(especially those with Neurofibromatosis type 2).
New Study Initiatives
189
Contributing Authors
Advanced Bionics gratefully acknowledges the following authors whose contributions to this research
bulletin appear on the pages listed.
A
Abdi, Susan 176, 182
Adelman, Cahtia 62
Adler, Miriam 62
Almario, Jorge E. 188
Armstrong, Shelley 34
Arnold, Laure 50, 68, 76, 164
B
Ballantyne, Deborah 36, 178
Ballay, Charles 108
Barbara, M. 60
Barrow, Kelly R. 80
Basta, Dietmar 70
Battmer, Rolf-Dieter 20, 66, 98, 156, 166
Bonnet, Raymond M. 94, 136
Bosco, Ersilia 36, 104, 178
Boyle, Patrick 68, 76, 164
Bracke, Peter 134
Brendel, Martina 66, 78, 122
Briaire, Jeroen J. 26, 84, 94, 136
Bromberg, Betsy 58
Brown, Carolyn J. 112
Buchman, Craig A. 112, 189
Büchner, Andreas 20, 66, 78, 98, 122, 156, 166
Buckler, Lisa 64, 152
C
Caner, Gül 50
Catterall, Nancy S. 172
Chan, Jenny C.Y. 144, 158
Chen, Haiming 128
Chen, Xue-qing 52, 106
Chénier, Josée 34
Chinnici, Jill 110
Chun, Young-Myoung 186
Collet, Lionel 68, 76
Couloigner, Vincent 164
190
Advanced Bionics® Auditory Research Bulletin 2005
D
D’Agosta, Luciana 104, 178
D’Elia, Chiara 36, 60, 104, 178
Dahme, Andreas 70
Dawson, Kristen 64, 152
De Clerck, Nora 28
Dinnesen, Antoinette G. 39
Donaldson, Gail S. 120
Dorman, Michael F. 114
Dunn, Camille C. 143
Dykmans, Philippe 134
E
Eddington, Donald K. 148
Eisen, Marc D. 82
Emadi, Gulam 124, 189
Ernst, Arne 70
Ezrati-Vinacour, Ruth 180
F
Faltys, Michael 189
Filipo, Roberto 36, 60, 104, 178
Finley, Charles C. 189
Firszt, Jill B. 72, 74
Fitzpatrick, Douglas C. 189
Fitzpatrick, Elizabeth 34
Franck, Kevin H. 82
François, Martine 164
Freed, Daniel J. 144, 158
Fridman, Gene 130
Frijns, Johan H.M. 26, 30, 84, 86, 94, 136
Frohne-Büchner, Carolin 20, 66, 78, 98, 122, 156,
166
G
Gantz, Bruce J. 143
Gärtner, Lutz 78, 156
Geleijns, Jakob 30
Guiraud, Jeanne 68, 76
—continued on next page
191
Contributing Authors—continued from previous page
H
M
Habermann, Corinna 122
Han, De-min 52, 106, 186
Hildesheimer, Minka 180
Hong, Robert S. 126, 128
Hughes, Michelle L. 80
Mancini, Patrizia 36, 60, 104, 178
Manis, Paul B. 189
Manolidis, Spiros 108
Marcinkevich, Paula B. 172
Marco, Jaime 57
Martinez, Paz 57
McDonald, Sonelle 133
McLaren, Sheena 133
Mertes, Jennifer 110, 174
Meyer, Ted A. 128
Mo, Ling-yan 52
Montey, Karen L. 138
Moon, Sung-Kyun 186
Morant, Antonio 57
Munevar, Fanny 188
I
Iwaki, T. 186
K
Kasulis, Heather 48
Kessler, Dorcas 144
Kishon-Rabin, Liat 180
Klop, W. Martin C. 94
Knief, Arne 39
Koch, Dawn Burton 173
