Manual 21362137

Manual 21362137
University of Pretoria etd – Marais, J J (2004)
Transducer influence on Auditory Steady State Evoked
Potentials
BY
JACOBUS JOHANNES MARAIS
April 2004
IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE
M COMMUNICATION PATHOLOGY
IN THE DEPARTMENT OF COMMUNICATION
PATHOLOGY
FACULTY HUMANITIES
UNIVERSITY OF PRETORIA
I
University of Pretoria etd – Marais, J J (2004)
Acknowledgements
I would like to thank the following people and institutions for their contributions
during the course of this project:
1. Prof. Rene Hugo: for all the support and input.
2. Dr. Dunay Schmulian: for all the support, input and understanding.
3. Mr. De Wet Swanepoel: for all the support and input.
4. Dr. Odette Guy: for all the laughs and for being there when I need you.
5. The University of Pretoria: for an achievement bursary.
6. John Ho & Brett Robertson: for their financial contribution.
7. Michelle Pokorny: for all the support, as well as the financial contribution.
8. My Family: for love and support throughout, as well as your financial
contributions.
9. My Grandfather: for being the best role model I could hope for.
10. My friends: for friendship, love, support and good advise.
11. Ina Lombard: for all the love, support and for being there for me –always!
12. Riëtte Bosman; for making the journey a little more enjoyable. Thanks for your
support throughout this project.
To our Heavenly Father: for guidance, love and life!
II
University of Pretoria etd – Marais, J J (2004)
ABSTRACT
Title:
Transducer Influence on ASSEP
Name:
Jacobus Johannes Marais
Promoter:
Dunay Liesel Schmulian
Co-promoter:
De Wet Swanepoel
Department:
Communication Pathology
Degree:
M Communication Pathology
Preliminary studies have stirred the hope that sound-field stimulation through auditory
steady state evoked potentials can be used to assess aided thresholds in the difficultto-test population. Before the introduction of ASSEP into the clinical field, as a
technique for the prediction of aided thresholds in the difficult-to-test population, a
question arises concerning its clinical validation. The application of ASSEP through
sound field stimulation, in the determination of aided thresholds and for the evaluation
of amplification fittings, is dependent on the determination of unaided responses.
Subsequently the estimation of unaided thresholds in the hearing impaired population
is dependent on the establishment of normative data from the normal hearing
population.
The aim of this study was to determine the influence of insert earphones and sound
field speaker presentation on threshold estimations using monotic auditory steady
state evoked potentials, in a group of normal hearing adults. To achieve the aim of the
study, a comparative, within-group experimental design was selected. The results of
the current study indicated that the monotic single ASSEP technique under both insert
earphone- and sound field conditions provided a reasonable estimation (25-35 dB HL
for inset earphones; 20-33 dB HL for sound field speaker presentation) of the
behavioural pure tone thresholds. The minimum response levels obtained under insert
earphone conditions differed significantly from those obtained under sound field
conditions for all the frequencies tested except 2 kHz (p < 0.01). Subsequently, the
current study indicates that minimum response levels obtained using a specific
III
University of Pretoria etd – Marais, J J (2004)
transducer should serve as the basis of comparison with behavioural thresholds
obtained under the same transducer. Therefore, behavioural pure tone thresholds
obtained under insert earphone conditions will not suffice as a basis of comparison for
minimum response levels obtained for the ASSEP technique under sound field
conditions, and vice versa.
This research endeavour concluded that the monotic ASSEP technique under both
insert earphone and sound field conditions provide useful information for the
estimation of frequency specific thresholds, but that the results are transducer specific
and that comparison across transducers should be avoided.
Key terms: Objective audiometry, minimum response levels, transducers, stimulus
presentations, auditory steady state evoked potential, estimated pure tone thresholds,
sound field, insert earphones, sound field speakers
IV
University of Pretoria etd – Marais, J J (2004)
OPSOMMING
Titel:
Die Invloed van Omvormers op OSOP
Naam:
Jacobus Johannes Marais
Leier:
Dunay Liesel Schmulian
Mede Leier:
De Wet Swanepoel
Departement:
Kommunikasiepatologie
Graad:
M Kommunikasie Patologie
Voorlopige studies het die hoop laat ontstaan dat Ouditief Standhoudende Ontlokte
Potensiale (OSOP) in vrye-veld aanbieding gebruik sal kan word vir die evaluering
van versterkte drempels in die moeilik-toetsbare populasie. Voordat OSOP egter as ‘n
drempel bepalende tegniek vir die moeilik toetsbare populasie beskou kan word, moet
sekere vrae ten opsigte van die kliniese toepaslikheid beantwoord word. Die
toepaslikheid van OSOP in die vrye-veld toets omgewing, aangewend vir die
evaluasie van versterkte drempels en gehoor apparaatpassings, word bepaal deur die
tegniek se vermoë om onversterkte drempels akkuraat te bepaal. Gevolglik hou die
vermoë van OSOP om onversterkte drempels te bepaal verband met die vermoë om
gehoordrempels in normaalhorende individue te bepaal.
Die doel van hierdie studie was om te bepaal wat die invloed van insteek-oorfone en
vrye-veld aanbieding van stimuli is op die drempel bepalingsvermoë van monogotiese
OSOP in ‘n groep normaalhorende proefpersone. Om hierdie doel te bereik, is ‘n
vergelykende in-groep eksperimentele navorsingsontwerp geselekteer.
Die resultate van die huidige studie het getoon dat tydens die aanbieding van OSOP
stimuli, beide die insteek oorfone en vrye-veld aanbieding redelike akkurate
gedragsdrempels voorspel het (25-35 dB HL vir insteek oorfone; 20-33 dB HL vir
vrye-veld aanbieding). Die minimum response vlakke wat bepaal is tydens die
aanwending van insteek oorfone het beduidend verskil van die minimum respons
vlakke in die vrye-veld kondisies behalwe vir 2kHz (p < 0.01). Gevolglik dui die
huidige studie daarop dat minimum respons vlakke aangebied deur ‘n spesifieke
omvormer afhang van die gedragsdrempels bepaal deur dieselfde omvormer. Dus kan
V
University of Pretoria etd – Marais, J J (2004)
drempels wat gerkry is deur insteek oorfone nie dien as vergelykende basis vir
minimum respons vlakke verkry tydens vrye-veld aanbieding nie, en omgekeerd.
Die gevolgtrekking waartoe hierdie huidige studie gekom het, is dat monogotiese
OSOP tydens insteek oorfoon- en vrye-veld aanbieding bruikbare inligting verskaf tot
frekwensie spesifieke drempel bepaling, maar dat die vergelyking tussen omvormers
eerder vermy moet word.
Sleutel Terme: Objektiewe oudiometrie, minimum respons vlakke, omvormers,
stimulus aanbieding, ouditief standhoudende ontlokte potensiaal, geskatte suiwertoon
drempels, vrye-veld, insteek oorfone, vrye-veld luidsprekers
VI
University of Pretoria etd – Marais, J J (2004)
LIST OF CONTENTS
CHAPTER ONE: BACKGROUND AND RATIONALE OF THE STUDY .................................... 1
1.1 INTRODUCTION .................................................................................................................................... 1
1.2 PERSPECTIVES ON THRESHOLD ESTIMATION: DIAGNOSIS OF HEARING LOSS WITHIN THE DIFFICULTTO-TEST POPULATION............................................................................................................................... 3
1.3 PERSPECTIVES ON AMPLIFICATION FITTING: BRIDGING THE GAP BETWEEN DIAGNOSIS AND
REHABILITATION ....................................................................................................................................... 6
1.4 RATIONALE ......................................................................................................................................... 9
1.5 PROBLEM STATEMENT ...................................................................................................................... 11
1.6 DIVISION OF CHAPTERS ..................................................................................................................... 12
1.7 CONCEPTUAL ORIENTATION AND DESCRIPTION OF DEFINITIONS ...................................................... 13
1.8 SUMMARY ......................................................................................................................................... 15
CHAPTER TWO: PERPSECTIVES ON SOUND-FIELD AUDIOMETRY AND POSSIBLE
APPLICATION THEREOF WITHIN THE ASSEP DOMAIN ...................................................... 16
2.1 INTRODUCTION .................................................................................................................................. 16
2.2 PERSPECTIVES ON SOUND-FIELD AUDIOMETRY ................................................................................ 17
2.2.1 Application of Sound-field Audiometry ........................................................................................ 17
2.2.2 Advantages of Sound-field Audiometry ........................................................................................ 19
2.2.3 Limitations of Sound-field Audiometry ........................................................................................ 20
2.2.4 Recommendations when testing in the Sound-field....................................................................... 21
2.2.4.1 Optimal Stimulus Characteristics ............................................................................................... 23
2.2.4.2 Required Modulation Waveform and Modulation Rate for Frequency Modulated Tones ......... 24
2.2.4.3 Direct versus Reverberant-field Testing ..................................................................................... 26
2.2.4.4 Important Test Room Considerations ......................................................................................... 28
2.2.4.5 Importance of calibration………………………………………………………………………29
2.3 PERSPECTIVES ON AUDITORY STEADY STATE EVOKED POTENTIALS ................................................ 33
2.3.1 Definition of Auditory Steady State Evoked Potentials................................................................. 33
2.3.1.1 The Physiology underlying ASSEPs…………………………………………………………...34
2.3.1.2 The Anatomy underlying ASSEPs……………………………………………………………..35
2.3.2 Current Clinical Applications of ASSEP....................................................................................... 36
2.3.3 Research Applications ................................................................................................................... 38
2.3.4 Current Research Available........................................................................................................... 39
2.3.5 Audiometric Variables................................................................................................................... 42
2.3.6 Participant Related Variables ........................................................................................................ 45
2.3.6.1 Influence of Age ......................................................................................................................... 45
2.3.6.2 Influence of State of Consciousness ........................................................................................... 46
2.3.6.3 Influence of Gender .................................................................................................................... 47
2.3.6.4 Hearing Sensitivity ..................................................................................................................... 47
2.3.7 Calibration…………..…………………………………………………………………………….49
2.3.8 Possible Advantages of ASSEP over other Techniques in the Estimation of Aided Thresholds... 50
2.4 SOUND-FIELD AS A POSSIBLE APPLICATION WITHIN THE ASSEP DOMAIN........................................ 55
I
University of Pretoria etd – Marais, J J (2004)
2.4.1 Stimulus Characteristics ................................................................................................................ 56
2.4.2 Modulation Characteristics............................................................................................................ 56
2.4.3 Evoked Responses ......................................................................................................................... 58
2.4.4 Automated Response Detection..................................................................................................... 59
2.5 CONCLUSION ..................................................................................................................................... 60
2.6 SUMMARY ......................................................................................................................................... 62
CHAPTER THREE: RESEARCH METHODOLOGY ................................................................... 63
3.1 INTRODUCTION .................................................................................................................................. 63
3.2 AIMS OF RESEARCH ........................................................................................................................... 64
3.2.1 Main Aim ...................................................................................................................................... 64
3.2.2 Sub-aims ........................................................................................................................................ 64
3.3 RESEARCH DESIGN ............................................................................................................................ 65
3.4 VOLUNTEERS..................................................................................................................................... 70
3.4.1 SELECTION CRITERIA: .................................................................................................................... 70
3.4.1.1 Normal-hearing........................................................................................................................... 70
3.4.1.2 Normal Middle Ear Functioning................................................................................................. 70
3.4.1.3 Age Distribution ......................................................................................................................... 71
3.4.1.4 Gender Distribution .................................................................................................................... 71
3.4.2 Selection Procedures...................................................................................................................... 71
3.4.3 Selection Apparatus....................................................................................................................... 73
3.5 PARTICIPANTS ................................................................................................................................... 73
3.5.1 Description of Participants ............................................................................................................ 73
3.6 PRELIMINARY STUDY ........................................................................................................................ 74
3.6.1 Aim:............................................................................................................................................... 74
3.6.2 Procedures ..................................................................................................................................... 74
3.7 DATA COLLECTION ........................................................................................................................... 79
3.7.1 Data Collection Apparatus............................................................................................................. 79
3.7.2 Data Collection Procedures ........................................................................................................... 80
3.8 DATA ANALYSIS PROCEDURES.......................................................................................................... 85
3.9 DATA PROCESSING PROCEDURES ...................................................................................................... 88
3.10 SUMMARY ....................................................................................................................................... 88
CHAPTER FOUR: RESULTS AND DISCUSSIONS....................................................................... 89
4.1 INTRODUCTION .................................................................................................................................. 89
4.2 A comparison between pure tone-, frequency modulated (FM) and mixed modulated (AM/FM) tone
behavioural thresholds at 0.5, 1, 2 & 4 kHz under insert earphone and sound-field speaker conditions 91
4.2.1 A comparison across Stimulus Presentations: ............................................................................... 91
4.2.1.1 Thresholds obtained under Insert Earphone Conditions ............................................................. 91
4.2.1.2 Thresholds obtained under Sound-field Conditions ................................................................... 93
II
University of Pretoria etd – Marais, J J (2004)
4.2.2 Comparison across Transducers .................................................................................................... 94
4.2.2.1 Thresholds obtained through Behavioural Pure Tone and Behavioural FM Tone Testing......... 95
4.2.2.2 Thresholds obtained through Behavioural Mixed Modulated Tone Testing .............................. 96
4.2.3 Discussion of Behavioural Thresholds obtained, when compared across Stimulus Presentations
and Transducers...................................................................................................................................... 98
4.2.3.1 Discussion of the Comparison across Stimulus Presentation (controlled for Transducers)........ 99
4.2.3.2 Discussion of the Comparison across Transducers (controlled for Stimulus Presentation):..... 101
4.3 A comparison between minimum response level detection at 0.5, 1, 2 & 4 kHz using a monotic
ASSEP technique under insert earphone and sound field speaker conditions ...................................... 104
4.3.1 A Comparison across Stimulus Presentations: ............................................................................ 104
4.3.1.1 Thresholds an Minimum Response Levels obtained under Insert Earphone Conditions.......... 105
4.3.1.2 Thresholds and Minimum Response Levels obtained under Sound-field Conditions .............. 109
4.3.2 Comparison across transducers.................................................................................................... 112
4.3.2.1 ASSEP Minimum Response Levels obtained at 0.5, 1, 2 and 4 kHz, under Insert Earphone and
Sound-field Conditions......................................................................................................................... 112
4.3.3 Comparison across Stimulus Presentations and Transducers ...................................................... 114
4.3.4 Discussion of the Comparison across Stimulus Presentation, Transducer and a combination of
Stimulus Presentation and Transducer………………………………………………………………...112
4.4 A Comparison between the actual Behavioural Thresholds (for both pure tone and modulated tone
stimulation) obtained and the GSI AUDERA System’s accuracy in predicting these Thresholds by
means of an Algorithm, for Insert Earphone and Sound-field Conditions ........................................... 125
4.4.1 The accuracy of the GSI AUDERA System’s Algorithm in predicting Behavioural Thresholds
obtained under Insert Earphone Conditions.......................................................................................... 125
4.4.2 The accuracy of the GSI AUDERA System’s Algorithm in predicting Behavioural Thresholds
obtained under Sound-field Conditions ................................................................................................ 127
4.4.3 Discussion on the accuracy of the GSI AUDERA System’s Algorithm in predicting Behavioural
Thresholds obtained under both Insert Earphone and Sound-field Conditions .................................... 128
4.5 SUMMARY ....................................................................................................................................... 130
CHAPTER FIVE: CONCLUSIONS ................................................................................................ 131
5.1 INTRODUCTION ................................................................................................................................ 131
5.2 CONCLUSIONS: THEORETICAL AND CLINICAL IMPLICATIONS IDENTIFIED ....................................... 133
5.2.1 A comparison between pure tone-, frequency modulated (FM) and mixed modulated (AM/FM)
tone behavioural thresholds at 0.5, 1, 2 & 4 kHz under insert earphone and sound-field speaker
conditions ............................................................................................................................................. 133
5.2.2 A comparison between minimum response level detection at 0.5, 1, 2 & 4 kHz using a monotic
ASSEP technique under insert earphone and sound field speaker conditions ...................................... 137
5.2.3 A Comparison between the actual Behavioural thresholds (for both pure tone and modulated tone
stimulation) obtained and the GSI AUDERA System’s accuracy in predicting these thresholds by
means of an Algorithm, for Insert Earphone and Sound-field Conditions ........................................... 140
5.3 CRITICAL EVALUATION OF THE CURRENT STUDY ........................................................................... 141
5.4 RECOMMENDATIONS FOR FUTURE RESEARCH ................................................................................. 145
III
University of Pretoria etd – Marais, J J (2004)
5.5 FINAL COMMENTS ........................................................................................................................... 149
5.6 SUMMARY ....................................................................................................................................... 149
References
Appendices
LIST OF FIGURES
Chapter 2:
Figure 2.1 Illustration of the Electrical Activity involved in Evoked Responses
(Adapted from Picton, Dimitrijevic & John, 2001)…………………………………..53
Figure 2.2 Representation of Automated Response Detection using the F-Test
(Adapted from Lins et al., 1996)……………………………………………………..54
Chapter 3:
Figure 3.1 Chapter Outline...………………………………………………………...58
Figure 3.2 Relationship between Independent-, Dependent- and Controlled
Variables……………………………………………………………………………...64
Figure 3.3 Age Distribution of the Participants according to age intervals (in years)
and percentages………………………………………………………………………69
Figure 3.4 Plan of the test room showing the lay out of the room, the position of the
speakers as well as the test positions for behavioural and electro-physiological
measures……………………………………………………………………………...72
Chapter 4:
Figure 4.1 Research Process: The sub-aims and their relation to the main aim of the
research project……………………………………………………………………….84
Figure 4.2 Behavioural Mixed Modulated Tone Thresholds, obtained under Insert
Earphone and Sound-field conditions………………………………………………..91
Figure 4.3 Electro-physiological Data, obtained using the ASSEP Technique across
Transducers………………………………………………………………………….108
Figure 4.4 Mean Pure Tone, and FM Tone Thresholds and Mean ASSEP Minimum
Response Levels obtained using both Insert Earphone and Sound-field Speaker
Presentation…………………………………………………………………………109
Figure 4.5 Frequency Distribution of Minimum Response Levels (in dB SPL) under
Insert Earphone conditions………………………………………………………….111
IV
University of Pretoria etd – Marais, J J (2004)
Figure 4.6 Frequency Distribution of Minimum Response Levels (in dB SPL) under
Sound-field conditions……………………………………………………………...111
Figure 4.7 Mean Predicted Thresholds versus Actual Thresholds for insert earphone
presentation (top) and Sound-field Speaker Presentation (bottom) in dB SPL……..123
LIST OF TABLES
Chapter 1:
Table 1.1 Definition of terminology…………………………………………………14
Chapter 2:
Table 2.1 Limitations of Sound-field Audiometry and recommendations for
addressing these limitations…………………………………………………………..20
Table 2.2 Summary of the Limitations of Sound-field Audiometry and
Recommendations for addressing these Limitations…………………………………32
Table2.3 A Summary of the advantages of ASSEP over other techniques………….56
Chapter 3:
Table 3.1 Counterbalancing of Practice- and Carry-over effects……………………63
Table 3.2 Approximate time required for the assessment of a single participant……71
Table 3.3 Monotic ASSEP Recording Procedure……………………………………79
Table 3.4 Data Analysis Procedure and the Rationale for Analysis Method of
Choice………………………………………………………………………………...80
Table 3.5 Data preparation, -Analysis and the Rationale for Analysis Method of
Choice………………………………………………………………………………...81
Chapter 4:
Table 4.1 Mean and Standard Deviation Values for Behavioural Pure Tone and
Behavioural Mixed Modulated Tone thresholds under insert earphone conditions
(n = 50)…………………………………………………………………………..…...86
Table 4.2 Mean and Standard Deviation Values for Behavioural FM Tone and Mixed
Modulated Tone thresholds under sound-field conditions (n = 50)………………….87
Table 4.3 Mean and Standard Deviation Values for Behavioural Pure Tone and
Behavioural FM Tone thresholds under insert earphone and sound-field conditions
(n = 50)……………………………………………………………………………….89
V
University of Pretoria etd – Marais, J J (2004)
Table 4.4 Mean and Standard Deviation Values for Behavioural Mixed Modulated
Tone thresholds under insert earphone and sound-field conditions (n = 50)...………90
Table 4.5 Summary of the Mean and Standard Deviation values across Stimulus
Presentations and Transducers (n = 50)…………………………………………...…93
Table 4.6 Mean and Standard Deviation values for Behavioural Pure Tone thresholds
and ASSEP minimum response levels under insert earphone conditions (n = 50)…..99
Table 4.7 Mean and Standard Deviation values for Behavioural FM Tone thresholds
and ASSEP minimum response levels under sound-field conditions (n = 50)……..103
Table 4.8 Mean and Standard Deviation values for ASSEP minimum response levels
under insert earphone and sound-field conditions (n = 50)…………………………106
Table 4.9 Mean and Standard Deviation values for Behavioural Pure Tone thresholds,
Mixed Modulated Tone thresholds and the Predicted Thresholds under insert
earphone conditions (n = 50)………………………………………………………..120
Table 4.10 Mean and Standard Deviation values for Behavioural FM Tone thresholds,
Mixed Modulated Tone thresholds and the Predicted Thresholds under sound-field
conditions (n = 50)………………………………………………………………….121
VI
University of Pretoria etd – Marais, J J (2004)
LIST OF APPENDICES
Appendix A Approval letter from the Research and Ethics Committee
Appendix B Informed Consent for Adults
Appendix B Continued
Adult Consent Reply Slip
Appendix B Continued
Informed Consent for Minors
Appendix B Continued
Minor Consent Reply Slip
Appendix C Normative data: Tympanometry
Categorization: degree og hearing loss
Appendix D Nonparametric statistics
Appendix D Continued
Non Parametric Statistics
Appendix E Gender Effects
Ear Effects
Appendix F Gender Effects
Appendix G Ear Effects
Appendix H Comparison across Transducers: Wilcoxon Paired T-test
Appendix I
Gender Effects
Appendix J
Ear Effects
Appendix K Ear effects
Appendix L Correction for Earphone Testing
Appendix M Permissible ambient noise levels
VII
University of Pretoria etd – Marais, J J (2004)
LIST OF ABBREVIATIONS:
ABR(s)
- Auditory Brainstem Response(s)
AEP(s)
- Auditory Evoked Potential(s)
AM
- Amplitude Modulation
ANSI
- American National Standards Institute
ASSEP(s)
- Auditory Steady State Evoked Potential(s)
cm
- Centimeter
cm
3
- Cubic Centimeter
dB
- Decibel
DWTs
- Dampened Wavetrains
EEG
- Electro-Encephalo-Gram
EP
- Evoked Potential
ER
- Evoked Response
FFT
- Fast Fourier Transform
FM
- Frequency Modulated
HL
- Hearing Level
Hz
- Hertz
ISO
- International Standards Organization
kHz
- Kilo Hertz
m
- meter
MASTER
- Multiple Auditory Steady State Evoked Response
MRI
- Magnetic Resonance Imaging
MRL
- Minimum Response Level
ms
-Millisecond
mV
- Micro Volt
nHL
- Normal Hearing Level
NIH
- National Institutes of Health
OAE(s)
- Oto-Acoustic-Emission(s)
REIG
- Real Ear Insertion Gain
SABS
- South African Bureau of Standards
SD
- Standard Deviation
sec
- Second
VIII
University of Pretoria etd – Marais, J J (2004)
SPL
- Sound Pressure Level
SSEP(s)
- Steady State Evoked Potential(s)
SSR(s)
- Steady State Response(s)
Wave V
- Wave V of the Auditory Brainstem Response
IX
University of Pretoria etd – Marais, J J (2004)
CHAPTER ONE: BACKGROUND AND RATIONALE OF
THE STUDY
This chapter aims to introduce the problem this study
confronts, to provide the orientation, to describe the
terminology used, and to present an overview of the
content and organization of the study.
1.1 Introduction
Audiology, in itself, can be broadly defined by two interdependent components,
namely diagnosis1 (which consists of detection and differential diagnosis) and
rehabilitation2 (Roeser, Buckley & Stickney, 2000a). These two components cannot
be seen as two separate entities, as they function as a synergistic unit within the field
of audiology. According to Katz (1994) the primary goal of any diagnostic procedure
is rehabilitation, and successful rehabilitation of any auditory impairment is therefore
dependent on reliable and valid diagnosis of such impairment.
The logical starting point for any diagnosis, and subsequent rehabilitation, is early
identification (detection3) of hearing impairment. Researchers and clinicians alike
have proposed several methods such as auditory brainstem responses (Durieux-Smith,
Picton, Bernard, MacMurray & Goodman, 1991; Galambos, Wilson & Silva, 1994),
oto-acoustic emissions (Smurzynski, Jung, Lafreniere, Kim, Vasudeva-Kamath,
Rowe, Holman & Leonard, 1993) and, most recently, the auditory steady state evoked
potential (Rickards, Tan, Cohen, Wilson, Drew & Clark, 1994; Rance, Rickards, Beer
& Clark, 1998) to facilitate earlier identification. The identification of hearingimpaired infants within the first few months of life, as well as the subsequent
treatment of these infants, in effect, has created a larger difficult-to-test population.
1
“The distinctive characterization in precise terms of a genus…” (Oxford Dictionary, 1990:321). In
this case hearing impairment. Diagnosis can be subdivided into detection and differential diagnosis.
2
“Restore to effectiveness or normal life by training…” (Oxford Dictionary, 1990:1012).
Rehabilitation within the hearing-impaired population has its focus on the normalizing or minimizing
of problems associated with hearing loss.
3
For the aim of this study detection is seen as the identification of a hearing impairment, and does not
involve the characterization of the specific hearing loss (e.g. a sloping high frequency sensorineural
hearing loss).
1
University of Pretoria etd – Marais, J J (2004)
This implies greater responsibility on the part of the audiologist to ensure appropriate
intervention. The first of these responsibilities has its focus on the diagnosis4 of the
hearing impairment. For most people with a hearing impairment who are able to
comply with voluntary audiometry, behavioural response audiometry is the least
expensive and most definitive of all approaches in obtaining hearing thresholds. For
certain difficult-to-test populations, of which infants form a very large group, hearing
thresholds, can only be obtained using electro-physiological measures that do not
require any voluntary responses from the individual (Goldstein & Aldrich, 1999).
The initial treatment (habilitation) of the hearing-impaired individuals will involve
the selection and fitting of amplification devices (Picton, Durieux-Smith, Champagne,
Whittingham, Moran, Giguére & Beauregard, 1998), as well as the subsequent
adaptation programme, and aims to bridge the gap between diagnosis and
rehabilitation. Measures, such as functional gain, as well as objective measures, such
as real-ear probe measurements, have been developed and specifically adapted to
ensure appropriate amplification of these difficult-to-test individuals (Harford, 1980;
Hawkins & Haskell, 1982; Moodie, Seewald & Sinclair, 1994; Seewald, Ross, &
Spiro, 1985). In this population who cannot reliably respond to behavioural
audiometry the problem surrounding the appropriate selection and fitting of
amplification originates from the fact that both functional gain measures and real-ear
probe measures are reliant on accurate unaided thresholds (Picton et al., 1998;
Stelmachowicz & Lewis, 1988).
Research within the field of Audiology has, up to now, had its primary focus on two
of its responsibilities, namely diagnosis and rehabilitation. Thus, there is an
abundance of information regarding detection and differential diagnosis, as well as
rehabilitation. There is, however, a definite lack of information when it comes to the
interdependent nature between diagnosis and rehabilitation, and this process has been
described as being dependent on “a matter of luck and intuition” (Picton et al.,
1998:329).
4
For the purpose of this study diagnosis involves the characterization of a hearing impairment, which
has already been defined.
2
University of Pretoria etd – Marais, J J (2004)
Recent studies (Cone-Wesson, Parker, Rickards, Rance, Liburti-Persi, Poulis, May,
Tan & Pollard, 2001; Picton et al., 1998) have stirred the hope that Auditory Evoked
Potentials (AEPs), and specifically, Auditory Steady State Evoked Potentials
(ASSEPs) might have an application in bridging the gap between diagnosis and
rehabilitation (appropriate selection and fitting of hearing aids). Although its focus at
present has been the establishment of accurate and objective hearing thresholds as an
assessment tool in the fitting of hearing aids, ASSEP may also play a role in
demonstrating the function of hearing aids (Picton, Dimitrijevic, Van Roon et al.,
2002b). Picton et al (1998) demonstrated that ASSEPs can be recorded when multiple
stimuli are presented simultaneously through a sound-field speaker and amplified
using a hearing aid. Therefore this procedure seems more useful than those procedures
using transient stimuli. Transient stimuli, such as clicks, are more likely to be
distorted by either the sound-field speaker or the hearing aid amplifier (Picton, John &
Dimitrijevic, 2002a).
Because these potentials can be recorded down to near-threshold intensities in persons
with no or unreliable thresholds, whether it be behavioural or electro-physiological
(Picton et al., 2002a), it seems that ASSEPs can be used to demonstrate that the
hearing aid is working. However, its application is limited as ASSEPs cannot provide
information regarding how well the aid is working, seeing that most of the current
hearing aid technology employs non-linear amplification (Picton et al., 2002a). Also,
ASSEPs might be utilized to assess supra threshold hearing, as other measures
(within-the-canal measurements) does not provide information regarding what the
aided sound is perceived as, following processing in the brain (Picton et al., 2002a).
The next section of this chapter will focus on the application of ASSEPs as a method
to bridge the gap between diagnosis and rehabilitation.
1.2 Perspectives on Threshold Estimation: Diagnosis of Hearing Loss
within the Difficult-To-Test Population
A global trend towards earlier identification and intervention concerning infants with
hearing impairments has led to an increase in difficult-to-test populations. These
infants are unable to part take in voluntary behavioural audiometry and are easily
fatigued by conditioning techniques (Fulton & Lloyd, 1969). The importance of
3
University of Pretoria etd – Marais, J J (2004)
obtaining frequency specific hearing thresholds for these infants is not only limited to
accurate diagnosis, but also plays a critical role in the appropriate selection and fitting
of hearing aids.
The logical procedural follow-up on behavioural audiometry in co-operative patients
with hearing loss entails the selection and fitting of amplification devices. Appropriate
selection and fitting of amplification in persons who can respond to behavioural
audiometry, is achieved by incorporating the individual’s subjective responses (Picton
et al., 1998). Since behavioural testing of infants, very young children and
developmentally delayed children is unreliable, alternative assessment techniques had
to be developed in order to obtain reliable responses to auditory stimuli (Rance et al.,
1998).
Similarly, the process can be traced in the difficult-to-test population, with the
primary distinction being the techniques used to obtain information when the
conventional methods are inconclusive. Specific procedures have been developed to
evaluate the difficult-to-test population in order to ensure accurate diagnosis of
hearing thresholds. These procedures include oto-acoustic emissions (OAE), auditory
brainstem responses (ABR) and most recently the ASSEP. A short discussion of the
role of these techniques in differential diagnosis will subsequently follow.
Oto-acoustic emissions (White & Behrens, 1993) and auditory brainstem responses
(Hyde, Riko & Malizia, 1990; Durieux-Smith et al., 1991) provide complementary
information from which inferences regarding the hearing status of the infant can be
made. The presence of OAEs, as a response to sound, has shown to be consistent with
normal or near-normal-hearing, but is limited in its application because of its
sensitivity to middle ear pathologies and greater than mild to moderate hearing losses
(Rance et al., 1998).
The auditory brainstem response (ABR), as an evoked potential technique, appeared
to be the most useful instrument for estimating hearing level in very young children
(Picton, Durieux-Smith & Moran, 1994). Although a close relationship between
hearing level and click-ABR threshold level has been established (Gorga,
Whorthington, Reiland, Beauchaine & Goldar et al., 1985) the procedure has very
4
University of Pretoria etd – Marais, J J (2004)
definite shortcomings. Maximum intensity levels for stimulation are limited and the
threshold advantage, seen for pure tones, is not obtained (Rance et al., 1998). This
advantage is related to the fact that short duration stimuli does not allow for the
temporal integration seen for pure tones. This implies that the possibility of residual
hearing at profound levels cannot be investigated thoroughly. Furthermore, click-ABR
thresholds correlate most strongly with behavioural thresholds in the 1 kHz to 4 kHz
frequency range (Durieux-Smith et al., 1991) and as a result useful low frequency
thresholds are not detected (Rance et al., 1998).
Reasonably frequency specific stimuli, such as brief tone bursts or tone pips, have
been used in a number of studies to elicit the ABR (Stapells, Gravel & Martin, 1995;
Stapells, Picton, Durieux-Smith, Edwards & Moran, 1990). Although these measures
offer some insight pertaining to low frequency hearing functioning, they are still
limited in their clinical application due to restrictions in presentation levels (Rance et
al., 1998). Both these procedures are time consuming and at present the commercial
availability of ABR equipment does not allow for objective frequency specific
response detection (Stürzebecher, Cebulla & Pschirrer, 2001).
A recent addition to objective threshold estimation in young children has been the
auditory steady state evoked potential (Rance et al., 1998). These responses arise from
the scalp in response to regularly repeating stimuli such as sinusoidal amplitude
and/or frequency modulated tones (Picton et al., 1998; Rance et al 1998). The ASSEP
has several advantages to other auditory evoked potentials (AEP). It is reliably
recordable in sleeping or sedated individuals, when modulation rates of 70 - 110 Hz
are used, (Cohen, Rickards & Clark, 1991) and reliably present in children of all ages,
including neonates (Rickards et al., 1994). Detection of the ASSEP is objective in that
no interpretation is required from the tester.
Furthermore, the continuous nature of the ASSEP stimuli offers presentation level
advantages over other short duration stimuli. These advantages are related to the fact
that the modulated tones used during ASSEP testing are similar to warble tones used
in behavioural testing (Rance et al., 1998). This implies that the corrections associated
with tone burst and clicks are not required and the stimuli can, therefore, be presented
at levels that extend to 120 dB HL. The ASSEP technique is therefore able to
5
University of Pretoria etd – Marais, J J (2004)
investigate minimal amounts of residual hearing (Rance et al., 1998). Lastly, because
the threshold estimates obtained from ASSEP testing are frequency specific (Lins,
Picton, Boucher, Durieux-Smith, Champagne, Moran, Perez-Abalo, Martin, & Savio,
1996) it allows for testing across the audiometric range and for the generation of
evoked potential audiograms (Rance et al., 1998).
Although measures have been developed to accurately diagnose hearing levels in the
difficult-to-test population, previous attention regarding the measurement of aided
thresholds within this population has been limited to functional gain and real-ear
probe measures (Picton et al., 1998). Except for its dependency on accurate unaided
thresholds, real-ear probe measures, in neonates and others within the difficult-to-test
population can present certain challenges as far as probe tube placement and its
maintenance is concerned (Picton, 1998). These measures are extremely important
since the exact amount of amplification required is dependent on the unaided
thresholds and therefore on the degree, configuration and type of hearing loss
(Yoshinaga-Itano, 2001). Aided thresholds are used to verify the amplification fitting,
as well as for the performance evaluation of the amplification.
Behavioural audiometry incurs problems associated with inaccuracy and reliability
(Rance et al., 1998), in the detection and differential diagnosis of hearing impairment
within the difficult-to-test population. It logically follows then, that it will be
inconclusive in the determination of the initial aided thresholds for this population.
Rehabilitation therefore is dependent, as in the co-operative population, on measures
such as functional gain, insertion gain and real ear measurements.
1.3 Perspectives on Amplification Fitting: Bridging the gap between
Diagnosis and Rehabilitation
Audiology as a profession dates back as far as the Second World War. Its
responsibility in the prescription of appropriate amplification seems to have originated
around this time. The earliest attempts to prescribe amplification advocated by
Knudson and Jones (1935) did not involve evaluation of the amplification fitting; it
simply aimed to mirror the audiogram. Mirroring of the audiogram involved
matching the gain of the hearing aid, dB-for-dB, with the amount of the hearing loss
6
University of Pretoria etd – Marais, J J (2004)
(Punch, 2000). This amount of gain proved to be much more than persons with
sensorineural hearing loss could tolerate. The concept of listening comfort was
introduced when Watson & Knudson (1940) suggested providing a comfortable
amount of gain that produced equal loudness across all frequencies. According to
Lybarger (1944) the appropriate amount of gain need not be more than one half of the
hearing loss. His half-gain rule became the basis for many subsequent prescriptive
fitting formulas, several of which are still used today (Punch, 2000).
One method for evaluating amplification is functional gain. Measurement of the
amount of functional gain is calculated as the difference between aided and unaided
thresholds at each specific frequency obtained through free-field testing (Mueller,
Hawkins & Northern, 1992) and is defined as the relative decibel difference between
the aided and unaided thresholds. Many audiologists, during hearing aid selection and
fitting commonly use functional gain to evaluate whether (a) the amount of
amplification meets the requirements of the hearing loss characteristics, (b) to
demonstrate improved hearing and (c) informally to demonstrate the listening
advantages of amplification (Mueller et al., 1992). Although it supplies the
audiologist with critically important information regarding the rehabilitation process,
the functional gain hearing aid evaluations have certain limitations.
Since this technique is based on voluntary behavioural procedures, the inherent degree
of variability to which behavioural threshold measurements are subjected (Kidd,
2002) will also influence functional gain measurements (Mueller et al., 1992). As this
technique is conducted in a sound-field situation, careful attention should be paid to
accurate calibration of stimuli to limit the influence of standing waves (and masking
of the non-test ear). It is, however, important to realize that functional gain can only
be calculated once the unaided audiometric thresholds are known (Picton et al., 1998)
and that without this data the hearing aid cannot be adjusted to match the prescriptive
targets (Stelmachowicz & Lewis, 1988).
Since functional gain measures are usually made at octave levels, inter-octave peaks
and troughs are easily overlooked (Northern, 1992). Thus, the technique does not
supply the audiologist with sufficiently comprehensive frequency specific
information. Furthermore, small changes in the electro-acoustic output of the hearing
7
University of Pretoria etd – Marais, J J (2004)
aid, or acoustic modifications created by the manipulation of the acoustic coupling
system may create alterations in the frequency response and gain characteristics of the
hearing aid that will not be noted in the functional gain measurement.
The preferred procedure introduced to verify hearing aid fitting was real-ear probemicrophone
measurements
or
real-ear
insertion
gain.
Sound
pressure
measurements are taken with, and without, the fitted hearing aid in place. Insertion
gain is then determined as the difference, in decibels, between the two response
curves. Real-ear measurements reveal not only those aspects of hearing aid
performance that require electro-acoustic adjustments, but also the immediate effects
of those adjustments (Punch, 2000). The ability of real-ear equipment to reveal output
or gain at fine frequency gradations is one of its greatest advantages. The same
information cannot be found, for example, with measures of functional gain, which
are made at selected audiometric frequencies. Once again advances in technology
provided a considerable advantage over previous efforts, but as Hawkins (1987)
noted, the use of real-ear measures alone does not result in a better hearing aid fitting.
Within a clinical set-up, and with specific reference to the neonatal population,
clinicians experience difficulty with probe tube placement (Picton et al., 1998). It is,
furthermore, important to note that real-ear measurements are electro-acoustic
measures, and only provides information regarding the output of the hearing aid at the
eardrum. These measures provide no information regarding neuro-physiological
functioning of the auditory system (Picton et al., 2002a).
Hearing aid selection and fitting have undergone a systematic evolution in the past
five decades. The technological sophistication of hearing aids continues to grow
rapidly. In many ways, these advances in technology have outpaced our ability to fit
hearing aids to individual needs. Increasingly sophisticated fitting methods are needed
to close the gap (Punch, 2000). The need for a technique that offers accurate,
objective estimation of frequency specific hearing thresholds within this population,
who in most instances cannot comply with supra-aural headphone or insert earphone
testing (Picton et al., 1998), becomes critical for the effective rehabilitation through
amplification.
8
University of Pretoria etd – Marais, J J (2004)
Auditory evoked potentials (AEP) have been used in the past to assess the functioning
of hearing aids. The technique that received most attention in the evaluation of
hearing aids was the ABR (Mahoney, 1985; Beauchaine & Gorga, 1988). This
procedure involved the adjustment of the hearing aid until the latency of wave V of
the ABR has decreased to within normal limits (Picton et al., 1998). The brief nature
of stimulation led to various technical problems. Due to the specific stimulation
characteristics the signal can undergo distortion in both the sound-field speaker and
the hearing aid. This will result in stimulus artifacts that can make accurate diagnosis
problematic (Hall & Ruth, 1985). The most significant limitation concerning this
technique stems from the fact that hearing aids react differently to rapidly changing
stimuli than to more continuous stimuli (Beauchaine & Gorga, 1988; Mahoney, 1985;
Picton et al., 1998). The ABR technique only supplies information regarding the
rapidly changing stimuli and no information regarding more continuous tones such as
speech stimuli.
Interest in ASSEP (Galambos, Makeig & Talmachoff, 1981; Rickards & Clark, 1984;
Rodriguez, Picton, Linden, Hamel & Laframboise, 1986; Stapells, Linden, Suffield,
Hamel & Picton, 1984) has led to the identification of several advantages over
transient evoked potentials such as the ABR. While the ASSEP has experienced
limited exploration concerning its application in terms of amplification (initial
habilitation), as was the case initially with ABR, these limitations are now being
addressed through research.
The advantages of the ASSEP technique form the
rationale of this study. Its application within early management and aided threshold
testing, forms the impetus of the current study, and will be explored in more detail in
the following section.
1.4 Rationale
In an attempt to clarify the rationale behind the current research endeavour, the
following section aims to delineate the suitability of the ASSEP technique in the
exploration of aided thresholds. Steady state evoked potentials (SSEP) have long been
used in the evaluation of visually evoked potentials (Regan, 1989). More recently
these SSEPs have been used more extensively to record auditory steady state evoked
potentials (Aoyagi et al., 1994; Aoyagi et al., 1996; Azzena et al., 1995; Galambos &
Makeig, 1998; John et al., 1998; Lins & Picton, 1995; Lins et al., 1995; Lins et al.,
9
University of Pretoria etd – Marais, J J (2004)
1996; Pantev et al., 1996; Picton et al., 1998; Rance et al., 1998; Rance et al., 1995;
Suzuki et al., 1994; Rickards et al., 1994. ASSEPs occur when the frequency
constituents of a response are stable in amplitude and phase (Regan, 1989). These
responses are usually elicited using periodic stimuli and measured at the frequency of
stimulation or one of its harmonics (John, Dimitrijevic & Picton, 2001).
Recently ASSEP have become available as a clinical objective hearing test option
(Rance et al., 1998). Several researchers have denoted the current applications of the
ASSEP technique within the scope of audiology. The technique has been shown to be
useful in the estimation of hearing thresholds with normal and impaired hearing
(Aoyagi et al., 1994; Aoyagi, Suzuki, Yokota, Furuse, Watanabe & Itoh, 1999; Lins,
Picton, Picton, Champagne & Durieux-Smith, 1995; Lins & Picton, 1995; Lins et al.,
1996; Picton et al., 1998; Rance et al., 1995; Rance et al., 1998; Rickards et al., 1994).
It has been reliably recorded in sleeping or awake infants, as well as in sleeping adults
(Aoyagi, Yoshinori, Suzuki, Fuse & Koike, 1993; Maurizi, Almadori, Paludetti,
Ottaviani, Rosignoli & Luciano, 1990; Stapells, Galambos, Costello & Makeig, 1988;
Suzuki & Kobayashi, 1984). The ASSEP technique has furthermore been identified as
a reliable tool for newborn hearing screening (Aoyagi et al., 1994; Lins et al., 1996;
Picton et al., 1998; Rance et al., 1995), as it provides reliable results in infants with
normal and impaired hearing. The introduction of the multiple auditory steady state
evoked response (MASTER) technique, with multiple carrier frequencies, each
modulated by its own signature modulation frequency, evokes multiple steady state
responses and may significantly reduce test-time (John et al., 1998).