Kong, Ying 52, 106
Kreft, Heather A. 120
Kretzmer, Erika A. 138
Kruger, Pamela 189
Kruger, Tracey 130
Kubo, T. 186
Kuzma, Janusz 26
L
Lake, Jennifer 150
Lee, Hanah 186
Lee, Jee-Yeon 186
Lee, Seung-Chul 186
Lenarz, Thomas 20, 66, 78, 98, 122, 156, 166
Lesinski-Schiedat, Anke 156, 166
Levi, Haya 62
Li, Yong-xin 52, 106
Litvak, Leonid 96, 120, 130
Liu, Bo 52
Liu, Sha 52, 186
Luetje, Charles 152
Luntz, Michal 146
192
N
Noel, Victor 148
Noël-Petroff, Nathalie 164
Novak, Michael A. 46, 149
Nuñez, Maria Piedad 188
Nunn, Terry 133
O
Offeciers, Erwin 28
Olgun, Levent 50
Olivier, Elrietha 160
Osberger, Mary Joe 102, 173
Overstreet, Edward 46, 54, 96
P
Pantev, Christo 39
Park, Hong-Joon 186
Peeters, Stefaan 28, 86, 133, 134, 136
Peters, B. Robert 150
Pillsbury III, Harold C. 189
Pitarch, Maria I. 57
Platero, Amparo 57
Polite, Colleen 162
Pongstaporn, Tan 138
Poon, Becky 148
Postnov, Andrei 28
Presley, Regina 170
Prieto, Jose Alberto 188
Advanced Bionics® Auditory Research Bulletin 2005
R
V
Rivas, Adriana 188
Rivas, Jose Antonio 188
Robbins, Amy McConkey 102
Rodríguez, Vicente 188
Roland, J. Thomas 24
Roland, Peter S. 22, 189
Ross, Bernhard 39
Rotz, Lee Ann 46
Rubinstein, Jay T. 126, 128
Runge-Samuelson, Christina L. 72, 74
Ryugo, David K. 138
Vanpoucke, Filiep J. 28, 86, 133, 134
van Buchem, Mark A. 30
Van den Abbeele, Thierry 164
Van Immerseel, Luc 134
Verbist, Berit M. 26, 30
Vermiglio, Andrew J. 144, 158
Voelkel, Andrew 96
Vujanovic, Irena 173
S
Sampson, Margaret 110
Sasaki, T. 186
Schramm, David R. 32, 34
Séguin, Christiane 34
Shapiro, William H. 58
Shiroma, M. 186
Shpak, Talma 146
Soli, Sigfrid D. 144, 158
Spahr, Anthony J. 114, 124
Stille, Lisa J. 80
Stinnett, Sandra S. 32
Stöver, Timo 78, 122
T
Taitelbaum-Swead, Riki 180
Tavakoli, Hassan 176, 182
Thomas, Jean 46
Todt, Ingo 70
Tomás, Manuel 57
Tonini, Ross 108
Traisci, Gabriella 104
Truy, Eric 68, 76
Tyler, Richard S. 143
W
Wackym, P. Ashley 72, 74
Wang, Liang 106
Weiss, Hadas 146
Wiener-Vacher, Sylvette 164
Willcox, Thomas O. 172
Witt, Shelley A. 143
Wolfe, Jace 48
Wollbrink, Andreas 39
Wright, C. Gary 22, 189
X
Xi, Xin 186
Z
Zarowski, Andrzej 26, 28
Zhao, Xiao-tian 52, 106
Zwolan, Teresa 54
U
Uhler, Kristin 162
Contributing Authors
193
Participating Research Centers
Advanced Bionics gratefully acknowledges the contributions of the following institutions to this research
bulletin. These research centers and clinics provided either a project summary, participated in a multicenter study, or are involved in an ongoing new study initiative.
ASIA
EUROPE
JAPAN
BELGIUM
Osaka University Graduate School of Medicine
Ghent University
Osaka
Ghent
International University of Health and Welfare
Medish Instituut St. Augustinus
Tochigi
Wilrijk
University of Antwerp
PEOPLE’S REPUBLIC OF CHINA
Tongren Hospital
Beijing
Antwerp
University of Leuven
Leuven
General Hospital of Chinese People’s
Liberation Army
FRANCE
Beijing
Avicenne Hospital
Bobigny
SOUTH KOREA
Soree Ear Clinic, Soree Hearing Center
Seoul
Ajou University School of Medicine
Suwon
Beaujon Hospitalier
Universitaire de Rouen
Rouen
Centre Hospitalier Universitaire de Toulouse
Toulouse
Gui de Chauliac Hospital
Montpellier
Hôpital Robert Debré
Paris
Hôpital Edouard Herriot
Lyon
Hôpital Charles Nicolle
Rouen
L’Ecole Normale Superieure
Paris
Pelegrin Hospital
Bordeaux
Purpan Hospital
Toulouse
Université Claude Bernard Lyon 1
Lyon
194
Advanced Bionics® Auditory Research Bulletin 2005
GERMANY
SPAIN
Hearing-Therapy-Center
Hospital Clinico Universitario Valencia
Potsdam
Medizinische Hochschule Hannover
Hannover
Unfallkrankenhaus Berlin
Berlin
Valencia
Hospital Son Dureta
Palma de Mallorca
Universitätskrankenhaus Eppendorf Hamburg
TURKEY
Hamburg
Hacettepe University
University of Münster
Ankara
Münster
SSK Izmir Hospital
Izmir
ITALY
Azienda Ospedaliera di Padova
UNITED KINGDOM
Padova
Guy’s and St. Thomas’ Hospital
Ospedale Civile di Rovereto
London
Rovereto
Royal National Throat, Nose & Ear
Hospital
Policlinico 1 Rome
Rome
London
U.O. Audiologia
The Ear Foundation
Ferrara
Nottingham
University of Rome La Sapienza
University College London
Rome
London
NETHERLANDS
Erasmus MC
Rotterdam
Leiden University Medical Center
Leiden
—continued on next page—
University Hospital Nijmegen
Nijmegen
SERBIA
Belgrade University
Belgrade
195
Participating Research Centers—continued from previous page
MIDDLE EAST
NORTH AMERICA
INDIA
CANADA
Desa’s Hospital
Mumbai
Alberta
Glenrose Rehabilitation Hospital
IRAN
Edmonton
Amir Alam Hospital
Tehran University of Medical Sciences
Manitoba
Tehran
Central Speech and Hearing Center
Winnipeg
ISRAEL
Québec
Bnai Zion Medical Center
L’ Hôtel-Dieu de Québec
Haifa
Chaim Sheba Medical Center
Québec City
Tel Hashomer
Ontario
Hadassah University Hospital
Children’s Hospital of Eastern Ontario
Jerusalem
Ottawa
Israel Institute of Technology
Ottawa Hospital (Civic Campus)
Haifa
Ottawa
Schneider Children’s Medical Centre
Rotman Research Institute
Baycrest Centre for Geriatric Care
Petah Tikva
Tel Aviv University
Tel Aviv
MOROCCO
Clinique Rachidi
Toronto
Sunnybrook & Women’s College
Health Sciences Center
Toronto
University of Ottawa
Ottawa
Casablanca
196
Advanced Bionics® Auditory Research Bulletin 2005
UNITED STATES
Arizona
Georgia
Arizona State University
Medical College of Georgia
Tempe
Mayo Clinic
Augusta
Scottsdale
Illinois
California
Carle Clinic Association
Carle Foundation Hospital
Let Them Hear Foundation
California Ear Institute
Urbana
Palo Alto
Indiana
House Ear Clinic
Communication Consulting Services
Los Angeles
Indianapolis
House Ear Institute
Indiana University
Los Angeles
Indianapolis
University of California
Los Angeles
Iowa
University of California
University of Iowa
San Francisco
Iowa City
Colorado
Maryland
Denver Ear Associates
Greater Baltimore Medical Center
Denver
Baltimore
University of Colorado Health Sciences Center
Johns Hopkins University
Aurora
Baltimore
Florida
Massachusetts
Tampa Bay Hearing & Balance Center
Massachusetts Eye and Ear Infirmary
Tampa
Boston
University of South Florida
Massacchusetts Institute of Technology
Tampa
Cambridge
University of Miami
New England Medical Center