Studies on the use of amplitude-modulated tones to stimulate responses (Lins et al.,
1996; Hood, 1998) have highlighted significant advantages over other objective
techniques. The computer program interprets the presence of responses. ASSEP
equipment has utilized averaging techniques that make the recording of responses,
more objective than that of the ABR for example (Stürzebecher et al., 2001). In the
auditory system these responses can be recorded using amplitude modulated tones that
are frequency specific and stable over a period of time (Picton et al., 1998). These
responses are consequently less likely to be distorted during sound-field stimulation
(Picton et al., 2002a).
10
University of Pretoria etd – Marais, J J (2004)
The considerable advantages of the ASSEP technique over other AEP techniques
during the detection and differential diagnosis of hearing impairment have stirred the
hope that it will gain clinical application within the scope of aided threshold detection
and evaluation of amplification fittings. Pilot and preliminary studies (Cone-Wesson
et al., 2001; Picton et al., 1998) noted the unlikely distortion of the stimuli by both the
sound-field speaker and the hearing aid itself and were able to record aided ASSEPs
at near threshold intensities (Picton et al., 1998). It is, however, important to validate
preliminary findings (Cone-Wesson et al., 2001; Picton et al. 1998) concerning the
application value of ASSEP, using sound-field presentation, in the estimation of
frequency specific aided thresholds and the evaluation of amplification fittings.
1.5 Problem Statement
In an attempt to determine the validity of any audiometric procedure, it is important to
establish the procedure’s ability to perform as intended (Roeser, Valente & HosfordDunn, 2000b). In the case of sound-field stimulation by means of auditory steady state
evoked potentials (ASSEP) it is necessary to determine whether thresholds can be
estimated and if so how they compare with the data obtained from the same procedure
using insert earphones. Thus a need arises for normative studies to validate soundfield presentation results.
Preliminary studies (Picton et al., 1998) and pilot studies (Cone-Wesson et al., 2001)
have stirred the hope that sound-field stimulation through auditory steady state evoked
potentials can be used to assess aided thresholds in the difficult-to-test population.
Validating these results through further research will establish a frequency specific,
objective procedure for estimating aided thresholds in difficult-to-test populations.
Before the introduction of ASSEP into the clinical field, as a technique for the
prediction of aided thresholds in the difficult-to-test population, a question arises
concerning its clinical validation. Reliability and validity is paramount to the
establishment of the ASSEP technique within the clinical arena. In other words, the
application of ASSEP through sound-field stimulation, in the determination of aided
thresholds and for the evaluation of amplification fittings, is dependent on the
determination of unaided responses. Subsequently the estimation of unaided
11
University of Pretoria etd – Marais, J J (2004)
thresholds in the hearing-impaired population is dependent on the estimation of
thresholds within the normal-hearing population. The question now arises:
How accurate is monotic auditory steady state evoked potentials in the prediction
of thresholds using sound-field stimulation, when compared to insert earphone
stimulation, in a group of normal-hearing adults?
1.6 Division of Chapters
In order to answer this question a research endeavour consisting of both an empirical
and theoretical component was conducted. In order to delineate the process behind the
research endeavour, the following section will delineate the division of chapters and
will furthermore provide a short summary of the contents of each chapter.
Chapter one: Orientation and Statement of the Problem
This chapter provides an overview of the importance of diagnostic procedures and
rehabilitation, as well as the interdependent nature of these two components. It
describes the responsibilities faced by the global trend towards earlier identification,
as the initial step of rehabilitation, and subsequent evaluation of aided thresholds in
hearing-impaired infants. It introduces the ASSEP as a technique for the estimation of
aided thresholds in the difficult-to-test population and delineates the purpose of this
study: to determine the clinical application value of the monotic ASSEP technique in
the estimation of aided thresholds and the evaluation of amplification fitting, through
sound-field presentation.
Chapter two: Theoretical perspectives on Auditory Steady State Evoked
Potentials using Sound-field presentation
This chapter is divided into three sections, namely Perspectives on Sound-field
Audiometry, Perspectives on ASSEPs and Sound-field Testing as a possible
application within the ASSEP domain. It will discuss sound-field audiometry, its
applications and limitations and recommendations for valid and reliable testing under
sound-field conditions. Further discussion concentrates on the current clinical
applications of ASSEP and research applications. The existing measures available for
the estimation of aided thresholds, as well as their limitations are discussed. The
critical evaluation of these procedures serves as the background to the discussion of
12
University of Pretoria etd – Marais, J J (2004)
the monotic ASSEP. Thorough theoretical and clinical advantages of how the ASSEP
might address the limitations of existing procedures is provided along with the effect
of sound-field stimulation and the results of other preliminary studies.
Chapter Three: Research Methodology
This chapter provides the operational framework implemented to conduct the
empirical research. This framework dictates the scientific process implemented to
determine the clinical validity of the monotic ASSEP to estimate thresholds through
sound-field presentation.
Chapter Four: Results and Discussion
This chapter presents the results obtained through statistical analysis. The results are
presented according to the sub-aims stipulated in chapter three. An interpretation and
discussion of its value in relation to the literature follows each result.
Chapter Five: Conclusions and Implications
This chapter summarizes the results obtained and provides an outline of the
significant results and the extent to which they contribute to current literature. Future
research recommendations are provided and a conclusion regarding the current study
is formulated.
1.7 Conceptual Orientation and Description of Definitions
Table 1.1 provides definitions for terminologies that are used throughout the empirical
component of this research endeavour.
13
University of Pretoria etd – Marais, J J (2004)
Table 1.1 Definition of Terminology
Stimulus
EP
Signal
• Stimulus: A Stimulus is that which excites or
produces a temporary increase of vital action,
either in the whole organism or in any of its
parts; especially any substance or agent capable
of evoking the activity of a nerve, or capable of
producing an impression upon a sensory organ
or more specifically upon its end organ (Oxford
Dictionary, 1990).
• Evoked Potential (EP): “Coming into action…”
(Oxford Dictionary, 1990:932). Within the
auditory system, potentials consist of alternating
positive and negative waves (electrical current)
that were evoked by a stimulation of the
auditory system.
• Signal: “A prearranged sign conveying
information…” (Oxford Dictionary, 1990:1129).
Within the human body cells communicate with
one another by signals. These signals include
small organic molecules (e.g. adrenaline), larger
molecules (e.g. proteins), and electrical- and
chemical signals. Within the setting of the
current study, namely auditory evoked
potentials, the focus will be on electrical signals.
• Minimum Response Level (MRL): The
minimum level at which a response is present. In
other words the intensity where the last
significant response was recorded (Herdman &
Stapells, 2001).
• Evoked Response (ER): An evoked response is
the recorded average of various signals.
• Threshold: “The minimum effective sound
pressure level of an acoustic signal producing an
auditory sensation within a specified section of
clinical trails (ANSI, 1986), is defined as the
threshold of audibility” (Yantis, 1994:97a)
EP
ER
MRL
Algorithm
Threshold
14
University of Pretoria etd – Marais, J J (2004)
1.8 Summary
This chapter aimed to provide relevant background information to introduce the focus
of the current research endeavour and to provide insight on the importance of the
rationale underlying the study. The importance of assessing auditory thresholds within
the difficult-to-test population and especially with regard to infants, who cannot
comply with behavioural test measures, was highlighted in this chapter. The influence
of insert earphones and sound-field stimulation on monotic auditory steady state
evoked potentials, in normal-hearing adults, as a first step towards validating the
estimation of aided thresholds through sound-field stimulation was discussed. The
discussion was based on a comparison with behavioural pure tone audiometry, which
forms the gold standard within the clinical audiometric test arena. Attention was
focused on the need for an objective test measure with the potential of bridging the
gap between diagnosis and rehabilitation.
Other objective techniques that have been developed to estimate hearing thresholds
within the difficult-to-test have been discussed. These included electro-physiological
measures such as OAEs, ABRs and most recently ASSEPs, as well as electro-acoustic
measures such as functional gain and insertion gain. Known advantages of ASSEPs
were highlighted in relation to these other objective techniques, with regard to insert
earphone stimulation. Finally, results of preliminary and pilot studies, with regard to
sound-field stimulation and its possible application within aided threshold
determination were emphasized in the light of limited research validation. Therefore,
the need to determine the influence of insert earphones and sound-field stimulation on
monotic auditory steady state evoked potentials was made apparent.
15
University of Pretoria etd – Marais, J J (2004)
CHAPTER TWO: PERPSECTIVES ON SOUND-FIELD
AUDIOMETRY AND POSSIBLE APPLICATION
THEREOF WITHIN THE ASSEP DOMAIN
This chapter aims to provide a theoretical framework as a support for the
empirical research component. It aims to specify concepts and constructs in
such a way that controllable observations can be made about them. It aims,
furthermore to provide a critical evaluation and interpretation of the relevant
literature, so as to identify the empirical referents of theoretical terminology.
2.1 Introduction
This chapter explores the suitability of Auditory Steady State Evoked Potentials
(ASSEPs) through sound-field presentation, as an objective procedure in the diagnosis
of hearing impairments. It is divided into three main sections, namely perspectives on
Sound-field audiometry, perspectives on ASSEP and lastly Sound-field presentation
as a possible application within the ASSEP domain.
In the first section, Perspectives on Sound-field Audiometry, the current application
of sound-field audiometry is discussed, in an attempt to delineate its scope within the
field of audiology. A discussion on the advantages and limitations that apply to
sound-field audiometry will introduce several comprehensive recommendations aimed
at addressing the optimal characteristics for accurate threshold detection. Focus will
be on stimulus related effects, optimal characteristics of such a stimulus, the relevant
merits of testing in the direct field versus the reverberant field, important test room
characteristics, importance of calibration, other important testing aspects and test
room arrangements. This section concludes with a summary of the above-mentioned
recommendations.
The second section of this chapter, Perspectives on ASSEP, aims to provide a logical
continuation of the first. Once again, the current application of ASSEP will introduce
this section. The physiological basis underlying ASSEP is discussed, in order to
facilitate a functional understanding of the ASSEP definition. The audiometric
16
University of Pretoria etd – Marais, J J (2004)
variables associated with evoked potential audiometry are highlighted with specific
reference to ASSEP, in an attempt to postulate recommendations, based on available
literature, relating to ASSEP testing.
The last section of this chapter, Sound-field Presentation as a Possible Application
within the ASSEP Domain, has its focus on stimulation and modulation
characteristics, evoked responses and automated response detection. This section,
furthermore, aims to delineate the possibilities of sound-field presentation within the
ASSEP domain, as a means for objectively assessing aided thresholds and will
conclude with a justification of the research question proposed in chapter one of this
study.
2.2 Perspectives on Sound-field Audiometry
This section is discussed according to the application, advantages and limitations of
sound-field audiometry. This section concludes with recommendations, identified
from the literature, for conducting sound-field audiometry including stimulus related
effects; optimal stimulus characteristics; required modulation waveform and rate for
frequency modulated tones; direct versus reverberant field testing; important test
room characteristics and the importance of calibration. The section will conclude with
a summary of the recommendations that have been discussed (Table 2.3).
2.2.1 Application of Sound-field Audiometry
Selecting the most appropriate transducer during audiometric testing depends entirely
on what the clinician or researcher wants to measure (Wilber, 2002). To ensure the
appropriate selection of an air conduction transducer, certain factors pertaining to
each transducer must be kept in mind. Initially, supra aural headphones appeared
(until recently) to be the most popular transducer for routine pure tone testing
(Arlinger & Jerlvall, 1987; Wilber 2002). The use of supra-aural headphones,
however, have certain limitations, which include narrow frequency responses, leakage
of sound, the possibility of collapsing ear canals, inability of reproducing extremely
short duration stimuli accurately, contribution to reduced inter-aural attenuation and
creating an occlusion effect (Wilber, 2002). Currently, the recommended transducer
for routine behavioural testing, namely insert earphones, show the same narrow
frequency response, but in comparison have less problems associated with sound
17
University of Pretoria etd – Marais, J J (2004)
leakage, collapsing ear canals, occlusion problems and reduced inter-aural attenuation
is significantly decreased (Wilber, 2002). However, both insert earphones and supraaural headphones have certain practical disadvantages within the difficult-to-test
population, and specifically pertaining to the testing of infants and very young
children.
Earphone placement, and maintenance of the placement, presents challenges when
testing infants, very young children and uncooperative patients (Picton et al., 1998).
Furthermore, some children are afraid of supra-aural headphones and/or insert
earphones and will not tolerate them (Magnusson, Börjesson & Axelsson, 1997). As
time is a constraining factor when testing infants and small children, it is very
important to minimize all distractions that might have an influence on test time
(Magnusson et al., 1997).
Due to these difficulties, the emphasis now shifts towards sound-field presentation.
Although diagnostic audiometry is, in typical conditions, associated with earphones
(either supra aural or insert earphones) as the transducer6 of choice (Arlinger &
Jerlvall, 1987), there are certain situations where it is impossible, or highly
inconvenient, to use earphones (Wilber, 2002). The rationale for using earphones as
the transducer of choice when performing audiometric test procedures is based on the
fact that it eliminates all influence of the acoustic properties of the test room on the
stimulus. When earphones cannot be used, however, the interaction between the test
stimulus and the test room becomes critically important (Arlinger & Jerlvall, 1987).
These situations include hearing assessments in the pediatric and mentally
handicapped populations where earphone testing is inappropriate, as it might lead to
distraction or fear (Arlinger & Jerlvall, 1987; Beynon & Munro, 1995), and in some
cases the individual simply won’t tolerate them (Magnusson et al., 1997). Sound-field
speaker presentation is also used when the participant will not tolerate earphones,
because of fitting problems such as unusually large head size (Wilber, 2002).
Furthermore, it is also used when functional gain measures of hearing aids are made
(Arlinger & Jerlvall, 1987; Beynon & Munro, 1995; Wilber, 2002) and lastly the
6
“Simply put a transducer is a device that converts input energy into output energy (or conversely
output energy into input energy)” (Wilber, 2002)
18
University of Pretoria etd – Marais, J J (2004)
evaluation of the aided thresholds (Arlinger & Jerlvall, 1987; Beynon & Munro, 1995;
Walker, Dillon & Byrne, 1984). Subsequently, sound-field audiometry becomes
viable when one cannot obtain measurements using earphones.
Sound-field audiometry, in itself, incurs many problems primarily because it is
impractical to keep the position of the participant’s ear absolutely constant. Even
when testing a co-operative adult participant, it is difficult to be certain that the
participant’s ear remains in exactly the same position, assumed during calibration,
throughout the test. The rationale behind maintaining the test position at all times
stems from the fact that the sound pressure level of the stimulus varies considerably
with distance from the speaker (Walker et al., 1984). In other words, in order to
ensure reliable threshold estimations, the participant’s head has to remain as close to
the test position assumed during calibration as possible. This problem is much greater
in the difficult-to-test population, due to difficulties in restraining movement without
jeopardizing co-operation or inhibiting capacity to respond naturally (Walker et al.,
1984).
2.2.2 Advantages of Sound-field Audiometry
Behavioural sound-field audiometry has been criticized as being an invalid and
unreliable means of determining the hearing status in the difficult-to-test population
and especially with reference to young children (Berlin & Hood 1993). However,
awareness of the pitfalls, as well as the use of procedures that increase validity and
reliability, makes the use of sound-field audiometry an efficient method for assessing
hearing status within the clinical arena (Diefendorf & Gravel, 1996; Renshaw &
Diefendorf, 1998). These “pitfalls” and how they have been addressed will be
discussed under limitations of sound-field audiometry in 2.2.3 of this section.
There are two main advantages relating to the use of sound-field audiometry. Firstly,
it provides a means of assessing hearing sensitivity for the difficult-to-test populations
such as infants, very young children, mentally handicapped and others within the
difficult-to-test population, when other behavioural techniques have failed (Dillon &
Walker 1981; Magnusson et al., 1997; Walker et al., 1984). It also provides a measure
to assess specific tasks such as hearing aid evaluations, and specifically the aided
thresholds of amplification users, so as to determine functional gain for these clients
19
University of Pretoria etd – Marais, J J (2004)
(Arlinger & Jerlvall, 1987; Beynon & Munro, 1995; Walker et al., 1984; Wilber,
2002).
2.2.3 Limitations of Sound-field Audiometry
Summarized in Table 2.1 below, are the main limitations of sound-field audiometry
and how these limitations have been addressed.
Table 2.1 Limitations of Sound-field Audiometry and recommendations for
addressing these limitations
Limitations
The interaction
between test
stimuli, the
participant
interface and the
test room
acoustics influence
the accurate
determination of
thresholds.
No ear specific
information can
be obtained.
How it has been addressed
Stimuli identified to limit the influence of interaction between test
stimuli, and the participant interface and test room acoustics:
Use of Complex Stimuli
Use of appropriate modulation of the stimulus
Use of appropriate bandwidth of the modulated stimulus
(Dillon & Walker, 1982a; Dillon & Walker, 1982b; Walker & Dillon,
1983; Walker et al., 1984)
Participant characteristics identified to limit the influence of the
interaction between the participant interface, and the test stimuli and
test room acoustics:
Limiting participant movement during threshold determination
Participant placement in relation to the speaker
Direct field placement versus reverberant field placement
(Walker et al., 1984)
Test room characteristics identified to limit the influence of
interaction between the test room acoustics, and the test stimuli and
participant interface:
Absorptivity of the rooms’ surfaces
Placement of the speaker
Dimensions of the test room
Suitable seating arrangements
Identifiable test position
(Walker et al., 1984)
By making use of direct field testing the advantage of close speaker
positioning seems to be that greater inter-aural attenuation values,
from 10 dB at the lower frequencies to about 25 dB at the highest
frequencies, are obtained. Though this is not sufficient to accurately
estimate monaural thresholds, it is sufficient for providing useful
hints as to monaural hearing loss.
(Magnusson et al., 1997)
These recommendations are discussed in depth later on in the following sections
where they will provide insight into the selection of appropriate stimulus and
recording parameters discussed in chapter three of this study.
20
University of Pretoria etd – Marais, J J (2004)
2.2.4 Recommendations when testing in the Sound-field
To ensure accurate estimation of hearing thresholds in the sound-field, it is important
to determine what type of stimulus is most suitable for sound-field audiometry and,
furthermore, what characteristics such a stimulus should possess. Walker et al (1984)
identified four stimulus characteristics required for accurate estimation of hearing
thresholds in the sound-field:
a) A stimulus should be sufficiently frequency specific to permit an accurate
description of the client’s hearing status.
b) The sound pressure level (SPL) should be sufficiently uniform over the whole area
in which the head of the participant is liable to move during the test procedure.
c) The SPL in the sound-field should be stable for small shifts in stimulus frequency.
d) The threshold obtained in the unaided condition in the sound-field should allow
for comparison to the thresholds obtained using earphones. This correlation is
important, as pure tone testing, using earphones, serves as the gold standard in
audiometric testing.
A comparison of different stimulus types, and specifically pure tones versus more
complex stimuli, in relation to the above mentioned characteristics should highlight
the appropriate stimulus type for the accurate estimation of thresholds in sound-field
audiometry. Using pure tones as the stimulus of choice in the sound-field meets the
first and last requirements set out in the above statement, however, standing waves are
created in the reverberant field7 (Walker et al., 1984). Standing waves, in turn, result
in significant changes in SPL at the participant’s ear, consequently making accurate
threshold determination problematic. Previous attempts to use graphic equalizers to
“flatten” the room’s frequency response, in order to minimize standing waves, were
limited to the use of more complex stimuli (Anderson, 1979) and did not tolerate head
movement by the participant (Walker et al., 1984). Direct-field testing8 is not greatly
influenced by the acoustic characteristics of the room (Dillon & Walker, 1981;
Magnusson et al., 1997; Walker et al., 1984) and therefore the use of pure tones in
7
In most audiometric test rooms the reverberant field constitutes the space exceeding 0.6 m from the
speaker.
8
In most audiometric test rooms the direct field constitutes the first 0.6 m from the speaker.
21
University of Pretoria etd – Marais, J J (2004)
sound-field audiometry is reasonable, provided that all testing is done in the direct
field (Walker et al., 1984).
Although it is widely recognized that the traditional audiometric stimulus, the pure
tone, is not optimal for testing in the reverberant field (Dillon & Walker, 1981;
Walker et al., 1984; Wilber, 2002), there has been some discussion on what the
optimal stimulus characteristics should be. Suggested stimuli that have come under
discussion include frequency modulated (warble) tones, narrow bands of random
noise, damped wavetrains (DWTs) and amplitude modulated (AM) tones.
The rationale for using more complex stimuli during sound-field presentation is based
on the assumption that increased uniformity can be attributed to the averaging of
intensity that occurs across the frequency range covered by the stimulus (Walker et
al., 1984). Therefore, a more complex stimulus will be less affected by a null9 in the
test room’s response, provided that it has sufficient energy within a frequency band,
which is wider than the null (Walker et al., 1984). The variability of the reverberant
field should be related to bandwidth as some stimuli create a more uniform field than
others. In other words, as the bandwidth increases, the variability will decrease. In the
direct field, the SPL is not influenced by the test room acoustics and thus there are no
advantages for using complex rather than pure tone stimuli (Walker et al., 1984).
There are, however, certain populations where the placement of the individual in
relation to the speaker has an influence on placement decisions.
When speakers are placed in close proximity of the individual, and especially in the
case of infants and children and others within the difficult-to-test population, the
speaker can act as a distraction (Magnusson et al., 1997). It is important to note that
speakers are used when testing infants and small children, who are usually afraid of
earphones. When speaker placement becomes a distraction, prolonged test time and
ultimately completion of the test is uncertain. It seems therefore that there are certain
practical disadvantages to testing in the direct field, which have to be considered
when selecting a stimulus type.
9
The aim of this procedure is to identify a listening position where the SPL within a sphere of radius
15cm does not vary by more than 2 dB from the value at the centre of the sphere. The smaller the
variation (in decibel) within the sphere, the greater the accuracy of these measures will be.
22
University of Pretoria etd – Marais, J J (2004)
The above-mentioned variables (Walker et al., 1984) and practical disadvantages
(Magnusson et al., 1997) associated with the use of pure tones, combined with the fact
that warble tones are more easily identifiable (Arlinger & Jerlvall, 1987), serves as
rationale for the use of more complex stimuli. There are, however, various complex
stimulus types, which have to be evaluated in order to determine the most appropriate
stimulus type for conducting sound-field audiometry.
2.2.4.1 Optimal Stimulus Characteristics
Having established the relevance of using more complex stimuli, it is now important
to evaluate the different types of stimuli available. Amplitude modulated tones create
a less uniform field than any other complex stimuli (Dillon & Walker, 1982a),
because of the limited advantage gained from averaging the sinusoidally amplitude
modulated tone (Walker et al., 1984). Although the use of more complex amplitude
modulated tones would improve uniformity in the field, the improvement in field
uniformity is still inferior to that gained from the use of frequency-modulated stimuli
(Walker et al., 1984). Narrow bands of noise are comparable to frequency-modulated
stimuli for larger bandwidths, with reference to the resulting field uniformity.
Whilst using dampened wavetrains creates superior uniformity in the field10
(Victoreen, 1974) the thresholds obtained using dampened wavetrains cannot be
related to those obtained through pure tone testing using earphones. This is because
the duration of DWT is shorter than the integration time of the normal ear and
subsequently the threshold becomes a function of stimulus duration (Walker et al.,
1984). Because thresholds obtained with dampened wavetrains, as stimulus
presentation cannot be generalized to thresholds obtained under earphone conditions it
has received little support within research and especially within the clinical arena. It
seems therefore that the use of frequency modulated tones or narrow bands of noise,
with appropriate bandwidths, are the most suitable stimuli for reverberant-field
testing.
10
Sound pressure levels (SPL) in a typical test room vary with distance from the speaker. Uniformity in
the field, therefore, relates to a uniform SPL in the field. Field uniformity is dependent on the distance
between the participant and the speaker.
23
University of Pretoria etd – Marais, J J (2004)
The selection of suitable bandwidths for behavioural testing under sound-field
conditions involves a compromise. Large bandwidths are necessary to establish
uniformity in the field, and conversely small bandwidths are necessary to ensure
frequency specificity (Walker et al., 1984). Furthermore, when it is not possible to
restrict head-movements to a 15-cm radius sphere, as is the case when testing infants,
larger bandwidths have to be employed (Walker et al., 1984) in order to minimize
errors associated with field variability.
There are, however, certain disadvantages to using stimuli with broad bandwidths11.
To reduce the errors arising from field variability, the hearing loss at a specific
frequency will be underestimated to some degree unless the loss is flat along the
frequency spectrum (Walker et al., 1984). Whenever there is a sloping hearing loss or
decrease in sensitivity along the frequency spectrum, some of the stimulus energy will
fall in regions with better hearing than the nominal frequency12. Substantial errors can
occur since hearing loss can change rapidly from frequency to frequency, particularly
in the high frequency regions (Walker et al., 1984).
In short, a stimulus with a broad bandwidth is less frequency specific. Thus to
minimize this effect, the desirable bandwidth must be kept as “small” as possible,
while still establishing uniformity in the field (Walker et al., 1984). This will then
ensure that the obtained threshold is frequency specific and that the individual is not
responding to sound energy that falls outside the nominal frequency being tested.
2.2.4.2 Required Modulation Waveform and Modulation Rate for Frequency
Modulated Tones
In the clinical setup, the validation for using modulated tones is that they are more
easily identifiable than steady pure tones (Arlinger & Jerlvall, 1987; Linden &
Kankkunen, 1969) and furthermore less influenced by reverberation (Arlinger &
Jerlvall, 1987; ISO 8253-2, 1985; Walker et al., 1984). Modulation entails altering the
amplitude or frequency of a wave, by another wave of a lower frequency to convey a
signal (Oxford Dictionary, 1990:763).
11
Stimuli with a broad spectral content, or broad bandwidth, may stimulate frequencies around the
nominal frequency, which leads to a decrease in frequency specificity.
12
Nominal frequency, in the text, will indicate the intended frequency of stimulation.
24
University of Pretoria etd – Marais, J J (2004)
The manner in which a signal sweeps back and forth around the centre/nominal
frequency defines the modulation waveform (Walker et al., 1984). Dillon & Walker
(1981) and Dillon & Walker (1982b) have investigated the merits of using sinusoidal,
triangular, ramp and rectangular modulation. Of these, the rectangular and ramp
modulation has been rejected due to considerable frequency “splatter” and a resultant
decrease in frequency specificity. There has, however, been some discussion on
whether to use sinusoidal or triangular modulation, as both offer certain advantages.
Sinusoidal modulation has the advantage of a more rapid falloff of energy outside the
nominal band, which in turn means improved frequency specificity, whereas using
triangular modulation results in a more uniform distribution of energy within the band
(Walker et al., 1984). Although both sinusoidal and triangular modulation is
satisfactory (Walker et al., 1984), most clinicians and researchers have definite
preferences. For the current study sinusoidal modulation would provide a stronger
correlate with the sinusoidal modulation used in the ASSEP technique.
Another variable that needs to be addressed is the rate of modulation, in other words
the rate at which the frequency range of the stimulus is swept (specified in Hz). Care
should be taken in the selection of modulation rate, as a too high rate can cause field
uniformity to suffer, as there might be insufficient spectral components to ensure
reasonable averaging within the band. While conversely, when the modulation rate is
too low, the participants’ ear will respond to intensity fluctuations (Dillon & Walker,
1982b; Walker & Dillon, 1983). This is due to its dependence on the test room
acoustics (e.g. the peaks and troughs in the rooms’ response). Therefore the measured
threshold will depend only on the highest peak occurring within the modulation cycle
(Walker et al., 1984).
Temporal integration properties of the listener’s ear will influence the modulation rate
at which individual peaks in the response begin to appear. Walker & Dillon (1983)
have indicated that the commonly used 5 Hz modulation rate may be adequate only
for the normal listener’s ear and those persons with abnormal temporal integration
required a rate of about 20 Hz to detect the presence of a stimulus and subsequently to
respond to it.
25
University of Pretoria etd – Marais, J J (2004)
When a single modulation frequency is required, as is mostly the case in the clinical
arena, it has been recommended that a modulation rate of around 20 Hz be used
(Walker et al., 1984). This relates to the fact that clinicians are bound to test both
individuals with normal-hearing, as well as those with hearing impairments, within
their scope of practice, and that they cannot always differentiate between these groups
prior to obtaining the individuals’ hearing thresholds. It is, however, important to bear
in mind that, in individuals with normal integration properties, slightly better results
would be obtained using a decreased modulation rate for the lower frequencies
(Walker et al., 1984).
In conclusion, modulated tones for use in sound-field testing should be triangularly or
sinusoidally modulated, rather than using rectangular or ramp modulation, and should
have a modulation rate of around 20 Hz to ensure frequency specificity. As the focus
of the current research endeavour is to provide normative data on normal-hearing
sensitivity, and subsequently normal temporal integration properties, modulation rates
around 5 Hz should be sufficient (Walker et al., 1984). However, as the data
collection was done within a clinical setup, the stimulus, during behavioural soundfield audiometry, was modulated closer to 20 Hz.
2.2.4.3 Direct versus Reverberant-field Testing
A further consideration that needs to be addressed when conducting sound-field
audiometry is the fact that the distance between the participant and the speaker takes
on considerable importance (Walker et al., 1984). Sound-field audiometry can be
divided into two main testing conditions, namely direct-field testing and reverberantfield testing. Distances beyond the near field and up to 0.6 m from the speaker are
known as direct-field testing. For these, the sound pressure level (SPL) is influenced
predominantly by the sound coming directly from the speaker. The SPL in direct-field
testing varies according to the inverse square law (Walker et al, 1984). This implies
that every doubling in distance from the speaker results in a 6 dB decrease in SPL
(Walker et al., 1984:14). Distances that exceed 0.6 m are known as reverberant-field
testing. For these, the SPL is influenced by reflected, as well as direct sound.
The main question in selecting the test position is whether the participant should be
placed in the direct or the reverberant field. Various factors such as the directivity of
26
University of Pretoria etd – Marais, J J (2004)
the speaker (Walker et al., 1984), the rooms’ dimensions (Anderson, 1979) and the
absorptivity of its surfaces will influence the extent of the direct field (Dillon &
Walker, 1981). In most audiometric test rooms, the direct field will extend only about
half a metre in front of the speaker. Subsequently, when testing in the direct field,
except for anechoic13 conditions, the participant needs to be placed in close proximity
to the speaker and meticulous control should be maintained over head- and body
movements, in order to maintain acceptable signal stability at the level of the ear. The
optimum positioning of the participant should, therefore, be as far away from the
speaker as possible, while still remaining in the direct field.
There are, however, certain audiometric applications, such as testing of infants, where
direct field testing is not feasible (discussed under 2.2.1) and others where the
advantages of direct field testing is questionable, such as measuring real ear insertion
gain (Walker et al., 1984). Although real-ear insertion gain (functional gain) of
hearing aids can be measured in the direct field, the functional gain and frequency
response will be influenced by the azimuth of the incident sound, because of head
baffle and –shadow effects (Walker et al., 1984). When testing in the reverberant field
it is of critical importance that the participants’ position is chosen so as to avoid
coinciding with a null at any of the audiometric frequencies. Procedural guidelines
will be stipulated in the methodological chapter of this study.
Within the adult population, the main application of sound-field audiometry is to
establish real ear insertion gain (Arlinger & Jerlvall, 1987; Dillon & Walker, 1981;
Walker at al., 1984), which in turn is a valid tool for verification of the target gain
(Mueller et al., 1992). When measuring insertion gain in the direct field, the azimuth
of the incident sound, because of head baffle or shadow will influence the functional
gain and frequency response (Walker et al., 1984). The choice of testing arrangement
(orientation of the participant in relation to the speaker) thus becomes an important
issue.
13
Sound-field testing can be most accurately carried out in an anechoic chamber since the sound
intensity varies smoothly and only gradually with distance from the speaker. Small random movements
by the participant away from the calibrated test position thus have little effect on the sound intensity at
the participant’s ear.
27
University of Pretoria etd – Marais, J J (2004)
For reverberant-field testing, due to the uniformity of the field, the azimuth of the
sound source will not influence functional gain, unless the direct component of the
field is substantial. According to Walker and colleagues (1984), a review of published
studies as well as some of their own unpublished data, revealed that the functional
gain and frequency response measured in the reverberant field would differ little from
the measurements made in the direct field with a signal incidence of zero degrees
(when the participant is directly facing the speaker). More current research (Muller et
al., 1992) suggests that a 0º or 45 º speaker azimuths is the preferred measurement
condition.
In conclusion, there is no simple answer when considering whether to test in the direct
or reverberant field when measuring the functional gain of a hearing aid. These
questions should be decided by considering the purposes of the measurement and the
age of the participant being tested.
2.2.4.4 Important Test Room Considerations
When testing via a speaker there should be no reflecting surfaces in the test room and
no objects between the speaker and the participants’ ear (Wilber, 2002). Whenever
possible, and presumably in all cases except when testing young children and
neonates, a head rest should be used in order to limit or minimize head movement.
Furthermore, no part of the headrest or the chair should protrude beyond the sides of
the participant’s head, as any such surface might disturb the sound-field. Preferably,
the chair should be adjustable in height, although making use of a cushion to adjust
the height of the participant is also satisfactory. The test position should be constant
and accurately identifiable. The centre of the participants’ head should be in line with,
and at the same height as, the centre of the speakers’ cone (Walker et al., 1984;
Wilber, 2002). Thus, the positioning of the participant in relation to the speaker seems
to be more important (Magnusson et al., 1997) than the positioning of the speaker
itself (Walker et al., 1984). Within the clinical setup this implies that when testing in
the direct-field, results are comparable to those obtained through testing in the
reverberant field as long as a signal incidence of 0° or 45° (i.e. with the participant
directly facing the speaker) is maintained (Walker et al., 1984).
28
University of Pretoria etd – Marais, J J (2004)
2.2.4.5 Importance of calibration
The standard for audiometers (ANSI 3.6, 1996) describes the primary characteristics
of sound-field testing. Special attention is paid to the test room, frequency response,
method for describing the intensity of the speech signal and the location of the
speaker. It also provides specific values for tonal stimuli, such as frequencymodulated tones and narrow bands of noise.
When calibrating the equipment it is important to place some sort of marker, so as to
easily identify the test position where the participant will be. When calibrating for
testing via a speaker there should, furthermore, be no reflecting surfaces in the room
and no objects between the speaker and the participant’s ear (Wilber, 2002).
Thus, when pre-test calibration is used, the selection and maintenance of the test
position, as well as procedural aspects, takes on critical importance. This implies that
a listening position should be identified where the SPL, within a sphere, will not vary
by more than 2 dB from the value at the centre of the sphere. This criterion might
have to be revised when testing infants, as their range of movement might exceed that
of the identified test position. The identification of the listening position or listening
sphere has an impact on the accuracy during testing. The smaller the variation, in
decibels, within the sphere, the more accurate the measurement will be (Walker et al.,
1984).
Calibration is important for two reasons. It establishes a relationship between the SPL,
measured at the same point, and the corresponding attenuator setting on the
audiometer or signal generator. Secondly, it establishes the SPL corresponding to
normal thresholds (audiometric zero) so that the individual participants’ hearing levels
can be compared to this standard (Walker et al., 1984). The second reason for
calibration is important for diagnostic audiology, but less so for hearing aid
evaluation. With hearing aid evaluation the emphasis is upon the absolute level or
difference in threshold brought about by the aid and upon the absolute level of sound,
which can be heard through the aid (Walker et al., 1984).
29
University of Pretoria etd – Marais, J J (2004)
Currently sound-field stimuli are pre-calibrated by measuring the SPL at a particular
point in the sound-field for various attenuator dial settings. This technique has the
advantage that head diffraction effects, caused by the introduction of the participants’
head into the sound-field, are automatically taken into account whenever a threshold
is determined (Walker et al., 1984). However, two potential sources of error are
introduced.
Firstly, the introduced head diffraction effects are, to some extent, dependant on the
particular test position, whereas one would preferably include only those systematic
components of head diffraction which the participant experiences in diverse acoustic
environments. Secondly, any movements by the participant (e.g. head movement)
from the pre-calibrated point cause a measurement error. Although stimuli are chosen
so that this error is acceptably small within a sphere radius of 15 cm, it may not
always be possible to ensure that the movement is restricted to this extent.
An alternative calibration technique involves the use of a microphone to continuously
monitor the SPL near the participants’ head. The output of the microphone is used to
automatically control the attenuator of the signal generator to maintain an invariant
signal level at the microphone (Walker et al. 1984). This technique has the advantage
that if the control microphone is attached to the participants’ head, the calibrated
position moves whenever he/she does. The major disadvantage associated with this
technique is that the SPL at the control microphone is not exactly that of the position
of interest (the hearing aid microphone or ear drum for aided and unaided assessment,
respectively). This discrepancy is both person and environment dependent.
Because the current study was performed within a clinical set up, the calibration
method employed was pre-calibration. Control microphone calibration has the
advantage of greater accuracy, especially within the infantile population, due to the
likelihood of considerable and unavoidable head movement (Walker et al., 1984).
Furthermore, when using any type of complex stimuli, the intensity at the
participant’s ear will vary with time. Although frequency modulated (FM) tones,
unlike other stimulus types, do not contain inherit intensity fluctuations the intensity
at the ear varies (in the reverberant field) because the peaks and troughs in the room’s
30
University of Pretoria etd – Marais, J J (2004)
frequency response are sequentially excited, as the tone sweeps the frequency (Walker
et al., 1984).
Previous research (Dillon & Walker, 1980) evaluated the measurement of intensity
fluctuating stimuli. Findings indicated that similar calibration figures of acceptable
accuracy can be obtained by reading peak deflections off a sound level meter set to
“route mean square-fast” mode. When calibration is done as indicated above,
thresholds obtained in the sound-field can be related to thresholds that would be
obtained by testing under earphones with pure tone stimuli. This is therefore the
method of calibration used during the current research endeavour.
The following summary of conclusions (Table 2.2) represents a reasonable set of
comprehensive recommendations based on the discussion provided so far.
31
University of Pretoria etd – Marais, J J (2004)
Table 2.2 Summary of the Limitations of Sound-field Audiometry and Recommendations for addressing these Limitations
Literature
Goal of Research Endeavour
Recommendation
1. Dillon &
Walker, 1982
2. Walker et al.,
1984
3. Dillon &
Walker, 1982
4. Dillon &
Walker 1982
Comparing stimuli used in soundfield audiometric testing
Recommendations for stimuli and
procedures in sound-field testing
Comparison of stimuli used in soundfield audiometric testing
The selection of modulation wave
form for frequency modulated soundfield stimuli
Recommendations for stimuli and
procedures in sound-field testing
The selection of modulation wave
form for frequency modulated soundfield stimuli
Complex stimuli, and not pure tone stimuli, should be used, as the type of
stimulus, for testing in the reverberant field. In most audiometric test
rooms this means any distance that extends beyond about half a metre
from the speaker.
FM Tones have been identified as the best type of stimulus, according to
its stimulus characteristics, for sound-field testing, having suitable
characteristics.
7. Dillon &
Walker 1981
8. Beynon &
Munro, 1995
9. Walker et al.,
1984
The effect of the acoustic
environment on the reliability of
sound-field audiometry
Measuring variability in sound-field
audiometry due to subject movement
Recommendations for stimuli and
procedures in sound-field testing
10. ANSI 3.6,
1996
11. Dillon &
Walker, 1980
12. Arlinger &
Jerlvall,
1987
Specifications for audiometers. ANSI
3.6-1996
The perceptions of normal-hearing
persons of intensity fluctuations in
narrow band stimuli and its
implications for sound-field
calibration procedures
Reliability in warble tone sound-field
audiometry
Test room considerations:
A height adjustable chair should be used whenever feasible.
The test room should be as non-reverberant as possible.
There should be no reflecting surfaces in the test room and no objects
between the speaker and the participants’ ear.
The test position should be constant and accurately identifiable.
The centre of the participants’ head should be in line with, and at the
same height as, the centre of the speakers’ cone.
Other important factors:
Traditional pre-calibration was chosen as a satisfactory method of
calibration.
Presentation level of a complex sound can be determined by reading
the peak deflection off the “route-mean-square- fast mode” on a sound
level meter. Therefore, sound-field measures can be described as dB
HL.
5. Walker et al.,
1984
6. Dillon &
Walker, 1982
FM tones should be sinusoidally or triangularly modulated. For this
particular study, and specifically the comparison between sound-field and
ASSEP, sinusoidally modulated tones were selected at a modulation rate
of around 20Hz. The modulation rate selected is sufficient for normal and
abnormal temporal integration.
Implication for Current
Study
Complex stimuli with large
bandwidths ensure uniformity
in the field, but have to be kept
as small as possible to ensure
frequency specificity.
Sinusoidal modulation ensures
improved frequency
specificity, whereas a
modulation rate of around 20
Hz will be sufficient for normal
and abnormal individuals
integration properties.
These considerations ensure
that the thresholds obtained
are accurate and that the
influence of the acoustic
environment is kept to a
minimum.
Calibration of the equipment
ensures that the thresholds
obtained under sound-field
conditions are comparable
with those obtained under
earphone testing.
32
University of Pretoria etd – Marais, J J (2004)
The final goal of the current research endeavour entails a more advanced application
of ASSEP, namely the estimation of hearing thresholds through sound-field
presentation. ASSEPs is a relatively new, objective technique in the AEP arsenal, that
has until recently been investigated as the method of choice in diagnostic hearing
threshold estimation. In an attempt to validate the possible application of ASSEPs
through sound-field presentation, it is necessary to discuss the current application of
steady state in order to delineate the current application status.
2.3 Perspectives on Auditory Steady State Evoked Potentials
The following section aims to provide the reader with information related to the
current application of ASSEPs, and argues for the use of sound-field speaker
presentation. The information is structured to provide a definition of auditory steady
state evoked potentials, followed by the current clinical applications as well as current
research applications. Thereafter the audiometric variables are discussed in depth,
according to transducer-, stimulus-, intensity- and latency-related variables. Attention
is paid to the influence of participant related variables, such as stimulus and response
characteristics, transducer influence, as well as the state of consciousness of the
individual. This section concludes with the possible advantages of ASSEP over other
techniques in the estimation of aided thresholds.
2.3.1 Definition of Auditory Steady State Evoked Potentials
ASSEPs are evoked by continuous tones (Lins et al., 1996). They are sinusoidally
amplitude and/or frequency modulated at frequencies between 3 and 200 Hz (Lins et
al., 1996), of which the best modulation rates for audiometric purposes appear to be
around 40 Hz and between 75-110 Hz (John et al., 1998). And, finally, they are
detectable at the frequency of modulation (Rance et al., 1998). This periodically
changing stimulus, although continuous in presentation, causes a shortened interstimulus interval. The resultant response detected is a compound response, of several
“overlapping” transient responses, and is referred to as Auditory Steady State Evoked
Potential for the purpose of this research.
33
University of Pretoria etd – Marais, J J (2004)
2.3.1.1 The Physiology underlying ASSEPs
ASSEPs can better be understood when considering the functioning of the cochlear
transducer (Lins & Picton, 1995). An auditory stimulus consists of acoustic energy
distributed over a frequency spectrum that can be processed by the cochlea
(Dimitrijevic et al., 2001). The basilar membrane and outer hair cells act as frequencyplace analyzers for the specific frequencies contained in incoming sound signals,
where higher frequencies activate the region of the basal membrane closer to the oval
window and lower frequencies activate regions further along the membrane (Lins &
Picton, 1995).
Any incoming auditory signal creates a “wave/traveling wave” that moves along the
basilar membrane of the cochlea. While mediating a frequency-to-place coding in the
auditory system (John & Picton, 2000), it also involves a delay since time is taken to
move along the basilar membrane. This serves as a possible explanation as to the fact
that physiological responses to auditory stimulation, generally occur at earlier
latencies for high frequencies, than responses to low frequency stimulation
The transducer function, associated with the depolarization of inner hair cells, is
asymmetric and non-linear, since larger potentials are evoked when the inner hair
cells are bent away from, rather than towards, the basal body (Lins & Picton, 1995).
Therefore the non-linear functioning of the cochlea can be attributed to the
preferential response of hair cell movement in one specific direction, namely away
from the basal body (John et al., 1998). In addition to the non-linear functioning, the
responsiveness of the cochlea saturates with increasing amplitude of the stimulus
(John et al., 1998), this in turn can be attributed to high levels of depolarization of the
inner hair cells (Lins & Picton, 2000).
Inner hair cells, then, synaptically activate dendrites of the ganglion cells that
compose the afferent fibers of the auditory nerve. Rectification of the signal occurs,
since only depolarization of the inner hair cells causes action potentials in the
ganglion cells. Larger depolarization’s causes faster firing rates in ganglion cells (Lins
& Picton, 2000). The cochlear transfer function can therefore be described as a
compressive rectification (John et al., 1998).
34
University of Pretoria etd – Marais, J J (2004)
A sinusoidal amplitude-modulated (AM) pure tone contains energy at three
frequencies (the carrier and two other frequencies separated from the carrier by the
modulation frequency). It is important to note that the signal contains no energy at the
modulation frequency. However when the signal undergoes non-linear distortion
resultant energy will by present at certain frequencies, and foremost among these
combined tones will be the modulation frequency (John et al., 1998).
Non-linearity’s in the output of the cochlea thus provide the basis for the detection of
responses (John et al., 1998). In other words the cochlea contains a rectified version of
the acoustic signal, which subsequently causes the output of the cochlea to contain a
spectral component at the frequency where the carrier signal was modulated. This
component, which is not present in the spectrum of the signal, can be used to assess
the response of the cochlea to the frequency of the carrier (Lins et al., 1996).
2.3.1.2 The Anatomy underlying ASSEPs
ASSEPs can be recorded by stimulus rates of up to several hundred hertz (Rickards &
Clark, 1984). The amplitude of these responses are larger at rates near 40 Hz
(Galambos et al., 1981) and 80 Hz (Lins et al., 1995). The nature of these potentials
has been studied with responses evoked with stimulus rates near 40 Hz (Galambos et
al., 1981; Stapells et al., 1984). Should these responses arise from the same neurons
that generate transiently evoked responses, then ASSEP could be predicted from the
super-positioning of multiple overlapping transient responses (Hari et al., 1989). It is,
however, also possible that the rapid stimulation might elicit a separate response from
neurons that are specifically responsible for rhythmic activity and that resonate at the
frequency of stimulation (Basar et al., 1987). These generators might also be
responsible for spontaneous cortical gamma rhythms (John & Picton, 2000).
Discrepancies between ASSEP responses and those predicted by super positioning
(Azzena et al., 1995) and the persistence of the response the time when the stimuli are
presented (Santarelli et al., 1995) support the idea of resonance. However by
modeling the refractory effects and considering the multiple generators, one can still
explain the most of the 40 Hz response findings on the basis of transient response
35
University of Pretoria etd – Marais, J J (2004)
waveforms (Gutschalk, et al., 1999). It is therefore unlikely that independent rhythmic
generators would account for the 40 Hz ASSEP, and more so for the 80 Hz response
(John & Picton, 2000). It seems more likely that these responses are generated by
neurons in the brainstem, and possibly the cortex, which respond to both transient
stimuli and become “locked” to the envelopes of amplitude modulated tones (John &
Picton 2000). The longer latencies of 80-110 Hz responses suggest either a generator
further along the auditory pathway or one lower in the auditory pathway, activated by
a multi-synaptic circuit (John & Picton, 2000).
To summarize it seems that many different regions of the auditory nervous system
from the auditory nerve to the auditory cortex generate responses that follow the
modulation signal of an amplitude-modulated tone. Each of these regions can create a
response that has a specific phase and latency relationship to the stimulus. All these
responses will overlap to make up the response recorded from the scalp (John &
Picton, 2000). Therefore the dominant response in an “awake” subject, at 40 Hz
probably derives from the auditory cortex, with the brainstem contributing only a
little. With sleep the cortical responses may attenuate and the scalp recorded response
may mainly reflect those potentials generated by the brainstem (John & Picton, 2000).
It seems that scalp recorded potentials in response to tones modulated around 80-110
Hz derives from a generator high in the brainstem (John & Picton, 2000).
2.3.2 Current Clinical Applications of ASSEP
The ultimate goal of objective audiometry is to generate a frequency specific
audiogram (Aoyagi et al., 1996), without any behavioural response from the
participant or subjective interpretation of the results by a clinician. ASSEP has
recently become available as an objective hearing test option (Rance et al., 1998).
This statement was confirmed by various researchers (Aoyagi et al., 1994; Cohen et
al., 1991; Lins et al., 1996; Rickards et al., 1994), having obtained reliable, and
frequency specific estimates of behavioural pure tone thresholds in adults, well babies
and hearing-impaired participants. Currently the ASSEP techniques have two major
applications within the clinical arena, namely detection (of hearing loss) and
diagnosis (characterization of the hearing impairment).
36
University of Pretoria etd – Marais, J J (2004)
With the current focus on neonatal hearing screening (NIH Consensus Committee,
1993; Joint Committee on Infant Hearing, 1994), and identification and treatment of
those infants with hearing impairments before the age of 6 months (Picton et al.,
1998), there has been an increase in the difficult-to-test populations. Researchers
(Perez-Abalo, Savio, Torres, Martin, Rodriguez & Galan, 2001; Picton et al., 1998;
Rance et al., 1998; Stürzebecher et al., 2001) have documented several advantages of
ASSEP, over other AEP techniques such as the ABR, as an objective screening
measure. The ASSEP technique is more time efficient, when the dichotic multiple
stimuli technique is used (Picton et al., 1998) and it has commercially available
response detection techniques, which allows for objective response detection
(Stürzebecher et al., 2001) not yet available in commercial Auditory Brainstem
Response (ABR) equipment. Furthermore, the technique is able to identify reliable
frequency specific thresholds in infants, regardless of the state of consciousness
(Aoyagi et al., 1993; Cohen et al., 1991; Lins & Picton, 1995; Rickards et al., 1994).
Recently, researchers (Picton et al., 2002a; Stürzebecher et al., 2001) have cautioned
against using the multiple stimuli technique at high intensities, as frequency
specificity might suffer due to overlapping of responses at high intensities. Picton and
colleagues (2002b) explained that at high intensities each stimulus would activate a
broader region of the basilar membrane than at lower intensities, causing subsequent
overlapping in the activation patterns of the different stimuli. Researchers (PerezAbalo et al., 2001) were, however, able to show that overlapping of responses and
subsequent decreased frequency specificity did not occur even at intensities around
110 dB. There are, however, still research questions regarding dichotic multiplestimulation at high intensities and the effect thereof on frequency specificity, the
effects of multiple stimuli on the maturation of infantile nervous systems and the
improved performance of statistical tests used for objective response detection to
minimize false-negative decisions.
The second, and most widely used, application of the ASSEP technique has its focus
on the diagnosis or characterization of the hearing impairment, especially within the
difficult-to-test population. As research improves the objective response detection
techniques available (Cebulla, Stürzebecher & Wernecke, 2001; John et al., 2001),
exciting possibilities arise in terms of improved objectivity and frequency specificity.
37
University of Pretoria etd – Marais, J J (2004)
Furthermore, the ability to employ increased signal presentation levels means that
residual hearing sensitivity, even in the severe to profound regions, can be accurately
estimated (Rance et al., 1995; Rickards et al., 1994). The accurate estimation of
residual hearing is crucial, especially with cochlear implant candidates (Rance et al.,
1994), and because ASSEP can be presented at high intensity levels without the
artifact complications associated with the ABR (discussed later on in this chapter), its
application within this population has been highly successful.
The ASSEP technique complies with all the requirements to allow for complete
characterization of the hearing impairment. The technique is able to accurately
determine degree of loss, predict the configuration of the loss and determine
frequency specific thresholds (Kuwada, Batra & Maher, 1986; Lins & Picton, 1995;
Lins et al., 1996; Picton et al., 1998). Results obtained can be presented as an
audiogram and the technique is, furthermore also time efficient (Rance et al., 1995;
Rickards et al., 1994).
2.3.3 Research Applications
Although the ASSEP technique has become available as an objective hearing test
option for infants and others in the difficult-to-test population (Rance et al., 1998),
there are still various possibilities regarding application within the clinical arena that
are yet to be established, through research. Research is still required to establish
whether single modulated tones offer higher frequency specificity at high stimulation
intensities. These results will offer insight into the application of ASSEP in cochlear
candidacy. Research is also required to establish whether aided thresholds can be
obtained from cochlear implant users, using an adapter cable, to maximize usage of
electrode configurations in the maps of difficult-to-test patients.
Although the current applications of ASSEP are aimed at providing the clinician with
an electro-physiological audiogram, to assess the hearing impairment, more
information is needed regarding supra-threshold discrimination (Picton et al., 2002a).
Although the ASSEP at this stage cannot directly evaluate speech perception it is able
to evaluate small changes in intensity and frequency, which provide the cortical
regions with information regarding speech perception. Further research is required,
where a single carrier is independently modulated through AM and FM, and with
38
University of Pretoria etd – Marais, J J (2004)
various combinations of AM/FM modulation (e.g. 50% AM and 25 % FM), to
establish how the cochlea processes information needed for speech perception.
Recently researchers (Lins & Picton, 1995) have endeavoured to further investigate
the physiology underlying ASSEP, using modulation rates between 150-190 Hz. At
these higher modulation rates equal contributions between the brainstem and cortical
areas were noted. With variation in modulation rate, insight might be gained into
pathology of the auditory system up to cortical level (Regan, 1989; Regan & Regan,
1993; Lins & Picton, 1995). Research in this regard has “not extended beyond
hypothesis in most instances” (Schmulian, 2002: 69).
2.3.4 Current Research Available
To date only two studies have investigated the objective evaluation of aided
thresholds using ASSEP. The first, a study by Picton et al (1998) used three different
groups of subjects. A normal hearing group of 10 subjects (6 female) with thresholds
less than 20 dB HL at frequencies between 0.5 and 4 kHz were selected. Subject age
varied between 13 and 40 (mean 29) years of age. The other normal hearing group
consisted of 10 female subjects aged between 23 and 43 (mean 31) years of age.
The second grouping consisted of 35 hearing-impaired children with moderate
hearing impairment. These children were using hearing aids, and were aged between
11 and 17 (mean 15) years of age. Only one ear of each subject was tested. The test
ear was the one with the best pure tone average, while random selection was carried
out for the children with symmetric hearing losses. The non-test was occluded. In all
15 left and 20 right ears were tested.
A small third group of subjects included 3 children who were being fitted with
hearing aids and who were unable to provide reliable responses to sound. Two of
these children were newly identified infants under 1 year of age. The first was
assessed at 9 months of age (X1), while the second was assessed at 8 months of age
(x2). The 8 month old was premature and developed respiratory distress syndrome.
The last child was a 9-year-old girl with severe developmental delays (x3).
39
University of Pretoria etd – Marais, J J (2004)
Picton and his colleagues (1998) used sinusoidally amplitude-modulated tones, with
the depth of modulation at 100% for each stimulus. The carrier frequencies included
were 0.5, 1, 2 and 4 kHz. These frequencies were modulated using 80.9, 88.9, 96.9
and 104.8 Hz respectively over the carrier frequencies 0.5-4 kHz. Simultaneous
presentation of stimulus was used in all cases, except in for some instances where the
normal hearing group was tested.
The timing of the stimulus and the recording was exactly synchronized for frequency
analysis to detect responses. The stimuli were therefore adjusted is that there was an
integer number of both the carrier frequency and the modulation frequency within the
754-msec buffer. A Madsen Micro 5 audiometer equipped with both TDH 39 supra
aural headphones and a Madsen FF-73 sound field speaker was used to present
stimuli. All subjects were tested in a single walled audiometric test room were
ambient noise levels complied with ANSI standards of 1991. Subjects were seated in
an armchair in the center of the room 1.25 meters away from the speaker at an
azimuth of 0º. The head position was in the center of the speaker axis. Stimuli were
calibrated according to the recommendations of Walker et al. (1984) and Beynon and
Munro (1995) as well as the ANSI S3.6 (1996).
Subjects were mostly asleep during recording of responses, with the exception of the
9-year-old girl in the third group. Ground electrode was connected to the lateral neck,
while reference electrodes were placed between the vertex and the posterior midline
neck, half way between the inion and the vertebra prominens. While infants presented
difficulties with electrode placement and maintenance and the reference electrode was
therefore placed on the ipsilateral mastoid to the test ear and the ground on the
forehead. Responses were amplified and filtered with a band pass of between 10 and
300 Hz. The signal was then converted from analogue to digital. Each recording
section included 512 samples and 16 sections were concatenated to form a full
recording sweep of 12.06 seconds. The amount of sweeps required was dependant on
the signal-to-noise ratio. This lasted between 6 and 15 minutes. The use of recording
sections allowed Picton and colleagues (1998) to reject artifacts over the section
rather than over the entire sweep. Artifact rejection was set at 40 µV for most
subjects, but where the artifacts were more frequent the rejection criteria was raised to
50 or 60 µV (depending on the subjects’ alpha rhythm). The average 12-second sweep
40
University of Pretoria etd – Marais, J J (2004)
was transformed to the frequency domain using a Fast Fourier Transform. The
amplitudes reported in their paper were baseline-to- peak amplitudes and the phases,
cosine-onset phase.
Threshold estimation for pure tones was using conventional audiometric techniques
(Carhart & Jerger, 1959). Stimuli were presented singly rather than in combination.
The voltage of the batteries of all hearing aid users was checked prior to the threshold
estimations. Physiological thresholds were determined by the presence or absence of
recognizable ASSEP responses.
In the normal hearing group the aim was twofold, namely to determine the influence,
if any of transducers and stimulus presentation. No significant difference was
observed when stimulus presentation was varied between single and multiple
conditions. However for both conditions physiological thresholds was 10 –20 dB
above behavioural threshold. When the influence of transducers was investigated
there was no significant difference between SPL thresholds obtained under earphones
and those obtained under sound field speakers.
The responses obtained from the second group (hearing aid users) revealed a
significant correlation between behavioural thresholds and ASSEP thresholds. In the
group with profound hearing losses aided thresholds were obtained at threshold levels
with the exception of 4 kHz where no responses were obtained even when the
stimulus was presented at levels that exceeded threshold level. A clear response was
obtained at 4 kHz when the stimulus presented singly instead of in combination.
The first experiment has significant influence on the current study, as it indicates that
reliable responses could be recorded close to behavioural thresholds (13-16 dB),
irrespective of the transducer used. Therefore it seems logical that should the
transducers be calibrated to present the same dB SPL at the eardrum for the different
stimuli used, the transducer effect will be negligible. A pilot study by Cone Wesson
and colleagues (2001) using 4 adult hearing aid users presented similar findings.
41
University of Pretoria etd – Marais, J J (2004)
2.3.5 Audiometric Variables
Transducer related variables
There is, currently, limited research available to document the influence of using
speakers as a transducer when recording ASSEP. Taking into account that an
ASSEP is evoked by regularly repeating stimuli, the response should stabilize
after the initial few stimuli (Picton et al., 1998). Furthermore, it should thereafter
contain constituent frequency components that remain constant in amplitude and
phase over time (Regan, 1989). When these responses are evoked, by amplitudemodulated tones, in the auditory nervous system, they are frequency specific and
stable over time. They should therefore not distort in the sound-field speaker or
the hearing aid (Picton et al., 1998).
Stimulus related variables
It is important to pay attention to the stimulus rate used to evoke these responses,
as the stimulation rates are influenced by the various factors (e.g. state of
consciousness) that will impact on the responses recorded. The 40 Hz potential,
initially described by Chatrian, Petersen and Lazarte (1960), as well as Schimmel,
Rapin and Cohen (1975), was first detailed according to its general characteristics
by Galambos and colleagues (1981). Although the 40 Hz potential has application
value to this day, the best modulation rates for audiometric purposes may be
between 70-110 Hz (Picton et al., 1998). The reason for this is that responses
evoked at these modulation rates are not significantly effected by state of
consciousness, or state of arousal (Cohen et al., 1991), and can be reliably
recorded in infants (Aoyagi et al., 1993; Rickards et al., 1994; Lins et al., 1996).
Intensity Effects of the Stimulus
Intensity is a well-documented variable in relation to other AEP measurements
such as the ABR and has an effect on the latency of evoked potential measures.
With a decrease in intensity the latency tends to increase (John & Picton, 2000).
John & Picton (2000) hypothesized that a shift in the cochlea, related to intensity
42
University of Pretoria etd – Marais, J J (2004)
variations, contributes significantly to a change in latency. They also mentioned
that short duration stimuli at high intensities evoke responses earlier on the rise
time of the synaptic activation pattern.
It is important, however, to remember that intensity fluctuations will have a more
adverse effect on transient stimulation. This is due to the short stimulus duration
and the subsequent silence following the stimulus, as the postsynaptic membrane
returns to a resting position (John & Picton, 2000). Although intensity variation
might have smaller latency effects on sustained stimulation, such as the ASSEP,
there might still be an influence, depending on the frequency of stimulation. This
is because low frequency carriers evoke an activation pattern on the basilar
membrane that covers a greater spatial extent than higher carrier frequencies.
According to John & Picton (2000) there might be a resultant latency jitter of the
responses, which in turn might attenuate the amplitude of the compound response.
This can be attributed to the fact that neurons along a broad area of the basilar
membrane all respond to the same low frequency carrier and some of these
responding neurons might be activated significantly earlier than others, and
especially at higher intensities (John & Picton, 2000).
Experiments that focused on the effect of stimulus intensity (Lins et al., 1995)
have indicated that for stimulus intensities above 70 dB SPL the amplitude of the
response grows much more rapidly than for stimulus intensities below 70 dB SPL.
Furthermore it also indicated that the difference in amplitude was attenuated for
lower frequencies, when compared to higher frequencies. Therefore, at threshold
levels only the fibers with characteristic frequencies near the carrier frequency
seems to be activated, whereas a larger number of fibers, with higher
characteristic frequencies, are activated when the intensity is increased above 6070 dB SPL. It seems that the intensity of the stimulus presented could possibly
impact on the estimation of lower frequencies, and more so for moderate to high
intensity presentation levels.
Latency Effects
The physiological interpretation of latencies of scalp-recorded ASSEP remains
difficult, as the responses evoked do not derive from a single generator. Because
43
University of Pretoria etd – Marais, J J (2004)
of the periodic nature of ASSEPs, however, the conventional measurement of
latency is prohibited in the frequency domain (Rickards et al., 1994). This
becomes evident when considering that there are several different regions of the
auditory nervous system that generate responses following the modulation signal
of an amplitude modulated tone. The compound responses recorded at the scalp
are made up of several individual responses along the auditory nervous system.
Each of these individual responses has their own phase- and latency-relationship
to the stimulus (John & Picton, 2000). There seems to be an indefinite relationship
between phase and latency, as phase cannot be directly translated to latency.
According to Regan (1966) this problem can be resolved through assessing phase
at different stimulus rates to arrive at clear estimates of apparent latencies, or
“group delay” as Goldstein, Baer and Kiang (1971) termed this phenomenon.
Frequency or latency related delays are an important consideration when
measurements are made in the frequency domain or in the time domain. These
latency-related delays have a definite influence on measurements that are obtained
in the frequency domain. There are five main contributors to the latency of neurophysiological responses to auditory stimulation, and each of these will be
discussed.
The acoustical delay relates to the delay between the stimulus being produced and
the stimulus arriving at the oval window after transmission through the air to the
tympanic membrane and through the middle ear (John & Picton, 2000). The
second contributor is related to transport time delays. This simply means that the
travelling wave takes time to move along the basilar membrane, and refers to the
amount of elapsed time between the arrival of the sound energy at the oval
window and the beginning of activity at the location on the basilar membrane.
Although most evidence suggests that the delay is in the millisecond range, and
that it increases exponentially with increasing distance along the basilar
membrane (John & Picton, 2000), there is still a lot of controversy surrounding
transport delays.
The third contributor is related to transduction or filter delays. This has to do with
the filtering of acoustic energy into electrical impulses, in other words, the time
44
University of Pretoria etd – Marais, J J (2004)
the acoustic energy takes to pass through the active filtering process of the
cochlear hair cells that are sensitive to the frequencies of sound in the stimulus.
The more sharply the filter is tuned, the longer it will take for the output of the
filter to reach maximum amplitude (John & Picton, 2000). This filter delay,
together with the transport delay, contributes within the normal cochlea, to the
frequency related delays of physiological responses.
Synaptic and conduction delays are the last two contributors to frequency delays.
It simply infers that there is a time delay as the sound energy is transmitted
between the inner hair cell and the afferent nerve fibre as well as the time it takes
to conduct the potential between the cochlea and the neural generator (of the
response). There seems to be no significant affects created by the frequency of the
signal (John & Picton, 2000). These delays could possibly create significant
differences between the responses obtained from male and female individuals
being tested, due to the difference in the length of their basilar membranes.
When stimulation rates around 70 Hz are used, apparent latencies ranging
between 11 and 14 ms are detected (Cohen et al., 1991; Lins & Picton, 1995;
Rickards et al., 1994) depending on the carrier frequency. Rickards et al (1994)
suggests that the use of ASSEPs in the assessment of the hearing sensitivity of
infants and young children is most successful when responses with latencies of
about 10 ms are used.
2.3.6 Participant Related Variables
Participant related variables are discussed according to age, gender and hearing
sensitivity:
2.3.6.1 Influence of Age
Although age seems to be a significant contributor to variability, research has
shown that when stimulation rates around 40 Hz are used, responses can be
consistently recorded (in the neonatal population) when modulation rates in
excess of 70 Hz are used (Rickards et al., 1994). Other related studies by
Boetcher, Madhotra, Poth and Mills (2002a) as well as Boetcher, Poth, Mills and
Dubno (2002b) on ASSEP have predicted that reduced amplitudes would be
45
University of Pretoria etd – Marais, J J (2004)
obtained for aged participants. This prediction is based on two main factors,
namely reduced frequency discrimination in geriatric participants, possibly due to
changes within the auditory system (including the effects of presbycusis). The
second factor relates to the fact that other AEPs (like the ABR) have reduced
amplitudes in geriatric participants.
2.3.6.2 Influence of State of Consciousness
Earlier studies (Galambos et al., 1981; Maurizi et al., 1990) have investigated the
reliability of the 40 Hz ASSEP, in the accurate estimation of frequency specific
hearing thresholds, in infants and other difficult-to-test populations. The robust
amplitudes observed near threshold level as well as their established relationship
to behavioural responses motivated these investigations. These investigators as
well as others have since established that the amplitudes, and subsequently
accurate threshold estimation capability, of the 40 Hz response are severely
affected by state of consciousness (Brown & Shallop, 1982; Linden, Campbell,
Hamel & Picton, 1985; Osterhammel, Shallop & Trkildsen, 1985; Plourde,
Stapells & Picton, 1991; Suzuki et al., 1994) and state of arousal (Plourde &
Picton, 1990). Thus, behavioural thresholds are over estimated.
Later research (Cohen et al., 1991) determined that when higher modulation rates
are employed, responses down to near-threshold levels could be recorded reliably
in infants regardless of state of consciousness (Aoyagi et al., 1993; Lins et al.,
1996; Rickards et al., 1994). This phenomenon is probably related to the neural
generators responsible for the evoked responses.
The 40 Hz ASSEP is thought to be generated by contributors from the ascending
auditory brainstem pathways and auditory regions of the thalamus and auditory
cortex (Pratt, Mittelman, Bleich & Zaaroor, 2002). According to these researchers
the relative contributions of cortical and sub-cortical regions are still unclear, but
appear to be modified by sleep (Makela & Hari, 1987; Spydell, Pattee & Goldie,
1985; Suzuki et al., 1994). John and Picton (2000) have hypothesized that the
dominant response at 40 Hz derives from the auditory cortex, with the brainstem
contributing only a little. This would explain the effect of state of consciousness,
as the cortical contribution may attenuate. This would mean that the scalp-
46
University of Pretoria etd – Marais, J J (2004)
recorded response would mainly reflect those potentials generated in the
brainstem. Thus there is a significant influence by state of consciousness on the
estimation of hearing sensitivity, when using the 40 Hz ASSEP. Therefore, as
long as strict control is exercised when monitoring the adult participants’ state of
arousal, accurate and reliable responses can be evoked by the 40 Hz ASSEP, due
to the larger response amplitudes.
Scalp potentials recorded in response to tones amplitude modulated at frequencies
between 80 and 110 Hz probably derives mainly from a generator in the high
brainstem (John & Picton, 2000; Mauer & Döring, 1999). The use of ASSEP in
the assessment of hearing in sleeping participants has shown to be most effective
when employing responses with latencies of around 10 ms (Rickards et al., 1994).
Results from a study done by Cohen and colleagues (1991) indicate that rates in
excess of 70 Hz are required in order to produce these latencies.
2.3.6.3 Influence of Gender
Don, Ponton, Eggermont and Masuda (1993) interpreted the distinct genderrelated latency differences in terms of two effects, firstly the faster cochlea delay
times (associated with the shortened length, and greater stiffness of the female
basilar membrane) and secondly the length of the cochlear nerve. There might,
however, be a third contributing factor relating to the gender differences.
According to Goldstein and Aldrich (1999) the inherit degree of variability in
head size, found between the genders, could possibly contribute to gender-related
latency effects. No conclusive evidence (to the knowledge of the researcher, at
date of publication) across different carrier frequencies, have been established to
confirm any of these hypotheses, as very limited research is available on the
effect of gender on ASSEPs.
2.3.6.4 Hearing Sensitivity
The 40 Hz response has been used to obtain reliable frequency specific measures
that correlate well with thresholds obtained through behavioural audiometry, in
normal-hearing and hearing-impaired adults (Lynn, Lesner, Sandridge &
Daddario, 1984). These responses can be obtained across the frequency range
(Lynn et al., 1984; Stapells et al, 1984). The response, however, is considerably
47
University of Pretoria etd – Marais, J J (2004)
affected by state of arousal and state of consciousness (Brown & Shallop, 1982;
Osterhammel et al., 1985).
One of the most comprehensive studies on frequency specificity of the ASSEP in
estimating hearing thresholds in both normal and hearing-impaired participants,
highlighted the variability related to the hearing sensitivity in threshold
estimation (Rance et al., 1995). Threshold estimation in normal-hearing adults
were about 10 dB higher than the ANSI standards for pure tones (on average 12
dB), whereas thresholds in “well” babies have been estimated to be 10-15 dB
higher than for the normal-hearing adults (with the difference being greater at
lower frequencies). They (Rance et al., 1995) related these smaller responses at
lower frequencies to inadequate coupling between the earphone and the infants’
ear.
Thresholds
in
the
adult
hearing-impaired
population
showed
smaller
discrepancies between physiological and behavioural responses, than those of the
normal-hearing population, and furthermore, showed improved threshold
detection when compared to transiently evoked ABR elicited by 1 kHz tone pips.
Thus it can be postulated that ASSEP responses accurately reflect behavioural
audiograms in normal-hearing and hearing-impaired participants, although it
seems that the technique favors pathology (Rance et al., 1995). Later studies by
Rance et al (1998) indicated that there is a close relationship between the
responses obtained during behavioural and physiological testing, and more
importantly, that this relationship is consistent across the frequency range.
According to Picton and colleagues (1998) as well as Rance and colleagues
(1998) this effect is probably related to recruitment. Due to the effects of
recruitment in the hearing-impaired population, the responses are likely to reach a
level where it is recognizable at a specific intensity closer to threshold than in
normal-hearing participants. In other words, hearing-impaired participants with
severe sensory-neural impairments can have a more pronounced increase in
ASSEP amplitudes near minimum response levels (Rance et al., 1998).
48
University of Pretoria etd – Marais, J J (2004)
A recent study (Picton et al., 2002a) has concluded that physiological responses
are usually between 10-20 dB above behavioural thresholds (Herdman &
Stapells, 2001; Lins et al., 1996; Perez-Abalo et al., 2001), and that these figures
are representative of both normal-hearing and hearing-impaired participants.
Having established the application of ASSEP as a reliable and valid tool in the
accurate estimation of frequency specific hearing thresholds, within a population
who cannot comply with behavioural testing, the focus now shifts to a more
advanced application within the ASSEP arsenal. Subsequently the suitability of
sound-field presentation as a method for obtaining objective, frequency specific
hearing thresholds, will be argued.
2.3.7 Calibration
Calibration is carried out for two reasons. First, it establishes the relationship between
the sound pressure level (measured at some point) and the corresponding attenuator
setting on the signal generator. Secondly, it establishes the sound pressure level
corresponding to normal thresholds (audiometric zero) so that an individual clients’
hearing levels may be compared to this standard.
For all complex stimuli, when use in the sound field, the intensity at the subjects ear
will vary with time (Walker et al., 1984), because peaks and troughs in the room’s
frequency response are sequentially excited as the tone sweeps across the frequency.
Calibration levels of acceptable accuracy by the simple method of reading peak
deflections of a sound level meter set to “route mean square-fast” mode. When
calibration is performed in this way thresholds/responses obtained in the sound field,
with any type of complex stimulus, can be related to stimulus levels obtained under
earphone presentation with pure tones (Walker et al., 1984). This eliminates the need
to establish normative data for individual test environments (Walker et al., 1984). It is,
however necessary to use conversion figures when comparing thresholds or responses
obtained under different transducers. Therefore thresholds obtained under sound field
conditions should correspond to thresholds obtained under earphone conditions.
49
University of Pretoria etd – Marais, J J (2004)
2.3.8 Possible Advantages of ASSEP over other Techniques in the Estimation of
Aided Thresholds
After the Second World War, Carhart (1946) described and applied a so-called
objective technique for hearing aid selection (Hall & Ruth, 1985). According to his
technique four parameters, namely effective hearing aid gain, tolerance for amplified
sound, speech discrimination in quiet as well as in background noise had to be
measured. According to Hall & Ruth (1985) his method can be described as objective
because it relies on quantitative audiometric data rather than depending entirely on
subjective impressions from the patient.
Various procedures have been developed to ensure appropriate selection and fitting of
amplification. As these procedures have been extensively reviewed by previous
studies (Harford, 1979; Libby, 1985), they will not be discussed in depth in the
current study. Rather, the current study will discuss the advantages and disadvantages
of these techniques and compare them with the characteristics of the ASSEP
technique in an attempt to highlight possible advantages.
Several fitting formulas (Berger, 1976; Fletcher, 1952), such as the half-gain rule and
Bragg converter (Bragg, 1977) and the selective amplification technique (Victoreen,
1973; Watson & Knudson, 1940) have been used in the past. Since the 2cm3 coupler
and the modified Kemar coupler (Romanov, 1942; Wisniewsky, 1970) only offers
approximate real-ear conditions (Hall & Ruth, 1985), hearing aid performance often
differs substantially from these measures.
Even functional gain measures, obtained under sound-field conditions, incur certain
problematic influences. The characteristics of the sound differ substantially due to
individual differences in head diffraction, head shadow effect (Hall & Mueller, 1997)
and “body baffle” (Hall & Ruth, 1985). These individual differences, with specific
reference to the ear canal volume, diameter and length of the canal, are unique for
each person and are markedly different between adults and infants (Hall & Ruth,
1985). Therefore, valid and reliable measurement of hearing aid function is dependent
on the specific amplification device being fitted, the specific mould attached to the aid
50
University of Pretoria etd – Marais, J J (2004)
and the specific hearing-impaired individual being fitted. Although most hearing aid
fitting procedures take this factor into account (McCandless & Lyregaard, 1983;
Moodie et al, 1994; Seewald et al., 1985), this process is demanding even when
testing co-operative individuals. However, when infants and others within the
difficult-to-test population undergo fitting procedures these demands are aggravated
(Seewald et al., 1985).
Taking into account that these measures are reliant on accurate behavioural thresholds
(Stelmachowicz & Lewis, 1988) and that behavioural results obtained from infants
and other individuals, of all ages, within the difficult-to-test population are often
unreliable (Picton et al., 1998) it becomes evident why this process has been described
as “a matter of luck and intuition” (Picton et al., 1998:329). When considering that
early identification of hearing loss is only valuable if it leads to aggressive (Hall &
Ruth, 1985) and appropriate management, it seems that an objective technique that is
able to measure the benefits of appropriate hearing aid fitting is required.
Subsequently the use of auditory brainstem responses, as well as the possible
application of ASSEPs will be evaluated, on a comparative basis, to establish the most
appropriate “objective” technique for the selection and fitting of amplification
devices. The comparison will be based on the stimulus used, the response obtained,
the transducer employed and the state of consciousness of the individual.
Stimulus Characteristics:
As mentioned earlier, the use of real-ear measures is reliant on accurate behavioural
thresholds. However, this is not the only complication that inhibits the use of real ear
measures, and especially in the difficult-to-test population. Probe tube placement and
the maintenance thereof in infants and uncooperative patients can be extremely
challenging (Picton et al., 1998). Thus, other factors rather than the broadband
stimulation employed during the recording of real ear measures precludes these
measures from providing reliable data within the clinical setup.
While subjective measures such as functional gain are reliant on behavioural
responses for both unaided and aided measures, objective measures such as, for
example the auditory brainstem response are reliant on aided and unaided electro51
University of Pretoria etd – Marais, J J (2004)
physiological thresholds that do not require any behavioural responses from the
individual (Beauchaine & Gorga, 1988). Previous investigations (Beauchaine &
Gorga, 1988; Cox & Metz, 1980), using the auditory brainstem response (ABR)
employed click and tone burst stimulation and evaluated component latencies,
thresholds and slope of latency-intensity functions under aided and unaided
conditions.
According to Cox & Metz (1980) the latency shift of wave V, of the ABR, could be
used to select appropriate amplification. Their results indicated that the hearing aid
response that produced the shortest wave V latency also provided the best speech
discrimination score. However, in order to evaluate the validation of this statement it
is necessary to critically evaluate the type of stimulus used by the ABR.
Because click-evoked ABRs originate from the basal regions of the cochlea, it is
highly unlikely that the technique would be able to differentiate between aids that
employ various amounts of low frequency gain (Beauchaine & Gorga, 1988). Also
low frequency stimulation may cause artifacts that prevent frequency specific
measurements (Hall & Ruth, 1985; Kileny, 1982; Mahoney, 1985). Both click- and
tone burst stimulation, because of their short duration, is more likely to distort in both
the sound-field speaker and the hearing aid amplifier (Picton et al., 1998).
Other studies investigating the appropriateness of tone burst stimulation (Stapells et
al., 1995; Stapells et al., 1990) have indicated that although this technique shows
some ability to investigate low frequency hearing, it is also restricted due to its
maximum presentation levels (Rance et al., 1998). Thus, hearing losses that exceed
105 dB nHL cannot be investigated with tone burst ABR measures (Rance et al.,
1998) and in cases of severe to profound losses the absence of ABR responses at high
intensities cannot exclude hearing aid benefit (Beauchaine & Gorga, 1988).
Furthermore, click stimulation is dependant on the integrity of the basal regions of the
cochlea, thus implying that hearing sensitivity can only be estimated for the higher
frequency region (Beauchaine & Gorga, 1988). While tone burst stimulation provides
more frequency specific measurements, these measurements can be rather time
consuming (Beauchaine & Gorga, 1988) and may cause feedback that may last as
52
University of Pretoria etd – Marais, J J (2004)
long as 20 milliseconds after stimulus onset (Kileny, 1982). Subsequent research by
Serpanos, O’Malley and Gravel (1997) demonstrated that the correlation between
wave V latency and loudness appears to be low. Lastly, hearing aids handle rapidly
changing stimuli (such as click and tone burst stimulation) differently from more
continuous stimuli (such as speech) and it is difficult, therefore, to predict the steady
state characteristics of hearing aids from responses evoked by short duration stimuli
(Gorga, Beauchaine & Reiland, 1987).
In contrast, ASSEP testing using modulated tones offers significant advantages over
techniques that employ brief duration stimuli. Because the tones are continuous they
do not suffer from spectral distortion associated with brief duration stimuli such as the
click stimuli employed by the ABR (Rance et al., 1998). The ASSEP offers frequency
specificity (Rance et al., 1998), which allows for the generation of evoked potential
audiograms, accurately reflecting the configuration of the hearing loss
(Lins et al.,
1996; Rance et al., 1995). The continuous nature of the stimuli also offers
presentation level advantages, which relates to presentation levels as high as 120 dB
HL. Thus, even minimal amounts of residual hearing can be investigated with the
ASSEP technique (Rance et al., 1995). Various studies (Herdman & Stapells, 2000;
Perez-Abalo et al., 2001; Picton et al., 1998) have found a close relation between the
actual hearing thresholds and the threshold estimations obtained through ASSEP
testing.
In summary, electro-physiological measures provide information regarding the
perception of sound (HL) and not only regarding the signal at the eardrum (SPL).
ASSEP, furthermore, provides several advantages over other evoked potential
techniques such as the ABR, limiting the amount of useless responses (Rance et al.,
1998).
Response Characteristics:
Auditory brainstem responses are measured in the time domain. Recording of
responses incurs some problems because the hearing aid acts as a filter or amplifier
and will thus alter the time waveform and amplitude spectra of the stimulus
(Beauchaine & Gorga, 1985). As mentioned earlier in this section the hearing aid
might introduce electro-magnetic artifacts that can distort the response waveform.
53
University of Pretoria etd – Marais, J J (2004)
This phenomenon will be aggravated if the hearing aid is in close proximity to the
recording electrode (Beauchaine & Gorga, 1985).
As regular repeating stimuli evoke the ASSEP the response stabilizes after the initial
few stimuli (Picton et al., 1998). Thereafter the response contains constituent
frequency components that remain constant in amplitude and phase over time
(Regan, 1989). It logically follows then that when these responses are evoked by
amplitude modulated tones that are frequency specific and stable over time, the
likelihood of distortion in either the sound-field speaker or the hearing aid amplifier is
limited (Picton et al., 1998). Also, the modulated tones used to elicit ASSEPs are
similar to the warble tones use in behavioural assessment (Rance et al., 1995).
Again, the ASSEP offers some advantages over the ABR technique, in that response
detection is “double” objective. In other words no response is required from the
individual being tested and, no interpretation is required from the clinician conducting
the test (Rickards et al., 1994). Responses are measured in the frequency domain at
the frequency of modulation and some of its harmonics (Lins & Picton, 1995).
Response detection is done mathematically through either an f-test or t-test. ASSEP is
limited in its application to hearing sensitivity assessment, as there are no latency
measures to indicate retro-cochlear lesions. Nevertheless, the application of ABR for
the investigation of retro-cochlear pathology is becoming less viable due to the
introduction of CT- and MRI-Scans (Jerger, Grimes, Jacobson, Albright & Moncrieff,
2000).
Transducer Influence:
As discussed previously the brief nature of the stimulus employed during ABR testing
show a high susceptibility to distortion in both the sound-field speaker and the hearing
aid amplifier (Hall & Ruth, 1985; Kileny, 1982). Picton and colleagues (1998)
hypothesized that the likelihood of distortion in either the sound-field speaker or the
hearing aid amplifier is limited due to the continuous nature of the stimuli used to
evoke ASSEPs. This theory has since been substantiated by preliminary research
(Cone-Wesson et al., 2001; Picton et al., 2002a).
State of Consciousness of the Individual:
54
University of Pretoria etd – Marais, J J (2004)
Probe tube placement and maintenance thereof, associated with electro-acoustic
measures (e.g. real-ear measures), is challenging in infants and uncooperative patients
(Picton et al., 1998). Electro-physiological measures allow for assessment regardless
of the state of consciousness of the individual (Aoyagi et al., 1994; Cohen et al., 1991;
Lins & Picton, 1995; Rickards et al., 1994). The ASSEP therefore apply to difficultto-test populations as it provides recordable responses at low sensation levels in
sleeping and sedated subjects when modulation rates in excess of 70 Hz are used
(Cohen et al., 1991; Lins & Picton, 1995). It is also reliably present in children of all
ages including the neonatal population (Aoyagi et al 1993; Rickards et al., 1994).
Finally, preliminary research (Cone-Wesson et al., 2001; Picton et al., 1998) has thus
far supported the advantages over other techniques that have been highlighted in this
section.
2.4 Sound-field as a Possible Application within the ASSEP Domain
As previously mentioned in the introduction of this study, the initial treatment of a
hearing-impaired individual will typically involve the fitting of appropriate
amplification (i.e. hearing aids). In adults as well as in older children amplification
can be selected and adjusted according to the participants’ subjective responses, as
they are able to comply with audiometric testing. One of the most important factors to
consider when fitting an amplification device is what gain is derived from the device.
The term functional gain can be defined as the difference between audiometric
thresholds with and without the device (Hawkins & Haskell, 1982). When testing
infants, very young children and others within the difficult-to-test population the
clinician is faced with several challenges. The most prominent of these challenges
seems to be how to obtain aided and unaided thresholds when the population that is
being testing, is unable to provide reliable behavioural responses to auditory
stimulation.
Objective techniques such as real-ear coupler differences (RECD) and real ear
insertion gain (REIG) have been employed in order to obtain measures that are
equivalent to functional gain (Dillon & Murray, 1987). It is important to mention that
these measurements are only of value when the actual unaided responses of the
participant are known (Picton et al., 1998). There are further practical challenges
when using these techniques. Probe-tube placement and low noise levels, necessary
55
University of Pretoria etd – Marais, J J (2004)
for the measurement of Real Ear Insertion Gain (REIG), is very difficult to sustain in
uncooperative and very young participants (Picton et al., 1998). Thus, an objective
method to measure the benefits of amplification in participants with a hearing
impairment, who are unable to reliably respond to behavioural audiometry, would be
of great value.
2.4.1 Stimulus Characteristics
The stimulus characteristics of amplitude modulated tones, which are frequency
specific and stable over time, seems to be much less likely to be distorted during
amplification in either the sound-field speaker or the hearing aid (Picton et al., 1998;
Picton et al., 2002a). Results from the aforementioned studies revealed that ASSEP
responses to amplitude modulated tones with modulation frequencies between 80 and
105 Hz can be recorded when stimuli are presented through a sound-field speaker and
amplified by a hearing aid (Picton et al., 2002a).
The complex nature of stimuli used during ASSEP testing allows frequency specific
description of the client’s hearing loss, while its bandwidth allows for sufficient
uniformity when testing in the reverberant field. Results obtained through preliminary
research on the topic furthermore indicates that responses obtained during the unaided
condition in the sound-field allows for a close comparison to behavioural thresholds
obtained using earphones (Picton et al., 2002a)
2.4.2 Modulation Characteristics
Initial interest around Steady State Responses (SSR) centered on amplitude
modulation rates of 40/sec (40 Hz) responses, since Galambos et al (1981)
demonstrated increased amplitudes (two to three times greater), of responses, when
modulation rates of 30-50/second were used instead of 10/second (Linden et al., 1985;
Stapells et al., 1984). Various researchers (Dauman, Szyfter, De Sauvage & Cazals,
1984; Kankkunen & Rosenhall, 1985; Lynn et al., 1984; Sammeth & Barry, 1985;
Stürzebecher, Kuhne & Berndt, 1985; Szyfter, Dauman & De Sauvage, 1984)
obtained threshold determinations, using the 40 Hz technique, which were in
concordance with pure tone audiometry thresholds. When comparing physiological
with behavioural responses, thresholds varied between 9-35 dB (Rodriguez et al.,
1986).
56
University of Pretoria etd – Marais, J J (2004)
There was however some problems associated with the use of the 40 Hz potential, of
which the greatest seemed to be that response amplitudes decreased dramatically
when participants drifted into sleep (Cohen et al., 1991). The 80 Hz response, in
contrast to the 40 Hz response, is not significantly affected by the participants’ state of
arousal (Aoyagi et al., 1993; Cohen et al., 1991; Lins & Picton, 1995), and has short
apparent latencies in both sleeping and awake participants (Lins & Picton, 1995).
Research (Stapells et al., 1984; Lins & Picton, 1995) has noted the differential effect
of sleep on latencies as well as amplitude changes, when using the 40 Hz response
compared to the 80 Hz response. These differences seem to indicate that the two
responses have two distinct neural generators.
The 40 Hz Steady State still has its applications within the clinical setup. It can be
used successfully for objective audiometry (Rance et al., 1995) in “awake” adults and
serves as a method to monitor state of arousal during general anesthesia (Plourde &
Picton, 1990). Furthermore, the involvement of the auditory corti holds exiting
research possibilities (Schmulian, 2002), focused on the auditory processing of sound
rather than hearing (supra-threshold levels rather than threshold levels).
Various researchers (Regan, 1989; Regan & Regan, 1993; Rickards et al., 1994) have
done comparative studies on the effects of a lower versus higher modulation rate.
Clinical utility of the 80 Hz ASSEP grew due to several factors. It can be reliably
recorded in most populations, including infants (Rickards et al., 1994) and the
response does not significantly change with changes in state of arousal or level of
consciousness (Aoyagi et al., 1993). It, furthermore, offers frequency specificity
(Lins et al., 1996) and easy, objective response detection above 70 Hz (Cebulla et al.,
2001; Cohen et al., 1991).
Although reliable responses have been recorded when modulation rates ranging
between 85 Hz and 105 Hz were used, when stimuli are presented through a soundfield speaker (Picton, 2002a), it is uncertain whether reliable responses could be
obtained using modulation rates around 40 Hz. However, if responses could be
evoked from 40 Hz stimulation, these responses due to their increased amplitudes and
57
University of Pretoria etd – Marais, J J (2004)
subsequent easy detection offer advantages in “awake” adult patients within the
difficult-to-test populations.
2.4.3 Evoked Responses
According to Lins et al. (1996) the transduction process of the hair cells and auditory
nerve fibers involves compressive rectification of the signal waveform. Therefore the
compound electrical activity recorded from the cochlear nerve contains a spectral
component at the rate of modulation (Lins et al., 1996), and at two side bands
separated from this frequency by the frequency of the modulation signal.
It is furthermore important to remember that auditory neurons do not fire
synchronously in response to high frequency sounds, as is the case with low
frequency sounds (Kuwada et al., 1986). Auditory neurons do, however, fire
synchronously to low frequency envelopes of high frequency signals that are
amplitude modulated (Rees & Moller, 1983).
The evoked response (scalp-recorded response) reflects a combination of activity of
the Steady State Response (SSR), at the modulation rate, and electrical noise
produced by the brain and scalp muscles (Lins et al., 1996). A visual representation of
the compound electrical activity recorded is provided in Figure 2.1.
Figure 2.1 Illustration of the Electrical Activity involved in Evoked Responses
(Adapted from Picton, Dimitrijevic & John, 2002c)
58
University of Pretoria etd – Marais, J J (2004)
Because of the distinct energy displacement at the modulation frequency, and its first
and second harmonics, response detection can be analyzed accurately, objectively and
rapidly.
2.4.4 Automated Response Detection
Steady state evoked potentials are essentially sinusoidal responses to auditory stimuli.
When the evoking stimuli are reduced in intensity or when the participants’ hearing is
impaired, or when the background noise (electrical or acoustical) increases, these
responses become difficult to detect. Subjective analysis, by examiners, of waveform
and spectrum may lead to confusion regarding the presence or absence of a response
(Dobie & Wilson, 1996). Truly objective methods of response detection, with known
statistical power and false-positive rates, then attain critical importance.
Evoked responses are converted from the time domain to the frequency domain
through Fast Fourier Transform and are automatically detected in the frequency
domain by the F-test (Schmulian, 2002). Previous research (Dobie & Wilson, 1996)
has indicated that F-test is preferable to t-tests based on either root-mean-square
amplitudes or dB values. When a Fast Fourier Transform (FFT) is employed it allows
for the conversion of amplitude-time domain waveforms to the frequency domain, as
a series of cosine waves. These waveforms are represented as a vector on a two
dimensional plane (John & Picton, 2000). The responses to each carrier frequency can
be assessed by the amplitude and phase of the FFT component and its correspondence
to the frequency of modulation of the carrier.
Amplitudes are calculated as baseline to peak, whereas the phase of the responses was
measured as the cosine onset phase of the recorded wave (Herdman & Stapells, 2001;
Lins et al., 1996) when analyzing responses. The F-technique allows for objective
response detection as it evaluates whether a response at a specific frequency of
stimulation can be differentiated from the noise in the adjacent frequencies (Wei,
1990; Zurek, 1992). Response detection criteria are based on the “phase coherence”
technique employed by Galambos et al (1981), Jerger, Chmiel, Frost & Coker (1986)
and Stapells, Makeig and Galambos (1987).
59
University of Pretoria etd – Marais, J J (2004)
Also, because the energy of the responses are found at the modulation frequency and
two of its harmonics, instantaneous response detection is possible (Levi, Folsom &
Dobie, 1993; Lins & Picton, 1995; Lins et al., 1996. Figure 2.2 provides a
diagrammatical representation of automated response detection.
Figure 2.2 Representation of Automated Response Detection using the F-Test
(Adapted from Lins et al., 1996)
2.5 Conclusion
The need for a technique that can bridge the gap between diagnosis and rehabilitation,
especially within the difficult-to-test population becomes increasingly evident as the
global trend towards earlier identification and rehabilitation intensifies. Although the
exploration of other AEP techniques has made several advances in addressing this
need, their implementation into the clinical arena has been unsuccessful due to
inherent limitations.
Sound-field presentation as a possible application within the ASSEP domain, offers
several advantages over other techniques previously explored, which holds the
promise of finally addressing this crucial need. A summary of the advantages
discussed is presented in Table 2.3
60
University of Pretoria etd – Marais, J J (2004)
Table 2.3 A Summary of the advantages of ASSEP over other techniques
Characteristics
of:
a.) Stimulus
b.) Response
c.) Transducer
d.) Participant
state of
consciousness
Electro-Acoustic Techniques
Implications for clinical Hearing testing
1. Real ear
insertion gain
Broad band
noise
2. ABR
3. ASSEP
Transient:
Click or Tone
pips
Sustained:
Sinusoidal
Mixed
Modulated tones
Frequency
domain
Time domain
Frequency
domain
Amplitude and
frequency
measurements
Sound-field
Speaker
Latency and
amplitude
measurements
Sound-field
Speaker
Amplitude and
Phase
measurements
Sound-field
Speaker
Awake/ Asleep Awake/Asleep
in infants
Awake/Asleep
Electrophysiological measures provide information regarding perception of
sound (HL) and not only regarding the signal at the eardrum (SPL). The spectral
jitter associated with brief duration stimuli (ABR) is eliminated with the use of
more complex stimuli (ASSEP), limiting the amount of useless responses (Rance
et al., 1998).
Response detection is double objective (ASSEP) as no response is required from
the participant and no interpretation by the clinician (Rickards et al., 1994), and
may be observed at the frequency of modulation and its harmonics (Lins &
Picton, 1995). Subsequently no subjective interpretation required as with ABR.
ASSEP provides information regarding perception and not only SPL values at
the eardrum.
ASSEP application is limited to hearing assessment, as there are no latency
measures to indicate retro-cochlear lesions. ABR application for retro-cochlear
lesions is becoming less viable due to CT- and MRI-Scans (Jerger, 2000).
Brief nature of the ABR stimuli shos high susceptibility to distortion in the
sound-field speaker and the hearing aid (Kileny, 1982; Hall & Ruth, 1985).
Complex stimuli of the ASSEP do not distort in the sound-field speaker or
hearing aid (Picton et al., 2002a).
Probe tube placement and maintenance of its position, associated with REIG, is
challenging in infants and uncooperative participants (Picton et al., 1998).
Electro-physiological measures allow for assessment regardless of state of
consciousness (Aoyagi et al., 1994; Cohen et al., 1991; Lins & Picton, 1995;
Rickards et al., 1994)
61
University of Pretoria etd – Marais, J J (2004)
2.6 Summary
This chapter aimed to provide a theoretical framework as support for the empirical
research component. To achieve this aim the chapter was subdivided into three
sections. The first section focused on the application of Sound-field Audiometry and
its advantages and limitations. Comprehensive recommendations for conducting
Sound-field Audiometry were described, critically evaluated and discussed. These
recommendations have their foundation based on research data available and highlight
the most appropriate method for conducting Sound-field Audiometry.
The second section focused on ASSEPs. Although the final goal of the current
research endeavour has its focus on a more advanced application of ASSEPs, it
remains important to delineate the current application value thereof. The technique
was defined and discussed according to its physiological basis. A brief description of
its stimulation and modulation characteristics, as well as evoked responses and the
automated detection thereof, was provided. These discussions formed the basis for the
critical evaluation of Sound-field Audiometry as a possible application within the
ASSEP domain, which formed the last section of this chapter.
Finally the application of Sound-field Audiometry within the ASSEP domain was
critically evaluated. The focus was on audiometric variables that apply to ASSEP
testing, the current clinical application of ASSEPs, research applications, current
research available regarding the application of Sound-field Audiometry within the
ASSEP domain and the advantages over other techniques currently available.
62
University of Pretoria etd – Marais, J J (2004)
CHAPTER THREE: Research Methodology
Aim: To provide the method used to conduct the empirical
research component of this study.
3.1 Introduction
In chapter one the problem surrounding the current research endeavour was
introduced. It provided an orientation, described the terminology and presented an
overview of the content and organization of the study. Chapter two aimed to provide a
theoretical framework, as a support for the empirical research component. It specified
concepts and constructs and furthermore provided a critical evaluation and
interpretation of the relevant literature available.
This chapter provides the methodological approach implemented in conducting the
empirical component of the current study. The chapter is discussed according to the
synopsis presented in figure 3.1 below.
Main Aim
Sub Aim
Sub Aim
Sub Aim
Sub Aim
RESULTS:
CHAPTER
4
Data
Analysis
Data
Analysis
Procedures
Data
Processing
Data
Collection:
Data
collection
apparatus
Data
collection
procedures
Volunteers:
Research
Design
Selection
criteria
Selection
procedures
Apparatus for
determining
participants
Research
Participants:
Participant
description
Preliminary Study
Figure 3.1 Chapter Outline
63
University of Pretoria etd – Marais, J J (2004)
3.2 Aims of Research
Although previous studies (Aoyagi et al., 1996; Lins et al., 1995; Perez-Abalo et al.,
2001; Rance et al., 1998) have already established ASSEP as an objective clinical
procedure for hearing threshold estimation, across a range of frequencies, less
attention has been paid to aided threshold estimation in the difficult-to-test
populations. Preliminary studies (Picton et al., 1998; Cone-Wesson et al., 2001) have
shown promising initial results, when recording ASSEP using sound-field stimulation.
The need now arises to validate these preliminary findings as accurate and reliable.
Validation of these preliminary findings depends on further investigation of threshold
estimation through sound-field stimulation using a monotic ASSEP technique. The
current study centered on normal-hearing adults in an attempt to provide normative
data for the monotic ASSEP technique, through sound-field stimulation. The
importance of this technique, should it gain clinical application, becomes evident
when we consider the advantage for assessment of amplification, in the difficult-totest population (Picton et al., 2002a). For this purpose it is of critical importance that
the technique has to be standardized. Standardization of any technique is dependent
on normative data as the validation thereof. Therefore the main aim of the current
study is:
3.2.1 Main Aim
To compare minimum response levels obtained from the monotic ASSEP technique
using insert earphones and sound-field speaker presentation, in a group of normalhearing adults.
The following sub-aims were formulated in order to realize the main aim of the study:
3.2.2 Sub-aims
To compare pure tone-, frequency modulated (FM) and mixed modulated
(AM/FM) tone behavioural thresholds at 0.5, 1, 2 & 4 kHz under insert
earphone and sound-field speaker conditions.
64
University of Pretoria etd – Marais, J J (2004)
To compare minimum response levels at 0.5, 1, 2 and 4 kHz obtained with a
monotic ASSEP technique under insert earphone and sound-field speaker
conditions.
To compare the actual behavioural thresholds (for pure tone, FM and mixed
modulated tone stimulation) with the GSI AUDERA system’s accuracy in
predicting these thresholds by means of an algorithm for insert earphone and
sound-field conditions.
3.3 Research Design
In order to achieve the aims of the study, a comparative, within-group experimental
design (Graziano & Raulin, 2000) was selected.
Within-group designs are true
experimental designs and are quite similar to the more common pretest-posttest
control group design (Graziano & Raulin, 2000; Leedy, 1997). The within-group
design has three basic characteristics: each participant is tested under each
experimental condition, scores in each condition is correlated with scores in other
conditions and the critical comparison is the difference between correlated groups on
the dependant variable (Graziano & Raulin, 2000). For within participant designs,
participants are sampled from a target or accessible population, and each participant is
exposed to all the experimental conditions: “In essence, each participant serves as his
or her own control” (Graziano & Raulin, 2000; 248). Therefore, sampling errors are
eliminated, as all participants are tested under all conditions. Hence, the suitability of
this specific research design for the current study is emphasized.
Strengths and weaknesses of the design:
There is a similarity between within-participant designs and single group pretestposttest designs, in that the same participant is tested under each condition (Graziano
& Raulin, 2000). However in the pretest-posttest design, participants respond first to
the pretest and then the posttest. The within-group designs, in contrast, allow for
participants to be tested under each condition but with no fixed order of presentation
(Graziano & Raulin, 2000).
This order effect is particularly important and it
contributes to reducing the possibility of a practice effect occurring (Graziano &
Raulin, 2000).
65
University of Pretoria etd – Marais, J J (2004)
This design also holds important advantages as the same participants are tested in
each condition and subsequently there can be no group differences due to sampling
errors. This suggests that participants are equal at the start of the study. If participants
were not equal to start with it would be impossible to deduce whether the differences
were due to the experimental manipulation or to the pre-existing differences within
the group (Graziano & Raulin, 2000). Therefore, the larger the individual differences
between participants, the greater the benefit derived from within-participant designs,
as a within-participant design not only controls, but also eliminates the variance due
to individual differences, thereby reducing error variance (Graziano & Raulin, 2000).
Because research participants in the current study were selected from a predetermined
“quota” (discussed under 3.4.2), and not through random selection, this design holds
an important advantage as it eliminates individual differences, thereby relating the
deductions to experimental manipulation rather than inter-participant variability.
Another advantage of this type of design further contributes to efficiency. Because the
same participants are tested under several conditions instructions can be given once
instead of at the beginning of each condition, or at least will require only slight
modification for each condition. When instructions are complicated or when a
practice period is part of the instructions, timesaving could be considerable (Graziano
& Raulin, 2000). As research participants will be tested with both behavioural, as well
as electro-physiological techniques, and under both insert earphone- and sound-field
conditions, the expected test time will be considerable. As a result, increased
efficiency will contribute to minimizing participant fatigue, as well as test time.
Despite the advantages offered by this design, there are also disadvantages relating to
the within-group design. For example, with regard to the sequence effect, the same
participants are exposed to a number of different conditions. Thus, the experience
participants have in one condition might affect how they respond to the subsequent
conditions (Graziano & Raulin, 2000). Hence, if differences were found between the
conditions of an experiment, they may not be due to the manipulation of the
independent variable but to the compounding effects of one condition on later
conditions (Graziano & Raulin, 2000).
66
University of Pretoria etd – Marais, J J (2004)
Sequence effects can occur in within-participant designs and should be controlled by
varying the presentation of conditions (Graziano & Raulin, 2000). There are two
important types of sequence effects that are relevant to this study, namely practice
effects and carry-over effects.
Practice effects are caused by participants’ practice and growing experience as they
move through the successive conditions. Disadvantages because of practice effects are
not related to any particular condition, but rather to the participants’ growing
familiarity with the conditions. Practice effects can be either negative (e.g. due to
fatigue) or positive (due to familiarity). Carry-over effects are sequence effects due
to the influence of a particular condition, or combination of conditions, on the
responses following in subsequent conditions. Presentations of conditions were varied
so that each condition was represented in all of the possible orders of presentation, the
same amount of times.
This random order of presentation constitutes the only sure way to counteract for
practice- and carry-over effects. Table 3.1 below provides a visual representation of
how practice and carry-over effects were counterbalanced. Letters of the alphabet
(e.g. A, B, C etc.) represent various conditions. Thus, each participant (e.g. 1,2,3 etc.)
was tested with conditions presented in a different order, as represented in Table 3.1
below.
67
University of Pretoria etd – Marais, J J (2004)
Table 3.1 Counterbalancing of Practice- and Carry-over effects
Participants
Order of Presentation:
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
A
B
C
D
E
F
G
F
E
D
C
B
A
G
B
C
D
E
F
G
A
E
D
C
B
A
G
F
C
D
E
F
G
A
B
D
C
B
A
G
F
E
D
E
F
G
A
B
C
C
B
A
G
F
E
D
E
F
G
A
B
C
D
B
A
G
F
E
D
C
F
G
A
B
C
D
E
A
G
F
E
D
C
B
G
A
B
C
D
E
F
G
F
E
D
C
B
A
Fundamental to the experimental design is the nature of the relationship between the
dependent and independent variables. It is important to be aware of the variables in a
particular experiment as they may impact on the results of a study.
Figure 3.2 below, provides a schematic representation of the relationship between the
Independent-, the Dependent- and the Controlled variables within the experimental
setting in this study.
68
University of Pretoria etd – Marais, J J (2004)
Key for Figure 3.2:
Independent Variables:
i.Behavioural pure tone testing under insert earphones
ii.Behavioural frequency modulated (FM) tones testing under sound-field speakers
iii.Behavioural mixed modulated (AM/FM) tone testing under insert earphones
iv.Behavioral mixed modulated (AM/FM) tone testing under sound field speakers
v.A monotic auditory steady state evoked potential technique under insert earphones
vi.A monotic auditory steady state evoked potential technique under sound-field speakers
Measured or dependent variables:
a.
b.
c.
d.
Pure tone thresholds obtained at 0.5, 1, 2 and 4 kHz during behavioural pure tone testing under insert earphones
FM tone thresholds obtained at 0.5, 1, 2 and 4 kHz during behavioural pure tone testing under sound-field speakers
Thresholds obtained at 0.5, 1, 2 and 4 kHz during behavioural mixed modulated tone testing under insert earphones
Thresholds obtained at 0.5, 1, 2 and $ kHz during behavioural mixed modulated tone testing under sound-field
speakers
e. Minimum response levels obtained at 0.5, 1, 2 and 4 kHz during monotic ASSEP recordings under insert earphones
f. Minimum response levels obtained at 0.5, 1, 2 and 4 kHz during monotic ASSEP recordings under sound-field
speakers
Controlled variables:
I.
II.
III.
IV.
Test environment
Hearing ability of the participants
Age of the participants
Gender of the participants
I
i
ii
IV
iii
Experimental Condition
II
iv
v
V
vi
IV
III
a
b
f
c
d
e
Figure 3.2 Relationship between Independent-, Dependent- and Controlled
Variables
69
University of Pretoria etd – Marais, J J (2004)
As can be seen from Figure 3.2 there are 4 co-variables (age, gender, hearing ability
and test environment) that have an influence on the experimental setting. Within the
experimental setting there are also 6 independent variables (behavioural pure tone
testing under insert earphones, behavioural FM tone testing under sound-field speaker
presentation, behavioural modulated tone testing under both insert earphone and
sound field speaker conditions, and the monotic ASSEP technique, under insert
earphone and sound-field speaker conditions). And, finally, there are 6 dependant
variables (the thresholds relating to pure tone audiometry and the behavioural
thresholds and minimal response levels relating to monotic ASSEP technique) that
result from the experimental setting.
3.4 Volunteers
A total of 35 people volunteered to take part in the current study. Selection criteria
were stipulated for the selection of participants. These criteria will be discussed below
with specific reference to: informed consent, otoscopic examination, middle ear
functioning and behavioural pure tone results.
3.4.1 Selection Criteria:
The University of Pretoria’s Research and Ethics Committee was approached for
permission to conduct this study. (Appendix A). Once permission was obtained
certain selection criteria was set for participant selection. These criteria are discussed
with reference to the rationale behind each.
3.4.1.1 Normal-hearing
Firstly, all participants were required to have hearing thresholds within the normal
range (0-25dB HL) across all the frequencies tested, namely 0.5, 1, 2 & 4 kHz
(Roeser, Valente & Hosford-Dunn, 2000b). This criterion was stipulated, as the aim
of the current research was to collect normative data.
3.4.1.2 Normal Middle Ear Functioning
Participants were required to have normal middle ear functioning. Conductive hearing
losses and/or components have an influence on the peripheral hearing (Hall &
Mueller, 1997) and thus also indirectly on the results obtained from ASSEP,
70
University of Pretoria etd – Marais, J J (2004)
specifically the amplitude of the steady state response (Yantis, 1994a; Yantis, 1994b;
Hall & Mueller, 1997). Furthermore, normal-hearing sensitivity does not exclude the
presence of middle ear pathology (Martin, 1997; Stach, 1998). Participants were
subsequently required to have normal otoscopic results as well as normal middle ear
functioning.
3.4.1.3 Age Distribution
Participants were required to be between the ages of 15 and 40 years of age. This
stipulation was made even though auditory evoked potentials can be elicited from
persons of any age, from neonates through to people in their ninth decade
(Abramovich, 1990; Hecox & Galambos, 1974). There are however some
characteristics that do change with age such as latencies, amplitudes and general
configuration (Goldstein & Aldrich, 1999). The lower age range was selected, as there
are certain differences between neonatal and adult AEPs, especially concerning low
and high frequency sensitivity (Wolf & Goldstein, 1980).
3.4.1.4 Gender Distribution
The last criterion was that equal numbers of male and female volunteer participants
were selected. This criterion was considered important as head size may affect
latencies and amplitudes recorded for evoked responses (Watson, 1996; Goldstein &
Aldrich, 1999). Some researchers believe that head size rather than intrinsic gender
differences account for the different latency values in men and women (Dempsey,
Censoprano and Mazor, 1986). Thus, in order to counteract the possibility of
differences due to head size (in particular, in women), participants were selected to
reflect a relatively even gender distribution.
3.4.2 Selection Procedures
Volunteer participants were selected, through non-probability quota sampling (Moore,
2000; Neuman, 1997). According to Moore (2000) quota sampling implies that people
will not have an equal chance at being selected, but instead that the researcher decides
which types of people he/she would like to select (Moore, 2000). Participants were
selected according to the selection criteria, as well as their availability (time
constraints on the part of the volunteers). All volunteer participants underwent the
71
University of Pretoria etd – Marais, J J (2004)
following participant selection procedures in order to establish their suitability for
participation in the study.
The researcher telephonically contacted participants volunteering for the project and a
suitable time was arranged for discussion of informed consent. All volunteers were
briefed on the non-invasive nature of the procedures and the time involved in the
execution of the procedure. The objective of the study was highlighted and volunteers
were assured that all biographical information would be treated confidentially.
Volunteers were furthermore informed that they would be able to obtain copies of the
results should they request them. Once informed consent was obtained, a suitable time
was made for the selection criteria testing, in order to establish which of the
volunteers would be selected as participants during the study itself. A copy of the
informed consent form can be found in Appendix B.
The first examination for participant selection criteria was an otoscopic examination
of the external meatus. This examination was conducted to identify any possible
pathology that could cause a conductive hearing loss (Ginsberg & White, 1994). In
order to pass the otoscopic examination all participants were required to have at least
an identifiable light reflex, while position, color and transparency of the tympanic
membrane was also taken into consideration (Silman & Silverman, 1991).
Immediately after meeting the minimum criteria for the otoscopic examination,
estimation of acoustic immittance at the tympanic membrane as a function of air
pressure was investigated. This procedure is known as tympanometry. Participants
were required to have type A tympanogram (Martin, 1994). The normative data for
type A Tympanograms can be found in Appendix C.
Finally, pure tone audiograms were obtained form all volunteers to determine whether
they had hearing sensitivity within the normal range. The normative data relating to
hearing sensitivity can also be found in Appendix C. Those volunteers that did not
meet the selection criteria were debriefed and referrals were made where appropriate.
72
University of Pretoria etd – Marais, J J (2004)
3.4.3 Selection Apparatus
The otoscopic examination of the external meatus and tympanic membrane
was performed with a Heine Minilux 2000 otoscope.
A tympanometric evaluation of the middle ear was performed using the GSI33 Middle Ear Analyzer. The apparatus met the requirements set out by the
South African Buro of Standards (SABS) and was calibrated in January 2002.
Pure tone behavioural thresholds were obtained using the GSI-61 Clinical
Audiometer. The audiometer is fitted with Bio-Logic Earlink Foam Eartips
for insert earphones. Behavioural pure tone stimuli were available in steady
tones for the insert earphones and in warble tones for the sound-field speaker
stimulation. The apparatus meets with the requirements set out by the SABS
and was calibrated in January 2002. The audiometer is housed in a singlewalled soundproof booth.
3.5 Participants
Following the selection criteria, 25 participants were selected according to nonprobability quota sampling. These participants took part in the research endeavour and
their behavioural pure tone thresholds, behavioural FM thresholds, behavioural mixed
modulated thresholds as well as their minimum response levels formed the data basis
of the current study. A description of the participants selected will subsequently
follow.
3.5.1 Description of Participants
The total sample (25 participants/50 ears) consisted of 13 females (52%) and 12 males
(48%). Participant age ranged between 15 and 36 with a mean of 23.4. The median
was 22.5 and the mode 22. The standard deviation was 4.3. All of the participants
met the requirements stipulated in the 3.4.1 (criteria for the selection of participants).
All the participants were selected according to non-probability quota sampling.
Participants were coded as N1-N25. In all 50 ears were tested in each of the test
conditions. A graphical representation of age distribution is provided in 3.3 below.
73
University of Pretoria etd – Marais, J J (2004)
30-34 years
4%
35-39 years
4%
15-19 years
20%
25-29 years
24%
20-24 years
48%
Figure 3.3 Age Distribution of the Participants according to age intervals (in
years) and percentages
3.6 Preliminary Study
Before the data for the current research endeavour was collected, a preliminary study
was performed in order to determine clinical accountability of the stimulus
parameters. Two participants volunteered to be part of the preliminary study. Both of
these participants met the criteria stipulated under criteria for the selection of
participants (3.4.1).
3.6.1 Aim:
To determine the appropriate stimulus parameters for the monotic ASSEP
technique
To determine the threshold criteria for electro-physiological testing
To determine the required time for the completion of the test procedures for a
single participant, in order to plan the data collection procedures.
To determine the appropriate placement of the test equipment and positioning
of the participants
3.6.2 Procedures
Determination of the Stimulus Parameters for the ASSEP Technique
Electrodes were fixed with electrolytic paste to the scalp at Fz (positive), Test
ear (negative) and non-test ear (ground). Impedance levels were maintained
below 7000 Ohms. The bio-electric activity was amplified with a gain of 100
000 and analogue filtered between 30 and 300 Hz. The notch filter was
activated at 50 Hz to avoid line interference. Sixty-four samples were
74
University of Pretoria etd – Marais, J J (2004)
averaged in a response, as per manufacturer specification. Response
significance was monitored with the vector view window and the probability
curve window. The presence of a response was determined using the F-test for
hidden periodicity in order to test the amplitude of the spectrum at each
modulation frequency against the 120 adjacent bins for significant amplitude
difference. Artifact rejection was performed with shorter epoch sections of 512
points. Amplitude was used as the criterion for rejection. A rejection level of
50 micro Volts was specified to reject any responses with amplitudes greater
than the specified value. A no-response was annotated after 40 epochs while
the minimum response level for each frequency in each ear was taken as the
threshold.
The recordings for both volunteers revealed minimum response levels that
closely approximated their behavioural thresholds (20-30 dB HL mean
difference to the behavioural threshold). Responses were readily identifiable
with both the 40 Hz and 80-110 Hz modulation techniques, when these
techniques were appropriately varied (depending on the volunteers’ state of
consciousness). No modifications were made to these parameters during the
data collection of the current study.
Determination of Threshold Criteria for Electro-Physiological Testing
All stimulation commenced at supra threshold level. Initial intensity was
presented at 50 dB HL, as prescribed by Swanepoel (2001). A descending
threshold seeking procedure was implemented, and intensities were decreased
in 10 dB steps. When a no-response was recorded the intensity level was
raised in 5 dB steps, until a response was recorded. The minimum response
level was taken as the intensity level where the last response was recorded.
Due to time constraints the lowest intensity level where a minimum response
level was recorded, was accepted as the electro-physiological threshold,
regardless of whether testing at lower intensity levels revealed a no-response
or a noise response, during the data collection of the current study. Also,
instead of commencing with stimulation at 50 dB HL, electro-physiological
testing commenced at 20 dB HL above the behavioural pure tone threshold
75
University of Pretoria etd – Marais, J J (2004)
(obtained for the specific transducer), during the data collection of the current
study.
Determination of the Time required for the Completion of all the Test
Procedures
In order to schedule appointments with the future participants a test battery
was performed during the preliminary study to determine the amount of time
that will be required for the assessment of each participant. As seen in Table
3.2 the complete assessment lasted for approximately two hours for each
participant.
Table 3.2 Approximate time required for the assessment of a single participant
Test Procedure
Otoscopic examination and Tympanometry
Behavioural pure tone Audiometry (using insert
earphones and sound-field speaker presentation)
Behavioural monotic ASSEP thresholds (using insert
earphones and sound-field speaker presentation)
Monotic ASSEP minimum response levels (using insert
earphones)
Monotic ASSEP minimum response levels (using soundfield speaker presentation)
Test time
5 minutes
15 minutes
TOTAL TESTING TIME
1 Hour and 55 minutes
15 minutes
40 minutes
40 minutes
From the above table (3.2) it is clear that the otoscopic examination,
tympanometry and all the behavioural thresholds were obtained in
approximately 35 minutes. The researcher recorded these times. Minimum
response levels obtained with the monotic ASSEP technique took
approximately 40 minutes when the insert earphones were used, as well as
when the sound-field speaker was used. The software recorded the time of the
recording (although not the time involved in participant preparation).
Determination of appropriate placement of the test equipment and
positioning of the participants
The preliminary study was conducted in an audiometric test room (the same
audiometric room was later used for the data collection). Figure 3.4 shows the
76
University of Pretoria etd – Marais, J J (2004)
plan of the test room, the placement of the speakers as well as the test
positions for both the behavioural measures and the electro-physiological
measures.
A1
Bed for electro-physiological testing
B2
C
ERA Test
B1
A2
Door into audio
test room
A1 – Sound-field speaker
A2 – Sound-field speaker
B1 – Test position during behavioural sound-field
testing
B2 – Test position during electro-physiological testing
C - Sound-field speaker for electro-physiological
testing, attached to height and directional
adjustable arm
Figure 3.4 Plan of the test room showing the lay out of the room, the position of
the speakers as well as the test positions for behavioural and electrophysiological measures
The test room was an audiometric test booth, “Controlled Acoustics Environment”
with internal floor dimensions 2.55m x 2.75m and height 2m. The booth is equipped
77
University of Pretoria etd – Marais, J J (2004)
with two sound-field speakers attached to the wall. The speakers are positioned at 45º
angles from the wall, facing the test position. The cone of each speaker is roughly at
head level of a seated adult, 1.2 m from the floor (Walker et al., 1984). The test
position during behavioural sound-field measures was exactly 1 m away from the
speaker. The position was identified for each participant individually, so that the test
point was at the centre of the participants’ head (Beynon & Munro, 1995).
Participants were required to minimize their head-movements as far as possible
(Walker et al., 1984) and these movements, if present, were monitored by the
audiologist. If there was any concern that the positioning of the participants’ head was
altered the test point was re-identified. The seat used, was height adjustable, to ensure
that the participants head was at the centre of the speakers’ cone.
Simple control measurements were made of the sound level of the FM tones at the
reference point and at points 10 cm from the reference point along longitudinal,
transverse and vertical lines at all test frequencies (0.5, 1, 2 & 4 kHz). The
measurements were made with a modulation frequency of 20 Hz and frequency
deviations 4 and 25%. The non-test ear was fitted with a wax plug during
measurements (Picton et al., 1998). Hearing thresholds were determined using an
ascending method of testing, with 10 dB increments until a response from the
participant was obtained and then decreased in a 5 dB increment. This procedure was
followed until a minimum response level was obtained 50% of the time. This level
was taken as the audiometric threshold for that specific frequency.
Electro-physiological measurements were made while the participant was lying,
comfortably on a bed. The sound-field speaker was mounted on a movable, heightadjustable arm. The speaker was placed directly in line with the centre of the
participants’ head, at exactly 0.5 m away. The speaker was then secured in place.
Participants were instructed to try and minimize head-movements (Walker et al.,
1984). The audiologist monitored the participant throughout the measurements to
ensure that the test position was accurately maintained.
Determination of appropriate calibration
Test stimuli were calibrated using procedures outlined by Walker et al. (1984), ISO
DP 8253/2 (1989), Beynon and Munro (1995), ANSI S3.6 and Picton et al. (1998) for
78
University of Pretoria etd – Marais, J J (2004)
quasi free field conditions. The sound field was determined for each FM stimulus with
the research participant and chair absent. The sound pressure level (SPL) was
measured at the reference test point and at six neighboring positions (15cm anterior,
posterior, to the left, to the right, above and below the test point). The stimuli was
measured using a Brüel and Kjaer Model 2209 sound level meter with a 1 inch Model
4145 condenser microphone. The mean of the seven SPL measurements was used as
the calibrated level of each stimulus. The 1 kHz and 2 kHz stimuli showed a level
5 dB greater than that of the 0.5 kHz and 4 kHz stimuli, for equivalent electrical input
to the speaker, as was the case in the study conducted by Picton et al. (1998).
The standard deviation values of the seven SPL measurements were used to estimate
the spatial variability of the sound field for each stimulus. Standard deviations were
similar to those obtained by Picton et al. (1998) and Beynon and Munro (1995). For
the stimuli used during the current research project standard deviations, as an estimate
of the error in the threshold measurement arising from the sound field variability was
+/- 3.9 dB (0.5 kHz), 3.8 dB (1 kHz), 2.5 dB (2 kHz) and 2.4 dB (4 kHz).
3.7 Data Collection
3.7.1 Data Collection Apparatus
•
Pure tone thresholds were obtained from the participants using a GSI 61 Clinical
Audiometer, calibrated January 2002. Pure tone stimuli were presented in steady
tones through Bio-Logic Earlink Foam Eartips for insert earphones and FM
(warble) tones for sound-field presentation, in a single- walled soundproof
booth.
•
Monotic ASSEP recordings were obtained with the GSI AUDERA system
(School of Audiology, University of Melbourne, Australia), an AUDERA
prototype. The equipment consists of a specialized Software component connected
to a Pentium Laptop Computer, a serial cable, an GSI AUDERA system Unit, a
fiber-optic cable, an EEG amplifier, tube phones and electrodes. The system is
operated by a software package specifically designed for the acquisition and
analysis of auditory evoked responses (AER) at varying frequencies and sound
levels as well as simultaneous analysis of patient EEG activity for evidence that an
79
University of Pretoria etd – Marais, J J (2004)
evoked potential has occurred. Calibration of the GSI AUDERA system was
performed in January 2002.
•
The AER measurements were obtained in a single-walled soundproof booth
using Bio-Logic Earlink Foam Eartips for insert earphones and a MS-697
BASS speaker for sound-field presentation to present acoustic signals while
participants were lying on a bed.
3.7.2 Data Collection Procedures
Six sets of data were collected from each participant, behavioural pure tone and mixed
modulated (AM/FM) tone thresholds under insert earphones, behavioural frequency
modulated (FM) tone and mixed modulated (AM/FM) tone thresholds under soundfield speakers, monotic ASSEP minimum response levels under insert earphones, and
lastly monotic ASSEP minimum response levels under sound-field speakers. Data for
each participant was collected on the same day. Behavioural pure tone thresholds
were obtained first, as part of the selection criteria, followed by the various ASSEP
recordings. Data collection was performed at the Department of Communication
Pathology at University of Pretoria.
Data Collection using Behavioural Pure and Behavioural FM Tone
Audiometry
These tests were conducted to determine the peripheral hearing acuity. The pure
tone and FM tone thresholds were used as a basis for comparison with the
thresholds obtained from mixed modulated stimuli. The comparison between
thresholds obtained under different stimulus presentations ultimately served as
the basis for the determination of the influence of insert earphones compared to
sound-field presentation on threshold estimation, when using a monotic ASSEP
technique. Thresholds were obtained for 0.5, 1, 2 and 4 kHz in each individual
ear, once normal middle ear functioning had been established through otoscopy
and tympanometry.
The frequencies were chosen to provide correlation with the monotic ASSEP
data, obtained at the same frequencies, and ultimately for data analysis.
Participants were classified as normal-hearing when all the thresholds at 0.5, 1, 2
and 4 kHz were equal to or less than 25 dB HL (Roeser et al., 2000a). Thresholds
80
University of Pretoria etd – Marais, J J (2004)
were obtained using an ascending intensity step, starting at –10 dB, of 10 dB and
a descending step of 5 dB until 50% accurate responses were obtained. This
procedure was followed for insert earphone presentation as well as sound-field
conditions.
Thresholds obtained through insert earphone presentation were obtained using
steady tones, while thresholds obtained through sound-field presentation were
obtained using FM (warble) tones, with the participant at 100 cm from the
speaker at an angle of zero degrees (0˚). The use of FM tones during sound-field
presentation limits the chance of free standing waves having an influence on the
results obtained.
Data Collection using Behavioural Mixed Modulated Tones
Mixed modulated (AM/FM) thresholds were obtained from the participants using
the GSI AUDERA (AUDERA prototype) system, calibrated in January 2002.
Stimuli were presented through insert earphones and a height adjustable sound
field speaker, in a single- walled soundproof booth. This test was conducted to
determine the peripheral acuity for modulated stimuli. The peripheral hearing
acuity for mixed modulated stimuli together with the peripheral hearing acuity for
pure tones and FM tones forms the basis of comparison form where the influence
of insert earphones compared to sound field presentation on threshold estimation
can be argued.
Thresholds were obtained at 0.5, 1, 2 and 4 kHz in each ear individually.
Thresholds were again obtained using an ascending intensity step method, starting
at -10 dB HL, in increments of 10 dB ascending and 5dB descending until 50%
accurate responses were present. This procedure was again followed for insert
earphone as well as sound field conditions.
Data Collection using the monotic ASSEP
Research participants who passed all the selection criteria were orientated as to
the procedure of the monotic ASSEP testing for this specific study. Definite
instructions as to the required behaviour during testing were discussed with each
participant.
81
University of Pretoria etd – Marais, J J (2004)
Preparation of participants involved cleaning the skin surface with NuPrep ECG
& EEG Abrasive Skin Prepping Gel, in order to clean the areas that was used
for placement of the electrodes (Mastoid and different scalp areas). A 15cm
cotton bud was used to apply the Nuprep. The skin was thereafter wiped with
gauze, after which Preptic-isopropyl-alchohol-blotters were applied, to clean
the skin areas again.
A conductive medium, namely Electro Gel, ECG Conductive Electrode Gel,
will be applied to the silver-chloride disk electrodes before application to the
scalp. After the placement of the electrodes, both ears were fitted with insert
earphones, namely Bio-Logic E-A-R Link Foam Ear Tips for 3A Insert
Earphones. These earphones served as the transducer through which the sound
stimuli will be presented.
The responses were elicited by using a Monotic
ASSEP technique, which incorporates the use of mixed modulated (AM/FM)
tones.
Once results were obtained for the monotic ASSEP technique using insert
earphone presentation, they were removed. The non-test ear was occluded and
sound- field presentation commenced with the speaker 50 cm from the
participants face at an azimuth of zero degrees (0˚). The same procedure was
followed in order to determine minimum response levels for both ears under
sound field conditions.
Determination of Stimulus Parameters for Pure tone and FM tone
Behavioural Audiometry
Four frequencies will be used as comparative reference points between the pure
tone behavioural data (insert earphone thresholds and sound-field thresholds) and
the monotic ASSEP estimated thresholds (for insert earphones and sound-field
presentation). Test stimuli included pure tones and FM tones for the behavioural
thresholds as well as for the carrier frequencies of the ASSEP at 0.5, 1, 2 and 4
kHz. These frequencies were chosen to ensure that the responses obtained from
the ASSEP testing, as well as the conclusions drawn from this study, had clinical
82
University of Pretoria etd – Marais, J J (2004)
value and were comparable, as they provide both high- and low frequency
information central to speech discrimination.
Specification of Stimulus Parameters for the Monotic Sequential Frequency
Auditory Steady State Response
The specifications of the stimulus parameters for the Monotic Sequential
Frequency Auditory Steady State Evoked potential technique will be discussed
below.
a.) Selection of carrier and modulation frequencies
Mixed-modulated tones with selected carrier frequencies, 0.5, 1, 2, and 4 kHz
were modulated at 40 Hz for “awake” and at 80-110 Hz for “sleeping”
participants. Multiple carrier frequencies were selected in order to provide, lowand high frequency information, central to speech discrimination, and also to
enable comparison between the ASSEP and the pure tone behavioural audiogram.
Carrier frequencies will be 100% amplitude modulated and 10% frequency
modulated at 40 Hz or between 80-110 Hz, depending on the state of
consciousness of the participant. Faster modulation rates were used for
participants that were awake, because of their resilience towards state of
consciousness (Lins et al., 1995).
b.) Selection of the stimulus intensity and threshold criteria
All stimulation for the estimation of behavioural thresholds commenced at –10
dB HL, while the estimation of monotic ASSEP commenced at 20 dB HL above
the behavioural threshold for that specific frequency. Steady state stimulus
intensity for the experimental study commenced at 20 dB HL above the
behavioural thresholds obtained during behavioural pure tone tests, as previous
research (Swanepoel, 2002), as well as subjective feedback from participant
during the execution of the preliminary study, showed uncomfortable loudness
levels at 70 dB HL. According to Rickards (2001) uncomfortable sound intensity
levels increase the EEG noise in the recording, probably due to muscle artifacts.
83
University of Pretoria etd – Marais, J J (2004)
An ascending threshold seeking procedure was used, during behavioural testing,
in steps of 10dB HL until a response was present; the intensity was then lowered
in steps of 5dB HL, until no response is present.
During electro-physiological testing, stimulation commenced at 20 dB HL above
the estimated behavioural threshold for that specific frequency. Stimulus intensity
was decreased in steps of 10 dB HL if a response to stimulation was recorded. If a
response to stimulation was not recorded, however, the intensity was increased
with 5 dB HL until a response to stimulation was present. The minimum response
level was identified as the lowest intensity where a response was recorded,
irrespective of whether a decrease in intensity revealed a no-response or a noiseresponse.
Table 3.3, below, contains a summary of the recording procedures for monotic
ASSEP through both sound-field speaker and insert earphone presentation.
84
University of Pretoria etd – Marais, J J (2004)
Table 3.3 Monotic ASSEP Recording Procedure
Recording procedure for the determination of minimum response levels
Using Insert Ear Phones
Electrode Placement
Impedance Values
Transducer Type
Participant
Positioning
Initial Stimulus
Intensity
Amplification and
Filtering
Using
Sound-field
Speaker
Presentation
Electrode discs of Ag/AgCI will be fixed with electrolytic paste to the scalp at Fz
(positive), Test ear (negative) and Non-test ear (ground)
Impedance values were kept below 7000 Ohms
Bio-Logic Earlink Foam Eartips for 3A Multi Media Sound-field Speakers MSinsert earphones will be used to present 697 was used to present stimuli
the stimuli
Participants were asked to lie on a bed,
Participants will be asked to lie in a bed in in a single walled soundproof booth.
a single walled soundproof booth and will The sound-field speaker was positioned
be encouraged to relax.
50cm away from the participant at 0º
azimuth.
Stimuli were presented monotically and sequentially at supra threshold intensities, as
determined in the preliminary study.
Bio-electric activity will be amplified at a gain of 100 000 and analogue filtered
between 30-300 Hz
Filter
The notch filter will be switched on at 50 Hz to eliminate line interference
Averaging
Sixty-four samples were averaged in a response. No less than 10, and no more than
40 epochs of 8 192 samples (digitized with a sampling period of 1.37 ms) each, will
be averaged in a response
A Fast Fourier Transformer (FFT) was calculated “online” for each long epoch
averaging the response spectra continuously. The presence of a response was
determined by using the F-test for hidden periodicity in order to test the amplitude of
the spectrum at each modulation frequency against the 120 neighbouring bins of
significant amplitude difference
Response Detection
Artifact Rejection
Minimum
Level
Artifact rejection was carried out with shorter epoch sections of 512 points. A
rejection level of 50 mV was specified to reject any response with amplitudes greater
than the specified value
Response Minimum response levels were established in descending intensity steps of 10 dB,
until no response was present. A no-response, however, could only be determined
after 40 epochs had been collected and averaged.
3.8 Data Analysis Procedures
The raw quantitative data was prepared and organized into a data set. All measures
will reflect results in dB HL. A statistician at the department of Information Sciences
analyzed the results. Data analysis procedures and the rationale for the analysis
method of choice are presented in Table 3.4 and 3.5 below.
85
University of Pretoria etd – Marais, J J (2004)
Table3.4 Data Analysis Procedures and the Rationale for Analysis Method of Choice
MAIN AIM OF THE RESEARCH ENDEAVOUR
The aim of this study was to determine the influence of insert earphones and sound-field speaker presentation on minimum response
level detection using a monotic steady state evoked potential technique, in a group of normal-hearing adults.
Literature
Sub-aims
Nature of Analysis Method
(http://ubmail.ubalt.edu/~harsham/stat-data/opre330.htm;
the Data
http://www.collegeboard.com/about/news_info/cbsenior/yr2002/htlm/define.html)
(*Oxford Dictionary, 1990:737)
1.) To determine
pure tone and
amplitude
modulated
behavioural
thresholds at 0.5, 1,
2 and 4 kHz using
insert earphone and
sound-field speaker
presentation.
Numerical
values in
dB
a.) Determine means, medians,
percentile ranks and normal
curves for each technique at
each frequency.
b.) Perform ANOVA analysis
on data obtained through
both techniques.
a.) Means provide the arithmetic averages, while the median value
indicates the middle value of a series of values arranged in order of
size *. The percentile rank indicates the number of participants who
fall below a particular scaled score. All these analysis methods
together provide information regarding the distribution.
b.) Analysis of variance allows the researcher to test the difference
between two or more means, by examining the ratio of variability
between conditions as well as within conditions
2.) To determine
minimum response
level detection at
0.5, 1, 2 and 4 kHz
using a monotic
ASSEP technique
using insert
earphone and soundfield speaker
presentation.
Numerical
values in
dB
a.) Determine means, medians,
percentile ranks and normal
curves for each frequency.
b.) Perform ANOVA analysis
on data obtained through
both techniques.
c.) Perform paired t-tests and
Pearson Correlation on data
obtained from both
techniques
a.) Means provide the arithmetic averages, while the median value
indicates the middle value of a series of values arranged in order of
size *. The percentile rank indicates the number of participants who
fall below a particular scaled score. All these analysis methods
together provide information regarding the distribution.
b.) Analysis of variance allows the researcher to test the difference
between two or more means, by examining the ratio of variability
between conditions as well as within conditions
c.) The t-test indicates whether the means are different while the
Pearson correlation indicates whether the judgments are otherwise
consistent
86
University of Pretoria etd – Marais, J J (2004)
Table 3.5 Data Preparation, -Analysis and the Rationale for Analysis Method of Choice
MAIN AIM OF THE RESEARCH ENDEAVOUR
The aim of this study was to determine the influence of insert earphones and sound-field speaker presentation on minimum response
level detection using a monotic steady state evoked potential technique, in a group of normal-hearing adults.
Sub-aims
Nature of the Analysis Method
Literature
Data
3.) To determine the
accuracy of the GSI
AUDERA system in
predicting the pure
tone thresholds
obtained from the
participants
Numerical
values in dB
(http://ubmail.ubalt.edu/~harsham/stat-data/opre330.htm;
http://www.collegeboard.com/about/news_info/cbsenior/yr2002/htlm/de
fine.html) (*Oxford Dictionary, 1990:737)
a) Determine means, medians, a) Means provide the arithmetic averages, while the median
percentile ranks and normal
value indicates the middle value of a series of values
curves for each frequency.
arranged in order of size*. The percentile rank indicates the
b) Perform ANOVA analysis on
number of participants who fall below a particular scaled
data obtained through both
score. All these analysis methods together provide
techniques.
information regarding the distribution.
c) Perform paired t-tests and b) Analysis of variance allows the researcher to test the
Pearson Correlation on data
difference between two or more means, by examining the
obtained from both techniques
ratio of variability between conditions as well as within
conditions
c) The t-test indicates whether the means are different while
the Pearson correlation indicates whether the judgments are
otherwise consistent
87
University of Pretoria etd – Marais, J J (2004)
3.9 Data Processing Procedures
A statistician at the Department Information Management of the University of
Pretoria was consulted during the planning of this study. Collected data was tabulated
and was made available to the statistician in the form of raw data. The following
procedure was followed:
Results from the audiological test battery was analyzed and interpreted as normal
or abnormal, so as to fit in with the selection criteria.
The results from the monotic sequential ASSEP was analyzed by the GSI
AUDERA system’s algorithm.
Data was summarized in a table, specifying participant numbers, ear tested, and
gender of the participant. Also included were the technique used, the frequency
tested and the thresholds and/or minimum response level obtained.
Data was delivered to the statistician, results were discussed and objectives for the
analysis of data were determined.
3.10 Summary
This chapter provided a description of the procedures implemented in the research
methodology to obtain the data according to the sub-aims of the current study. This
was done in order to achieve the main aim of the study. The need for a normative data
basis regarding the influence, if any, of insert earphones and sound-field speaker
presentation on threshold estimation when using a monotic ASSEP technique, was the
motivation behind the current study. The volunteers were discussed in terms of
criteria for selection, procedures involved in the selection and apparatus used for the
selection of participants. Subsequently a description of the selected participants was
provided. A discussion surrounding the preliminary study provided insight into the
data collection section of this chapter. Data collection was discussed in terms of the
apparatus used and the collection procedures used. The chapter concluded with an
overview of the data analysis and the processing thereof.
88
University of Pretoria etd – Marais, J J (2004)
FOUR: Results and Discussions
This chapter aims to create and present new meaning as a
contribution to the field of audiology. Meaning is derived
from findings through a process of interpretations. It
furthermore aims to establish validity of the findings, by
relating them to theoretical principles.
4.1 Introduction
In this chapter the results of the current study are presented and discussed. The main
aim of this study was to determine the influence of insert earphones and sound-field
speaker presentation on threshold estimations using a monotic auditory steady state
evoked potential technique, in a group of normal-hearing adults. In order to address
the main aim of this study, three closely related sub-aims were formulated. These
aims and their relation to the main aim of the study are represented in figure 4.1.
Presentation of these results addresses the main aim of this study. Results are
discussed based on the current body of knowledge (data obtained through the current
study) and the integration of this knowledge into the literature currently available. The
results are presented according to the three sub-aims and a discussion of these results,
alongside relevant literature, follows.
89
University of Pretoria etd – Marais, J J (2004)
Sub-Aim 1: To compare pure tone-,
frequency modulated (FM) and mixed
modulated (AM/FM) tone behavioural
thresholds at 0.5, 1, 2 & 4 kHz under
insert earphone and sound-field
speaker conditions.
MAIN AIM:
To determine the
influence of insert
earphones and soundSub-Aim 2: To compare minimum
response level detection at 0.5, 1, 2 &
4 kHz using a monotic ASSEP
technique under insert earphone and
sound field speaker conditions
field speaker
presentation on
minimum response level
detection using a
monotic auditory steady
state evoked potential
technique, in a group of
Sub-Aim 3: To compare the actual
behavioural thresholds (for pure tone,
FM and mixed modulated tone
stimulation) with the ERA system’s
accuracy in predicting these thresholds
by means of an algorithm for insert
earphone and sound-field conditions.
normal-hearing adults
Figure 4.1 Research Process: The sub-aims and their relation to the main aim of
the research project
90
University of Pretoria etd – Marais, J J (2004)
Figure 4.1, above, provides a visual representation of how each sub-aim,
independently relates to the main aim. The three sub-aims are discussed separately in
an attempt to address the main aim of the current study.
4.2 A comparison between pure tone-, frequency modulated (FM) and mixed
modulated (AM/FM) tone behavioural thresholds at 0.5, 1, 2 & 4 kHz under
insert earphone and sound-field speaker conditions
Results are presented in relation to two comparisons, namely a comparison across
stimulus presentation, and one across transducers. Both of these comparisons were
controlled for gender and ear effects. Following the presentation of results, a
discussion relating to current literature follows.
A comparison of the behavioural thresholds using the two different stimulus
presentations, for both insert earphones and sound-field speaker presentation,
establish a basis from where the accuracy of the calibration of the different stimulus
presentations and transducers can be argued. These comparisons also establish a basis
from which the accuracy of the minimum response levels obtained using the monotic
ASSEP technique under both insert earphone and sound-field conditions can be
argued.
4.2.1 A Comparison across Stimulus Presentations:
The comparison across stimulus presentation details the discussion of thresholds
obtained through behavioural pure tone, behavioural FM tone and behavioural mixed
modulated tone presentation, subdivided into insert earphone and sound-field
conditions.
4.2.1.1 Thresholds obtained under Insert Earphone Conditions
Table 4.1 below provides a summary of the mean values (in dB HL), and standard
deviations, obtained from the participants during behavioural pure tone and mixed
modulated tone presentations (using the GSI AUDERA equipment), at 0.5, 1, 2 & 4
kHz, when insert earphone presentation was used.
91
University of Pretoria etd – Marais, J J (2004)
Table 4.1 Mean and Standard Deviation Values for Behavioural Pure Tone and
Behavioural Mixed Modulated Tone thresholds under insert earphone conditions
(n = 50)
Thresholds obtained under
500 Hz
Insert Earphone Conditions
(in dB HL)
Pure Tone Thresholds Mean 7 (+/-7)
(SD)
Mixed Modulated
Mean 7 (+/-7)
Tone Thresholds
(SD)
1000 Hz
2000 Hz
4000 Hz
1 (+/-5)
1 (+/-6)
1 (+/-7)
3 (+/-5)
2 (+/-5)
3 (+/-7)
Behavioural pure tone thresholds, as well as behavioural mixed modulated thresholds,
were within 5 dB from one another when insert earphones were used. No difference
was found between the techniques at 0.5 kHz where the mean value for both
techniques was 7 dB.
The standard deviation for the behavioural pure tone thresholds varied between 5 dB
and 7 dB for all the frequencies tested. The standard deviation for the behavioural
mixed modulated tone thresholds, were exactly the same except at 2 kHz where there
was a 1 dB difference in standard deviation. Bearing in mind that behavioural
thresholds are estimated using 5 dB increments, these standard deviations were almost
identical. For both techniques the greatest standard deviations were present at 0.5 and
4 kHz, namely 7 dB.
Statistical analysis of the behavioural pure tone thresholds compared to the
behavioural mixed modulated tone thresholds, under insert earphone conditions
revealed no significant differences for any of the frequencies tested (p < 0.05). It is
clear from this analysis that the behavioural thresholds obtained from the different
stimulus presentations, under insert earphone conditions, are comparable based on a
5% level of significance. The numerical data obtained from the statistical analyses are
available in Appendix D.
This experiment was controlled for both gender- and ear effects. When the results
were analyzed using the Mann-Whitney paired T-Test with p < 0.01 no statistical
differences, for behavioural pure tone thresholds, were noted between male and
female research participants. The only statistical difference was obtained at 2 kHz
92
University of Pretoria etd – Marais, J J (2004)
when p< 0.05 was used to analyze the results. No statistical differences were
obtained between male and female research participants for either p< 0.01 or p<0.05
when the behavioural modulated tone thresholds were analyzed. Statistical analyses of
ear effects revealed no significant differences between the left and right ears analyzed
for both p<0.01 and p<0.05, regardless of the stimulus presentation employed in
obtaining the thresholds. The numerical data obtained from the statistical analyses for
both the gender and ear analyses are presented in Appendix E.
4.2.1.2 Thresholds obtained under Sound-field Conditions
Table 4.2 below provides a summary of the mean values and standard deviations (in
dB HL) obtained from the participants during behavioural pure tone and behavioural
modulated tone (using the GSI AUDERA equipment), at 0.5, 1, 2 & 4 kHz, when
sound-field speaker presentation was used.
Table 4.2 Mean and Standard Deviation Values for Behavioural FM Tone and
-Mixed Modulated Tone thresholds under sound-field conditions (n = 50)
Thresholds obtained under Soundfield Conditions (in dB HL)
Behavioural FM Tone
Thresholds
Behavioural Mixed
Modulated Tone Thresholds
Mean
(SD)
Mean
(SD)
500 Hz
1000 Hz
2000Hz
4000 Hz
4 (+/-6)
-2 (+/-4)
-1 (+/-5)
3 (+/-5)
1 (+/-6)
0 (+/-5)
0 (+/-4)
7 (+/-7)
Thresholds obtained through behavioural FM tone testing, as well as behavioural
mixed modulated tone testing, were again within 5 dB from each other, when soundfield speaker presentation was used. The smallest difference with regard to the mean
dB values was at 2 kHz where the behavioural FM tone stimulation showed –1dB,
whereas the mixed modulated tone stimulation showed a mean value of 0 dB. The
greatest difference in mean values was found at 4 kHz with a difference of 4 dB. The
standard deviations for the behavioural pure tones varied between 4 dB and 6 dB,
whereas the standard deviations for the behavioural mixed modulated tones varied
between 4 dB and 7 dB. The behavioural mixed modulated tones showed the greatest
standard deviation (at 4 kHz), namely 7 dB.
The greatest difference between the two stimulus presentations (FM tone -versus
modulated tone presentation) is found at 4 kHz. The greater mean values as well as
93
University of Pretoria etd – Marais, J J (2004)
the greater standard deviations at 4 kHz shown for the mixed modulated tones when
sound-field speaker presentation is used could probably be explained by the fact that
the field is less uniform for higher frequencies (Dillon & Walker, 1981). Furthermore
thresholds obtained at 4 kHz are more susceptible to head-movements (directionality),
as well as the head shadow effect (Dillon & Walker, 1981; Hall & Mueller, 1997).
Statistical analysis of the behavioural FM tone thresholds compared to the behavioural
mixed modulated tone thresholds, under sound-field conditions revealed no
significant differences for any of the frequencies tested (p < 0.05). It is clear from this
analysis that the behavioural thresholds obtained from the different stimulus
presentations, under sound-field conditions, are comparable based on a 5% level of
significance. The numerical data from the statistical analyses are available in
Appendix D.
This experiment was again controlled for gender and ear effects. Gender effects were
analyzed using the Mann-Whitney test and showed no significant differences between
male and female research participants for p < 0.05 or p < 0.01, regardless of the
transducer used. The numerical data from the statistical analyses are available in
Appendix F. There was a significant difference between left and right ears at 0.5 kHz,
when the FM tones were analyzed, with a p < 0.05, while analysis with a p < 0.01 did
not reveal a significant difference. For all the other frequencies tested with the
behavioural FM tones, namely 1, 2 & 4 kHz, there were no significant differences
between left and right ears. The numerical data from the statistical analyses are
available in Appendix F. With regard to ear effects, no significant differences were
noted at any of the frequencies tested (0.5 - 4 kHz) with the behavioural mixed
modulated tones regardless of which p-value was used. The results from the statistical
analysis are available in Appendix F.
4.2.2 Comparison across Transducers
The comparison across transducers will detail the discussion of thresholds obtained
under insert earphone and sound-field conditions. The discussion across transducers
will be subdivided for thresholds obtained through behavioural pure tone, behavioural
FM tone and behavioural mixed modulated tone testing.
94
University of Pretoria etd – Marais, J J (2004)
4.2.2.1 Thresholds obtained through Behavioural Pure Tone and Behavioural
FM Tone Testing
Table 4.3 provides a summary of the mean values and standard deviations (in dB HL)
obtained from the participants during behavioural pure tone and behavioural FM tone
testing, at 0.5, 1, 2 & 4 kHz, when insert earphones and sound-field speaker
presentation were used.
Table 4.3 Mean and Standard Deviation Values for Behavioural Pure Tone and
Behavioural FM Tone thresholds under insert earphone and sound-field
conditions (n = 50)
Pure Tone and FM Tone
Techniques (in dB HL)
Insert Earphone
Presentation (Pure Tone)
Sound-field Speaker
Presentation (FM)
Mean
(SD)
Mean
(SD)
500 Hz
7 (+/-7)
1000 Hz
1 (+/-5)
2000Hz
1 (+/-6)
4000 Hz
1 (+/-7)
4 (+/-6)
-1 (+/-4)
-1 (+/-5)
3 (+/-5)
The thresholds obtained through behavioural pure tone and FM tone testing, using
insert earphones and sound-field speaker presentation, were within 5 dB from one
another, across transducers. The greatest mean difference is found at 0.5 kHz where
the mean difference was 3 dB. The standard deviations for the pure tone behavioural
technique, using insert earphones, varied between 5 dB and 7 dB for all the
frequencies tested. Standard deviations for the behavioural FM tone technique, using
sound-field speaker presentation varied between 4 dB and 6 dB.
Statistical analysis of the behavioural pure tone and FM tone thresholds obtained
under insert earphone and sound-field conditions revealed significant differences for
all of the frequencies tested (p < 0.05).
It is clear from this analysis that the
behavioural thresholds obtained from the different stimulus presentations, under insert
earphone and sound-field conditions, are not comparable based on a 5% level of
significance. When the thresholds were analyzed on a 1% level of significance there
was no significant difference at 4 kHz between the thresholds obtained under insert
earphone and sound-field conditions. The numerical data obtained from the statistical
analyses are available in Appendix G.
95
University of Pretoria etd – Marais, J J (2004)
This experiment was controlled for gender- and ear effects. Gender effects were
analyzed using the Mann-Whitney test and showed no significant differences between
male and female research participants for p < 0.05, for all frequencies tested except 2
kHz, regardless of the transducer used. Measured against p < 0.01, there were no
significant differences between genders. There was a significant difference between
left and right ears at 0.5 kHz, when the FM tones were analyzed, with p < 0.05, while
analysis with p < 0.01 did not reveal a significant difference. The results are tabulated
in Appendix E and F.
4.2.2.2 Thresholds obtained through Behavioural Mixed Modulated Tone
Testing
Table 4.4 provides the summary of the mean values and standard deviations obtained
(in dB HL) from the participants during behavioural mixed modulated tone testing, at
0.5, 1, 2 & 4 kHz, when insert earphones and sound-field speaker presentation were
used.
Table 4.4 Mean and Standard Deviation Values for Behavioural Mixed
Modulated Tone thresholds under insert earphone and sound-field conditions
(n = 50)
Behavioural Mixed Modulated
Tone Thresholds (in dB HL)
Insert Earphone
Mean
Presentation
(SD)
Sound-field Speaker
Mean
Presentation
(SD)
500 Hz
7 (+/-7)
1000 Hz
3 (+/-5)
2000Hz
2 (+/-5)
4000 Hz
3 (+/-7)
1 (+/-6)
0 (+/-5)
-1 (+/-4)
7 (+/-7)
The thresholds obtained through behavioural mixed modulated testing, across
transducers, was within 5 dB from one another, except at 0.5 kHz (highlighted in the
Table) where there is a mean difference of 6 dB. The data contained in Table 4.5 will
subsequently also be presented in Figure 4.2 to highlight the difference. The smallest
difference, in mean value, was found at 2 kHz were the mean difference was 2 dB.
The standard deviations for the behavioural mixed modulated technique, using insert
earphones, varied between 5 dB and 7 dB for all the frequencies tested. Standard
deviations for the same technique, using sound-field speaker presentation varied
between 4 dB and 7 dB.
96
University of Pretoria etd – Marais, J J (2004)
Statistical analysis of the behavioural mixed modulated tone thresholds obtained
under insert earphone and sound-field conditions revealed significant differences for
all of the frequencies tested (p < 0.05/ p < 0.01), except at 2 kHz for p < 0.01. It is
clear from this analysis that the behavioural thresholds obtained from the modulated
tone presentations, under insert earphone and sound-field conditions, are not
comparable based on a 5% level of significance. The numerical data obtained from
the statistical analyses are available in Appendix G.
This experiment was again controlled for gender- and ear effects. No statistical
differences were obtained between male and female research participants for either a
p value of < 0.01 or 0.05 when the behavioural modulated tone thresholds were
analyzed, regardless of which transducer was used. With regard to ear effects, there
were no significant differences were noted at any of the frequencies tested (0.5 - 4
kHz) with the behavioural modulated tones regardless of which p-value was used,
regardless of which transducer was used. The numerical data obtained from the
statistical analyses are tabulated in Appendix E and F.
Figure 4.2, below, provides a visual interpretation of the behavioural mixed
modulated tone thresholds obtained using insert earphone- and sound-field speaker
presentation at 0.5, 1, 2 & 4 kHz.
Mean Vaules (in dB HL)
15
10
5
0
-5
-10
500
1000
2000
4000
Using Insert Earphone
Presentation
7
3
2
3
Using Sound Field Speaker
Presentation
1
0
-1
7
Frequencies Tested (in Hz)
Using Insert Earphone Presentation
Using Sound Field Speaker Presentation
Figure 4.2 Behavioural Mixed Modulated Tone Thresholds, obtained under
Insert Earphone and Sound-field conditions
97
University of Pretoria etd – Marais, J J (2004)
Mean thresholds obtained through behavioural testing across transducers, when using
a mixed modulated (AM/FM) stimulus, revealed no greater than 5 dB differences
across 1, 2 & 4 kHz. At 0.5 kHz a mean difference of 6 dB can be observed which is
marginally larger than the minimum standard error (5 dB) introduced for behavioural
testing. A statistical difference is noted for all the frequencies tested (0.5, 1, 2 & 4
kHz) for the comparison across transducers and is tabulated in Appendix G. In light of
the above, and especially at 0.5 kHz, it will therefore be pertinent to compare across
stimulus presentation rather than across transducer. This implies that, for this
particular study, behavioural data specific to the type of transducer employed will
have to be used as a baseline. Therefore, behavioural thresholds obtained through
mixed modulated stimuli (tones), using insert earphone presentation, will not suffice
as a baseline measure against which monotic ASSEP minimum response levels, using
sound-field speaker presentation, can be compared. The same principle applies to
behavioural thresholds obtained through mixed modulated stimuli (tones) under
sound-field conditions when compared with monotic ASSEP minimum response
levels obtained under insert earphone conditions.
4.2.3 Discussion of Behavioural Thresholds obtained, when compared across
Stimulus Presentations and Transducers
The results of the current study indicate that no greater than 5 dB mean differences
were obtained for the comparison across stimulus presentations (regardless of the
transducer used). However for the comparison across transducers, and specifically
relating to the behavioural thresholds obtained through mixed modulated stimulation
(modulated tones) at 0.5 kHz, a difference in excess of 5 dB was recorded. According
to Katz (1985) extrinsic factors (temperature, light and ambient noise) as well as
intrinsic factors (motivation, attention and familiarity with the listening task), within
the test environment play a significant role when the audibility thresholds of average
listeners are estimated.
Within the clinical setting this implies that certain compromises have to be made in
order to obtain hearing thresholds within a relatively short period of time (Katz,
1985). Because attenuator steps of 5 dB are employed, a minimum standard error of
+/- 5 dB is introduced (Katz, 1985) for behavioural testing of any individual. These
minimum standard errors together with the statistical analysis of the results form the
98
University of Pretoria etd – Marais, J J (2004)
basis from which the accurate calibration of the equipment is argued. Furthermore,
sets the basis for appropriate comparisons of the monotic ASSEP thresholds with
behavioural thresholds. Table 4.5, below provides a summary of the results presented,
when compared across stimulus presentations and transducers.
Table 4.5 Summary of the Mean and Standard Deviation values across Stimulus
Presentations and Transducers (n = 50)
Thresholds Obtained Under Behavioural
Testing
Behavioural Pure Tone Thresholds Mean
under Insert Earphone conditions
(SD)
Behavioural FM Tone Thresholds
Mean
under Sound-field conditions
(SD)
Behavioural Mixed Modulated
Mean
Tone Thresholds under Insert
(SD)
Earphone conditions
Behavioural Mixed Modulated
Mean
Tone Thresholds under Sound-field (SD)
conditions
0.5 kHz
7 (+/-7)
1 kHz
1 (+/-5)
2 kHz
1 (+/-6)
4 kHz
1 (+/-7)
4 (+/6)
-2 (+/-4)
-1 (+/-5)
3 (+/-5)
7 (+/- 7)
3 (+/-5)
2 (+/-5)
3 (+/-7)
1 (+/-6)
0 (+/-5)
0 (+/-4)
7 (+/-7)
4.2.3.1 Discussion of the Comparison across Stimulus Presentation (controlled
for Transducers)
It is therefore clear that for the comparison across stimulus presentation (regardless of
the transducer used) there were no mean differences that exceeded the minimum
standard error. Also, no statistical differences were obtained between the techniques
(within each specific transducer used). Therefore, the calibration of the equipment
relating to the comparison across stimulus presentations seems to be comparable. The
numerical data from the comparison across stimulus presentation is available in
Appendix D.
The results, relating to the comparison across technique, revealed pure tone
behavioural thresholds that closely approximated 0 dB HL, which according to Roeser
and colleagues (2001) represents the “perfect” normal-hearing sensitivity standard.
Mean thresholds were between 1 and 7 dB HL. A large percentage (78%) of the
behavioural pure tone thresholds was equal to or less than 5 dB HL and this increased
to 91% if 10 dB HL is included. Therefore the great majority of the behavioural pure
tone responses were less than 10 dB HL, even though the normal cut-off of 25 dB HL
was used during research participant selection. The standard deviations observed for
99
University of Pretoria etd – Marais, J J (2004)
the pure tone behavioural technique varied between 5-7 dB, which closely
approximates the prescribed standard variation of 6 dB across all the frequencies
tested (Haugton, 1980).
Although there is very little literature available to compare the behavioural mixed
modulated thresholds, using insert earphone presentation, Picton and colleagues
(1998) reported that a 100% modulated tone is clearly distinguishable from an unmodulated tone at
5 dB above the hearing threshold for the particular tone. This is
corroborated by earlier research (Bacon & Viemeister, 1985) reporting that
modulations of –6 dB (equivalent to 50%) can be detected at 5 dB above hearing
threshold, at modulation frequencies between 64-128 Hz. The mean thresholds
obtained closely approximated 0 dB HL and ranged between 1 and 7 dB HL. Of the
behavioural thresholds obtained through mixed modulated stimuli under insert
earphone conditions, 78% were equal to or less than 5 dB HL. This figure increased to
92% if 10 dB HL was included. Again the great majority of the behavioural mixed
modulated thresholds were less than 10 dB HL. Standard deviations observed for the
behavioural thresholds obtained through modulated stimuli under insert earphone
conditions were identical to the standard deviations for the behavioural pure tone
technique, and varied between 5 and 7 dB.
The same pattern emerged when the results relating to the comparison across
technique for behavioural FM tone and behavioural mixed modulated thresholds,
using sound-field speaker presentation, were evaluated. The thresholds obtained
closely approximated 0 dB HL. Of all the ears tested behaviourally under the soundfield conditions, 89 % revealed thresholds of 5 dB HL or less, while this figure grew
to 95% if 10 dB HL was included. Once again it is clear that for the majority of the
ears tested, thresholds were 10 dB HL or less. Standard deviations for the behavioural
pure tone technique, using sound-field speaker presentation, varied between 4 and 6
dB, which is slightly higher than those obtained by Arlinger & Jerlvall (1987), namely
2.5-3.4 dB, also obtained from adult research participants.
The higher standard deviations obtained in the current study can possibly be attributed
to the stipulated range of normality. In the study done by Arlinger & Jerlvall (1987)
research participants were required to have hearing thresholds that ranged between
0-20 dB HL, whereas the current study stipulated normal-hearing as ranging between
100
University of Pretoria etd – Marais, J J (2004)
0-25dB HL. Thus, the greater range (of normality) analyzed could have slightly
elevated the standard deviations obtained during the current study. However, in light
of the fact that 95% of the behavioural thresholds obtained under sound-field
conditions (in the current study) were equal to or less than 10 dB HL, this explanation
seems highly unlikely.
Another difference between the current study and that of Arlinger & Jerlvall (1987) is
that the modulation frequencies employed differed substantially. For their study,
Arlinger & Jerlvall (1987) used a modulation frequency of 5 and 20 Hz, whereas the
current study employed modulation frequencies that varied between 80 and 110Hz
depending on the carrier frequency being modulated. Further investigation is required,
as other variables (e.g. ambient noise levels in the test room) could also possibly have
attributed to the higher than reported standard deviations.
Again, there is limited literature available to serve as comparison between the
behavioural modulated thresholds obtained under sound-field conditions. Mean
thresholds varied between –1 and 4 dB HL and yet again closely approximated 0 dB
HL. Of these thresholds 85% were equal to or less than 5 dB HL. This figure grew to
93% when 10 dB HL was included. Once again the majority of the thresholds were
obtained at 10 dB HL or less. Standard deviations for the thresholds obtained with
modulated tones, under sound-field conditions, were similar to those obtained using
the behavioural FM tone technique, and varied between 4 and 7 dB. These results
correlate well in light of the range of normality (0-25 dB HL) of the research
participants involved.
4.2.3.2 Discussion of the Comparison across Transducers (controlled for
Stimulus Presentation):
As with the comparison across stimulus presentations, the comparison across
transducers with specific reference to the behavioural pure and –FM tone technique
again revealed no greater mean differences than 5 dB. However statistical analysis of
these thresholds revealed significant differences for all the frequencies tested using a
p < 0.05, thus compromising the comparison across transducers. When considering,
the comparison across transducers, specifically related to the behavioural thresholds
obtained when mixed modulated (AM/FM) stimuli were used, are considered, the
mean difference at 0.5 kHz exceeds the minimum standard error of 5 dB and
101
University of Pretoria etd – Marais, J J (2004)
furthermore shows statistical differences for all the frequencies tested. The numerical
data obtained from the comparison across transducers are available in Appendix G.
Thereby compromising the comparison based on the assumption of accurate
calibration.
In light of this phenomenon, further comparisons with behavioural modulated
thresholds will have to be considered on a transducer-specific basis. Stated differently,
behavioural mixed modulated thresholds obtained under insert earphone conditions
will have to serve as the basis for comparison with monotic ASSEP minimum
response levels obtained under the same conditions, and likewise with the comparison
obtained under sound-field conditions.
Both behavioural pure tone and behavioural FM tone data closely approximated 0 dB
HL. Of the pure tone thresholds obtained using insert earphone presentation, 91%
were equal to or less than 10 dB HL while 89% of the FM tone thresholds obtained,
using sound-field speaker presentation, were equal to or below 10 dB HL. The
standard deviations for the technique while using insert earphone presentation varied
between 5 and 7 dB, while the standard deviations for the same technique using
sound-field speaker presentation, varied between 4 and 6 dB HL. Furthermore there
were significant differences when non-parametric statistics (Wilcoxon Paired T-Test)
were investigated (p < 0.05). Therefore, the results obtained for behavioural pure tone
testing are not comparable across the transducer used. These numerical data from the
statistical analyses are tabulated in Appendix G.
The comparison across transducers, specifically related to the behavioural thresholds
obtained when mixed modulated tones were employed, revealed a greater than
expected variance (> 5 dB) of 6 dB at 0.5 kHz, bearing in mind the +/- 5 dB minimum
standard error introduced for behavioural testing. For all the other frequencies tested
(1, 2 & 4 kHz) the variance was no greater than 5 dB, and calibration of the
equipment for these frequencies can therefore be seen as accurate. The Wilcoxon
Paired T-Test revealed significant differences between these transducers for all the
frequencies when p < 0.05 was implemented and is tabulated in Appendix G.
There are several factors that could have influenced the results obtained. Participants
were tested in the reverberant field for behavioural FM tone testing and in the direct
102
University of Pretoria etd – Marais, J J (2004)
field for behavioural mixed modulated tone ASSEP testing. According to Walker et
al. (1984) the sound pressure level in the reverberant field is influenced by reflected as
well as direct sound. Mixed modulated thresholds, in the current study, were obtained
in the direct field and therefore there was less interaction of the acoustic environment,
while the behavioural pure tones were obtained in the reverberant field. Thus the
influence of the acoustic environment could possibly have affected the thresholds
obtained to a greater degree.
Other factors such as directivity of the speaker (Walker et al., 1984), the rooms’
dimensions (Anderson, 1979) and absorptivity of its surfaces will influence the extent
of the direct field. In the current study participants were positioned according to a zero
degree azimuth in relation to the speaker. It is possible that absorptivity of the
surfaces influenced the results obtained at least to a certain extent seeing as the
behavioural pure tone testing was done while participants were seated in a chair,
whereas the behavioural mixed modulated testing was done with participants lying on
a bed. The absorptivity of the bed would be significantly different from the
absorptivity of the room behind the seated participant. Also the dimensions of the test
room (discussed in the methodological chapter of this study) are not constant for the
two conditions.
Positioning of the participants as well as their head movements also influence the
thresholds obtained (Magnusson et al., 1997). Positioning of participants was
correlated with a marker at the position identified during calibration. The researcher
and another qualified audiologist, throughout the recording of thresholds and
minimum response levels, checked the maintenance of the test position. It is possible,
however, that small head movements could have been overseen and that these
movements subsequently had an effect on the results obtained.
Lastly, where sound-field presentation was the last presented condition, fatigue of the
participants could also have impacted on the results. As discussed in the
methodological chapter of this study the test conditions were randomized so that each
condition accounted for practice and carry-over effects. It is unlikely therefore that
fatigue would have adversely affected the results obtained, however the possibility
still exists. Although it is seemingly impossible to pinpoint the direct cause of the
103
University of Pretoria etd – Marais, J J (2004)
greater than expected variance and statistical differences, it is more likely than not a
combination of these variables that have contributed to the results obtained.
4.3 A comparison between minimum response level detection at 0.5, 1, 2 & 4 kHz
using a monotic ASSEP technique under insert earphone and sound field
speaker conditions
This sub-aim details and discusses the results from four related groups of data
pertaining to its focus. Three comparisons were analyzed, namely a comparison across
techniques, one across transducers and the last across both technique and transducer.
The comparison across techniques provides the basis from where the accuracy of the
threshold estimations using the different transducers can be argued. For this
comparison behavioural data are compared, as a basis, against the electrophysiological data to establish how closely the electro-physiological minimum
response levels approximate the actual behavioural thresholds. The same comparison
was made between behavioural mixed modulated tone thresholds and electrophysiological data to establish whether the electro-physiological minimum response
levels are more closely related to the behavioural pure tone thresholds or the
behavioural modulated tone thresholds.
The comparison across transducer provides critical information regarding the
influence of the different transducers on ASSEP minimum response level detection.
Lastly, the comparison across technique and transducer provides critical information
relating to whether minimum response levels, using sound-field speaker presentation
can be used, with normative data from both pure tone thresholds using insert earphone
presentation and pure tone thresholds using sound-field speaker presentation.
4.3.1 A Comparison across Stimulus Presentations:
The comparison across stimulus presentation will detail the discussion of thresholds
obtained through behavioural testing and ASSEP testing. The discussion of stimulus
presentations will be subdivided for thresholds obtained under insert earphone- and
sound-field conditions.
104
University of Pretoria etd – Marais, J J (2004)
4.3.1.1 Thresholds an Minimum Response Levels obtained under Insert
Earphone Conditions
Table 4.6 provides a summary of the comparison between the mean behavioural
thresholds, for both pure tone stimulation and modulated tone stimulation, and the
mean ASSEP minimum response levels at 0.5, 1, 2 and 4 kHz, using insert-earphone
presentation.
Table 4.6 Mean and Standard Deviation Values for Behavioural Pure Tone
thresholds and ASSEP minimum response levels under insert earphone
conditions (n = 50)
Thresholds obtained under Insert
Earphone conditions (in dB HL)
Behavioural Pure Tone
Mean
Thresholds
(SD)
Behavioural Mixed
Mean
Modulated Thresholds
(SD)
ASSEP Minimum
Mean
Response Levels
(SD)
500 Hz
7 (+/-7)
1000 Hz
1 (+/-5)
2000Hz
1 (+/-6)
4000 Hz
1 (+/-7)
7 (+/-7)
3 (+/-5)
2 (+/-5)
3 (+/-7)
35 (+/-12)
25 (+/-10)
26 (+/-9)
27 (+/-11)
The thresholds obtained through behavioural pure tone testing differ substantially
from the ASSEP minimum response levels obtained under insert earphone conditions.
The greatest mean difference for the comparison of ASSEP minimum response levels
with behavioural pure tones (28 dB) is found at 0.5 kHz where the mean value for the
electro-physiological threshold was 35 dB HL compared to the mean pure tone
threshold of 7 dB HL. The smallest difference (24 dB) in mean values was found at 1
kHz where the mean value for the ASSEP minimum response level was 25 dB HL
compared to the mean behavioural threshold of 1dB HL. The difference in mean
values between the behavioural thresholds and the electro-physiological minimum
response levels for 2 kHz was 25 dB, and 26 dB at 4 kHz.
Although elevated, these figures correlate well with previous studies investigating the
threshold estimation of the monotic ASSEP technique in normal-hearing adults.
Picton and colleagues (1998) reports a mean minimum response level, for monotic
ASSEP, of 34 dB SPL at 0.5 kHz compared to the 44 dB SPL (35 dB HL), corrected
for ER-3A insert earphones, of the current study. The 25 dB HL obtained at 1 kHz is
slightly better than the 29 dB HL reported by Aoyagi and colleagues (1994) at the
same frequency. This figure when corrected for supra aural headphones relates to 37
105
University of Pretoria etd – Marais, J J (2004)
dB SPL, which is higher than 34 dB SPL, again corrected for insert earphones,
obtained in the current study. Research by Herdman & Stapells (2001) on monotic
ASSEP revealed significantly lower MRLs of 20 dB SPL compared to the 33dB SPL
(26 dB HL), corrected for insert earphones, at 2 kHz. Again at 4 kHz the current
study revealed a generally lower minimum response level of 29 dB SPL (27 dB HL),
again corrected for insert earphones, compared to the 41dB SPL (30 dB HL),
corrected for TDH 49 supra-aural headphones obtained by Aoyagi et al (1994).
However, a very close similarity is obtained with the results that Picton and
colleagues (1998) obtained, namely 33 dB SPL.
Research done by Herdman & Stapells (2001) reported ASSEP minimum response
levels that ranged between 18 and 20 dB SPL across 0.5-4 kHz, which is significantly
better than the 29-44 dB SPL obtained in the current study. However when viewed in
the light of other studies the range obtained in the current study (29-44 dB SPL)
compares well with the 36-41 dB SPL (29-30 dB HL), corrected for TDH 49 supraaural headphones, obtained by Aoyagi and colleagues (1994), and the 33-34 dB SPL
obtained by Picton and colleagues (1998). Although the minimum response levels
obtained in the current study compare well with other literature, it is important to
mention that at present no universal criteria exists to define “acceptable difference in
threshold”.
When the ASSEP minimum response levels, obtained in the current study, are
compared with the behavioural thresholds obtained from modulated stimuli
(modulated tones), also obtained in the current study, slightly smaller differences were
observed between 1 and 4 kHz, while the mean difference at 0.5 kHz was exactly the
same. Both comparisons revealed a mean difference of 28 dB at 0.5 kHz, while the
mean difference between the behavioural modulated tone threshold and ASSEP
minimum response level was 22 dB at 1 kHz and 24 dB at 2 and 4 kHz.
There is limited literature available to compare these behavioural modulated tone
thresholds against. As discussed earlier, a 100% modulated tone is distinguishable
from a pure tone at 5 dB above hearing threshold (Bacon & Viemeister, 1985; Picton
et al., 1998). Furthermore, there were no significant statistical differences between the
behavioural pure tone and behavioural modulated tone data (discussed earlier in this
chapter). Finally these sets revealed no greater than 5 dB difference (minimum
106
University of Pretoria etd – Marais, J J (2004)
standard error) between them. Therefore in light of the aforementioned factors, these
differences seem marginal.
The standard deviations for the behavioural pure tone technique, using insert
earphones varied between 5 dB and 7 dB for all the frequencies tested. Standard
deviations for the ASSEP technique, using insert earphones, varied between 9 dB and
12 dB. The greatest standard deviation for the ASSEP technique was found at 0.5
kHz, namely 12 dB, while the smallest standard deviation was observed at 2 kHz,
namely 9 dB.
The standard deviations for both the behavioural pure tone thresholds as well as the
ASSEP minimum response levels were within 5 dB HL from each other for 0.5 & 4
kHz with a difference of 5 dB HL at 1 & 2 kHz. These standard deviations, obtained
in the current study, are marginally better than those obtained by Aoyagi and
colleagues (1994). The standard deviation at 1 kHz was 10 dB (current study)
compared with the 14 dB, obtained by Aoyagi and colleagues (1994), and 11 dB
(current study) compared with the 15 dB at 4 kHz. The greater standard deviations
obtained by Aoyagi and colleagues (1994) could possibly be attributed to the
transducers used for presentation of stimuli. According to Lins and colleagues (1996),
supra aural headphones attenuate ambient noise levels by approximately 10 dB for the
lower frequencies, while higher frequencies are attenuated by as much as 20 dB.
Additionally, the smaller standard deviations obtained in the current study can
possibly be attributed to the use of insert earphone stimulation, which has less
problems with sound leakage (Wilber, 2002).
It seems therefore that the ASSEP minimum response levels, under insert earphone
conditions, compare well with those reported in the literature (Aoyagi et al., 1994;
Picton et al., 1998), except possibly for those reported in a study by Herman &
Stapells (2001). The average SPL thresholds varied between 29 and 44 dB SPL
compared with the 36-40 dB SPL reported by Aoyagi and colleagues (1994), with
values converted from HL (at 1 & 4 kHz) and those obtained by Lins and colleagues
(1996), namely 29-36 dB SPL.
Statistical analyses of the thresholds obtained under behavioural testing (both pure
tone- and modulated tone) and those obtained under ASSEP testing differ
107
University of Pretoria etd – Marais, J J (2004)
significantly from each other, controlled for insert earphone presentation. For all the
frequencies tested (0.5-4kHz) both sets of behavioural data showed greater than 0.05
p-values when compared to the ASSEP minimum response levels. These sets of data,
therefore, differ significantly from each other on a 5% significance level. The results
of this analysis are available in Appendix D.
This experiment was, again, controlled for gender- and ear effects. On a 5% level of
significance the minimum response levels obtained from male and female research
participants did not show significant differences. However, at 2 kHz there was a
significant difference relating to the gender effect. When these results are viewed on a
1% level of significance there is no statistical difference between the genders. For all
the other frequencies there were no statistical differences on either a 1% or a 5% level
of significance. The numerical data from the statistical analyses are available in
Appendix H.
The statistical difference obtained between genders could possibly be attributed to a
difference in the length of the basilar membrane, between males and females.
According to John & Picton (2000) the representation of frequencies along the basilar
membrane are not accurately known. The length of the membrane differs from person
to person, and is longer for males compared to females (John & Picton, 2000). Thus,
longer traveling time associated with the increased length of the basilar membrane,
may account for the difference observed for male participants (when compared to
female research participants).
Goldstein & Aldrich (1999) reported that the differences in head size between genders
could create gender related latency delays. Although this statement refers to auditory
evoked potentials in general, the latency effects will almost certainly affect the phase
measures of the ASSEP. Thus, there is some support in literature to corroborate the
findings obtained in the current study, but more research is required before any
definite conclusions can be drawn. In light of the fact that the difference was not
noted when a stricter (p < 0.01) analysis was made, the influence of gender did not
seem to significantly affect the results obtained from the research participants. The
numerical data from the analysis is available in Appendix H.
108
University of Pretoria etd – Marais, J J (2004)
When the same comparison was made for ear effects there were no statistical
differences between the ears for either a 1% or 5% level of significance. These results
are available in Appendix I.
4.3.1.2 Thresholds and Minimum Response Levels obtained under Sound-field
Conditions
Table 4.7 below, again provides a summary of the behavioural FM tone technique’s
mean thresholds compared to the ASSEP techniques’ mean thresholds, using soundfield speaker presentation.
Table 4.7 Mean and Standard Deviation values for Behavioural FM Tone
thresholds and ASSEP minimum response levels under sound-field conditions
(n = 50)
Thresholds obtained under
Sound-field conditions
(in dB HL)
Behavioural FM Tone
Mean
Thresholds
(SD)
Behavioural Mixed
Mean
Modulated Tone
(SD)
Thresholds
ASSEP Minimum
Mean
Response Levels
(SD)
500 Hz
1000 Hz
2000Hz
4000 Hz
4 (+/-6)
-2 (+/-4)
-1 (+/-5)
3 (+/-5)
1 (+/-6)
0 (+/-5)
-1 (+/-4)
7 (+/-7)
26 (+/-12)
18 (+/-10)
21 (+/-10)
36 (+/-11)
Thresholds obtained, under sound-field conditions, through behavioural FM tone
testing, as well as those for behavioural mixed modulated tone testing, differ
substantially from the ASSEP minimum response levels. The comparison between
ASSSEP minimum response levels and behavioural FM tones, the greatest difference
(33 dB) is found at 4 kHz. The mean value for the electro-physiological minimum
response level was 36 dB HL, while the FM tone threshold mean was 3 dB HL. The
smallest difference (20 dB) in mean values was found at 1 kHz where the mean value
for the electro-physiological minimum response level was 18 dB HL compared to the
mean behavioural threshold of –2 dB HL. The difference in mean values between the
behavioural and electro-physiological minimum response levels was 22 dB both at 2
& 4 kHz.
109
University of Pretoria etd – Marais, J J (2004)
These figures are compared to those obtained by Picton and colleagues (1998) as
there is very limited published literature available. Similarities as well as differences
between the current study and that of Picton and colleagues (1998) will be highlighted
during the description of results as well as in the subsequent discussion. The ASSEP
minimum response level obtained at 0.5 kHz in the current study is 35 dB SPL (26 dB
HL), corrected for sound-field speaker presentation, which closely correlates to the
31dB SPL obtained by Picton and colleagues (1998) for the same frequency. The 24
dB SPL (18dB HL) again corrected for sound-field conditions, obtained in the current
study at 1 kHz is significantly lower than the 36 dB SPL obtained by Picton and
colleagues (1998). However, at 2 kHz Picton and colleagues (1998) reported an
ASSEP minimum response level of 26 dB SPL, which is lower than the 33 dB SPL
(21 dB HL) obtained in the current study. For the comparison at 4 kHz Picton and
colleagues (1998) obtained a mean ASSEP minimum response level of 26dB SPL,
which is significantly lower than the 51 dB SPL (36 dB HL) obtained in the current
study. This relates to a mean difference of 20 dB between behavioural thresholds and
electro-physiological minimum response levels obtained by Picton and colleagues
(1998) compared to a mean difference of 33 dB in the current study. The mean values
for both 2 and 4 kHz are highlighted in Table 4.7, above, in order to emphasize the
higher than reported ASSEP minimum response levels obtained in the current study.
As with the comparison between behavioural pure tone testing and ASSEP minimum
response levels, certain patterns emerged. Slightly smaller mean differences were
obtained at 2 kHz (22 dB) and 4 kHz (29 dB), while the mean differences at 1 kHz
were equal (18 dB) and the mean difference at 0.5 kHz was slightly bigger (25 dB)
when the behavioural modulated tone thresholds were compared to the ASSEP
minimum response levels.
As mentioned earlier in this section, there is very little literature available to compare
the behavioural mixed modulated tone thresholds against, whether it is under insert
earphone- or sound-field conditions. As with the behavioural ASSEP minimum
response levels obtained under insert earphone conditions, there were no greater than
5 dB differences between the behavioural pure tone thresholds and those behavioural
thresholds obtained with a modulated stimulus (modulated tone). Although for most
frequencies, the mean difference between the behavioural FM tones and the ASSEP
minimum response levels compared to the mean difference between the behavioural
110
University of Pretoria etd – Marais, J J (2004)
mixed modulated tone thresholds and the ASSEP minimum response levels, seem
marginal there are certain discrepancies.
The standard deviations for the behavioural FM tone technique, using sound-field
speaker presentation varied between 4 dB and 6 dB for all the frequencies tested.
Standard deviations for the ASSEP technique, using insert earphones, varied between
10 dB and 12 dB. The greatest standard deviation for the ASSEP technique was found
at 0.5 kHz, namely 12 dB, while the smallest difference was found at 1 and 2 kHz,
namely 10 dB.
Statistical analyses of the thresholds obtained under sound-field conditions, for
behavioural testing (both FM tone- and mixed modulated tone and those obtained
under ASSEP testing differ significantly from each other. For all the frequencies
tested (0.5 - 4 kHz) both sets of behavioural data showed greater than 0.05 p-values
when compared to the ASSEP minimum response levels. These sets of data, therefore,
differ significantly from each other on a 5% significance level. The numerical data
from the statistical analysis is available in Appendix D.
This experiment was, again, controlled for gender- and ear effects. At 5 kHz there was
a significant difference relating to gender effects between male and female research
participants, when ASSEP minimum response levels were analyzed. Therefore, on a
5% level of significance the minimum response levels obtained from male and female
research participants differed significantly from each other. However, when these
results are viewed on a 1% level of significance there is no statistical difference
between the genders. Although gender might have some influence on the results
obtained (discussed earlier in this section), the influence does not seem to be related to
specific frequencies. Stricter analysis (p < 0.01) again revealed no significant
difference, and further studies are required to address the possible influence of gender
on the minimum response levels obtained from ASSEP testing. For all the other
frequencies there were no statistical differences on either a 1% or a 5% level of
significance. These results are available in Appendix J.
When the same comparison was made for ear effects there were no statistical
differences between the ears for either a 1% or 5% level of significance. These results
are available in Appendix K.
111
University of Pretoria etd – Marais, J J (2004)
4.3.2 Comparison across transducers
The comparison across transducers will detail the discussion of thresholds obtained
through behavioural testing and ASSEP testing. The discussion of transducers will be
subdivided for thresholds obtained with pure tone stimulation and those obtained with
modulated tone stimulation.
4.3.2.1 ASSEP Minimum Response Levels obtained at 0.5, 1, 2 and 4 kHz, under
Insert Earphone and Sound-field Conditions
Table 4.10 provides a summary of the comparison between the mean ASSEP
minimum response levels obtained from the monotic ASSEP technique, using insert
earphones and sound-field speaker presentation at 0.5, 1, 2 & 4 kHz.
Table 4.8 Mean and Standard Deviation values for ASSEP minimum response
levels under insert earphone and sound-field conditions (n = 50)
ASSEP
Minimum
Levels (in dB SPL)
Using Insert Earphone
Presentation
Using Sound-field
Speaker Presentation
Response
Mean
(SD)
Mean
(SD)
500 Hz
1000 Hz
2000Hz
4000 Hz
44 (+/-12)
29 (+/-10)
34 (+/-10)
29 (+/-11)
35 (+/-12)
24 (+/-10)
32 (+/-10)
51 (+/-11)
ASSEP minimum response levels obtained, under insert earphone conditions, as well
as those obtained under sound-field conditions, differ substantially from one another
at 4 kHz while the mean differences across 0.5-2 kHz were marginal. The greatest
difference (22 dB) for the comparison was found at 4 kHz where the mean value for
the electro-physiological minimum response level, using insert earphones, was 29 dB
SPL, compared to the electro-physiological minimum response levels, using soundfield speaker presentation, of 51 dB SPL. The smallest difference (2 dB) in mean
values was found at 2 kHz where the mean value for the ASSEP minimum response
level using insert earphones was 34 dB SPL compared to the mean ASSEP minimum
response level value of 32 dB SPL obtained under sound-field conditions. The
difference in mean values across the transducers was 9 dB at 0.5 kHz and 5 dB at 1
kHz.
112
University of Pretoria etd – Marais, J J (2004)
Picton and colleagues (1998) found no significant differences between the SPL
minimum response levels with insert earphones or sound-field speakers. It is
important to note however that their study did not make use of repeated measures
since the subjects were different for the two measures (Picton et al., 1998). The
current study did make use of repeated measures within the same group of research
participants. There are however certain similarities when the two studies are
compared. As is the case with the current study, Picton and colleagues (1998) also
found the smallest mean difference across transducers to be at 2 kHz. Both studies
obtained a mean difference of 2 dB SPL across the transducers.
The 9 dB mean difference obtained in the current study at 0.5 kHz is marginally
higher than the 6 dB mean difference obtained by Picton and colleagues (1998).
Picton and colleagues (1998) obtained a mean difference of 4 dB across the
transducers at 1 kHz, which correlates very well with the 5 dB obtained in the current
study for the same comparison. For the comparison at 4 kHz, however, the study done
by Picton and colleagues revealed a mean difference of 4 dB across the transducers,
which is significantly lower than the 22 dB obtained during the current study. It is
important to note that multiple simultaneous stimuli were used during the data
collection of the study done by Picton and colleagues (1998), while the present study
employed single sequential stimuli.
Figure 4.3, below, provides a visual interpretation of the results obtained across
transducers (in dB SPL) at 0.5, 1, 2 & 4 kHz, in an attempt to clarify the subsequent
interpretation of the results.
113
University of Pretoria etd – Marais, J J (2004)
70
Mean Values (in dB SPL)
60
50
40
30
20
10
0
500
1000
2000
4000
Monotic ASSEP using Insert
Earphone Presentation
44
29
34
29
Monotic ASSEP using Sound
Field Speaker Presentation
35
24
32
51
Frequencies Tested (in Hz)
Monotic ASSEP using Insert Earphone Presentation
Monotic ASSEP using Sound Field Speaker Presentation
Figure 4.3 Electro-physiological Data, obtained using the ASSEP Technique
across Transducers
Mean minimum response levels obtained through electro-physiological testing across
transducers, when using a monotic ASSEP stimulus, revealed no greater than 5 dB
differences for 1 or 2 kHz. At 0.5 kHz a mean difference of 9 dB can be observed,
while the mean difference obtained at 4 kHz was 22 dB. The electro-physiological
data obtained using the different transducers seem to have a close resemblance, with
the exception at 4 kHz.
However, when the mean ASSEP minimum response levels obtained using the
different transducers were analyzed statistically certain statistical differences were
noted. All the frequencies tested did show significant differences for both p < 0.05
and p < 0.01, with the exception of 2 kHz which showed no significant difference for
p < 0.05. The results from the statistical analysis are presented in Appendix G.
Therefore, only the minimum response levels obtained at 2 kHz are comparable on a
5% level of significance.
4.3.3 Comparison across Stimulus Presentations and Transducers
The comparison across stimulus presentations and transducers will detail the
discussion of thresholds obtained through behavioural testing and the minimum
response levels obtained through monotic ASSEP testing.
114
University of Pretoria etd – Marais, J J (2004)
ASSEP Minimum Response Levels at 0.5, 1, 2 and 4 kHz compared to
behavioural thresholds, under Insert Earphone Presentation as well as
Sound-field Speaker Presentation
Figure 4.4, below provides a visual interpretation of the behavioural thresholds
and ASSEP minimum response levels obtained with the different transducers at
0.5, 1, 2 & 4 kHz.
40
35
30
Mean Values (in dB HL)
25
20
15
10
5
0
-5
500
1000
2000
4000
Behavioral pure tones using insert
earphone presentation
7
1
1
1
Behavioural FM tones using sound
field speaker presentation
4
-2
-1
3
Monotic ASSEP MRL's using insert
earphone presentation
35
25
26
27
Monotic ASSEP MRL's using sound
field speaker presentation
26
18
21
36
Frequencies Tested (in Hz)
Behavioral pure tones using insert earphone presentation
Behavioural FM tones using sound field speaker presentation
Monotic ASSEP MRL's using insert earphone presentation
Monotic ASSEP MRL's using sound field speaker presentation
Figure 4.4 Mean Pure Tone, FM Tone Thresholds and Mean ASSEP Minimum
Response Levels obtained using both Insert Earphone and Sound-field Speaker
Presentation
When the mean differences between the thresholds and minimum response levels are
compared, within the same transducer, the following figures are noted: At 0.5 kHz the
mean difference for insert earphone conditions was 28 dB compared to the 22 dB
115
University of Pretoria etd – Marais, J J (2004)
difference for the sound-field conditions. Mean differences obtained at 1 kHz
revealed a 24 dB difference for the insert earphone conditions, while the sound-field
conditions revealed a difference of 20 dB. Insert earphone conditions revealed a 25
dB a mean difference at 2 kHz, while sound-field conditions revealed a mean
difference of 22 dB for the same frequency. The greatest mean differences were
recorded at 4 kHz where a value of 26 dB HL was obtained for the insert earphone
conditions and a value of 33 dB HL for the sound-field speaker presentation.
As discussed earlier in this chapter, there were no greater than 5 dB differences
between the behavioural thresholds (across transducers) however statistical
differences (Appendix G) were noted for all the frequencies tested (p < 0.05).
Statistical analyses of the ASSEP minimum response levels obtained under insert
earphone- and sound-field conditions revealed significant differences (Appendix G)
for all the frequencies tested except 2 kHz.
Other researchers (Picton et al., 1998) have reported no significant differences
between results obtained under insert earphone conditions and those obtained under
sound-field conditions. However, in light of the aforementioned variation in mean
differences, as well as the fact that statistical analyses highlighted several significant
differences between the results obtained using the monotic ASSEP technique with
different transducers, the current study cautions against comparison across
transducers.
Figure 4.5 and 4.6 below provides a visual representation of the frequency
distribution. It represents the percentage of ASSEP minimum response levels acquired
at different intensity levels.
116
University of Pretoria etd – Marais, J J (2004)
50
45
40
35
30
Percentages 25
20
15
10
5
0
5.0-15
20-25
30-35
40-45
50+
Minimum Response Levels (in dB SPL)
0.5kHz
1 kHz
2 kHz
4 kHz
Figure 4.5 Frequency Distribution of Minimum Response Levels (in
dB SPL), under Insert Earphone Conditions
50
45
40
35
30
Percentages 25
20
15
10
5
0
0-15
20-25
30-35
40-45
50+
Mininmum Response Levels (In dB SPL)
0.5kHz
1 kHz
2 kHz
4 kHz
Figure 4.6 Frequency Distribution of Minimum Response Levels (in
dB SPL, under Sound-field conditions
117
University of Pretoria etd – Marais, J J (2004)
All the minimum response levels presented in Figure 4.5 and Figure 4.6 are
expressed in dB SPL. The correction factors for the conversion of minimum response
levels from dB HL to dB SPL, under insert earphone and sound field conditions are
available in Appendix L. The majority of minimum response levels, obtained under
insert earphone conditions (figure 4.5), were recorded at 30-35 dB SPL, across 1-4
kHz. However, the percentage of minimum response levels for 0.5 kHz, where the
equal percentages of the minimum response levels were recorded at 40-45 dB SPL
and 50+ dB SPL.
The minimum response levels obtained under sound-field conditions (Figure 4.6) did
not show a uniform distribution for all the frequencies tested (0.5-4 kHz). The
majority of the minimum response levels recorded for 0.5 kHz was between 30-45 dB
SPL. The minimum response levels recorded for 1 kHz revealed that the greatest
percentage of MRLs was recorded between 0-25 dB SPL. Thirty two percent of the
minimum response levels, recorded for 1 kHz were recorded between 0-15 dB SPL,
which was exactly the same as the percentage recorded for 20-25 dB SPL. The
greatest percentage of the minimum response levels recorded for 2 kHz were between
20-35 dB SPL (with equal percentages, namely 36% were recorded between 20-25
and 30-35 dB SPL). The analysis of the minimum response levels recorded at 4 kHz
revealed that the greatest percentage was recorded at 50+ dB SPL.
4.3.4 Discussion of the Comparison across Stimulus Presentation, Transducer
and a combination of Stimulus Presentation and Transducer
The comparison across stimulus presentation, transducer and the combination of
stimulus presentation and transducer will be discussed in terms of the influence of
certain variables, and not separately as was the case with the discussion of the first
sub aim. The variables that will be discussed are the influence of noise, the influence
of amplitude modulation on frequency specificity, the physiological influence on
response estimation and the influence of the acoustic environment.
Influence of noise:
The technique implemented in determining whether a response was present or not, in
this case the Fast Fourier Transform (FFT), might provide insight regarding the
influence of noise. The signal-to-noise ratio in the frequency domain depends equally
on the variability and the amplitude of the response in the FFT bins adjacent to the
118
University of Pretoria etd – Marais, J J (2004)
signal (Regan and Regan, 1989; Regan 1989). Furthermore the ASSEP signal has its
“power” concentrated on the modulation frequency or some harmonics thereof (Lins
et al., 1995). Because of the high resolution of the FFT, the power of the signal is
greatly confined to a single FFT bin, while the noise is distributed across a wider
frequency spectrum (Lins et al., 1995). By increasing the resolution of the FFT, the
signal-to-noise ratio also increases. According to Lins and colleagues (1995) the noise
levels can be decreased through both averaging and decreasing the width of the
frequency bin. However in the current study the researcher did not have the option to
alter the prescribed parameters and therefore it is necessary to further investigate the
influence of noise on the ASSEP minimum response levels that were obtained.
It is possible though that the presence of ambient noise levels within the test
environment could have acted as a background masking noise (Perez-Abalo et al.,
2001), thereby elevating the pure tone thresholds without affecting the ASSEP
minimum response levels (Picton et al., 1998). This phenomenon could possibly have
a more pronounced effect on the behavioural thresholds obtained under sound-field
conditions as these thresholds are also influenced by other factors within the acoustic
environment, such as direct versus reverberant field testing, absorptivity of the rooms’
surfaces, directivity of the speakers used and head movements of the participants
being tested (Walker et al., 1984). In the current study the presence of ambient noise
did not influence the sound-field conditions more adversely than those obtained under
insert earphones, with the possible exception of the minimum response levels
obtained for 4kHz. This can be seen when the results obtained (see Figure 4.4) are
considered.
According to Picton and colleagues (1998), the background noise could possibly
imitate a mild sensorineural hearing loss with recruitment. Other authors (Lins et al.,
1996; Rickards et al., 1994) have commented that the difficulty in obtaining ASSEP
minimum response levels at 0.5 kHz could partly be attributed to “the enhanced
masking effect” of ambient noise at lower frequencies. This correlates well with
results obtained under insert earphone conditions in the current study, as the greatest
mean values were obtained at 0.5 kHz (35 dB HL). Other studies, such as Aoyagi and
colleagues (1994) have noted the same difficulty in estimating low frequency hearing
sensitivity, where the lower frequencies (0.25 kHz) showed the greatest mean values
(34 dB HL). Rickards and colleagues (1994) also reported difficulty (50 dB SPL) in
119
University of Pretoria etd – Marais, J J (2004)
the estimation of low frequency thresholds (0.5 kHz). Valdes, Perez-Abalo, Martin,
Savio, Sierra, Rodriguez and Lins (1997) obtained similar results, which relate to the
difficulty in estimating low frequency thresholds.
The results obtained under sound-field conditions also show a greater than average
ASSEP minimum response level (26 dB HL) at 0.5 kHz. However the greatest mean
ASSEP minimum response level under sound-field conditions was obtained at 4 kHz.
Although the influence of the “enhanced masking effect” of ambient noise at lower
frequencies also holds true for results obtained under sound-field conditions, other
additional influences now comes into play. These influences (with specific reference
to 4 kHz) will be discussed under effects of the acoustic environment, later in this
section.
There is a significant difference between the actual behavioural pure tone thresholds
obtained under either insert earphone- or sound-field conditions, when compared with
the ASSEP minimum response levels obtained with the same transducers. Although
ambient noise levels were measured for this study (results available in Appendix M),
it seems that noise levels in the test room, although higher than the prescribed ANSI
standards, did not have a significant influence on the results obtained. The noise levels
did, however, comply with the standards set out by the South African Buro of
Standards (SABS). This is deducted from the fact that, for both insert earphone
presentation as well as sound-field speaker presentation, the behavioural pure tone
thresholds obtained did not differ significantly from the behavioural modulated tone
thresholds. This difference was within 5 dB HL for 0.5, 1, 2 & 4 kHz, which
according to Katz (1985) is acceptable seeing as a minimum standard error of +/- 5dB
is introduced for behavioural testing. Furthermore there was no significant difference
between these sets of data when they were statistically analyzed. It seems therefore
that other variants, rather than ambient noise had an influence on the ASSEP
minimum response levels obtained under insert earphone conditions, again with the
possible exception of 0.5 kHz. The possible influence of the acoustic environment
will be discussed later on (under effects of the acoustic environment) in this chapter.
The Influence of Amplitude Modulation on Frequency Specificity
When the results of the current study are considered in terms of frequency specificity,
the minimum response levels obtained provided reasonable estimates of the
120
University of Pretoria etd – Marais, J J (2004)
behavioural pure tones for all the frequencies tested. Amplitude modulated stimuli
are reasonably frequency specific, and therefore the minimum response levels
recorded should not be affected by “frequency splatter” observed in brief duration
stimuli such as clicks (Lins et al., 1996). The problem however is not related to
frequency specificity as much as to “place-specificity” (Starr & Don, 1988). When
low frequency stimuli are presented at moderate to high intensity levels they can
possibly cause activation of the more basal regions of the cochlea, which is concerned
with high frequency stimulation, and in so doing cause distortion of the response
(Lins et al., 1996). However, stimulation at higher intensity levels was not required
during the current study, as all the participants had normal-hearing sensitivity. It is
highly unlikely that the minimum response levels obtained in the current study were
“distorted” as the modulation rate employed already ensures frequency specificity
(Lins et al., 1996) and the presentation of stimuli was below moderate to high
intensities.
Modulation effects, with specific reference to results obtained under sound-field
conditions also have to be considered. According to Picton and colleagues (1998) the
depth of modulation is bound to decrease from 100%, because of acoustic reflections
within the room arriving at the participants’ head with different delays. This might
account for the statistical differences noted under different transducers (for all
frequencies tested except 2 kHz). Although a high probability exists that the
modulation might have decreased to some extent, the effect thereof is bound to be
minimal, since the response is stable at modulation depths that exceed 50 percent
(Lins et al., 1995). The possibility exists, that other influential variables, rather than
modulation effects attributed to the statistical differences observed across transducers,
in the current study.
The Physiological Influence on Response Estimation
When considering the results obtained in the current study it is evident that the MRLs
at 0.5 kHz reveal a much greater mean difference when compared to the actual
behavioural pure tone thresholds, than any other frequency except for 4kHz under
sound-field conditions. Other reports (Aoyagi et al., 1994; Lins et al., 1996; Rance et
al., 1995) mention similar findings when using either sequential (single) frequency
presentation or simultaneous (multiple) frequency presentation. Lins and colleagues
121
University of Pretoria etd – Marais, J J (2004)
(1996) related the difficulty in estimating 0.5 kHz thresholds to the influence of
ambient noise (previously discussed).
Prominent among other explanations is the explanation that higher frequencies, during
multiple frequency stimulation, would interact and most likely affect the estimation of
the 0.5 kHz response, either through suppression or masking (Perez-Abalo et al.,
2001). The fact that other researchers (Aoyagi et al., 1994; Valdes et al., 1997), while
using sequential frequency presentation, have experienced similar difficulties makes
this explanation unlikely.
A further explanation is related to group delays (Goldstein et al., 1971), which
postulates that phases of the ASSEP can be sensibly converted to latencies and that
these latencies are related to frequency-related delays (John & Picton, 2000).
According to John & Picton (2000) there are five main factors that contribute to the
latency of a neuro-physiological response, namely acoustic delays, transport delays,
filter built-up times, synaptic delays and conduction delays. The problem is that the
latency of a neural response is not simply a sum of these five delays and that the
recorded response may not derive from only a single generator. However, many
different regions of the auditory system generate responses following the stimulation
through a modulated tone. Each of these regions has a specific phase and latency
relationship to the stimulus and the recorded response on the scalp is made up from an
overlapping of all of these responses (John & Picton, 2000).
Studies performed on animals (Joris & Yin, 1992) have found group delays relating to
the carrier frequency used around 0.5 kHz. These results cannot be used to validate
present findings in human participants, as there are specie-related differences. Some
of the larger delays could, however, probably relate to frequency-related change in
conduction delay (John & Picton, 2000). The results obtained in the current study
substantiate reports in the literature, as mean minimum response levels were elevated
for 0.5kHz, and more so in males than females. These elevated minimum response
levels can possibly be attributed to differences in length of the basilar membrane
found between males and females (John & Picton, 2000).
Lastly, threshold estimation in the lower frequency region (i.e. 0.5 kHz) might be
problematic due to the intrinsic characteristics of the responses. Lins and colleagues
122
University of Pretoria etd – Marais, J J (2004)
(1996) mentioned that the responses to lower frequencies are attenuated because of
“jitter” in the transmission time between cochlear receptors and the neural generators
of these responses. This “jitter” would then go on to cause the neurons to
desynchronize in their generation of responses, and subsequently this will then lead to
a decrease in response amplitude (Lins et al., 1996). The “jitter” mentioned by Lins
and colleagues (1996) could be caused by a wider activation of the basilar membrane.
It is important to mention however, that their theory is specifically related to the
activation of the basilar membrane in infants and might not hold true for adult
subjects.
It seems therefore that conduction delays (John & Picton, 2000), related to the specific
carrier frequency, as well as the effect of ambient noise (Lins et al., 1996) might cause
problems when estimating thresholds in the lower frequency region. Research
conducted by Herdman & Stapells (2001) contradicted findings by other researchers
(Aoyagi et al., 1994; Rickards et al., 1994), reporting minimum response levels within
10 dB of the behavioural thresholds. They attributed their results to low levels of
ambient noise (10-12 dB SPL). Picton and colleagues (1998) suggested that
irrespective of the amount of averaging used, minimum response levels are highly
unlikely to be recorded within 10 dB of the hearing threshold, due to the latency
“jitter” at near threshold intensities. Even though the threshold estimations obtained
by Herdman & Stapells seem promising, proving that the ASSEP technique can
estimate thresholds within 10 dB of the actual behavioural threshold, some concerns
regarding clinical feasibility arise. Stringent threshold criteria, such as prolonged
averaging and low ambient noise levels, employed by Herdman & Stapells (2001)
translate to prolonged test time, which, within the clinical setup is not always feasible.
The question regarding the feasibility of prolonged test time is aggravated within the
difficult-to-test population. Further studies, however, are needed to investigate these
possibilities.
Influence of the Acoustic Environment
Although ambient noise levels in the test room might have an effect on the
behavioural pure tone thresholds obtained under insert earphone conditions, as
discussed earlier, the effects of the acoustic environment have a more pronounced
effect on results obtained under sound-field conditions. Arlinger & Jerlvall (1987)
commented that “when earphones cannot be used the interaction between the test
123
University of Pretoria etd – Marais, J J (2004)
stimuli and the test room acoustics assumes critical importance” (Arlinger &
Jerlvall, 1987:21). When the results obtained during monotic ASSEP testing under
sound-field conditions are considered, it seems likely that the acoustic environment
had at least some impact on the minimum response levels obtained at 4 kHz.
There is limited literature available regarding the effect of the acoustic environment
on the modulated stimuli used during monotic ASSEP testing. This becomes evident
when compared to the extensive research (Arlinger & Jerlvall, 1987; Beynon &
Munro, 1995; Dillon, 1982; Dillon & Walker, 1980; Dillon & Walker, 1982; Orchik
& Rintelman, 1978; Walker et al., 1984) available regarding the influence of the
acoustic environment on behavioural pure tone measures obtained in the sound-field.
According to Beynon & Munro (1995) there are several different directional
components, which show a more pronounced effect at higher frequencies. This might
serve as explanation as to why the minimum response level obtained at 4 kHz under
sound-field conditions was elevated (36 dB HL/ 51 dB SPL). Although ASSEP testing
was conducted in the direct field and control measures were taken to ensure minimal
head movement by the research participants, the effects of the acoustic environment
might still have influenced the ASSEP minimum response level obtained at 4 kHz.
Some of these factors include reflective surfaces, shape of the room, placement of
research participants and test arrangements.
During the current study the following measures were taken to ensure minimal
interference through test room acoustics. All reflective surfaces were covered with
material, with the exception of the walls of the test room. The problem is that the test
room wall, although sound treated, could possibly be somewhat reflective. According
to Walker, (1979) the influence of any reflective surface becomes more pronounced as
its size increases and the closer the proximity to the person being tested. As one of the
walls was uncovered and directly next to the research participants, reflectivity might
have impacted on the results obtained and especially for the higher test frequencies.
This could possibly account for the “right-ear-effect” (observed at some frequencies)
p<0.05 analyzed for ear effects. Other researchers have also observed this effect
(Picton et al., 2002), however it is worth mentioning that they used noise stimulation.
The test room complies with prescribed standards for audiometric testing (ANSI,
1996), rendering this explanation somewhat unlikely. The “right-ear-effect” has been
124
University of Pretoria etd – Marais, J J (2004)
described in the literature as the “right ear advantage” and postulates that stimuli
presented to the right ear are more quickly and accurately processed (Clark, 2002).
All subjects were placed at a zero degree azimuth in relation to the speaker for the
electro-physiological testing, done in the direct field. According to Walker and
colleagues (1984) the frequency response obtained in the reverberant field will differ
little from the measurements made in the direct field at an azimuth of zero degrees.
Research participants were, furthermore placed so that the centre of their heads were
in line with, and at the same height, as the centre of the loudspeaker’s cone. It seems,
therefore, that subject placement did not have a pronounced effect on the ASSEP
minimum response levels obtained. The elevated minimum response level obtained at
4 kHz under sound-field conditions could therefore most likely be attributed to
directional components affecting higher frequencies more adversely than lower
frequencies (Beynon & Munro, 1995).
4.4 A Comparison between the actual Behavioural Thresholds (for both pure
tone and modulated tone stimulation) obtained and the GSI AUDERA System’s
accuracy in predicting these Thresholds by means of an Algorithm, for Insert
Earphone and Sound-field Conditions
This sub-aim will detail and discuss the results from four related groups of data
pertaining to its focus. The results will be discussed as listed below:
The accuracy of the GSI AUDERA System’s Algorithm in predicting Behavioural
Thresholds obtained under Insert Earphone conditions
The accuracy of the GSI AUDERA System’s Algorithm in predicting Behavioural
Thresholds obtained under Sound-field conditions
Discussion on the accuracy of the GSI AUDERA System’s Algorithm in
predicting Behavioural Thresholds obtained under both Insert Earphone and
Sound-field Conditions
4.4.1 The accuracy of the GSI AUDERA System’s Algorithm in predicting
Behavioural Thresholds obtained under Insert Earphone Conditions
Table 4.9, below provides a summary of the behavioural pure tone thresholds,
obtained under insert earphone conditions, as well as the predicted behavioural
thresholds (based on the calculation by the GSI AUDERA system’s algorithm).
125
University of Pretoria etd – Marais, J J (2004)
Table 4.9 Mean and Standard Deviation values for Behavioural Pure Tone
thresholds, Behavioural monotic ASSEP minimum response levels and the
Predicted Thresholds under insert earphone conditions (n = 50)
Under Insert Earphone
Conditions
Behavioural Pure Tone
Thresholds
Behavioural Monotic
ASSEP Thresholds
Predicted Behavioural
Pure Tone Thresholds
Mean
(SD)
Mean
(SD)
Mean
(SD)
500 Hz
1000 Hz
2000Hz
4000 Hz
7 (+/-7)
1 (+/-5)
1 (+/-6)
1 (+/-7)
7 (+/-7)
3 (+/-5)
2 (+/-5)
3 (+/-7)
12 (+/-11)
9 (+/-10)
11 (+/-8)
7 (+/-9)
The predicted thresholds, based on the algorithm of the GSI AUDERA system, are
more predictive of the actual behavioural monotic ASSEP minimum response levels
across 1-2 kHz, when compared to the behavioural pure tone thresholds obtained
under insert earphone conditions. The only instance where the predicted threshold
resembled the actual mean behavioural pure tone threshold more closely was at 4
kHz, while identical mean differences (5 dB) were obtained at 0.5 kHz for both the
behavioural pure tones as well as the behavioural monotic ASSEP minimum response
levels. The smallest mean difference between the predicted and actual pure tone
thresholds was found at 0.5 kHz, namely 5 dB. The greatest mean difference was
recorded at 2 kHz, namely 10 dB. The mean difference obtained for 1 kHz was 8 dB,
while the mean difference at 4 kHz was 6 dB.
When the comparison was drawn between the predicted thresholds and the actual
behavioural monotic ASSEP minimum response levels, the following mean
differences were noted. The smallest mean difference was recorded for 4 kHz were
the mean difference was 4 dB. The greatest mean difference was found at 2 kHz,
namely 9 dB. The mean difference recorded for 0.5 kHz was 5 dB, while 1 kHz
recorded a mean difference of 6 dB. Standard deviations for the predicted thresholds
were slightly elevated (8-11 dB) in comparison to the actual behavioural thresholds
obtained using the pure tone (5-7 dB) and monotic ASSEP technique (3-7 dB).
Although most literature do not make mention of algorithms employed by the
equipment used, subsequently making comparison with relevant literature
problematic, it seems that the GSI AUDERA system is quite accurate in predicting
behavioural pure tone thresholds obtained under insert earphone conditions. Mean
126
University of Pretoria etd – Marais, J J (2004)
differences varied between 5-10 dB above the actual threshold. Slightly smaller
variations were observed when predicted thresholds were compared to the behavioural
monotic ASSEP minimum response levels, namely 4-9 dB. When the mean
differences from the two comparisons were analyzed on a frequency-to-frequency
basis there were no greater than 5 dB differences between them.
There were significant differences between the predicted pure tone thresholds (as
based on the GSI AUDERA system’s algorithm) and the actual pure tone thresholds,
obtained under insert earphone conditions, for all the frequencies tested, except at 0.5
kHz on a 5% level of significance. These results are available in Appendix D. It is
important to note, however that for 1-4 kHz the differences were within 10 dB from
each other, which in patients with no reliable behavioural thresholds are still
acceptable.
4.4.2 The accuracy of the GSI AUDERA System’s Algorithm in predicting
Behavioural Thresholds obtained under Sound-field Conditions
Table 4.10, below provides a summary of the behavioural FM tone thresholds,
obtained under sound-field conditions, as well as the predicted behavioural thresholds
(based on the calculation by the GSI AUDERA system’s algorithm).
Table 4.10 Mean and Standard Deviation values for Behavioural FM Tone
thresholds, Mixed Modulated Tone thresholds and the Predicted Thresholds
under sound-field conditions (n = 50)
Under Sound-field Conditions
Behavioural FM Tone
Thresholds
Behavioural Mixed
Modulated Tone Thresholds
Predicted Behavioural FM
Tone Thresholds
Mean
(SD)
Mean
(SD)
Mean
(SD)
500 Hz
1000 Hz
2000Hz
4000 Hz
4 (+/-6)
-2 (+/-4)
-1 (+/-5)
3 (+/-5)
1 (+/-6)
0 (+/-5)
-1 (+/-4)
7 (+/-7)
12 (+/-11)
9 (+/-10)
11 (+/-8)
7 (+/-9)
Identical mean differences (12 dB) were obtained at 2 kHz for both the behavioural
FM tones, as well as the behavioural mixed modulated tone thresholds. For the
comparison between the behavioural FM tone thresholds and the predicted thresholds
the smallest mean difference between the predicted and actual FM tone thresholds
was found at 0.5 kHz, namely 8 dB. The greatest mean difference was recorded at 1
kHz, namely 11 dB. The mean difference obtained 4 kHz was 4 dB.
127
University of Pretoria etd – Marais, J J (2004)
The smallest mean difference (0 dB) between the behavioural mixed modulated tone
thresholds and the predicted thresholds were recorded at 4 kHz, where the predicted
threshold was identical to the actual behavioural mixed modulated tone threshold. The
greatest mean difference (12 dB) was found at 2 kHz (as mentioned above), while the
mean difference for 0.5 kHz was 8 dB. The mean difference recorded at 1 kHz was 11
dB. Standard deviations for the predicted thresholds were slightly elevated (8-11 dB)
in comparison to the actual behavioural thresholds obtained using the FM tone (4 - 6
dB) and monotic ASSEP technique (4 - 7 dB). These standard deviations closely
resemble those obtained under insert earphone conditions.
There were significant differences between the predicted FM tone thresholds (as
based on the GSI AUDERA system’s algorithm) and the actual FM tone thresholds
obtained for all the frequencies tested, except at 0.5 kHz on a 5% level of significance
(when tested under sound-field conditions). These results are available in tabulated
form in Appendix D. It is important to note, however that for 1-4 kHz the differences
were within 10 dB from each other, which in patients with no reliable behavioural
thresholds are still acceptable.
4.4.3 Discussion on the accuracy of the GSI AUDERA System’s Algorithm in
predicting Behavioural Thresholds obtained under both Insert Earphone and
Sound-field Conditions
In order to facilitate the discussion of results a visual representation of the predicted
thresholds will be presented in Figure 4.7, below together with the actual behavioural
thresholds. The visual representation includes the predicted- and actual thresholds for
both insert earphone and sound-field speaker presentation.
128
Mean Threshold Values
(In dB SPL)
University of Pretoria etd – Marais, J J (2004)
40
35
30
25
20
15
10
5
0
0.5 kHz
1 kHz
2 kHz
4 kHz
Frequency (In kHz)
Mean Threshold Values
(In dB SPL)
Predicted Thresholds
Actual Thresholds
50
40
30
20
10
0
0.5 kHz
1 kHz
2 kHz
4 kHz
Frequency (In kHz)
Predicted Thresholds
Actual Thresholds
Figure 4.7 Mean Predicted Thresholds versus Actual Thresholds for insert
earphone presentation (top) and Sound-field Speaker Presentation (bottom) in
dB SPL
When the predicted thresholds (based on the GSI AUDERA System’s algorithm) are
compared across the transducers it is clear that for all the frequencies tested except 4
kHz they more closely approximate the thresholds obtained under insert earphone
conditions. However given the high standard deviations these figures are unlikely o
have a significant influence on predictive value of the algorithm employed by the GSI
AUDERA system. When the predicted thresholds are compared with the behavioural
modulated tone thresholds, they (predicted thresholds) again more closely
approximate the thresholds obtained under insert earphone conditions, with the
exception of 4 kHz. As the GSI AUDERA system is the first ASSEP test protocol to
include an algorithm for the prediction of actual hearing thresholds, the limited
amount of available research literature becomes evident.
129
University of Pretoria etd – Marais, J J (2004)
It is therefore almost impossible to validate these findings in terms of relevant
literature. It is, however, clear from the results obtained, that the system was quite
accurate in predicting both the behavioural pure and -FM tone thresholds as well as
the behavioural mixed modulated tone thresholds regardless of the transducer used. It
is important, though, to mention that in the majority of the cases analyzed the
predicted thresholds more closely approximated the thresholds obtained under insert
earphone conditions, regardless of the stimulus presentation employed. This is likely
to be because the algorithm is based on research that focused on threshold estimations
obtained under insert earphone conditions. Therefore, the effect of the acoustic
environment as well as transducer-related influences was most likely not taken into
account when the algorithm was developed. Further investigation into this matter will
be required before assumptions regarding the accuracy of the GSI AUDERA System’s
algorithm can be validated. Put simply, before clinicians can rely on the predictive
function (algorithm) of the GSI AUDERA system further research will have to be
done to validate these preliminary findings.
4.5 Summary
This chapter reported and discussed the results obtained in the current study according
to the three sub-aims specified in chapter three. These sub-aims were specifically
selected in an attempt to address the main aim of the study. Each of the sub-aims
provided results pertaining to its focus, and integrated these results with the current
literature to ascertain the validity thereof. Conclusions will subsequently be drawn in
relation to the specific sub-aims in the final chapter of the study.
130
University of Pretoria etd – Marais, J J (2004)
CHAPTER FIVE: CONCLUSIONS
This chapter aims to draw relevant conclusions and stipulate
implications obtained from the current study. It will, furthermore,
critically evaluate the findings obtained in the current study and
make recommendations for further research.
5.1 Introduction
Since the early 1980’s objective procedures such as the Auditory Brainstem Response
(ABR) have been used, not only to categorize, but also to assess appropriate
amplification fittings (Beauchaine & Gorga, 1988; Hecox, 1983; Kileny, 1982;
Mahoney, 1985). Since then the literature is replete with certain disadvantages
(discussed in the second chapter of this study) relating to the technique and this has
subsequently led to limited application of the ABR in the assessment of amplification
within the clinical arena. However with the arrival of Auditory Steady State Evoked
Potentials on the diagnostic scene, the old question of whether it will be a suitable
alternative in the aided Auditory Evoked Potential test battery has resurfaced.
Preliminary investigations (Cone-Wesson et al., 2001; Picton et al., 1998) have stirred
interest in the validation of Auditory Steady State Evoked Potentials, as a valuable
instrument during the initial habilitation of hearing-impaired individuals. Picton and
colleagues (1998) found that the minimum response levels obtained in their study
correlated well with previous research using normal-hearing subjects. The average
minimum response levels from their study (Picton et al., 1998) ranged between 26-36
dB SPL. Other authors seem to be in agreement (Lins et al., 1996; Aoyagi et al.,
1994). These SPL minimum response levels were significantly higher than the normal
HL levels from the research participants tested (Picton et al., 1998). They continue to
report the difference between electro-physiological minimum response levels and
behavioural thresholds to range between 18-26 dB. When compared to their previous
study (Lins et al., 1996), this is a greater difference. They (Picton et al., 1998)
contributed the higher differences obtained to the acoustic environment in which the
131
University of Pretoria etd – Marais, J J (2004)
data was collected. Unlike their (Picton et al., 1998) previous study (1994), the latter
(1998) was conducted in a sound-attenuated chamber. Therefore, the presence of lowlevel background noise in the earlier study (1994) could have elevated behavioural
thresholds without affecting the minimum response levels (Picton et al., 1998). In
other words, the background noise imitated a mild sensorineural hearing loss with
recruitment.
Their study (Picton et al., 1998) also reported that there were no differences between
stimuli presented simultaneously and those presented sequentially. The implication
for the current study would be that the minimum response levels obtained, through
sequential stimulation can be related to both the minimum response levels obtained
through sequential and simultaneous stimulation obtained in their study.
Lastly, Picton and colleagues (1998) reported that there were no significant
differences between the minimum response levels obtained under supra aural
earphone and sound-field conditions. Their measures made use of SPL measurements
and did not correct for small differences in HL levels between supra aural headphones
and sound-field conditions (Picton et al., 1998). This implies that the minimum
response levels obtained under earphone conditions would have been slightly lower,
as 0 dB HL correlates with 7 dB SPL for earphone testing and 4 dB SPL for soundfield testing (with an azimuth of 0º). According to their (Picton et al., 1998)
calculations, these differences would not have affected the results obtained.
In light of the conclusions drawn from their study, the aim of the current study was to
contribute to the existing body of knowledge by addressing the research question set
out in chapter one, namely:
How accurate is monotic auditory steady state evoked potentials in the prediction
of thresholds using sound-field stimulation, when compared to insert earphone
stimulation, in a group of normal-hearing adults?
This chapter aims to logically conclude the current research endeavour by providing
answers to the respective sub-aims in an attempt to answer the main aim. It will
subsequently highlight and discuss the strengths and weaknesses of the current study
and through this critical evaluation introduce recommendations for future research.
132
University of Pretoria etd – Marais, J J (2004)
5.2 Conclusions: Theoretical and Clinical Implications Identified
This section delineates the respective sub-aims set out in an attempt to answer the
main aim of the study. Each sub-aim is concluded in terms of the results obtained, the
theoretical and clinical implications, and finally their relation to the main aim.
5.2.1 A comparison between pure tone-, frequency modulated (FM) and mixed
modulated (AM/FM) tone behavioural thresholds at 0.5, 1, 2 & 4 kHz under
insert earphone and sound-field speaker conditions
This sub-aim was structured to obtain results, presented in relation to two
comparisons, namely a comparison across stimulus presentation, and one across
transducers. Both of these comparisons were controlled for gender- and ear effects.
The comparison of the behavioural thresholds using the two different stimulus
presentations, for both insert earphones and sound-field speaker presentation, aimed
to establish a basis from where the accuracy of the calibration of the different stimulus
presentations and transducers could be argued. These comparisons also aimed to
establish a basis from where the accuracy of the minimum response levels obtained
using the monotic ASSEP technique under both insert earphone and sound-field
conditions could be argued.
When the behavioural thresholds obtained were compared across stimulus
presentation, the pure tone-, FM tone- and mixed modulated tone thresholds
they were comparable with one another, for both insert earphone and soundfield conditions.
This finding is supported by the statistical analyses (Friedmann Test), as none of the
thresholds obtained (at 0.5, 1, 2 & 4 kHz) revealed significant differences for p values
< 0.05. The findings of the statistical analysis were supported by the fact that the
mean difference between the thresholds (obtained at each specific frequency) under
the different stimulus presentations was < 5 dB. Therefore it seems that the calibration
across pure tone and modulated stimuli are comparable. Put simply, if the thresholds,
obtained from behavioural pure tone-, FM tone- and mixed modulated tone testing,
are comparable within a specific transducer, the calibration for the two stimulus
presentations are comparable as well. Keeping in mind those factors, other than
stimulus presentation, influencing the minimum response levels obtained, and this
comparison lays the foundation for comparison of behavioural thresholds with
133
University of Pretoria etd – Marais, J J (2004)
minimum response levels. This correlates with available research (Bacon &
Viemeister, 1985; Picton et al., 1998) stating that a modulated tone should be
distinguishable from an un-modulated tone at 5dB above the hearing threshold.
The results, across stimulation presentation, were furthermore analyzed to determine
whether the gender of the research participants had a significant impact on the
thresholds obtained. The statistical analyses revealed that gender did not impact
significantly on the thresholds obtained across stimulus presentation, and within
a specific transducer, when the results were analyzed with p < 0.01. When the
results were analyzed on a 5% level of significance, most of the thresholds obtained
still did not reveal any significant differences. However, when insert earphones were
used, there was a statistical difference at 2 kHz (p < 0.05) between male and female
research participants. Thresholds obtained under sound-field conditions also revealed
a significant difference when a 5% level of significance was used to analyze the
results. However, this time the differences were observed at 0.5 kHz. Put simply, it
seems that gender does not affect the thresholds obtained under different stimulus
presentations, when viewed on a 1% level of significance.
Lastly, the results across stimulus presentation were analyzed to determine whether
ear effects had a significant impact on the thresholds obtained. When the stimulus
presentation was compared under insert earphone conditions there were no significant
differences on either a 5% or 1% level of significance. Results from stimulus
presentation under sound-field conditions revealed a significant difference at 0.5 kHz
for p < 0.05 between left and right ears for pure tone stimulation. No statistical
differences were noted for thresholds obtained under sound-field conditions between
the left and right ears for modulated tone stimulation. Thus, the statistical analyses
revealed that ear effects did not significantly impact on the thresholds obtained
across stimulus presentation, and within a specific transducer, when the results
were analyzed with p < 0.01.
134
University of Pretoria etd – Marais, J J (2004)
When compared across transducer, the behavioural thresholds obtained
under insert earphone- and sound-field conditions were not comparable with
one another, for neither pure tone-, FM tone- nor mixed modulated tone
stimulation on a statistical level.
This is based on the fact that statistical analyses revealed significant differences
between the thresholds using p < 0.05. These differences were obtained between
insert earphone- and sound-field conditions, for all the frequencies tested, irrespective
of the stimulus presentation used. When a significance level of 1% was used to
analyze the results relating to the comparison across transducers, for pure tone
stimulation, significant differences were obtained at all the frequencies tested except
at 4 kHz. When the same comparison was made for modulated tone stimulation, all
the frequencies tested showed significant differences except at 2 kHz.
Within the clinical arena the comparison across transducer, and within a specific
stimulus presentation, would be comparable for most of the frequencies tested. This is
based on the fact that the thresholds, during mixed modulated tone presentation, were
within 5 dB from one another for all the frequencies tested, except at 0.5 kHz.
Furthermore, during pure tone and FM tone stimulation the thresholds, for all the
frequencies tested, were within 5 dB from each other. Thus, the 5 dB minimum
standard error introduced for behavioural testing would still allow for comparison at
most of the frequencies tested, while the statistical test showed a greater sensitivity
and prohibits comparison across transducers.
Therefore, within the clinical arena comparison across transducers is permissible. The
behavioural thresholds obtained under insert earphone- and sound-field conditions
were within 5 dB from one another which complies with the 5 dB minimum standard
error. Put simply, it seems that although the calibration across transducers is most
likely accurate (thresholds for most frequencies were within 5 dB HL of each other)
other influences within the test environment as well as the greater sensitivity of the
statistical measures highlighted significant differences for the comparison across
transducers. Therefore, behavioural thresholds obtained under a specific transducer
could serve as the basis for comparison with minimum response levels obtained using
either insert earphones or sound-field speakers as the transducer.
135
University of Pretoria etd – Marais, J J (2004)
The results, across transducers, were furthermore analyzed to determine whether the
gender of the research participants had a significant impact on the thresholds obtained.
The statistical analyses revealed that gender did not impact significantly on the
thresholds obtained across transducers, and with specific reference to stimulus
presentation, when the results were analyzed with p < 0.01. When the results were
analyzed on a 5% level of significance, most of the thresholds obtained still did not
reveal any significant differences. However, when pure tone and FM tone stimulation
was used, there was a statistical difference at 2 kHz (p < 0.05) between male and
female research participants. Thresholds obtained through mixed modulated tone
stimulation did not reveal any significant differences for either a 1% or 5% level of
significance. Put simply, it seems that gender does not affect the thresholds obtained
under different transducers, when viewed on a 1% level of significance.
Lastly, the results across transducers were analyzed to determine whether ear effects
had a significant impact on the thresholds obtained. When the transducers were
compared under pure tone and FM tone stimulation no significant differences were
noted for any of the frequencies tested on either a 5% or 1% level of significance,
with the exception of 0.5 kHz (p < 0.05). Results from mixed modulated tone
presentation under insert earphone- and sound-field conditions revealed no significant
differences at any of the frequencies tested, for p < 0.05 or p < 0.01 between left and
right ears. Thus, the statistical analyses revealed that ear effects did not
significantly impact on the thresholds obtained across stimulus presentation, and
within a specific transducer, when the results were analyzed with p < 0.01.
To summarize, thresholds obtained under different stimulus presentations are
comparable, while thresholds obtained under different transducers are not. When
analyzed on a 1% level of significance neither gender, nor ear effects had a significant
impact on the thresholds obtained, irrespective of stimulus presentation and/or
transducers.
These findings correlate well with other reports in the literature (Bacon & Viemeister,
1985; Picton et al., 1998), stating that a 100% modulated tone is distinguishable from
a pure tone at 5 dB above the hearing threshold. The comparison of modulated tone
thresholds under the different transducers, and with specific reference to 0.5 kHz, is
the only comparison that delivered unexpected results. The 6 dB difference for
136
University of Pretoria etd – Marais, J J (2004)
modulated tone stimulation between the different transducers exceeds the minimum
standard error (Katz, 1985) of 5 dB. Within the clinical arena, where behavioural
testing is usually associated with 5 dB increments, this finding will compromise testretest reliability, regardless of whether the statistical analysis of the results found
significant differences or not.
The implication for the clinical arena is that data obtained from behavioural
tone testing is comparable across transducers. Therefore, unaided thresholds
obtained through behavioural pure tone testing, under insert earphone
conditions, are comparable with aided minimum response levels obtained under
sound-field conditions (and vice versa).
It is possible that the influence of the acoustic environment (Anderson, 1979;
Magnusson et al., 1997; Walker et al., 1984) on thresholds obtained under sound-field
conditions could have distorted the comparison with thresholds obtained under insert
earphone conditions. Picton and colleagues (1998) reported no significant differences
between results obtained under insert earphone and sound-field conditions. It is
important to note that their experiment did not make use of the same research
participants in the repeated measures, whereas the current study did. Although there is
extensive research available regarding sound-field audiometry, the only other
published literature investigating transducer influences on ASSEP was by Picton and
colleagues (1998). Further research is needed to establish if the thresholds, and
subsequently, the minimum response levels obtained under different transducers are
comparable.
5.2.2 A comparison between minimum response level detection at 0.5, 1, 2 & 4
kHz using a monotic ASSEP technique under insert earphone and sound field
speaker conditions
This sub aim was structured to obtain results, presented in relation to three
comparisons, namely a comparison across stimulus presentation, one across
transducers and the last across both stimulus presentation as well as transducers.
All of the comparisons were controlled for gender- and ear effects. The stimulus
presentation comparison aimed to establish the relationship between behavioural
thresholds and monotic ASSEP minimum response levels (controlled separately for
insert earphone and sound-field speaker presentation). The comparison relating to
137
University of Pretoria etd – Marais, J J (2004)
transducers aimed to establish the relationship between behavioural thresholds and
monotic ASSEP minimum response levels (controlled separately for pure tone and
modulated tone stimulation). From these two assessments a basis of comparison is
established from where the experiment across both stimulus presentation and
transducers can be argued. The last assessment provides the vital comparison between
the minimum response levels obtained under insert earphone and sound field
conditions. Based on the results from sub aim one, these thresholds were converted to
dB SPL in order to make the comparison valid.
The comparison relating to stimulus presentation, controlled for insert earphone
presentation, revealed that the minimum response levels obtained more closely
approximated the modulated thresholds for all the frequencies tested, except 0.5 kHz.
The difference between the mean minimum response level and the mean behavioural
thresholds for both pure tone- and mixed modulated tone stimulation at 0.5 kHz was
identical (28 dB mean difference). Considering that modulated tones are
distinguishable from un-modulated tones at 5 dB above the hearing threshold (Bacon
& Viemeister, 1985; Picton et al., 1998), the closer correlation between minimum
response levels and modulated thresholds, under insert earphone conditions, seems
accurate.
The comparison relating to stimulus presentations, controlled for the sound-field
condition, revealed mixed results with the mixed modulated thresholds correlating
more closely at 1 and 4 kHz, while FM tone thresholds correlated better at 0.5 kHz.
Both FM tone thresholds and mixed modulated tone thresholds showed a mean
difference of 22 dB at 2 kHz. The acoustic environment could possibly have affected
results obtained under sound-field stimulation. The comparison across stimulus
presentation, and controlled for sound-field speaker presentation between pure tone
and modulated tone thresholds did not reveal significant differences (p< .05) as
discussed in 5.2.1. Furthermore, the same participants were tested in each of the
conditions. The only variant during the comparison was the influence of the test room
acoustics. Participants were tested in the reverberant field during behavioural FM tone
testing and in the direct field during behavioural mixed modulated tone testing.
Reflectivity of the rooms’ surfaces were different under the different stimulus
presentations, as participants were seated on a chair during FM tone testing, while
they were lying on a bed on during mixed modulated tone testing.
138
University of Pretoria etd – Marais, J J (2004)
The comparison across transducers revealed that the minimum response levels
obtained using the different transducers closely approximated each other at 1 kHz (5
dB SPL mean difference) and 2 kHz (2 dB mean difference). The correlation at 0.5
kHz is somewhat elevated (9 dB mean difference) between the two transducers, while
the greatest mean difference between the minimum response levels obtained from the
two transducers was found at 4 kHz (22 dB). Statistical analysis of the results
obtained revealed significant differences for all the frequencies tested, except 2 kHz
on both a 5% and 1% level of significance.
The last experiment relating to the comparison across both stimulus presentation and
transducers revealed that both the thresholds obtained under insert earphone- and
sound-field conditions more closely approximated the minimum response levels
obtained under sound-field conditions, in the majority of cases. The closer correlation
is apparent between the behavioural pure tone and FM tone thresholds (obtained
under insert earphone and sound-field conditions) and the minimum response levels
(obtained under sound-field conditions) across 0.5-2 kHz. Both sets of pure tone data
more closely approximated the minimum response level obtained at 4 kHz, under
insert earphone presentation.
By implication the minimum response levels obtained under sound-field conditions
more closely approximated the actual behavioural thresholds in the majority of cases,
and with the exception of 4 kHz. Therefore behavioural data should be used as a
basis for comparison with minimum response level data on a transducer specific
basis. In other words unaided thresholds, as well as unaided minimum response
levels, obtained under a specific transducer should serve as the basis of
comparison with the aided minimum response levels. Put more simply, if a
clinician requires aided thresholds estimations using the monotic ASSEP technique,
he/she will have to ensure that the unaided thresholds are obtained under sound-field
conditions.
The findings of the current study contradict previous literature available (Picton et al.,
1998), which found no significant differences between minimum response levels
obtained under insert earphone- and sound-field conditions. Although the results of
the current study do not support the findings of Picton and colleagues (1998), it is
139
University of Pretoria etd – Marais, J J (2004)
worth mentioning that in the current study the same research participants were used
in repeated measures during the course of the study. This was not the case with the
study done by Picton and colleagues (1998). Although the minimum response levels
obtained in the current study, for both insert earphone- and sound-field speaker
presentation correlate well with the literature available, further research is
needed to validate the comparison of minimum response levels across different
transducers.
5.2.3 A Comparison between the actual Behavioural thresholds (for both pure
tone and modulated tone stimulation) obtained and the GSI AUDERA System’s
accuracy in predicting these thresholds by means of an Algorithm, for Insert
Earphone and Sound-field Conditions
This sub aim was structured to obtain results relating to the GSI AUDERA system’s
accuracy in predicting the actual behavioural thresholds. In order to address the sub
aim, two experiments were conducted. The first focused on the system’s accuracy in
predicting actual behavioural thresholds obtained under insert earphone conditions
and the second on predicting the actual thresholds obtained under sound-field
conditions.
For the comparison across stimulus presentation and transducers the predicted
thresholds obtained under sound-field conditions more closely approximated the
actual thresholds obtained under insert earphone conditions for all the frequencies
tested except 4 kHz. For the same comparison, with predicted thresholds obtained
under insert conditions the closer approximation was again with the actual thresholds
obtained under insert earphone conditions. Nevertheless the predicted thresholds
showed a strong correlation with the actual thresholds obtained under sound-field
conditions. None of the mean differences exceeded 5 dB regardless of which
transducer was used. Therefore the predicted thresholds (based on the algorithm
of the GSI AUDERA equipment) are closely related to the actual behavioural
thresholds. While the predicted thresholds more closely approximate the actual
thresholds obtained under insert earphone conditions, the correlation between
predicted thresholds and the actual thresholds obtained under sound-field
conditions are also strong. At present there limited literature available to compare
the GSI AUDERA system’s predicted thresholds with. This is because the GSI
140
University of Pretoria etd – Marais, J J (2004)
AUDERA system is one of very few ASSEP test protocols to include an algorithm
for the prediction of actual thresholds.
Finally, the results of the sub-aims, when viewed as a whole provides the following
answer to the main aim of the study. Behavioural thresholds obtained from different
stimulus presentations correlate strongly with one another. When the comparison was
drawn across transducers significant differences were obtained, indicating that
comparison across transducers is not viable. The thresholds obtained for the different
transducers show a strong correlation, with the MRLs obtained under sound-field
speaker presentation being slightly better in the majority of cases (0.5-2 kHz). Both
the minimum response levels obtained under insert earphone and sound-field
conditions provide a reasonable estimation of thresholds when compared to the
current body of knowledge. The estimation of actual thresholds through the predictive
function (algorithm) of the GSI AUDERA equipment revealed excellent correlations,
with the actual thresholds. The mean differences between predicted- and actual
thresholds ranged between 0-10dB HL. The findings of the current study relating to
the predictive function of the GSI AUDERA equipment will have to be validated by
subsequent research, as the GSI AUDERA system was the first ASSEP test protocol
to include an algorithm for the prediction of thresholds from minimum response levels
obtained.
5.3 Critical Evaluation of the Current Study
A critical evaluation of the empirical research component of any research endeavour
is essential in the determination of the value of the results obtained therein. Reliability
and validity of the results as well as the influence of identified limitations, inherent to
the study, is required to ensure appropriate interpretation thereof (Dane, 1990).
One factor that needs to be considered is the sampling method employed, in the
current study, during the selection of research participants. The research design,
namely a within group comparative experimental design, minimizes the effect of
variability relating to research participants, in that all the participants are tested in all
of the conditions of the experiment (Graziano, 2000). Furthermore the sample size can
be described as being sufficient for making inferences regarding the results obtained,
as a total of 50 ears (25 research participants) were analyzed. However, although the
gender distribution was approximately equal (13 female vs. 12 male), the
141
University of Pretoria etd – Marais, J J (2004)
disproportion could have affected the results obtained from the statistical tests during
analyses of gender effects. In order to determine the effect, if any, of gender on the
thresholds obtained, an equal distribution of male and female participants would
probably provide more reliable results.
Furthermore the age distribution of the current study represents an adult population
and inferences as well as conclusion from the results are therefore not applicable to,
for example a pediatric population. Previous research (Olsho, Koch, Carter, Halpin, &
Spetner, 1988; Schneider & Trehub, 1992) reported minimum response levels
obtained from infants and adult research participants’ that revealed mean differences
ranging between 10-30 dB. According to Werner, Folsom and Manel (1993) the
higher minimum response levels obtained from infants could possibly attributed to the
fact that neural response systems develop independently from sensory systems.
Another possibility is that the minimum response levels to low frequency stimulation
could be attenuated because of a “jitter” in the transmission between cochlear
transducers and neural receptors, thereby activating a wider range of activation on the
basilar membrane of infants. Put simply, the cochlear- as well as neural development
of infants could possibly cause an elevation in minimum response levels obtained
when compared with those obtained for adults. Thus it is clear that the thresholds
obtained in the current study would not suffice as normative data for pediatric testing.
Due to the age distribution of the current study, inferences and conclusions drawn
from the results obtained will only apply to populations showing similar age
distributions. The age range represented in the current study varied, primarily between
15-25, with reasonable distribution between 26-45 years of age.
Another factor, the influence of the test environment has to be considered. Two
factors could possibly have influenced the results obtained. Firstly the ambient noise
levels in the test room complied with the standards set out by the South African Buro
of Standards, but exceeded the acceptable noise levels used by other researchers (e.g.
Picton et al., 1998) and prescribed by ANSI standards (1996). According to PerezAbalo and colleagues (2001) as well as Picton and colleagues (1998) the ambient
noise levels within the test room could affect behavioural thresholds without affecting
the minimum response levels obtained through ASSEP testing. Although the
behavioural thresholds for both pure tone- and modulated tone stimulus presentation
142
University of Pretoria etd – Marais, J J (2004)
generally correlated well with each other across transducers, a discrepancy is noted at
0.5 kHz for the comparison of modulated tone presentation across transducers. It is
furthermore possible that the behavioural thresholds were slightly elevated by the
ambient noise levels. As this elevation would have occurred for both insert earphone
and sound-field conditions it is unlikely that a discrepancy would be noted between
thresholds obtained under different transducers. However, seeing as the ambient noise
levels do not affect the minimum response levels obtained it is possible that the results
obtained revealed distorted estimations of the actual thresholds.
In other words, if the ambient noise levels were within the prescribed international
range (ANSI, 1996), it is possible that the behavioural thresholds for both pure tone
and modulated tone stimulation would have been slightly lower. This would imply
that the mean differences obtained between behavioural data and electro-physiological
data would be greater than those reported in the current study.
Also, the influence of the acoustic environment could possibly have distorted the
results obtained in the current study. According to Beynon and Munro (1995) the
directional components, during sound-field testing, are more pronounced for higher
frequencies. This factor could possibly have attributed to the elevation of the
threshold estimations obtained under sound-field conditions at 4 kHz. Minimum
response levels obtained at 4 kHz revealed a mean value of 51 dB SPL under soundfield conditions compared to the 29 dB SPL obtained under insert earphone
conditions. Although the fact that directional components are more pronounced for
high frequencies, under sound-field testing, this inherent factor cannot be seen as a
limitation of the current research endeavour on its own, although it might have
attenuated the elevated threshold estimation (at 4 kHz) in conjunction with other
factors.
These include placement of the research participants and maintenance of the
placement in relation to the speaker (Walker et al., 1984). The test position was
clearly identifiable and research participants were placed on a height-adjustable chair.
Participants were asked to sit as still as possible during the test procedure, after the
chair had been adjusted so that the head of the participant was in line with the centre
of the speakers’ cone (at an azimuth of zero degrees). During the test procedures the
researcher as well as another qualified audiologist monitored the maintenance of the
143
University of Pretoria etd – Marais, J J (2004)
test position. It is debatable whether small head movements were overseen and what
the influence of these movements would have been on the results obtained. The
critical question is how viable is the technique (under sound-field conditions) if these
movements have such a profound impact on the results obtained. And more so, if the
technique is aimed at assessing hearing sensitivity in the difficult-to-test population
who is generally less cooperative than other populations.
Further influences of the test environment on the MRLs obtained under sound-field
conditions include the influence of reflective surfaces within the test room (Walker,
1979), directivity of the speaker, shape of the room, as well as test arrangements. Of
these influences there are two more influences that deserve mention. During the data
collection of the current study, and specifically during the recording of ASSEP
minimum response levels under sound-field conditions, research participants were
placed in close proximity to a large reflective surface, namely the wall of the test
room. Although the test room is sound treated, some degree of reflectivity cannot be
helped.
According to Walker (1979) reflective surfaces can influence responses obtained
under sound-field stimulation, and the extent of the influence increases as the size of
the surface increases. The study (Walker, 1979) does not however mention the exact
extent to which the responses are influenced. Thus it is clear that the reflectivity of the
rooms’ surfaces could have impacted on the minimum responses levels obtained,
although the degree to which the MRLs were influenced, if at all, is not known.
Lastly, participants were tested in the reverberant field during behavioural testing and
in the direct field during electro-physiological testing, due to calibration of the
equipment prior to the conduction of the current study. This is a major limitation of
the current study as reverberant field-testing is influenced by direct, as well as the
reflected characteristics of the stimulus (Walker et al., 1984).
Yet another factor that has to be considered is the recording procedure utilized
during minimum response level detection. Although both behavioural data and
electro-physiological data were obtained using 5dB intensity increments, the
minimum response levels for each research participant were taken as the lowest
recorded response. When the signal to noise ratio rendered reliable response detection
144
University of Pretoria etd – Marais, J J (2004)
impossible the stimulus was presented again at the same intensity, however due to
the prolonged test time involved in the current study, if a response was still not
recorded due to the influence of noise, it was considered as a no-response. It is
therefore possible that lower than recorded minimum response levels could have been
obtained if repeated measures were made for intensity levels where the signal to noise
ratio was not favourable. Thus, further averaging could have shown minimum
response levels at intensities closer to the actual behavioural thresholds. Within the
clinical arena this would mean that if ASSEP is applied under sound-field conditions
to obtain aided thresholds a closer approximation would relate into a better
performance evaluation of the specific aid.
The critical evaluation of the current study, as well as the implications thereof has
revealed several research implications that are discussed in the following section of
this chapter.
5.4 Recommendations for Future Research
“Whoever, in the pursuit of science, seeks after immediate practical utility, may
generally rest assured that he will seek in vain…” (von Helmholz, 1863). Results from
the current study showed that minimum response levels provided accurate threshold
estimations, recorded through sound-field stimulation. A previous paper by Picton and
colleagues (1998) also reported promising results. However, like Picton and
colleagues (1998) the researcher admits that there is still much to be done, and
“admits to cautious optimism” (Picton et al., 1998:329).
The first recommendation is to compare the accuracy of the monotic
ASSEP technique in the estimation of aided and un-aided hearing
thresholds in hearing-impaired individuals. This type of study will validate
the estimation of aided thresholds. While the current study aimed to determine
the influence of transducers on ASSEP minimum response level detection,
further investigation into the estimation of aided thresholds will serve to validate
the findings obtained by Picton and colleagues (1998). Although their (Picton et
al., 1998) findings indicated that smaller mean differences (13-17 dB) were
obtained for aided testing conditions when compared to those for normal
145
University of Pretoria etd – Marais, J J (2004)
subjects (26-36dB), these findings have to be corroborated by further research
in order to be validated.
More information regarding the effect of stimulus presentation on aided
threshold estimation is required. Results from the study done by Picton and
colleagues (1998), furthermore indicated that the threshold estimations obtained
from single stimulus presented alone was much more accurate than when the
stimulus was presented as part of a combined stimulus. A possible explanation
would be that when multiple stimuli are presented simultaneously the low
frequency stimuli might interfere with the response to high-frequency
stimulation. Although this would only occur at high intensity levels for normalhearing individuals, it occurs close to threshold in individuals with hearing loss
(Picton et al., 1998).
Picton and colleagues (1998) postulated that minimum response levels at high
frequencies were elevated when presented as part of a combined stimulus. This
might reflect on the “effective” hearing sensitivity of the individual being tested.
In other words, although a stimulus in the absence of competing stimuli might
be audible, the frequency information might not be available in the stimuli of
everyday environmental sounds. Therefore, research aimed at determining the
ability of the monotic ASSEP technique to estimate aided thresholds would
provide additional information to the limited body of knowledge currently
available.
The second recommendation would be to compare the test-retest-reliability
of ASSEP minimum response levels obtained under sound-field conditions
for normal-hearing as well as hearing-impaired individuals. This type of
study would validate the ASSEP technique as a valid and reliable tool in the
estimation of thresholds under sound-field conditions.
In order to determine whether any technique provides valid and reliable results,
it has to be able to produce similar results for the same individual in test-retest
measures. If thresholds estimations obtained from the same individual differs
significantly from one test to the next, the technique does not provide valid or
reliable results. Within the clinical arena over-estimation of the amount of gain
146
University of Pretoria etd – Marais, J J (2004)
provided by the hearing aid, might lead to an individual receiving insufficient
amplification.
Also, if the technique reveals a significant range of variance during test-retest
measures for normal-hearing individuals, differentiation between normalhearing and a mild hearing loss could be unreliable. Previous research has
established that thresholds obtained through pure tone stimulation under soundfield conditions correlated well for test-retest reliability. The need now arises to
determine whether factors other than the acoustic environment could possibly
influence the minimum response levels obtained from the ASSEP technique.
The third recommendation would be to determine whether a correlation
between ASSEP minimum response level amplitudes and loudness exists. If
such a correlation exists it could provide information regarding optimal hearing
aid settings. Previously auditory brainstem responses were used in an attempt to
assess the optimal hearing aid settings. The rationale was to adjust the hearing
aid settings until wave V of the auditory brainstem response (obtained with click
stimulation) fell within normal limits (Beauchaine et al., 1988; Kileny, 1982;
Mahoney, 1985). The application of this technique was limited, as click evoked
auditory brainstem responses correlate mainly to high frequency gain (Serpanos,
1997). The technique also revealed a low correlation between wave V latency
and loudness, which was aggravated for sloping hearing losses (Serpanos,
1997).
Current literature available (John et al., 1998) suggests that the ASSEP
technique would provide information mainly about the low range intensities, as
interactions occur between stimuli at high intensities. Research using a monotic
sequential ASSEP technique, under sound-field stimulation, is required to
confirm this hypothesis. It might therefore be possible to provide objective
assessment of comfort levels, by presenting a stimulus close to threshold level
and adjusting the hearing aid to provide optimal amplitudes without substantial
harmonics (indicating distortion).
Another advantage of the ASSEP technique is that minimum response levels
obtained is sufficiently frequency specific, so that masking is not required, as is
147
University of Pretoria etd – Marais, J J (2004)
the case with the auditory brainstem response. This does not however, exclude
the possibility that the minimum response levels obtained will not be placespecific, for low frequency stimulation and even more so than for high
frequency stimulation (Picton et al., 1998).
Lastly, ASSEP can be presented at sufficiently high intensity levels to detect
hearing losses that would not have been detected by auditory brainstem response
testing, due to the severity of the loss (Rance et al., 1998). Therefore not only
does the technique provide a method for accurate diagnosis, but it also puts in
place a means of assessing the initial habilitation of individuals with identified
losses (Picton et al., 1998).
The fourth recommendation would be to assess supra threshold hearing in
the selection and fitting of amplification devices in order to ensure
optimally discernable sound input. Another possible contribution of the
ASSEP technique, under sound-field conditions would be to assess supra
threshold hearing. Gain measurements of hearing aids can be assessed with in
the ear measures. These measures only provides information regarding sound
pressure levels at the eardrum, while ASSEP minimum response levels recorded
for aided conditions would incorporate the perception of the aided sound after
being processed in the brain (Picton et al., 2002a).
In order to determine if the amplification provided is sufficient for comfortable
speech discrimination, older children and adults are evaluated with word
recognition tests. Supra threshold assessments through the ASSEP technique,
under sound-field conditions, might provide valuable information in the
pediatric- and difficult-to-test populations who would not be able to comply
with word recognition testing. Preliminary results (Picton, 2002b) allowed for
optimism, however these results will have to be validate through further
research.
The fifth recommendation would be to establish an optimal range relating
to test time and the amount of averages required to obtain reliable and
accurate threshold estimations. Picton, Dimitrijevic, John and Van Roon
(2001) reported that test times in excess of 20 minutes are required to
148
University of Pretoria etd – Marais, J J (2004)
comprehensively evaluate representative thresholds. This type of study could
delineate what the test time and amount of averaging required would be and
whether this range would be clinically applicable. Thereby defining stimulus-,
acquisition- and recording parameters that are, both optimal for accurate
threshold estimation, as well as time-efficient, and subsequently clinically
applicable.
5.5 Final Comments
This study has investigated the influence of transducers on the monotic Auditory
Steady State Evoked Potential. The use of these potentials to measure the audiogram
of individuals identified as hearing-impaired, has been justified by previous research
(Picton et al., 2002a). However, even concerning the diagnostic application of
ASSEPs there is a need for more research regarding the accuracy of the various
techniques, to improve stimulation techniques as well as recording parameters (Picton
et al., 2002a). Auditory Steady State Evoked Potentials are currently entering the
realm of hearing screening, but yet again there is a need for ongoing investigation into
its validity, as well as its reliability.
Research regarding the management of the hearing-impaired individual through
Auditory Steady State Evoked Potentials is in its “infancy”. Future research on these
and other techniques are required to validate the use of evoked potentials in the
identification, evaluation and management of infants, and others within the difficultto-test population. As Picton and colleagues stated: “If we do not do so, we shall
never know that our present techniques are indeed the best. We must not lock
ourselves in 20-year old technologies…” without the means to collect data to improve
our management of hearing loss (Picton et al., 2002a:33).
5.6 Summary
This chapter drew conclusions to each of the sub-aims formulated in chapter three.
The conclusions were discussed to address the main aim of the study. Following the
conclusions a critical evaluation of the current study followed. Recommendations for
future research were presented based on the results of the current study as well as
literatureavailable.
149
University of Pretoria etd – Marais, J J (2004)
Reference List
Abramovich, S.J. (1990). Electric Response Audiometry in Clinical Practice. New
York: Churchill Livingstone Inc.
American National Standards Institute. (1996). American National Standard
Specification for Audiometers (ANSI S3.6-1996). New York: National Standards
Institute.
American National Standards Institute. (1999). Maximum permissible ambient
noise levels for audiometric test rooms. (ANSI S3.1 – 1991). New York: Nationale
Standards institute.
American National Standards Institute. (1986). Criteria for permissible ambient
noise during audiometric testing. (ANSI S3.6-1977) (R1986, N.Y.)
Anderson, C.D. (1979). New ideas in sound field systems. Hearing Instruments, vol.
30: 12-13,43.
Aoyagi, M., Kiren, T., Furuse, H., Fuse, T., Suzuki, Y., Yokota, S. & Koike, S.
(1994). Pure-tone threshold prediction by 80-Hz amplitude modulation following
response. Acta Otolaryngology Suppl. (Stockholm), vol. 511:7-14.
Aoyagi, M., Suzuki, Y., Yokota, M., Furuse, H., Watanabe, T. & Ito, T. (1999).
Reliability of 80-Hz amplitude-modulation following-response detected by phase
coherence. Audiology & Neuro-Otology, vol. 4:2-37.
Aoyagi, M., Yamazaki, Y., Yokota, M., Fuse, T., Suzuki, Y., Itoh, S. & Watanabe, T.
(1996). Frequency specificity of 80Hz amplitude modulation following response.
Acta Otolaryngologica, Supplement, 552:6-10.
150
University of Pretoria etd – Marais, J J (2004)
Aoyagi, M., Yoshinori, K., Suzuki, Y., Fuse, T. & Koike, Y. (1993). Optimal
modulation frequency for amplitude-modulation following response in young
children during sleep. Hearing Research, vol. 65:253-261.
Arlinger, S.D. & Jerlvall, L.B. (1987). Reliability in warble-tone sound field
audiometry. Scandinavian Audiology, vol.16: 21-27.
Azzena, G.B., Conti, G., Santarelli, R., Ottaviani, F., Paludetti, G. & Maurizi, M.
(1995). Generation of human auditory steady-state responses (SSRs). I: Stimulus
rate effects. Hearing Research, vol. 83:1-8.
Bacon, S.P., & Viemeister, N.F. (1985). Temporal modulation transfer functions in
normal hearing and hearing impaired listeners. Audiology, vol. 24:117-134.
Basar, E., Rosen, B., Basar-Eroglu, C. & Greitschus, F. (1987). The associations
between 40 Hz-EEG and the middle latency response of the auditory evoked
potential. International Journal of Neurophysiology. Vol. 62:465-467.
Beauchaine, K.A. & Gorga, M.P. (1988). Applications of the Auditory Brainstem
Response to pediatric hearing aid selection. Seminars in Hearing, vol. 9:61-74.
Berger, K.W. (1976). Prescription of hearing aids: a rationale. Journal of the
American Auditory Society, vol. 2: 71-78.
Berlin, C. & Hood, I. (1993). Pseudo central hearing loss: a confusion of central
for peripheral hearing loss caused by faulty conditioning techniques and lax
criteria. Seminars in Hearing, vol. 14:215-223.
Beynon, G.J. & Munro, K.J. (1995). Measurement of the variability in sound field
audiometry due to subject movement. British Journal of Audiology, vol. 29: 285291.
Boettcher, F.A., Madhotra, D., Poth, E.A. & Mills, J.H. (2002a). The frequencymodulation following response in young and aged human subjects. Hearing
Research, vol. 165: 10-18.
151
University of Pretoria etd – Marais, J J (2004)
Boettcher, F.A., Poth, E.A., Mills, J.H. & Dubno, J.R. (2002b). The amplitudemodulation following response in young and aged human subjects. Hearing
research, vol. 153: 32-42.
Bragg, V.C. (1977). Toward a more objective hearing aid fitting procedure.
Hearing Instruments, vol. 28: 6.
Brown, D. & Shallop, J. (1982). A clinically useful 500 Hz evoked response. Nicolet
Potentials, vol. 1: 9-12.
Carhart, R. (1946). Selection of hearing aids. Archives of Otolaryngology, vol. 44:
1-18.
Cebulla, M., Stürzebecher, E. & Wernecke, K. (2001). Objective detection of the
amplitude modulation following response. Audiology, vol. 245-252.
Chatrian G.E., Petersen, M.C. & Lazarte, J.A. (1960). Responses to clicks from the
human brain: some depth electrographic observations. Electroencephalography
and Clinical Neurophysiology, vol. 110: 62-70.
Clark, J.J. (2002). The Brain. Available: http://www.uoregon.edu/~jjclark/Brain.html
Cohen, L.T., Rickards, F.W. & Clark, G.M. (1991). A comparison of steady-state
evoked potentials to modulated tones in awake and sleeping humans. Journal of
the Acoustical Society of America, vol. 90: 2467-2479.
Cone Wesson, B., Parker, J., Rickards, F., Rance, G., Liburti-Persi, M., Poulis, C.,
May, E., Tan, L. & Pollard, J. (2001). Auditory steady-state evoked potentials as a
diagnostic tool for early identification and intervention programs. Oral
Presentation, 17th Biennial Symposium. International Evoked Response Audiometry
Study Group (IERASG), Vancouver.
Cox, L.C. & Metz, D.A. (1980). ABER in the prescription of hearing aids. Hearing
Instruments, vol. 31: 12-15,55.
152
University of Pretoria etd – Marais, J J (2004)
Dauman, R.W., Szyfter, R., De Sauvage, R.C. & Cazals, Y. (1984). Low frequency
thresholds assessed with 40 Hz MLR in adults with impaired hearing. Archives of
Oto-Rhino-Laryngology, vol. 240: 85-89.
Dane, F.C. (1990). Research Methods. Brooks Cole Company, San Diego.
Dempsey, J.J., Censoprano, E. & Mazor, M. (1986). Relationship between head size
and latency of the auditory brainstem response. Audiology, vol.25: 258-262.
Diefendorf, A.O. & Gravel, J.S. (1996). Visual reinforcement and behavioral
observation audiometry. In Handbook of Pediatric Audiology, Gerber, S.E. (ed.).
Gallaudet University Press, Washington, Dc. pp. 55-83.
Dillon, H. (1982). The use of control microphone in reverberant soundfield
audiometric testing. Paper presented at the 104th meeting, Acoustical Society of
America, Orlando, F.L. Journal of the Acoustical Society of America, vol. 72 Suppl
(1): S108.
Dillon, H. & Murray, N. (1987). Accuracy of twelve methods for estimating the
real ear gain of hearing aids. Ear & Hearing, vol. 8: 2-11.
Dillon, H. & Walker, G. (1980). The perception of normal hearing persons of
intensity fluctuations narrow band stimuli and its implication for sound field
calibration procedures. Australian Journal of Audiology, vol., 2: 72-82.
Dillon, H. & Walker, G. (1981). The effect of the acoustic environment on the
reliability of sound field audiometry. Australian Journal of Audiology, vol. 3:67-72.
Dillon, H. & Walker, G. (1982a). Comparison of stimuli used in sound field
audiometric testing. Journal of the Acoustic Society of America, vol. 7:161-172.
Dillon, H. & Walker, G. (1982b). The selection of modulation waveform for
frequency modulated sound field stimuli. Australian Journal of Audiology, vol.
4:56-61.
153
University of Pretoria etd – Marais, J J (2004)
Don, M., Ponton, C.W., Eggermont, J.J. & Masuda, A. (1993). Gender differences in
cochlear response time: an explanation for gender amplitude differences in the
un-masked auditory brain-stem response. Journal of the Acoustical Society of
America, vol. 94: 2135-2148.
Dobie, R.A. (1993). Objective Response Detection. Ear & Hearing, vol. 14:31-35.
Dobie, R.A., & Wilson, M.J. (1993). Phase weighting: a method to improve
objective response detection of steady state evoked potentials. Hearing Research,
vol. 79:94-98.
Dobie, R.A., & Wilson, M.J. (1996). A comparison of t test, F test, and coherence
methods of detecting steady-state auditory-evoked potentials, distortion product
otoacoustic emissions, or other sinusoids. Journal of the Acoustical Society of
America, vol. 100:2236-2246.
Durieux-Smith, A., Picton, T.W., Bernard, P., MacMurray, B., & Goodman, J.T.
(1991). Prognostic validity of brainstem electric response audiometry in infants
of a neonatal intensive care unit. Audiology, vol. 30:249-265.
Eggermont, J.J. (2001). The Neurophysiology of Perception: Single Units versus
Evoked Potentials. 17th Biennial Symposium, International Evoked Response
Audiometry Study Group (IERASG). Oral presentation. University of British
Columbia, Vancouver, B.C., Canada.
Fletcher, H. (1952). The perception of speech sounds by a deafened person.
Journal of the Acoustic Society of America, vol. 24:490-497.
Fulton, T.R. & Lloyd, L.L. (1969). Audiometry for the retarded. The Williams &
Wilkins Co., Baltimore, U.S.A.
154
University of Pretoria etd – Marais, J J (2004)
Galambos, R. & Makeig, S. (1988). Dynamic changes in steady-state responses. In
Dynamics of sensory and cognitive processing by the brain. Basar, E. (Ed.), Springer,
Berlin, pp103-122.
Galambos, R., Makeig, S. & Talmachhoff, P.J. (1981). A 40 Hz auditory potential
recorded from the human scalp. Proceedings of the National Academy of Sciences
of the United States of America, vol. 78:2643-2647.
Galambos, R., Wilson, M.J. & Silva, D.P. (1994). Identifying hearing loss in the
intensive care nursery: a twenty year summary. Journal of the American Academy
of Audiology, vol. 5:151-162.
Graziano, A.M. & Raulin, M.L. (1993). Research Methods: A Process of Inquiry.
2nd ed., HarperCollins College Publishers, New York, USA.
Graziano, A.M. & Raulin, M.L. (2000). Research Methods: A Process of Inquiry.
4th ed., Allan & Bacon, Massachusetts, USA.
Ginsberg, I.A. & White, T.P. (1994). Otologic Disorders and examination:
Evaluation of Cochlear and Retrocochlear Functions. In Handbook of Clinical
Audiology, J.Katz (ed.) 4th ed, Williams & Wilkins, Baltimore.
Goldstein, R. & Aldrich, W.M. (1999). Evoked Potential Audiometry:
Fundamentals and Applications. Allyn and Bacon, Boston.
Goldstein, J.L., Baer, T. & Kiang, N.Y.S. (1971). A theoretical treatment of
latency, group delay, and tuning characteristics for auditory nerve responses to
clicks and tones. In Physiology of the Auditory System; Based on the Proceedings of
a Workshop, Sachs, M.B. (ed.). National Educational Consultants, Baltimore: 133141.
Gorga, M.P., Beauchaine, K.A. & Reiland, J.K. (1987). Comparison of onset and
steady state responses of hearing aids: implications for the use of the auditory
brainstem response in the selection of hearing aids. Journal of Speech and Hearing
Research, vol. 30;130-136.
155
University of Pretoria etd – Marais, J J (2004)
Gorga, M.P., Worthington , D.W., Reiland, J.K., Beauchaine, K.A. & Goldgar, D.E.
(1985). Some comparisons between auditory brainstem response thresholds,
latencies and the pure-tone audiogram. Ear & Hearing, vol. 6:105-112.
Hall, J.W. & Mueller, H.G. (1997). Audiologists’ Desk Reference, Volume 1:
Diagnostic Principles, Procedures, and Applications. Singular Publishing Group
Inc. San Diego, California.
Hall, J.W. & Ruth, R.A. (1985). Acoustic reflexes and auditory evoked responses
in hearing aid selection. Seminars in hearing, vol.6: 251-277.
Harford, E.R. (1979). Hearing aid amplification for adults. Monographs in
contemporary audiology, vol. 1:1-37.
Harford, E.R. (1980). The use of miniature microphone in the ear canal for the
verification of hearing aid performance. Ear & Hearing, vol. 1:329-337.
Hari, R., Hamalainen, M., Joutsiniemi, S.L. (1989). Neuromagnetic steady-state
responses to auditory stimuli. Journal of Acoustic Society of America, vol. 86:10331039
Haughton, P.M. (1980). Basic principles of auditory assessment. College-Hill Press,
San Diego.
Hawkins, D. (1987). Clinical ear canal probe tube measurements. Ear & Hearing,
vol. 8 (Supplement 5), 74s-81s.
Hawkins, D. & Haskell, G. (1982). A comparison of functional gain and 2 cubic
centimeter coupler gain. Journal of Speech and Hearing Disorders, vol. 47:71-76.
Hecox, K. & Galambos, R. (1974). Brainstem auditory evoked responses in human
infants and adults. Archives of Otolaryngology, vol. 99:30-33.
Hecox, K.E. (1983). Role of auditory brainstem responses in the selection of
hearing aids. Ear & Hearing, vol. 4:51-55.
156
University of Pretoria etd – Marais, J J (2004)
Herdman, A.T. & Stapells, D.R. (2001). Thresholds determined using the monotic
and dichotic auditory steady-state response technique in normal hearing
subjects. Scandinavian Audiology, vol. 30: 41-49.
Hood, L.J. (1998). Clinical applications of the auditory brainstem response.
Singular Publishing Group, San Diego.
Hyde, M.L., Riko, K. & Malizia (1990). Audiometric accuracy of the click ABR in
infants at risk for hearing loss. Journal of the American Academy of Audiology,
vol. 1:59-66.
ISO/DIS 8253. (1985). Acoustics. Basic Pure-tone Audiometric Test Methods.
Draft International Standard. International Organization for Standardization, Geneva.
ISO. (1989). Audiometric Test Methods. Part 2. Sound Field Audiometry with
Pure Tone and Narrow-Band Signals. DP 8253/3. Geneva: International
Electrotechnical Commission.
Jerger, J., Chmiel, R., Frost, J.D. & Coker, N. (1986). Effect of sleep on the auditory
steady state evoked potential. Ear & Hearing, vol. 7:240-245.
Jerger., J.F., Grimes, A.M., Jacobson, G.P. Albright, K.A. & Moncrieff, D. (2000).
The future of diagnostic Audiology. In Audiology Diagnosis, Eds Roeser, R.J.,
Valente, M & Hosford-Dunn, H. Thieme Medical Publishers, New York: 615-626.
John, M.S., Dimitrijevic, A. & Picton T.W. (2001). Weighted averaging of steadystate responses. Clinical Neurophysiology, vol. 112:555-562.
John, M.S., Lins, O.G., Boucher, B.L. & Picton, T.W. (1998). Multiple auditory
Steady State Responses (MASTER): stimulus and recording parameters.
Audiology, vol.37: 59-82.
John, M.S. & Picton, T.W. (2000). Human auditory steady-state responses to
amplitude-modulated tones: phase and latency measurements*. Hearing
Research, vol. 141:57-79.
157
University of Pretoria etd – Marais, J J (2004)
Joint Committee on Infant Hearing. (1994). 1994 Position Statement. ASHA vol.
36:38-42.
Joris, P.X. & Yin, T.C. (1992). Responses to amplitude-modulated tones in the
auditory nerve of a cat. Journal of the Acoustical Society of America, vol. 91: 215232.
Kankkunen, A. & Rosenhall, U. (1985). Comparison between thresholds obtained
with pure tone audiometry and the 40-Hz middle latency response. Scandinavian
Audiology, vol. 14:99-104.
Katz, J. (1985). Handbook of Clinical Audiology. 3rd Ed, Katz, J. (ed.) Williams &
Wilkins, Baltimore, United States of America.
Katz, J. (1994). Handbook of Clinical Audiology. 4th Ed, Katz, J. (ed.) Williams &
Wilkins, London.
Kidd, G. (2002). Pshycoacoustics. In Handbook of Clinical Audiology. 5th Ed. Katz,
J. (ed.) Lippincott Williams & Wilkins. Baltimore, United States of America.
Kileny, P. (1982). Auditory brainstem responses as indicators of hearing aid
performance. Annals of Otology, vol. 91:61-64.
Knudsen, V.O., & Jones, I.H. (1935). Artificial aids to hearing. Laryngoscope, vol.
45: 48-69.
Kuehl, R.O. (2000). Design of Experiments: Statistical principles of Research
Design. 2nd ed., Brooks/Cole Publishing Company, California, USA.
Kuwada, S., Batra, R. & Maher, V.L. (1986). Scalp potentials of normal and
hearing impaired subjects in response to sinusoidally amplitude-modulated
tones. Hearing Research, vol. 21:179-192.
158
University of Pretoria etd – Marais, J J (2004)
Leedy, P.D. (1997). Practical Research: Planning and Design. 6th edn., Prentice
Hall, New Jersey.
Levi, E.C., Folsom, R.C. & Dobie, R.A. (1993). Amplitude-modulation following
response (AMFR): effects of modulation rate, carrier frequency, age and state.
Hearing Research, vol. 68:42-52.
Libby, E.R. (1985). State-of-the-art of hearing aid selection procedures. Hearing
Instruments, vol. 36: 30,34,36,38,62.
Lin, J.M., Lesner, S.A., Sandridge, S.A. Daddario, C.C. (1984). Threshold predicon
from the auditory 40-Hz evoked potential. Ear & Hearing, vol., 5: 366-370.
Linden, R.D., Campbell, K.B., Hamel, G. & Picton, T.W. (1985). Human auditory
steady state evoked potentials during sleep. Ear & Hearing, vol. 6: 167-174.
Lindén, G. & Kankkunen, A. (1969). Visual reinforcement audiometry. Archives
Otolaryngology, vol. 89: 865-872.
Lins, O.G. & Picton, T.W. (1995). Auditory steady-state responses to multiples
simultaneous stimuli. Electro-encephalograph Clinical Neurophysiology, vol.
96:420-432.
Lins, O.G., Picton, T.W., Boucher, B.L., Durieux-Smith, A., Champagne, S.C.,
Moran, L.M., Perez-Abalo, M.C., Martin, V. & Savio, G. (1996). Frequency Specific
Audiometry Using Steady-State Responses. Ear & Hearing, vol. 17:81-96.
Lins, O.G., Picton, P.E., Picton, T.W., Champagne, S.C. & Durieux-Smith, A. (1995).
Auditory steady-state responses to tones amplitude-modulated at 80-110 Hz.
Journal of the Acoustic Society of America, vol. 97:3051-3063.
Lybarger, S.F. (1944). U. S. Patent Application SN 543: 278.
159
University of Pretoria etd – Marais, J J (2004)
Lynn, J.M., Lesner, S.A., Sandridge, S.A. & Daddario, C.C. (1984). Threshold
prediction from the auditory 40-Hz evoked potential. Ear & Hearing, vol. 5: 366370.
McCandless, G.A. & Lyregaard, P.E. (1983). Prescription of gain/output (POGO)
for hearing aids. Hearing Instruments, vol. 34: 16.
Magnusson, L., Börjesson, E. & Axelsson, A. (1997). Visual reinforcement
audiology. Scandinavian Audiology, vol. 26:247-251.
Mahoney, T.M. (1985). Auditory brainstem response hearing aid application. In
The auditory Brainstem Response. Jacobson, J.T., (ed.) San Diego: College Hill Press,
349-370.
Makela, J.P. & Hari, R. (1987). Evidence for cortical origin of the 40 Hz auditory
evoked response in man. Electroencephalography and Clinical neurophysiology, vol.
66: 539-546.
Martin, F.N. (1994). Introduction to Audiology. 5th ed., Prentice-Hall Inc, New
Jersey.
Martin, F.N. (1997). Introduction to Audiology. 6th ed., Prentice-Hall Inc, New
Jersey.
Mauer, G. & Döring, W.H. (1999). Generators of amplitude modulation following
response (AMFR). Paper presented at the June 1999 meeting of the International
Evoked Response Audiometry Study Group, Tromso.
Maurizi, M., Almadori, G., Paludetti, G., Ottaviani, F., Rosignoli, M. & Luciano R.
(1990). 40-Hz steady-state response in newborns and in children. Audiology, vol.
29: 322-28.
Moodie. S., Seewald, R. & Sinclair, S. (1994). Selecting amplification
characteristics for young hearing-impaired children. American Journal of
Audiology, vol. 3:23-31.
160
University of Pretoria etd – Marais, J J (2004)
Moore, N. (2000). How to do research: the complete guide to designing and
managing research projects. 3rd edn. Library Association Publishing, London.
Mueller, H.G., Hawkins, D.B. & Northern, J.L. (1992). Probe Microphone
Measurements: Hearing aid selection and assessment. Singular Publishing Group,
San Diego, California.
Neuman, W.L. (1997). Social Research Methods: Qualitative and Quantitative
Approaches. 3rd edn. Allyn & Bacon, Boston.
National Institutes of Health Consensus Committee. (1993). Early identification of
hearing impairment in infants and young children. NIH Consensus Statement, vol.
11: 1-24.
Northern, J.L. (1992). Introduction to computerized probe-microphone real-ear
measurements in hearing aid evaluation procedures. In Probe Microphone
Measurements: Hearing Aid Selection and Assessment, Mueller, G.H., Hawkins,
D.B. & Northern, J.L. (eds.). Singular Publishing Group, San Diego, California: 1-20.
Olsho, L.W., Koch, E.G., Carter, E.A., Halpin, C.F. & Spetner, N.B. (1988). Pure
tone sensitivity of human infants. Journal of the Acoustical Society of America, vol.
84: 1316-1324.
Orchik, D. & Rintelman, W. (1979). Comparison of pure-tone, warble-tone and
narrow band noise thresholds of young normal hearing children. Journal of the
American Audiological Society, vol. 3: 214-220.
Osterhammel, P.A., Shallop, J.K. & Terkildsen, K. (1985). The effect of sleep on the
auditory brainstem response (ABR) and the middle latency response (MLR).
Scandinavian Audiology, vol. 14: 47-50.
Oxford Dictionary (1990). The Oxford Paperback Dictionary. 3rd Edn. Oxford
University Press, Walton Street, New York.
161
University of Pretoria etd – Marais, J J (2004)
Pantev, C., Roberts, L.E., Elbert, T., Rob, B. & Wienbruch, C. (1996). Tonotopic
organization of the sources of human auditory steady-state responses. Hearing
Research, vol. 10: 62-74.
Perez-Abalo, M.C., Savio, G., Torres, A., Martin, V., Rodriguez, E. & Galan, L.
(2001). Steady state responses to multiple amplitude modulated tones: an
optimized method to test frequency specific thresholds in hearing impaired
children and normal subjects. Ear & Hearing, vol. 22(3): 200-211.
Picton, T.W. (1991). Clinical Usefulness of Auditory Evoked Potentials: A Critical
Evaluation. Journal of Speech Language Pathology and Audiology, vol. 15: 3-26.
Picton, T.W., Dimitrijevic, A, John M.S. & Van Roon, P., (2001). The use of phase
in the detection of auditory steady-state responses. Clinical Neurophysiology, vol.
112: 1692-1711.
Picton, T.W., Dimitrijevic, A. & John, M.S. (2002c). Multiple auditory steady-state
responses. Annals of Otolgy, Rhinology and Larangology, vol. 111: 16-21.
Picton, T.W., Dimitrijevic, A. & Van Roon, P., et al. (2002b). Possible roles for the
auditory steady-state responses in fitting hearing aids. In A sound foundation
through early amplification. Seewald, R.C. & Gravel, J.S. (eds.) Basel:Phonak,
2002, pp. 63-73.
Picton, T.W., Durieux-Smith, A., Champagne, S., Whittingham, J., Moran, L.,
Giguére, C. & Beauregard, Y. (1998). Objective evaluation of aided thresholds
using auditory steady state responses. Journal of the American Academy of
Audiology, vol. 9: 315-331.
Picton, T.W., Durieux-Smith, A. & Moran, L. (1994). Recording auditory
brainstem responses from infants. International Journal of Paediatric
Otolaryngology, vol. 28: 93-110.
162
University of Pretoria etd – Marais, J J (2004)
Picton, T.W., John, M.S. & Dimitrijevic, A. (2002a). Possible roles for the auditory
steady-state response in identification, evaluation and management of hearing
loss in infancy. Audiology Today, October 2002:29-34.
Plourde, G. & Picton, T.W. (1990). Human auditory steady-state responses during
general anesthesia. Aneasthiology Analogs, vol. 71: 460-468.
Plourde, G., Stapells, D.R. & Picton T.W. (1991). The human auditory steady-state
evoked potentials. Acta Otolaryngologica, Supplement 491: 153-160.
Pratt, H, Mittelman, N., Bleich, N. & Zaaroor, M. (2002). Band-pass specific
contributors of multiple generators to the auditory 40-Hz steady state potentials.
Ear & Hearing, vol. 23; 41-48.
Punch, J.L. (2000). Contemporary issues in hearing aid selection and fitting.
Available at http://www.kingpax.com/~jlpunch/contemp/issues/hear/aid/selec/html
Rance, G., Rickards, F.W., Beer, D.E. & Clark, G.M. (1998). Steady state evoked
potential and behavioral hearing thresholds in a group of children with absent
click evoked auditory brainstem response. Ear & Hearing, vol. 19: 48-61.
Rance, G., Rickards, F.W., Cohen, L.T., De Vidi, S. & Clark G.M. (1995). The
automated prediction of hearing thresholds in sleeping subjects using auditory
steady-state evoked potentials. Ear & Hearing, vol. 16:499-507.
Rees, A. & Moller, A.R. (1983). Responses of the neurons in the inferior colliculus
of the rat to AM and FM tones. Hearing Research, vol. 10:301-330.
Regan, D. (1966). Some characteristics of average steady-state and transient
responses evoked by modulated light. Electroencephalography and Clinical
Neurophysiology, vol. 20: 238-248.
Regan, D. (1989). Human brain electrophysiology: Evoked potentials and evoked
magnetic fields in science and medicine. Elsevier Science Publishing Co., Inc.,
New York.
163
University of Pretoria etd – Marais, J J (2004)
Regan, M.P. & Regan D. (1993). Nonlinear terms produced by passing amplitude
modulated sinusoids through a hair cell transduction function. Biol. Cybern., vol.
69:439-446.
Renshaw, J.J. & Diefendorf, A.O. (1998). Adapting the test battery for the child
with special needs. In Children with Hearing Loss: Contemporary Trends. Bess,
F.H. (ed.). Bill Wilkerson Center Press, Nashville: 83-103.
Rickards, F.W. (2001). Oral presentation. The 17th Biennial Symposium,
International Evoked Response Audiometry Study Group (IERASG), Vancouver,
from the Department of Learning and Educational Assessment, University of
Melbourne, Australia.
Rickards, F.W. & Clark, G.M. (1984). Steady-state evoked potentials to amplitudemodulated tones. In Evoked Potentials II. Nodar, R.H. & Barber, C. (eds.)
Butterworth, Boston: 163-168.
Rickards, F.W., Tan, L.E., Cohen, L.T., Wilson, O.J., Drew, J.H., & Clark, G.M.
(1994). Auditory steady state evoked potentials in newborns. British Journal of
Audiology, vol. 28: 327-337.
Rodriguez, R., Picton T.W., Linden, D., Hamel, G. & Laframboise, G. (1986).
Human auditory steady state responses: Effects of intensity and frequency. Ear &
Hearing, vol. 7:300-313.
Roeser, R.J., Buckley, K.A. & Stickney, G.S. (2000a). Pure Tone Tests. IN
Audiology Diagnosis. R.J. Roeser, M. Valente & H. Hosford-Dunn (eds.). Thieme
Medical Publishers, New York, pp227-252.
Roeser, R.J., Valente, M. & Hosford-Dunn, H. (2000b). Diagnostic Procedures in
the Profession of Audiology. In Audiology Diagnosis. R.J. Roeser, M. Valente & H.
Hosford-Dunn (Eds.). Thieme Medical Publishers, New York, pp.1-18.
164
University of Pretoria etd – Marais, J J (2004)
Romanov, F.F. (1942). Methods for measuring the performance of hearing aids.
Journal of the Acoustical Society of America, vol. 13: 294-304.
SABS 00182 (1998). South African Buro of Standards: Maximum permissible
Ambient Noise Levels, SABS 0182 – 1998.
Sammeth, C.A. Barry, S.J. (1985). The 40-Hz event-related potential as a measure
of auditory sensitivity in normals. Scandinavian Audiology, vol. 14: 51-55.
Schmulian D.L. (2002). The prediction of hearing thresholds with multiple
frequency steady state evoked potential compared to an auditory brainstem
response protocol. Unpublished doctoral dissertation, University of Pretoria.
Schneider, B.A. & Trehub, S.E. (1992). Sources of developmental change in
auditory sensitivity. In Developmental Pschycoacoustics. Werners, L.A. & Rubel,
E.W. (eds.). Washington DC: American Pschycological Association.
Schimmel, H., Rapin, I. & Cohen, M.M. (1975). Improving evoked response
audiometry: results of normative studies for machine scoring. Audiology, vol. 14:
466-479.
Seewald, R.C., Ross, M. & Spiro, M.K. (1985). Selecting amplification
characteristics for young hearing-impaired children. Ear & Hearing, vol. 6:48-53.
Serpanos, Y.C., O’Malley, H. & Gravel, J.S. (1997). The relationship between
loudness intensity functions and the click-ABR wave V latency. Ear & Hearing,
vol. 18: 409-419.
Shallop, J.K. & Osterhammel, P.A. (1983). A comparative study of measurements
of SN-10 and the 40/sec middle latency responses in newborns. Scandinavian
Audiology, vol. 12:91-95.
Silman, S. & Silverman, C.A. (1991). Auditory Diagnosis: Principles and
Applications. Academic Press Inc., San Diego, California
165
University of Pretoria etd – Marais, J J (2004)
Smurzynski, J., Jung, M.D., Lafreniere, D., Kim, D.O., Vasudeva-Kamath, M.,
Rowe, J.C., Holman, M.C. & Leonard, G. (1993). Distortion-product and clickevoked otoacoustic emissions of preterm and full-term infants. Ear & Hearing,
vol14: 258-274.
Spydell, J.D., Pattee, G. & Goldie, W.D. (1985). The 40 hertz auditory eventrelated potential: Normal values and effects of lesions. Electroencephalography
and Clinical Neurophysiology, vol. 62: 193-202.
Stach, B.A. (1998). Clinical Audiology as an Introduction. Singular Publishing
Group, San Diego, California.
Stapells, D.R., Galambos, R., Costello, J.A. & Makeig, S. (1988). Inconsistency of
auditory middle latency and steady-state responses in infants.
Electroencephalography Clinical Neurophysiology, vol. 7:289-295.
Stapells, D.R., Gravel, J.S. & Martin, B.A. (1995). Thresholds for auditory
brainstem responses to tones in notched noise from infant and young children
with normal hearing and sensorineural hearing loss. Ear & Hearing, vol. 16:361371.
Stapells, D.R., Linden, D., Suffield, J.B., Hamel, G. & Picton, T.W. (1984). Human
auditory steady state potentials. Ear & Hearing, vol. 5:105-113.
Stapells, D.R., Makeig, G. & Galambos, R. (1987). Auditory steady-state response
threshold prediction using phase coherence. Electroencephalography and Clinical
Neurophysiology, vol. 67: 260-270.
Stapells, D.R., Picton, T.W., Durieux-Smith, A., Edwards, C.G. & Moran, L.M.
(1990). Thresholds for short-latency auditory-evoked potentials to tones in
notched noise in normal-hearing and hearing impaired subjects. Audiology,
vol.29: 262-274.
166
University of Pretoria etd – Marais, J J (2004)
Starr, A. & Don, M. (1988). Brain Potentials Evoked by Acoustic Stimuli. In
Human event-related potentials. Handbook of electroencephalography and
clinical neurophysiology, Picton, T.W. (ed.). Elsevier, Amsterdam; 97-157.
Stelmachowicz, P.G. & Lewis, D.E. (1988). Some theoretical considerations
concerning the relation between functional gain and insertion gain. Journal of
Speech and Hearing Research, vol. 39: 491-496.
Stürzebecher, E., Cebulla, M. & Pschirrer, U. (2001). Efficient stimuli for recording
of the amplitude modulation following response. Audiology, vol. 40:63-68.
Stürzebecher, E., Kuhne, W. & Berndt, H. (1985). Delectability of acoustically
evoked composite response (40 Hz potential) near threshold. Scandinavian
Audiology, vol. 14: 214-219.
Suzuki, T. & Kobayashi, K. (1984). An evaluation of 40-Hz event-related
potentials in young children. Audiology, vol. 23:599-604.
Suzuki, T., Kobayashi, K. & Yuri, U. (1994). Effect of natural sleep on auditory
steady state responses in adult subjects with normal hearing. Audiology, vol.
33:274-279.
Swanepoel D.C.D. (2001). Estimating pure tone thresholds with the dichotic
multiple frequency auditory steady state response compared to an ABR protocol
in normal hearing subjects. Unpublished Masters thesis, University of Pretoria.
Szyfter, W., Dauman, R. & De Sauvage, R.C. (1984). 40 Hz middle latency
responses to low frequency tone pips in normal hearing adults. Journal of
Otolaryngology, vol. 13: 275-280.
Valdes, J.L., Perez-Abalo, M.C., Martin, V., Savio, G., Sierra, C., Rodriguez, E. &
Lins, O. (1997). Comparison of statistical indicators for the automatic detection
of 80 Hz auditory steady state responses. Ear & Hearing, vol. 18: 420-429.
167
University of Pretoria etd – Marais, J J (2004)
Victoreen, J. (1973). Basic principles of otometry. Springfield, Illinois: Charles C
Thomas.
Victoreen, J. (1974). Equal loudness pressure determined with a decaying
oscillatory waveform. Journal of the Acoustical Society of America, vol. 55: 309312.
Von Helmholtz, H.L.F. (1863). Die Lehre von dem Tonempfindungen als
psichologische grundlage für die theorie der musiek. Brunswick: Vieweg-Verlag.
Walker, G. (1979). The pure tone in sound field testing: experimental results and
suggested procedures. Australian Journal of Audiology, vol. 1: 49-60.
Walker, G. & Dillon, H. (1983). The selection of modulation rates for frequency
modulated sound field stimuli. Scandinavian Audiology, vol. 12:151-156.
Walker, G., Dillon, H. & Byrne, D. (1984). Sound field audiometry: recommended
stimuli and procedures. Ear & Hearing, vol. 5:13-21.
Watson, D.R. (1996). The effects of cochlear hearing loss, age and sex on the
auditory brainstem response. Audiology, vol. 35: 246-258.
Watson, N. & Knudson, V. (1940). Selective amplification in hearing aids. Journal
of the Acoustical Society of America, vol. 11:406-419.
Wei, W.W.S. (1990). Estimation of the spectrum. In Time Series Analysis:
Univariate and Multivariate Methods, Redwood City, California, Addison-Wesley:
256-287.
Werner, L.A., Folsom, R.C. & Mancl, L.R. (1993). The relationship between
auditory brainstem response and behavioral thresholds in normal hearing
infants and adults. Hearing Research, vol. 68: 131-141.
168
University of Pretoria etd – Marais, J J (2004)
White, K.R. & Behrens, T.R. (1993). The Rhode Island hearing assessment
project. Implications for universal newborn hearing screening. Seminars in
Hearing, vol. 14: 1-119.
Wilber, L.A. (2002). Transducers for audiologic testing. In Handbook of Clinical
Audiology. 5th edn, Katz, J. (ed.). Lippincott Williams & Wilkins, Baltimore
(U.S.A.), pp. 88-95.
Wisniewsky, W.H. (1970). Measurement of the electroacoustical characteristics of
hearing aids. Journal of Auditory Technology, vol. 9: 152-170.
Wolf, K.E. & Goldstein, R. (1980). Middle component AER’s from neonates to
low-level tonal stimuli. Journal of Speech and Hearing Research, vol. 23: 185-201.
Yantis, P.A. (1994a). Pure Tone Air Conduction Threshold Testing. In Handbook
of Clinical Audiology. 4th edn, Katz, J. (ed.). Williams & Wilkins, London, pp. 97108.
Yantis, P.A. (1994b). Puretone air-conduction threshold testing. In Handbook of
Clinical Audiology. 4th edn, Katz, J. (ed.). Williams & Wilkins, London, pp. 375-386.
Yoshinaga-Itano, C. (2001). Universal newborn hearing screening (UNHS).
Phonak International Pediatric Conference 2001. Oral presentation, Phonak
International Pediatric Conference, Pretoria.
Zurek, P.M. (1992). Delectability of transient an sinusoidal otoacoustic emissions.
Ear & Hearing, vol. 13: 307-310.
169
University of Pretoria etd – Marais, J J (2004)
APPENDIX A
170
University of Pretoria etd – Marais, J J (2004)
APPENDIX B
June 2002
Dear Sir / Madam
Thank you, so much for showing interest in this research project being
conducted at the Hearing Clinic, Department of Communication
Pathology at the University of Pretoria. The Audiology Department is
very exited about the arrival of the GSI AUDERA Auditory Steady State
Evoked Potential equipment, a state-of-the-art objective hearing
evaluation procedure. We are currently undertaking clinical trails of the
equipment and we need your assistance. We need to establish norms for
normal-hearing ears using both insert earphones and sound-field speaker
presentation with this technique, and we kindly ask for your participation
in this study.
Participation in this study will involve the following:
You will undergo a standard hearing evaluation (pure tone behavioural
audiometry), where you are required to respond to sound stimulation.
You will need to repeat this test using insert earphones and sound-field
speaker presentation. This procedure takes approximately 15 min. The
same (behavioural) procedure will then be followed using the auditory
steady state evoked potential equipment. This procedure also takes about
15 min.
An Auditory Steady State Evoked Potential (ASSEP) test, using insert
earphones, will follow. No response is required during this procedure.
You will simply be asked to lie down on a bed, with three electrodes
attached to your head. This test takes approximately 90 min.
171
University of Pretoria etd – Marais, J J (2004)
The last procedure, again, involves the ASSEP test, this time using a
sound-field speaker. Again, no response is required during this procedure.
You will lie down on a bed, with three electrodes attached to your head.
This time the sound stimulation will come from the speaker and not from
the insert earphones as in the second procedure. Again this test will take
approximately 90 min.
All the procedures (tests) are non-invasive and only the behavioural
procedures require responses from you. The test battery will be divided
into two sessions of no longer than two hours each. All acquired
information will be treated as confidential and no names will be used. A
copy of your results will be made available to you, should you request it.
Thank you for your assistance.
Should you require any further information, you are welcome to contact
us.
Yours sincerely
Prof. René Hugo
Head of Department
Mr De Wet Swanepoel
Research Supervisor
Mr Cobus Marais
Researcher
172
University of Pretoria etd – Marais, J J (2004)
APPENDIX B CONTINUED
University of Pretoria
Department Communication Pathology: Audiology
Name:
Occupation:
Age:
Contact numbers:
Date of birth:
.
.
(mm/dd/yy/)
Please complete the following reply slip:
I
hereby agree to participate in this
project. I am aware that I can withdraw from this project, at any time,
should I want to.
Signature
Date
173
University of Pretoria etd – Marais, J J (2004)
APPENDIX B CONTINUED
June 2002
Dear Parents
Thank you, so much for showing interest in this research project being
conducted at the Hearing Clinic, Department of Communication
Pathology at the University of Pretoria. The Audiology Department is
very exited about the arrival of the GSI AUDERA Auditory Steady State
Evoked Potential equipment, a state-of-the-art objective hearing
evaluation procedure. We are currently undertaking clinical trails of the
equipment and we need your child’s assistance. We need to establish
norms for normal-hearing ears using both insert earphones and soundfield speaker presentation with this technique, and we kindly ask for your
child’s participation in this study.
Participation in this study will involve the following:
Your child will undergo a standard hearing evaluation (pure tone
behavioural audiometry), where he/she will be required to respond to
sound stimulation. He/she will need to repeat this test using insert
earphones and sound-field speaker presentation. This procedure takes
approximately 15 min. The same (behavioural) procedure will then be
followed using the auditory steady state evoked potential equipment. This
procedure also takes about 15 min.
An Auditory Steady State Evoked Potential (ASSEP) test, using insert
earphones, will follow. No response is required during this procedure.
Your child will simply be asked to lie down on a bed, with three
electrodes attached to his/her head. This test takes approximately 90 min.
The last procedure, again, involves the ASSEP test, this time using a
sound-field speaker. Again, no response is required during this procedure.
174
University of Pretoria etd – Marais, J J (2004)
Your child will lie down on a bed, with three electrodes attached to
his/her head. This time the sound stimulation will come from the speaker
and not from the insert earphones as in the second procedure. Again this
test will take approximately 90 min.
All the procedures (tests) are non-invasive and only the behavioural
procedures require responses from your child. The test battery will be
divided into two sessions of no longer than two hours each. All acquired
information will be treated as confidential and no names will be used. A
copy of your child’s results will be made available to you, should you
request it.
Thank you for your assistance.
Should you require any further information, you are welcome to contact
us.
Yours sincerely
Prof. René Hugo
Head of Department
Mr De Wet Swanepoel
Research Supervisor
Mr Cobus Marais
Researcher
175
University of Pretoria etd – Marais, J J (2004)
APPENDIX B CONTINUED
University of Pretoria
Department Communication Pathology: Audiology
Name:
Occupation:
Age:
Contact numbers:
Date of birth:
.
.
(mm/dd/yy/)
Please complete the following reply slip:
I
parent/guardian of
hereby give permission that he/she
may participate in this project. I am aware that my child can withdraw
from this project, at any time, should he/she or I want to.
Signature
Date
176
University of Pretoria etd – Marais, J J (2004)
APPENDIX C
Type A Tympanogram: Normative Data
Tympanometric parameters Physical parameters References
Ear canal volume
1.0-1.4 cc
Silman & Silverman, 1991
Compliance
0.50-1.75 ml
Silman & Silverman, 1991
Pressure (daPa)
-50 - +50 daPa
Silman & Silverman, 1991
Categorization of Degree of Hearing Loss
Categorization of Degree of Hearing Loss
0-25 dB
Normal
Goodman, 1965
26-40 dB
Mild
Goodman, 1965
41-55 dB
Moderate
Goodman, 1965
56-70 dB
Moderately-Severe
Goodman, 1965
71-90 dB
Severe
Goodman, 1965
> 91 dB
Profound
Goodman, 1965
177
University of Pretoria etd – Marais, J J (2004)
APPENDIX D
(X-No significant difference √-Significant difference)
Non Parametric Statistics
Comparison
Frequencies
Tested
Sign Test:
Z-Stat
P<0.05
Insert
Behavioural Pure Tone vs.
Earphone
Behavioural Mixed
Presentation Modulated Tone Thresholds
0.5 kHz
1 kHz
2 kHz
4 kHz
0.12
1.05
0.4
1.55
X
X
X
X
Sound-field Behavioural FM Tone vs.
Presentation Behavioural Mixed
Modulated Tone Thresholds
0.5 kHz
1 kHz
2 kHz
4 kHz
1.59
0.89
0.04
2.21
X
X
X
X
Insert
Behavioural Pure Tone
Earphone
Thresholds vs. ASSEP
Presentation Minimum Response Levels
0.5 kHz
1 kHz
2 kHz
4 kHz
0.5 kHz
1 kHz
2 kHz
4 kHz
7.94
8.56
9.06
9.02
7.28
8.83
9.02
9.68
√
√
√
√
√
√
√
√
Insert
Behavioural Mixed
Earphone
Modulated Tone vs. ASSEP
Presentation Minimum Response levels
0.5 kHz
1 kHz
2 kHz
4 kHz
8.06
7.51
9.10
7.47
√
√
√
√
Sound-field Behavioural Mixed
Presentation Modulated Tone vs. ASSEP
Minimum Response levels
0.5 kHz
1 kHz
2 kHz
4 kHz
0.5 kHz
1 kHz
2 kHz
4 kHz
0.5 kHz
1 kHz
2 kHz
4 kHz
8.87
7.94
8.99
7.47
1.47
3.10
4.61
2.90
0.66
3.60
4.57
9.683.60
√
√
√
√
X
√
√
√
X
√
√
√
Sound-field Behavioural FM Tone
Presentation Thresholds vs. ASSEP
Minimum Response Levels
Insert
Behavioural Pure Tone
Earphone
Thresholds vs. Predicted
Presentation Thresholds
(ASSEP Algorithm)
Sound-field Behavioural FM Tone
Presentation Thresholds vs. Predicted
Thresholds
(ASSEP Algorithm)
178
University of Pretoria etd – Marais, J J (2004)
APPENDIX D CONTINUED
(X-No significant difference √-Significant difference)
Non Parametric Statistics
Sound-field
Presentation
Insert Earphone
Presentation
Sound-field
Presentation
ASSEP Minimum Response Levels vs.
Predicted Thresholds (ASSEP Algorithm)
Insert Earphone
Presentation
Behavioural Mixed Modulated Tone vs.
Predicted Thresholds
(ASSEP Algorithm)
Comparison
Frequencies
Tested
Sign Test:
Z-Stat
P<0.05
0.5 kHz
1.59
X
1 kHz
2.05
X
2 kHz
4.65
√
4 kHz
1.36
X
0.5 kHz
2.25
X
1 kHz
2.71
√
2 kHz
4.53
√
4 kHz
1.39
X
0.5 kHz
6.47
√
1 kHz
5.46
√
2 kHz
4.45
√
4 kHz
6.12
√
0.5 kHz
6.62
√
1 kHz
5.23
√
2 kHz
4.45
√
4 kHz
6.08
√
179
University of Pretoria etd – Marais, J J (2004)
APPENDIX E
(X-No significant difference √-Significant difference)
Analysis of Gender Effects: Mann-Whitney
Behavioural Mixed
Behavioural Pure Tone
Modulated Tone
Thresholds obtained
Thresholds under Insert under Insert Earphones
Earphones
Frequency
Tested
Mean Value
for Male
Participants
8.9583
Mean Value
for Female
Participants
5.1923
P-Value
P<0.05
P<0.01
0.2092
X
X
1 kHz
2.0833
0.7692
0.1770
X
X
2 kHz
2.9167
-0.3846
0.0340
√
X
4 kHz
3.1250
-1.1538
X
X
0.5 kHz
7.0833
6.9231
0.9921
X
X
1 kHz
2.9167
3.0769
0.9513
X
X
2 kHz
2.2917
0.7692
0.2089
X
X
4 kHz
5.2083
1.7308
0.1642
X
X
0.5 kHz
0.0915
Analysis of Ear Effects: Mann-Whitney
Behavioural Mixed
Modulated Thresholds
obtained under Insert
Earphones
Behavioural Pure Tone
Thresholds obtained
under Insert Earphones
Frequency
Tested
Mean Value
for Left Ears
0.5 kHz
6.2
Mean Value
for Right
Ears
7.8
P-Value
P<0.05
P<0.01
0.5369
X
X
1 kHz
1.6
1.2
0.7714
X
X
2 kHz
2.2
0.2
0.3548
X
X
4 kHz
2.4
-0.6
0.2375
X
X
0.5 kHz
7.6
6.4
0.6767
X
X
1 kHz
3.2
2.8
0.8788
X
X
2 kHz
2.4
0.6
0.1913
X
X
4 kHz
4.2
2.6
0.6700
X
X
180
University of Pretoria etd – Marais, J J (2004)
APPENDIX F
(X-No significant difference √-Significant difference)
Analysis of Gender Effects: Mann-Whitney
Behavioural Mixed
Behavioural FM Tone
Modulated Tone
Thresholds obtained
Thresholds obtained under under Sound-field
Sound-field Conditions Conditions
Frequency
Tested
Mean Value
for Male
Participants
3.5417
Mean Value
for Female
Participants
3.8461
P-Value
P<0.05
P<0.01
0.5968
X
X
1 kHz
-0.8353
-0.1154
0.2620
X
X
2 kHz
-1.8750
-0.5769
0.2450
X
X
4 kHz
3.1250
2.3077
0.9587
X
X
0.5 kHz
-0.2083
2.5000
0.1117
X
X
1 kHz
-0.8333
0.3846
0.6804
X
X
2 kHz
-1.6667
-0.1923
0.1827
X
X
4 kHz
7.2917
6.5385
0.9840
X
X
0.5 kHz
Analysis of Ear Effects: Mann-Whitney
Behavioural Mixed
Modulated Tone
Thresholds obtained
under Sound-field
Conditions
Behavioural FM Tone
Thresholds obtained
under Sound-field
Conditions
Frequency
Tested
Mean Value
for Left
Ears
2
Mean Value
for Right
Ears
5.4
P-Value
0.0463
√
X
1 kHz
-2
-1
0.3984
X
X
2 kHz
-1.2
-1.2
0.9250
X
X
4 kHz
3
2.4
0.9176
X
X
0.5 kHz
1.8
0.6
0.3212
X
X
1 kHz
-0.6
0.2
0.6360
X
X
2 kHz
-0.4
-1.4
0.5123
X
X
4 kHz
7.4
6.4
0.9920
X
X
0.5 kHz
P<0.05 P<0.01
181
University of Pretoria etd – Marais, J J (2004)
APPENDIX G
(X-No significant difference √-Significant difference)
Wilcoxon: Paired T-Test
Comparison
Pure Tone (inserts): FM Tone (sound-field) at 0.5kHz
Sign
Test
0.001
Pure Tone (inserts): FM Tone (sound-field) at 1kHz
P<0.05 P<0.01
√
√
0.0002
√
√
Pure Tone (inserts): FM Tone (sound-field) at 2kHz
0.003
√
√
Pure Tone (inserts): FM Tone (sound-field) at 4kHz
0.03
√
X
Wilcoxon: Paired T-Test
Comparison
Mixed Modulated Tone (inserts): Mixed Modulated Tone
(sound-field) at 0.5kHz
Mixed Modulated Tone (inserts): Mixed Modulated Tone
(sound-field) at 1kHz
Mixed Modulated Tone (inserts): Mixed Modulated Tone
(sound-field) at 2kHz
Mixed Modulated Tone (inserts): Mixed Modulated Tone
(sound-field) at 4kHz
Sign
Test
0.00001
P<0.05
P<0.01
√
√
0.0002
√
√
0.01
√
X
0.0014
√
√
Wilcoxon: Paired T-Test
Comparison
ASSEP MRLs (inserts): ASSEP MLRs (sound-field) at
0.5kHz
ASSEP MRLs (inserts): ASSEP MRLs (sound-field) at
1kHz
ASSEP MRLs (inserts): ASSEP MRLs (sound-field) at
2kHz
ASSEP MRLs (inserts): ASSEP MRLs (sound-field) at
4kHz
Sign
Test
0.00001
P<0.05
P<0.01
√
√
0.0002
√
√
0.0744
X
√
0.0005
√
√
182
University of Pretoria etd – Marais, J J (2004)
APPENDIX H
(X-No significant difference √-Significant difference)
Analysis of Gender Effects: Mann-Whitney
ASSEP MRLs obtained
under Insert Earphone
conditions
Behavioural Mixed
Modulated Thresholds
obtained under Insert
Earphone Conditions
Behavioural Pure Tone
Thresholds obtained
under Insert Earphone
Conditions
Frequency
Tested
Mean Value
for Male
Participants
8.9583
Mean Value
for Female
Participants
5.1923
P-Value
P<0.05
P<0.01
0.2092
X
X
1 kHz
2.0833
0.7692
0.1770
X
X
2 kHz
2.9167
-0.3846
0.0340
√
X
4 kHz
3.1250
-1.1538
0.0915
X
X
0.5 kHz
7.0833
6.9231
0.9921
X
X
1 kHz
2.9167
3.0769
0.9513
X
X
2 kHz
2.2917
0.7692
0.2089
X
X
4 kHz
5.2083
1.7308
0.1642
X
X
0.5 kHz
37.0833
33.4614
0.1539
X
X
1 kHz
26.8749
23.0768
0.2406
X
X
2 kHz
28.1249
23.4615
0.0155
√
X
4 kHz
28.5416
25.9614
0.3472
X
X
0.5 kHz
183
University of Pretoria etd – Marais, J J (2004)
APPENDIX I
(X- No significant difference √-Significant difference)
Analysis of Ear Effects: Mann-Whitney
ASSEP MRLs obtained Behavioural Mixed
under Sound-field
Modulated Tone
conditions
Thresholds obtained
under Insert Earphone
Conditions
Behavioural Pure Tone
Thresholds obtained
under Insert Earphone
Conditions
Frequency
Tested
Mean
Value for
Left Ears
6.2
Mean
Value for
Right Ears
7.8
P-Value
P<0.05
P<0.01
0.5369
X
X
1 kHz
1.6
1.2
0.7714
X
X
2 kHz
2.2
0.2
0.3548
X
X
4 kHz
2.4
-0.6
0.2375
X
X
0.5 kHz
7.6
6.4
0.6767
X
X
1 kHz
3.2
2.8
0.8788
X
X
2 kHz
2.4
0.6
0.1913
X
X
4 kHz
4.2
2.6
0.6700
X
X
0.5 kHz
34.1999
36.1999
0.5294
X
X
1 kHz
24.1999
25.5999
0.5543
X
X
2 kHz
25.9999
25.3999
0.7287
X
X
4 kHz
26.7999
27.5999
0.6883
X
X
0.5 kHz
184
University of Pretoria etd – Marais, J J (2004)
APPENDIX J
(X- No significant difference √-Significant difference)
Analysis of Gender Effects: Mann-Whitney
ASSEP MRLs obtained Behavioural Mixed
under Sound-field
Modulated Thresholds
conditions
obtained under Soundfield Conditions
Behavioural FM Tone
Thresholds obtained
under Sound-field
Conditions
Frequency
Tested
Mean Value
for Male
Participants
3.5417
Mean Value
for Female
Participants
3.8461
P-Value
P<0.05
P<0.01
0.5968
X
X
1 kHz
-0.8353
-0.1154
0.2620
X
X
2 kHz
-1.8750
-0.5769
0.2450
X
X
4 kHz
3.1250
2.3077
0.9587
X
X
0.5 kHz
-0.2083
2.5000
0.1117
X
X
1 kHz
-0.8333
0.3846
0.6804
X
X
2 kHz
-1.6667
-0.1923
0.1827
X
X
4 kHz
7.2917
6.5385
0.9840
X
X
0.5 kHz
29.5832
22.8845
0.0325
√
X
1 kHz
16.4583
19.6153
0.3077
X
X
2 kHz
19.5833
22.3076
0.2460
X
X
4 kHz
36.4582
34.6153
0.6167
X
X
0.5 kHz
185
University of Pretoria etd – Marais, J J (2004)
APPENDIX K
(X-No significant difference √-Significant difference)
Analysis of Ear Effects: Mann-Whitney
ASSEP MRLs obtained Behavioural Mixed
under Sound-field
Modulated Tone
conditions
Thresholds obtained
under Sound-field
Conditions
Behavioural FM Tone
Thresholds obtained
under Sound-field
Conditions
Frequency
Tested
Mean
Value for
Left Ears
6.2
Mean
Value for
Right Ears
7.8
P-Value
P<0.05
P<0.01
0.5369
X
X
1 kHz
1.6
1.2
0.7714
X
X
2 kHz
2.2
0.2
0.3548
X
X
4 kHz
2.4
-0.6
0.2375
X
X
0.5 kHz
7.6
6.4
0.6767
X
X
1 kHz
3.2
2.8
0.8788
X
X
2 kHz
2.4
0.6
0.1913
X
X
4 kHz
4.2
2.6
0.6700
X
X
0.5 kHz
26.3999
25.7999
0.7023
X
X
1 kHz
17.5999
18.5999
0.7687
X
X
2 kHz
21.5999
20.3999
0.8906
X
X
4 kHz
33.9999
36.9999
0.2240
X
X
0.5 kHz
186
University of Pretoria etd – Marais, J J (2004)
APPENDIX L
Correction Factors Earphone Testing
Frequency
Tested
TDH 49 -Supra
Aural Headphones(Martin, 1994)
ER-3A – Insert
Earphones(Martin, 1994)
Sound Field
Presentation (SABS
0182-1998)
0.5 kHz
13.5
8.5
9.0
1 kHz
7.5
3.5
5.5
2 kHz
11.0
6.5
11.5
4 kHz
10.5
1.5
15.0
187
University of Pretoria etd – Marais, J J (2004)
APPENDIX M
PERMISSIBLE AMBIENT NOISE LEVELS
Frequencies
Test Room
Tested
(Actual Noise
Levels)
SABS 0182-1998
(Permissible
Sound Pressure
Levels for Insert
Earphone Testing
SABS 0182– 998
(Permissible
Sound Pressure
Levels for soundfield testing)
ANSI
Standards
(Prescribed
Acceptable
Noise Levels)
0.5 kHz
11.8 dB SPL
20.5 SPL
7.5 SPL
14.5 dB SPL
1 kHz
<10 dB SPL
24.0 SPL
6.5 SPL
12.5 dB SPL
2 kHz
10.8 dB SPL
31.0 SPL
6.0 SPL
8.5 dB SPL
4 kHz
12.7 dB SPL
37.0 SPL
2.0 SPL
9.0 dB SPL
188
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