Miami
Boston
University of Massachusetts
Amherst
—continued on next page—
Participating Research Centers
197
Participating Research Centers—continued from previous page
UNITED STATES (continued)
Michigan
North Carolina
Spectrum Health
Duke University
Grand Rapids
Durham
University of Michigan
University of North Carolina
Ann Arbor
Chapel Hill
Minnesota
Ohio
Mayo Clinic
Lippy Group for Ear, Nose and Throat
Rochester
Warren
University of Minnesota
University Hospitals of Cleveland
Minneapolis
Cleveland
Mississippi
Oklahoma
Jackson Ear Clinic
Integris Health
Jackson
Oklahoma City
Missouri
Oregon
Midwest Ear Institute
Oregon Health Sciences University
Kansas City
Portland
Washington University
St. Louis
Nebraska
Pennsylvania
The Children’s Hospital of Philadelphia
Philadelphia
Boys Town National Research Hospital
Thomas Jefferson University
Omaha
Philadelphia
New York
Beth Israel Medical Center
University of Pennsylvania
Philadelphia
New York
South Carolina
New York Presbyterian Hospital
Columbia University Medical Center
Medical University of South Carolina
New York
Charleston
New York University Medical Center
Tennessee
New York
Vanderbilt University
Nashville
198
Advanced Bionics® Auditory Research Bulletin 2005
SOUTH AMERICA
Texas
Baylor College of Medicine
Houston, Texas
Dallas Otolaryngology
Dallas
Ear Medical Group
San Antonio
COLOMBIA
Rivas Clinic
Bogota
San Rafael Clinic
Bogota
San Ignacio Hospital
Bogota
Houston Ear Research Foundation
Houston
Otology Group of San Antonio
San Antonio
Texas Children’s Hospital
Houston
University of Texas at Southwestern
Medical Center
Dallas
Virginia
Medical College of Virginia
Richmond
Washington
Spokane Ear, Nose & Throat Clinic
Spokane
University of Washington
Seattle
Wisconsin
Medical College of Wisconsin
Milwaukee
Participating Research Centers
199
Research Staff Worldwide
North America
Clinical Research
Mary Joe Osberger, Ph.D.
Global Director, Clinical Research
Cynthia J. Bergan, M.A.
Clinical Research Specialist
Jennifer Del Villar
Clinical Research Associate
Ann Kalberer, M.S.
Clinical Research Specialist
Dawn Burton Koch, Ph.D.
Senior Clinical Research Scientist
Elizabeth MacDonald, M.Ed.
Clinical Research Scientist
Edward H. Overstreet, Ph.D.
Senior Research Scientist
Sue Zimmerman-Phillips, M.S.
Field Manager, Auditory Clinical Studies
Research and Development
Mike Faltys, B.S.
Director, Systems Development
Gulam Emadi, Ph.D.
Research Scientist
Tracey Kruger, M.S.
Manager, Clinical Development and Product Assessment
Abhijit Kulkarni, Ph.D.
Staff Scientist
Leonid Litvak, Ph.D.
Principal Scientist
Lakshim Mirsha, M.S.
Principal Systems Engineer
Aniket Saoji, Ph.D.
Research Audiologist
Andrew Voelkel
Consultant, Systems Engineering
200
Advanced Bionics® Auditory Research Bulletin 2005
Europe
Clinical Research
Patrick Boyle, M.Sc.
Director, Clinical Research
Laure Arnold, M.Sc.
Clinical Research Engineer
Carolin Frohne-Büchner, Ph.D.
Clinical Research Scientist
Barbara Kienast, Ing.
Clinical Research Scientist
Fiona Robinson
Clinical Studies Coordinator
Deborah Vickers, Ph.D.
Manager, Clinical Studies
European Research Center
Antwerp, Belgium
Filiep Vanpoucke, Dr. Ir.
Head, European Research Center
Peter Bracke, M.Eng.
Software Engineer
Stefaan Peeters, Dr. Ir.
Consultant
Luc Van Immerseel, Dr. Ir.
Senior Research Engineer
Asia and Latin America
Clinical Research
Jianing Wei, Ph.D.
Manager, Clinical and Technical Services, Asia Pacific
Franco Portillo, Ph.D.
Manager, Technical and Scientific Affairs, Latin America
201
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement