Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Comparison of wideband energy reflectance and tympanometric measures obtained with Reflwin Interacoustics,… Shaw, Jeffrey 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2009_fall_shaw_jeffrey.pdf [ 3.54MB ]
Metadata
JSON: 24-1.0070857.json
JSON-LD: 24-1.0070857-ld.json
RDF/XML (Pretty): 24-1.0070857-rdf.xml
RDF/JSON: 24-1.0070857-rdf.json
Turtle: 24-1.0070857-turtle.txt
N-Triples: 24-1.0070857-rdf-ntriples.txt
Original Record: 24-1.0070857-source.json
Full Text
24-1.0070857-fulltext.txt
Citation
24-1.0070857.ris

Full Text

Comparison of Wideband Energy Reflectance and Tympanometric Measures Obtained With Reflwin Interacoustics, Mimosa Acoustics and GSI Tympstar Systems by Jeffrey Shaw Honours B.SC., The University of Toronto, 2007  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Audiology and Speech Sciences) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) September 2009  © Jeffrey Shaw 2009  i  Abstract In this study the effects of gender, instrument, and ethnicity on measures of wideband energy reflectance, wideband reflectance tympanometry and standard tympanometry were analyzed. Measures of energy reflectance (ER) and power absorption (PA) were made using the Mimosa Acoustics and Reflwin Interacoustics middle ear analyzer systems. Tympanograms were generated using the Reflwin Interacoustics system and GSI Tympstar. There were a total of sixty normal hearing participants (113 ears), with an equal number of Chinese and Caucasian males and females. There were five primary purposes to the following study: (1) To determine whether the Mimosa and the Reflwin systems yield similar measures of ER and PA; (2) To determine whether pressurization method (static versus dynamic) has an effect on ER and PA measurements obtained using the Reflwin system; (3) To determine whether the Mimosa and Reflwin systems are capable of detecting ER and PA differences that exist between Chinese and Caucasian young adults; (4) To determine whether Reflwin and GSI tympstar generate comparable tympanograms; and (5) To determine how effectively the norms from each instrument can be used to identify otosclerosis. Analysis of purpose (1) showed that the ER values were larger for the Mimosa system at frequencies below 630Hz. Analysis of purpose (2) showed that the dynamic pressurization technique resulted in lower PA at the low frequencies compared to the PA values obtained using the static pressurization technique. Analysis of purpose (3) showed that ER, PA, and tympanometric parameters obtained in this study were consistent with findings of previous studies in which ER, PA, Vea, Ytm and TW varied as a function of ethnicity. Analysis of purpose (4) showed that Ytm was larger and more variable for GSI Tympstar, TPP was more negative for the Reflwin system compared to the GSI Tympstar, ii  and Vea was larger for the Reflwin system for the probe tone frequency of 1000 Hz. Analysis of purpose (5) showed that both instruments are equally capable of detecting otosclerosis when static ER values at 500 Hz, 630 Hz, and 800 Hz are used. This suggests that instrument-specific norms are not warranted for the detection of otosclerosis.  iii  TABLE OF CONTENTS  Abstract ..........................................................................................................................................ii Table of Contents .......................................................................................................................... iv List of Tables ................................................................................................................................ viii List of Figures..................................................................................................................................x List of Abbreviations .................................................................................................................... xiii Acknowledgements ..................................................................................................................... xiv Dedication .................................................................................................................................... xv  CHAPTER 1 INTRODUCTION .......................................................................................................... 1 1.1 Basic principles........................................................................................................... 1 1.2 Wideband energy reflectance in clinical practice ...................................................... 4 1.3 Factors affecting middle ear parameters .................................................................. 9 1.3.1 Effects of ethnicity on WBER and tympanometric parameter ................. 9 1.3.2 Effects of gender on WBER and tympanometric parameters ................. 13 1.3.3 Effects of instrument (Mimosa versus Reflwin) on WBER ...................... 15 1.3.4 Effects of instrument (Reflwin versus GSI Tympstar) on tympanometric parameters .............................................................................................. 21 1.4 Clinical decision analysis for the detection of otosclerosis ..................................... 25 1.5 Significance of the following study .......................................................................... 29 1.5.1 Clinical significance.................................................................................. 29 1.5.2 Purposes .................................................................................................. 31  iv  CHAPTER 2 METHODS ................................................................................................................ 32 2.1 Participants .............................................................................................................. 32 2.1.1 Description of participants ...................................................................... 32 2.1.2 Inclusion and exclusion criteria ............................................................... 33 2.2 Instrumentation ...................................................................................................... 33 2.2.1 Extended high-frequency audiometry instrumentation ......................... 33 2.2.2 WBER instrumentation ............................................................................ 35 2.2.2.1 Mimosa (RMS v4.0.4.4)............................................................ 35 2.2.2.2 Reflwin (Eclipse v1) ................................................................. 36 2.2.3 Tympanometric instrumentation ............................................................ 38 2.3 Procedures .............................................................................................................. 39 2.3.1 Consent.................................................................................................... 39 2.3.2 Distortion-Production otoacoustic emissions ......................................... 39 2.3.3 Standard tympanometry ......................................................................... 40 2.3.4 Conventional and extended high-frequency audiometry ....................... 40 2.3.5 WBER and wideband reflectance tympanometry .................................. 41 2.4 Statistical analyses .................................................................................................. 43  CHAPTER 3 RESULTS ................................................................................................................... 46 3.1 Wideband energy reflectance and power absorption ............................................ 46 3.1.1 Wideband energy reflectance ................................................................. 46 3.1.2 Power absorption .................................................................................... 51 3.1.3 Static versus dynamic pressure power absorption measurements using the Reflwin system ................................................................................................ 53 3.2 Tympanometric parameters................................................................................................. 56 3.2.1 Ear canal volume ................................................................................... 59 v  3.2.1.1 226 Hz .............................................................................. 59 3.2.1.2 1000Hz ............................................................................. 62 3.2.2 Static admittance .................................................................................. 66 3.2.2.1 226 Hz .............................................................................. 66 3.2.2.2 1000Hz ............................................................................. 70 3.2.3 Tympanometric peak pressure ............................................................ 72 3.2.3.1 226 Hz .............................................................................. 72 3.2.4 Tympanometric width ......................................................................... 74 3.2.4.1 226 Hz .............................................................................. 74 3.3 Efficacy of Mimosa and Reflwin in identifying otosclerosis ................................................. 78 3.3.1 ANOVA and post-hoc analyses ............................................................ 78 3.3.1.1 Mimosa system ............................................................... 78 3.3.1.2 Reflwin system................................................................ 80 3.3.2 ROC analysis ........................................................................................ 81 3.3.3 Dot plot analysis ................................................................................. 84  CHAPTER 4 DISCUSSION ............................................................................................................. 89 4.1 Sources of differences for WBER measurements................................................................. 90 4.1.1 Comparison of Mimosa results to previous studies ................................................. 90 4.1.2 Comparsion of Reflwin results to previous studies .................................................. 95 4.1.3 Comparison between Mimosa and Reflwin in the current study ........................... 99 4.1.4 Ethinicity as a source of variation in ER.................................................................. 101 4.2 Clinical significance of WBER variability between instruments ......................................... 102 4.3 Tympanometric parameters............................................................................................... 104 4.3.1 Ear canal volume................................................................................................. 107 4.3.1.1 226 Hz ...................................................................................................... 107  vi  4.3.1.2 1000Hz ..................................................................................................... 108 4.3.2 Static admittance .............................................................................................. 110 4.3.2.1 226 Hz ..................................................................................................... 110 4.3.2.2 1000Hz .................................................................................................... 111 4.3.3 Tympanometric peak pressure........................................................................................ 111 4.3.4 Tympanometric width ..................................................................................................... 112  CHAPTER 5 CONCLUSIONS ....................................................................................................... 114 5.1 Summary............................................................................................................................. 114 5.2 Limitations of the current study and areas for future research......................................... 117 5.3 General conclusions ........................................................................................................... 118 References ................................................................................................................................. 119 Appendix I Statistical analysis tables ........................................................................................ 124 Appendix II Consent form for normal hearing subjects ........................................................... 142  vii  List of Tables Table. 1.1. Comparisons of Ytm, TW and TPP between Caucasian and Chinese young adults.. 13 Table. 3.1. Descriptive statistics, including the mean, standard deviation, and 90 percent range at each of the fifteen frequencies for measurements made at ambient pressure using the Mimosa system and the Reflwin Interacoustics system .............................................................................................. 47 Table 3.2. Descriptive statistics including the mean, median, standard deviation, 5th percentile, and 95th percentile for the parameters of Vea+, Vea-, Ytm+, Ytm-, TPP, and TW made using the Mimosa Acoustics and the Reflwin Interacoustics system for both ethnicities........................................................ 58 Table. 4.3. Summary of AUROC plots and 95% CI along with pair-wise comparison of AUROC between Mimosa and Reflwin for ER at 500Hz, ER at 630Hz, and ER at 800Hz......................................... 82 Table. 4.1. Descriptive statistics for Ytm, TW and TPP obtained using both the GSI and the Reflwin system. Some other published normative data are also included for comparison. C = Caucasian; A = Chinese; M = male; F = female .................................................................................................. 105 Table. A1. ANOVA table for ER with ethnicity, gender, and ear as within-subject factors and instrument and frequency as between-subject factors ............................................................................... 124 Table. A2. Greenhouse-Geiser Correction for the ANOVA of ER using the Mimosa versus Reflwin system ................................................................................................................................................... 125 Table. A3. Tukey’s HSD analysis for the ethnicity by frequency interaction from the ANOVA of ER using the Mimosa versus Reflwin system ........................................................................................... 126 Table. A4. Tukey’s HSD analysis for the gender, instrument, and frequency interaction from the ANOVA of ER using the Mimosa versus Reflwin system ........................................................................ 127 Table. A5. ANOVA table for PA with ethnicity, gender, and ear as within-subject factors and instrument and frequency as between-subject factors ............................................................................... 128 Table. A6. Greenhouse-Geiser Correction for the ANOVA of ER using the Mimosa versus Reflwin system ................................................................................................................................................... 129 Table. A7. Tukey’s HSD analysis for the ethnicity by frequency interaction from the ANOVA of PA using the Mimosa versus Reflwin system ........................................................................................... 130 Table. A8. Tukey’s HSD analysis for the instrument by frequency interaction from the ANOVA of PA using the Mimosa versus Reflwin system ................................................................................. 131 Table. A9. ANOVA table for PA with ethnicity, gender, and ear as within-subject factors and pressurization method and frequency as between-subject factors ......................................... 132 Table. A10. Greenhouse-Geiser Correction for the ANOVA of PA using the dynamic versus static pressurization technique for Reflwin ........................................................................................ 133 Table. A11. Tukey’s HSD analysis for the pressurization technique by frequency interaction from the ANOVA of PA using the dynamic versus static pressurization technique for Reflwin .............. 134 viii  Table. A12. ANOVA table for Vea at 226 Hz with ethnicity, gender, and ear as within-subject factors and instrument and tail as between-subject factors ....................................................................... 135 Table. A13. ANOVA table for Vea at 226 Hz with ethnicity, gender, and ear as within-subject factors and instrument and tail as between-subject factors ....................................................................... 135 Table. A14. ANOVA table for Ytm at 226 Hz with ethnicity, gender, and ear as within-subject factors and instrument and tail as between-subject factors ....................................................................... 136 Table. A15. ANOVA table for Ytm at 1000Hz with ethnicity, gender, and ear as within-subject factors and instrument and tail as between-subject factors ....................................................................... 136 Table. A16. ANOVA table for TPP at 226 Hz with ethnicity, gender, and ear as within-subject factors and instrument as a between-subject factor ................................................................................... 137 Table. A17. ANOVA table for TW at 226 Hz with ethnicity, gender, and ear as within-subject factors and instrument as a between-subject factor ................................................................................... 137 Table. A18. ANOVA table for the comparison between otosclerotic and normal ears assessed on the Mimosa system. Condition (normal versus otosclerosis) served as a between-subject factor and frequency served as a within-subject factor ............................................................................. 138 Table. A19. Greenhouse-Geiser Correction for the ANOVA of the comparison between otosclerotic and normal ears assessed on the Mimosa system........................................................................... 138 Table. A20. Tukey’s HSD analysis for the condition by frequency interaction from the ANOVA of the comparison between otosclerotic and normal ears assessed on the Mimosa system ............ 139 Table. A21. ANOVA table for the comparison between otosclerotic and normal ears assessed on the Reflwin system. Condition (normal versus otosclerosis) served as a between-subject factor and frequency served as a within-subject factor ............................................................................. 139 Table. A22. Greenhouse-Geiser Correction for the ANOVA of the comparison between otosclerotic and normal ears assessed on the Reflwin system............................................................................ 140 Table. A23. Tukey’s HSD analysis for the condition by frequency interaction from the ANOVA of the comparison between otosclerotic and normal ears assessed on the Reflwin system ............. 140 Table. A24. Comparison of ER data between the current study (N = 60; 18 – 38 yr.), Shahnaz & Bork (2006) (N = 128; 18 – 32 yr.), and Voss et al. (1994) (N = 10; 18 – 24). The current study and Shahnaz & Bork (2006) use the same version of the Mimosa system, while the study by Voss et al. (1994) uses an older version of the Mimosa system ......................................................................................... 141 Table. A25. Comparison of ER data between the current study (N = 60; 18 – 38 yr.), Sanford & Feeney (2008) (N = 20; 22 – 30 yr.), Feeney et al. (2004) (N = 40; 18 – 28 yr.), and Keefe et al. (1993) (N = 10; 20 – 50 yr.). The current study and Sanford & Feeney (2008) used the same version of the Reflwin system, while the studies by Feeney et al. (2004) and Keefe et al. (2004) used an older version of the Reflwin system ....................................................................................................................................... 141  ix  List of Figures Figure. 1.1. Normative data obtained for measures of energy reflectance (bottom panels) and power absorption (top panels). Yellow region represents the 90th percentile range. The most energy is absorbed at the mid-frequencies and the least energy is absorbed at the low (250Hz) and high (8000Hz) frequencies .................................................................................................................................... 6 Figure. 1.2. Comparison of WBER profiles in healthy and otosclerotic ears. Note the greater reflectance at low frequencies in otosclerotic ears. Taken from Shahnaz et al., 2009 .................................. 7 Figure. 1.3. Illustration of the statistical significance of the interaction of the variables of frequency and ethnicity in previous studies. At the lower pitches the Chinese group has greater reflectance and at the higher pitches the Caucasian group has lower reflectance. Taken from Shahnaz, & Bork (2006) ..................................................................................................................................................... 11 Figure. 1.4. Left: ADC waveforms recorded for the long tube 1 (top) and the short tube 2 (bottom). The incident ADC waveform (bottom) is defined by separating the incident signal in the tube 1 waveform. Right: ADC waveform from tube 1 is redrawn by clipping the amplitude. Taken from Keefe, & Simmons (2002) .......................................................................................................................................... 19 Figure. 1.5. The SPL spectra are plotted for the long tube (tube 1) as a solid gray curve, the short tube (tube 2) as a dashed black curve, and the incident signal as a solid black curve. Taken from Keefe &, Simmons (2002)........................................................................................................................... 20 Figure. 1.6. Three dimensional plot of pressure (P), frequency (f), and energy absorbance as measured during wideband energy reflectance tympanometry. Figure taken from Reflwin Interacoustics from an ear with otosclerosis ................................................................................................................... 23 Figure. 2.1. Illustration of the Mimosa Acoustics Instrumentation. The probe, four-wall cavity, PIC, DSP card, and computer screen are depicted .................................................................................... 35 Figure. 2.2. Illustration of the Reflwin Interacoustics WBER system. Note the AT235h audiometer used for changing the pressure in the ear canal and the probe tip (bottom right), which are connected to a PC ..................................................................................................................................................... 37 Figure. 3.1. Ethnicity by frequency interaction: ER at each of the fifteen frequencies for Caucasian and Chinese Young Adults. Data from the Reflwin Interacoustics and Mimosa Middle Ear Analyzer systems are Pooled together .................................................................................................................... 49 Figure. 3.2. Gender by frequency by instrument interaction: ER at each of the fifteen frequencies for females and males. Data from Caucasian and Chinese ethnicities are pooled.......................... 50 Figure. 3.3. Frequency by instrument interaction: PA at each of the fifteen frequencies for the Mimosa and Reflwin systems. Data for Caucasian and Chinese ethnicities are pooled .......................... 52 Figure. 3.4. Frequency by ethnicity interaction: PA at each of the fifteen frequencies for Caucasian and Chinese ethnicities. Data from both instruments are pooled .................................................... 53 Figure. 3.5. Frequency by pressurization method interaction: PA at each of the fifteen frequencies using static and dynamic pressurization for the Reflwin system. Data from ethnicities are pooled together ..................................................................................................................................................... 55 x  Figure. 3.6. Mean and 0.95 confidence intervals (vertical bars) for PA between Caucasian and Chinese young adults using static and dynamic pressurization techniques ............................................. 56 Figure. 3.7. Mean and 0.95 confidence intervals (vertical bars) for Vea (in mmho) in the female and male groups .......................................................................................................................................... 60 Figure. 3.8. Mean and 0.95 confidence intervals (vertical bars) for Vea (in mmho) in the Caucasian and Chinese groups ............................................................................................................................ 61 Figure. 3.9. Mean and 0.95 confidence intervals (vertical bars) for Vea (in mmho) between the positive and negative tails in the female and male groups ...................................................................... 62 Figure. 3.10. Mean and 0.95 confidence intervals (vertical bars) for Vea (in mmho) in the female and male groups ................................................................................................................................. 63 Figure. 3.11. Mean and 0.95 confidence intervals (vertical bars) for Vea (in mmho) using the GSI Tympstar and Reflwin middle ear analyzer ................................................................................. 64 Figure. 3.12. Mean and 0.95 confidence intervals (vertical bars) for ear canal volume Vea (in mmho) at positive (Vea+) and negative (Vea-) tails..................................................................................... 65 Figure. 3.13. Mean and 0.95 confidence intervals (vertical bars) for ear canal volume Vea (in mmho) between males and females in the positive (Vea+) and negative (Vea-) tail groups.................. 66 Figure. 3.14. Mean and 0.95 confidence intervals (vertical bars) for Ytm for Caucasian and Chinese young adults ................................................................................................................................ 67 Figure. 3.15. Mean and 0.95 confidence intervals (vertical bars) for Ytm using the GSI Tympstar and the Reflwin Interacoustics system ..................................................................................................... 68 Figure. 3.16. Mean and 0.95 confidence intervals (vertical bars) for Ytm between Caucasian and Chinese young adults in the female and male groups .............................................................................. 69 Figure. 3.17. Mean and 0.95 confidence intervals (vertical bars) for Ytm between negative (Ytm-) and positive (Ytm+) tails in the female and male groups................................................................... 70 Figure. 3.18. Mean and 0.95 confidence intervals (vertical bars) for Ytm calculated using negative and positive tails................................................................................................................................. 71 Figure. 3.19. Mean and 0.95 confidence intervals (vertical bars) for static admittance Ytm between males and females for the Caucasian and Chinese groups ......................................................... 72 Figure. 3.20. Mean and 0.95 confidence intervals (vertical bars) for TPP measurements made using the GSI Tympstar and the Reflwin Interacoustics system ................................................................. 73 Figure. 3.21. Mean and 0.95 confidence intervals (vertical bars) for TPP using the GSI and Reflwin systems for Caucasian and Chinese young adults between females and males ........................ 74 Figure. 3.22. Mean and 0.95 confidence intervals (vertical bars) for TW for female and male groups ..................................................................................................................................................... 75 Figure. 3.23. Mean and 0.95 confidence intervals (vertical bars) for TW for Caucasian and Chinese groups .......................................................................................................................................... 76 xi  Figure. 3.24. Mean and 0.95 confidence intervals (vertical bars) for TW between Caucasian and Chinese young adults for the female and male groups ............................................................................ 77 Figure. 3.25. Mean and 0.95 confidence intervals (vertical bars) for TW using the GSI and Reflwin systems for the Caucasian and Chinese young adult groups between males and females........ 78 Figure. 3.26. Condition by Frequency Interaction (Mimosa): ER at each of the fifteen frequencies for Mimosa normal ears (N = 60) and Mimosa otosclerotic ears (N = 28) ....................................... 79 Figure. 3.27. Condition by Frequency Interaction (Reflwin): ER at each of the fifteen frequencies for Reflwin normal ears (N = 60) and Mimosa otosclerotic ears (N = 28) ........................................ 81 Figure. 3.28. ROCs showing plots of the sensitivity versus specificity for the detection of otosclerosis using the ER norms for Reflwin and Mimosa systems at 500 Hz, 630 Hz, and 800 Hz ............... 84 Figure. 3.29. Dot plot showing the sensitivity and specificity for the Mimosa norm at detecting otosclerosis for the criterion ER > 0.80 at 500 Hz ....................................................................... 85 Figure. 3.30. Dot plot showing the sensitivity and specificity for the Reflwin norm at detecting otosclerosis for the criterion ER > 0.80 at 500 Hz ....................................................................... 86 Figure. 3.31. Dot plot showing the sensitivity and specificity for the Mimosa norm at detecting otosclerosis for the criterion ER > 0.79 at 630 Hz ....................................................................... 86 Figure. 3.32. Dot plot showing the sensitivity and specificity for the Reflwin norm at detecting otosclerosis for the criterion ER > 0.79 at 630 Hz ....................................................................... 87 Figure. 3.33. Dot plot showing the sensitivity and specificty for the Mimosa norm at detecting otosclerosis for the criterion ER > 0.66 at 800Hz ........................................................................ 87 Figure. 3.34. Dot plot showing the sensitivity and specificity for the Reflwin norm at detecting otosclerosis for the criterion ER > 0.66 at 800 Hz ....................................................................... 88 Figure. 4.1. Comparison of pooled ER data for frequencies obtained at ambient pressure using the Mimosa system for the current study in dots (N = 60) and the data obtained by Shahnaz, &, Bork (2006) using the same instrument in squares (N=126). Both studies were balanced by gender and ethnicity ..................................................................................................................................................... 92 Figure. 4.2. Comparison of ER data between the current study (N = 60; 18 – 38 yr.), Shahnaz & Bork (2006) (N = 128; 18 – 32 yr.), and Voss et al. (1994) (N = 10; 18 – 24). The current study and Shahnaz & Bork (2006) use the same version of the Mimosa system, while the study by Voss et al. (1994) uses an older version of the Mimosa system ........................................................................................... 94 Figure. 4.3. Comparison of pooled ER data for frequencies obtained at ambient pressure using the Reflwin system for the current study in dots (N=60) and the data obtained by Sandford and Feeney (2008) using the a similar instrument in squares (N=20) ............................................................ 96 Figure. 4.4. Comparison of ER data between the current study (N = 60; 18 – 38 yr.), Sanford & Feeney (2008) (N = 20; 22 – 30 yr.), Feeney et al. (2004) (N = 40; 18 – 28 yr.), and Keefe et al. (1993) (N = 10; 20 – 50 yr.). The current study and Sanford & Feeney (2008) used the same version of the Reflwin system, while the studies by Feeney et al. (2004) and Keefe et al. (2004) used an older version of the Reflwin system ......................................................................................................................................... 98 xii  List of Abbreviations Abbreviation  Definition  ANOVA AUROC daPa DPOAEs ER HL HSD Hz OAEs PA Ps ROC SPL TPP TW Vea Vea+ VeaWBER Ytm Ytm+ YtmZs  Analysis of variance Area under the receiver operating curve dekaPascal Distortion-product otoacoustic emissions Energy reflectance Hearing level Honestly significant difference Hertz Otoacoustic emissions Power absorption Source pressure Receiver operating characteristic Sound pressure level Tympanometric peak pressure Tympanometric width Ear canal volume Positive compensated ear canal volume Negative compensated ear canal volume Wideband energy reflectance Static admittance Positive compensated static admittance Negative compensated static admittance Source impedance  xiii  Acknowledgements I would like to thank Dr. Navid Shahnaz for his direction and supervision as my Thesis advisor. His wisdom and enthusiasm have made for an enjoyable journey in the completion of this Thesis. I trust that the knowledge he has imparted upon me throughout my time at UBC will be invaluable as I begin my work as a clinician. I am also grateful for the suggestions made by Dr. Susan Small and Sandra Baker who served on my Thesis committee. Furthermore, I need to thank my colleagues (particularly Sarah Barnes) for their support in helping me face many challenges during the program and my instructors who have prepared me for the road ahead. Thank you to everyone who participated in my research study. In particular, I’d like to thank the overwhelming number of residents from Green College for their participation. I can assure you that your commitment to research will help improve our understanding of the middle ear and the quality of audiology services in the years to come. Finally, I would like to thank my parents (David and Carol Shaw) and my sisters (Alyssa, Margaret, Jenny and Cathy). Without your moral and financial support I would never have made it this far in my education. I am also grateful to have fantastic friends who are always willing to lend me their ears – both figuratively and literally.  xiv  Dedication  For Margaret who continues to inspire me.  xv  Chapter 1 Introduction 1.1 Basic principles The transmission of sound from the environment to the middle ear is a complex and efficient process that can easily be disrupted if there are any abnormalities in the structure of the middle ear. Given that sound must first travel through the middle ear to reach any other portion of the auditory system, the functioning of the middle ear plays a crucial role in virtually all audiological tests. For instance, while otoacoustic emission (OAE) testing primarily targets the functioning of the outer hair cells of the cochlea, any problems at the level of the middle ear will prevent the sounds of the cochlear amplifier from being measured (Sang, Park, Park & Do, 2003). It is possible to make OAE measurements because so little energy is lost in the transmission of retrograde sounds (Allen, Jeng & Levitt, 2005). As sound reaches the eardrum, the incus, malleus, and stapes must convert the waveform energy into a mechanical energy that is capable of travelling through the fluid medium of the inner ear. This process depends on overcoming the impedance mismatch that exists between the air in the ear canal and the fluid within the cochlea (Kemp, 1979). Despite the challenge of overcoming the impedance mismatch, the transmission of sound is remarkably efficient as less than 3 dB of sound is lost as sound travels from the ear canal to the cochlea (Allen et al., 2005) and a minute vibration of the ear drum can result in the perception of a sound (Tonndorf & Khanna, 1970). Standard tympanometry, acoustic reflex testing, and OAEs are some of the tools that are most widely used in clinic for assessing the status of the middle ear. Recent research, 1  however, suggests that measures of wideband energy reflectance (WBER) (eg. Shahnaz & Bork, 2006; Sanford & Feeney, 2008), WBER tympanometry (eg. Margolis, Saly & Keefe, 1999), and multi-frequency tympanometry (eg. Hunter & Margolis, 1992) may have some advantages over standard tympanometry. The accuracy of these tests in detecting middle-ear pathologies lie largely in the assumptions they make and how they compensate for the assumptions that are made. For instance, tympanometry and WBER both rely upon making assumptions about the distance from the probe to the ear drum, the changes in the cross-sectional area of the ear canal as sounds go deeper into the ear canal, and the complex geometry of the ear drum (Voss & Allen, 1993). Standard tympanometry, however, makes a further assumption that admittance is a scalar quantity, rather than a vector quantity that relies on the rectangular components of acoustic susceptance and acoustic conductance (Hunter & Margolis, 1992). A major theme of the current study is the differences in the ways in which assumptions are addressed for instruments that measure the same parameters. Middle ear assessment techniques are dependent on the basic principles of simple harmonic motion. The middle ear is analogous to a mass attached to a spring in which the compressed ear drum (compressed spring) is a source of potential energy that can be released in the form of kinetic energy (Allen et al., 2005). The term admittance (Y) refers to the acoustic energy that flows into the middle-ear system and is the parameter most commonly assessed by immittance instruments (Shahnaz & Davies, 2006). The alternative is to measure impedance (Z), which is defined as the total opposition to the flow of energy within the system. The acoustic impedance is discontinuous because it varies with frequency as the relative contributions of acoustic resistance (R) and acoustic reactance (X) change (Feeney, Grant & 2  Marryott, 2003). Acoustic impedance, resistance and reactance are related through the following equation in rectangular notation: II  ˪II - ˞I  ˪{I˭I - IJI{ - {˞˭I - ˞JI{  (1.1)  Where acoustic reactance (jXa) is the imaginary part of impedance and acoustic resistance (Ra) is the real part of acoustic impedance. Ra remains consistent, but Xa changes as a function of frequency, which changes relative to the contribution of mass and stiffness (Allen et al., 2005). At frequencies below 1000 Hz, the impedance of the ear drum is due mostly to the stiffness of the annular ligament (Lynch, Nedzelnitsy & Peake, 1982). For frequencies less than 800 Hz, the impedance is mostly due to a stiffness-based reactance (Xsa) (Puria & Allen, 1998). At 100 Hz, reactance is almost ten times larger than resistance, but at 1000 Hz their relative contributions become equal (Allen et al., 2005). At the higher frequencies, above 6000 Hz, mass-based reactance (Xma) becomes more important and the overall contribution of reactance is greater than resistance (Allen et al., 2005). From 1-5 KHz, the relative contributions of reactance and resistance are more comparable (Allen et al., 2005). The relative contribution of the mass reactance and stiffness reactance at different frequencies is important to keep in mind when measuring frequency-specific impedance. Resistance, however, is independent of frequency, as it is inherent in the system based on its physical and mechanical properties. Measures of WBER are dependent on how much energy is prevented from being absorbed into the middle ear at each frequency and greater overall impedance results in greater ER (Keefe & Simmons, 2003).  3  1.2 Wideband energy reflectance in clinical practice WBER is a middle ear assessment technique in which complex sounds ranging from 200 to 10 000 Hz or higher are presented into the ear canal and the amount of energy reflected back from the middle ear is calculated. Power reflectance, R(f) refers to the ratio of the incident (forward) and retrograde (backward) pressure waves, while /R(f)/2 is the energy reflectance (ER) (Liu et al., 2008). A value of 0 means that all of the sound energy is absorbed by the middle ear, while a value of 1 means that all of the energy is reflected back from the middle ear (Stinson, 1990). The reciprocal of ER is known as power absorption (PA). Reflectance is mathematically defined as the ratio of 1 minus the product of the admittance (Y) and characteristic impedance (Zc) and 1 plus the product of the admittance and characteristic impedance at different frequencies and static pressures (Keefe & Levi, 1996).  ˞{˦ ˜{  #  {  {  #  {  {  (1.2)  Where Zc and Y are defined by the following equations: II  I{˦{    (1.3)  ŵÈI {˦{  (1.4)  In equation 1.3, ρ is density, c is the speed of sound, and S is the cross-sectional area of the ear canal. Measures of WBER can also be made at static pressure (p) or at dynamic pressures (P). In the field of audiology, static pressure (also known as ambient pressure) refers to the pressure that exists within the ear canal when no external forces are applied and dynamic pressures refer to pressure exerted from a pump, similar to tympanometry. 4  Measures of WBER can be very useful for detecting middle-ear problems and have several advantages over tympanometry. One advantage is the fact that WBER can provide clinical information regarding the impedance across a broad range of frequencies, while standard tympanometry only gives the clinician information about the admittance at a particular frequency (Vander Werff, Prieve & Geogantas, 2007). Another advantage of WBER over standard tympanometry is the fact that the positioning of the probe tip in the ear canal will have no effect on the measurements that are obtained because the same proportion of sound energy will be reflected back (Keefe & Levi, 1996). Meanwhile, it has been shown that tympanometric measures are greatly influenced by the depth of the probe tip insertion (eg. Shanks & Lilly, 1981). Furthermore, assessments using tympanometry in infants are often faulted for being inaccurate due to large movements of the tympanic membrane when the ear canal is pressurized (eg. Holte, Cavanaugh & Margolis, 1990). WBER also has an advantage over multifrequency tympanometry because frequency-specific information can be obtained more rapidly and WBER covers a much broader range of frequencies (Vander Werff, Prieve & Geogantas, 2007). WBER is also capable of predicting the degree of conductive hearing loss (Keefe & Simmons, 2003; Piskorski, Keefe, Simmons & Gorga, 1999). Normative data on measures of WBER suggest that the most energy is reflected at the low frequencies, there are regions of lower reflectance in the mid frequencies, and moderate reflectance at high frequencies (eg. Shahnaz & Bork, 2006; Margolis, Paul, Saly & Schachern, 2001). WBER is generally expressed as a graph of how much energy was reflected in relation to 90% range of the normative data as shown in the bottom panels of Figure 1.1. Any ER values that lie outside of this range are indicative of a middle-ear disorder. For instance, excessively  5  high reflectance at the low frequencies is indicative of otosclerosis, while excessively low reflectance at the low frequencies would suggest ossicular discontinuity (Allen et al., 2005).  Figure. 1.1. Normative data obtained for measures of ER (bottom panels) and PA (top panels). Yellow region represents the 90th percentile range. The most energy is absorbed at the mid frequencies and the least energy is absorbed at the low (250 Hz) and high (8000 Hz) frequencies. Screenshot from Hear ID+MEPA RMS v4.0.4.4, Mimosa Acoustics, Inc.  Current testing protocols used in clinics are not always ideal, as not all middle-ear problems are readily detectable. For instance, patients with otosclerosis often have a conductive hearing loss and a normal or low compliance tympanogram. This profile can be confused with other disorders, such as middle ear effusion or an ear drum with low compliance (Browning, Swann & Gatehouse, 1985). Research suggests that WBER could potentially be effective in the diagnosis of otosclerosis. A study by Shahnaz and colleagues showed that in otosclerosis, a greater 6  amount of energy is reflected back from the ear drum at the low and mid frequencies (250 – 1500 Hz) compared to normal ears (Shahnaz, Bork, Polka, Longridge, Bell & Westerberg, 2009) as shown in Figure 1.2. Meanwhile, reflectance patterns above 2000 Hz are similar between normal and otosclerotic ears (Figure. 1.2). While the reflectance profiles of various middle-ear pathologies are not well-established, data suggest that a variety of middle-ear disorders, including otosclerosis and tympanic membrane perforations can be detected using WBER (Allen et al., 2005; Shahnaz et al., 2009). For instance, in one case study for otitis media the ER was 0.9 in the mid frequencies, which is much higher than 0.3 which is typically found in normal ears (Allen et al., 2005). This suggests that otitis media results in poorer transmission of sound energy into the middle ear. Since the energy loss within the adult ear canal is very small, it is the middle ear that absorbs most of the energy.  Figure. 1.2. Comparison of WBER profiles in healthy and otosclerotic ears. Note the greater reflectance at low frequencies in otosclerotic ears. Taken from Shahnaz et al., 2009. Reprinted with permission from Wolters Kluwer Health. 7  Further evidence suggests a useful role for ER in the diagnosis of ears with otitis media (Keefe & Levi, 1996; Feeney et al., 2003). Hunter, Bagger-Sjoback and Lundberg (2008) found improved test performance of ER in correctly identifying middle ear effusion in an infant population compared to conventional and high-frequency tympanometry, with a lower incidence of inconclusive results. A case study by Hunter and Margolis (1997) revealed that the presence of middle ear effusion resulted in abnormal ER when conventional tympanometry indicated normal middle-ear admittance. Several studies have reported significantly higher ER in children and adults with otitis media with effusion compared to age-matched control subjects and that elevated ER occurs over a wide range of frequencies (Jeng, Levitt, Lee & Gravel, 1999; Feeney, Grant & Marryott, 2003; Hunter et al., 2008). In addition to distinguishing between normal and pathological middle ears, WBER can provide information about the type of middle-ear pathology (Allen et al., 2005). Unlike tympanometry, Piskorski and colleagues found that WBER is able to predict conductive hearing loss. The authors reported that ER between 2 and 4 kHz is a sensitive indicator of the middle-ear status, and is a more accurate predictor of conductive impairment at 0.5 kHz (Piskorski et al., 1999). The reliability of comparing ER measurements to the 90% range of the normative data for the purpose of detecting a middle-ear disorder has been analyzed. For instance, Vander Werff, Prieve, and Georgantas (2007) examined the test-retest reliability of measuring WBER. In this study, three measures of WBER were recorded in each ear and the consistency between these three trials was assessed. It was found that test-retest reliability was poorest in the low and high frequencies and WBER measures were very consistent for the mid frequencies. Furthermore, infants who failed their OAE screening had significantly higher reflectance at the 8  mid frequencies, thereby lending further support to the idea that WBER may be useful for the detection of conductive hearing losses (Vander Werff, Prieve & Georgantas, 2007). Consistency between trials is particularly poor at the low frequencies because they are the most greatly affected by noise in the testing environment (Vander Werff et al., 2007). ER in the low frequencies is also affected by the seal of the probe tip in the ear canal, which is shown by very low ER values at the low pitches when acoustic leaking occurs (Allen et al., 2005).  1.3 Factors affecting middle ear parameters 1.3.1 Effects of ethnicity on WBER and tympanometric parameters The use of norms in evidence-based practice is a critical concept that has been used for decades. By knowing the range of results that one would expect in people with normal hearing or normal middle ears, it is possible to more accurately assess whether somebody has a hearing loss or a middle-ear problem. Clinicians generally use the same normative data for the entire adult population despite the fact that there is evidence to suggest that high-frequency hearing thresholds, tympanometric parameters, and WBER differ in different ethnic groups (eg. Dreisbach, Kramer, Cobos & Cowart, 2007; Shahnaz & Davies, 2006; Wan & Wong, 2002; Roup et al., 1998; Shahnaz & Bork, 2006). Once different norms are established for Caucasian and Chinese individuals, it will then be possible to determine if similar norms exist between Chinese and other ethnicities in Eastern Asia and whether separate Asian and Caucasian norms are warranted.  9  In order to determine whether ethnic-specific norms are warranted, comparisons of WBER have been made between Caucasian and Chinese individuals. This research has shown that Chinese young adults have higher reflectance at the low frequencies in comparison to Caucasian young adults. Meanwhile, Caucasians have more reflectance at the high frequencies compared to the Chinese counterparts (Shahnaz & Bork, 2006) as shown in Figure. 1.3. These results are consistent with the idea that the middle ears of Chinese individuals are stiffer and therefore more likely to reflect energy at the lower frequencies. It has also been suggested that the larger mass of the Caucasian ossicular chain could result in poorer high-frequency response of the middle ear, resulting in greater reflectance (Shahnaz & Bork, 2006). One confounding factor when comparing the differences between Caucasian and Chinese young adults is the effect of body size. On average, Caucasians are taller and heavier than Chinese individuals. Males are taller and heavier than females (Bell, Adair & Popkin, 2002). In order to account for these differences, comparisons can be made between Caucasian females and Chinese males who have comparable body sizes. Shahnaz and Bork (2006) made such a comparison and found that the effects of ethnicity were no longer significant when comparisons were made between Caucasian females and Chinese males, thereby suggesting that further research is needed to determine the effects of body size on WBER (Shahnaz & Bork, 2006). It was found, however, that hit rates for the detection of otosclerosis improved when ethnic-specific norms were applied, thereby lending support to the use of ethnic-specific norms for WBER (Shahnaz & Bork, 2006).  10  Figure. 1.3. Illustration of the statistical significance of the interaction of the variables of frequency and ethnicity in previous studies. At the lower pitches the Chinese group has greater reflectance and at the higher pitches the Caucasian group has lower reflectance. Taken from Shahnaz and Bork (2006). Reprinted with permission from Wolters Kluwer Health.  Measures of standard tympanometry have also been found to differ in Caucasian and Chinese individuals. The tympanogram can be analyzed based on the parameters of ear canal volume (Vea) from the positive tail (Vea+) and negative tail (Vea-); static admittance (Ytm) measured from the positive tail (Ytm+) and negative tail (Ytm-), tympanometric width (TW) and tympanometric peak pressure (TPP). Ytm refers to acoustic admittance and has the limitation of not taking into account the vector components of B and G, as discussed earlier (Hunter & Margolis, 1992). Another controversial issue is whether estimates of Ytm should be made by subtracting the positive tail (Vea+) or the negative tail (Vea-) from the peak. Calculations are generally made by subtracting the positive tail because these calculations have better testretest reliability (Margolis & Goycoolea, 1993) and are more accurate at estimating tympanometric width (Margolis & Shanks, 1991). Subtracting the negative tail from the peak, 11  however, has the advantage of being a more accurate estimation of Vea (Vea) (Shanks & Lilly, 1981). As shown in Table 1.1, Shahnaz and Davies (2006) found that Chinese individuals had significantly smaller Ytms, wider TWs, smaller Veas, and more positive TPPs than their Caucasian counterparts. When comparisons were made between Chinese males and Caucasian females, however, the effect of ethnicity was no longer significant at the 226 Hz probe-tone frequency for any tympanometric parameters (Shahnaz & Davies, 2006). Ethnicity was still significant at higher probe-tone frequencies for the parameter of Ytm and applying Caucasian norms to Caucasians with otosclerosis resulted in improved hit rates (Shahnaz & Davies, 2006). These findings illustrate the need to develop ethnic-specific norms and that norms based on body size may also be useful. The next logical course of action would be to determine whether similar norms exist between Chinese and other East Asian groups in order to determine whether Caucasian and Asian norms are warranted.  12  Table 1.1. Comparison of Ytm, TW and TPP between Caucasian and Chinese young adults. Table taken from Shahnaz and Davies, 2006. Reprinted with permission from Wolters Kluwer Health.  1.3.2 Effects of gender on WBER and tympanometric parameters The effects of gender on WBER have not been studied in much detail. The only study which analyzed the effects of gender on WBER was by Shahnaz and Bork (2006) in which no effects of gender were reported. Margolis, Saly and Keefe (1999) analyzed the effects of gender on acoustic impedance, which, as discussed earlier, is related to WBER. They reported that males had greater resistance at the frequencies below 1000 Hz, but had less resistance in the higher pitches of 2000 – 4000 Hz. Relative to females, males had ear drums that were less stiffness-dominated at frequencies below 1000 Hz. The effects of gender on tympanometric parameters have been inconsistent across different studies, with some studies showing gender effects (Wiley, Cruickshanks, Nondahl & Tweed, 1999; Roup, Wiley, Safady & Stoppenbach, 1998), another showing a gender effect  13  within an ethnic group (Shahnaz & Davies, 2006), and others showing no gender effects (Holte, 1996; Margolis & Goycoolea, 1993). There is some suggestion that females have a stiffer middle ear than males, much like Chinese young adults appear to have a stiffer middle ear compared to Caucasian young adults. This is reflected in the fact that some studies have shown that males have a larger Ytm and a narrower TW (Wiley et al., 1996; Roup et al., 1998). The parameter of Vea has been consistently larger in males than females regardless of whether measurements were made using negative or positive compensation (Shahnaz & Davies, 2006; Wan & Wong, 2002; Roup et al., 1998). This is a reflection of the fact that males generally have a larger overall body size compared to females, and body size is correlated with ear canal size (Shahnaz & Davies, 2006). It has also been shown that Vea+ is larger than Vea-; and given that Ytm is a measure of the peak admittance minus the Vea, it follows that Ytm- is larger than Ytm+ admittance (Wiley et al., 1999; Holte, 1996; Shahnaz & Davies, 2006; Wan & Wong, 2002; Roup et al., 1998). The parameter of TPP is generally consistent across studies and is close to 0 daPa, regardless of gender or ethnicity (Shahnaz & Davies, 2006; Shahnaz & Bork, 2006; Wan & Wong, 2002). Some studies have also reported different effects of gender as a function of ethnicity and/or instrument. For instance, Shahnaz and Davies (2006) reported that young Caucasian males had larger Veas than females, but there were no effects of gender for any other parameters. Meanwhile, Chinese females had significantly lower Veas and Ytms compared to Chinese males. Shahnaz and Bork (2008) also reported a three-way interaction of the variables gender, ethnicity and instrument (GSI Tympstar versus Virtual 310) for TW. It was found that TW was wider in the Virtual system compared to the GSI system for the Caucasian group; but 14  that TW was narrower in the Chinese group for the Virtual system compared to the GSI Tympstar.  1.3.3 Effects of instrument (Mimosa Versus Reflwin) on WBER Another important issue in the field of audiology is that results be consistent across different measurement systems. Can the norms obtained from one system be applied to the norms obtained from another system? One commercially available system is the Mimosa Acoustics system, which is a portable system in which a 24-bit proprietary PC card used for digital-signal processing (DSP) is inserted into a laptop computer and a probe tip system is attached to the DSP card. The probe tip is calibrated using four different-sized cavities of known volume and measurements are made in each of the four cavities. Proper calibration is dependent on the quality of the probe tip and the level of noise in the room (Mimosa acoustics, 2002). Measures of ER made using the Mimosa Acoustics system depend on obtaining the unknowns for an equation of WBER during the calibration procedure, and combining the values of these unknowns for each frequency with values obtained while making a measurement within the ear canal. In this procedure, the input acoustic admittance of the middle ear (Ym) is calculated by means of the following equation (Withnell, Jeng, Waldvogel, Morgenstein & Allen, 2009):  I˭  ˡJ È ˜˭ – IJ  (1.5)  15  Where Us is the velocity of the sound from the source and is equal to Ps/Zs, which are the sound pressure of the sound source (Ps) and the acoustic impedance at the source (Zs). Ps and Zs are the quantities that are measured during the calibration procedure, which makes use of a sound pressure technique (Withnell et al., 2009; Voss & Allen, 1994; Douglas & Levi, 1996). The quantities Pm and Ys refer to the sound pressure of the microphone source and the admittance of the source respectively. Pm and Ys are the quantities that are obtained directly during the measurement within the ear canal. The input admittance (Ym) is related to reflectance through the following equation (Withnell et al., 2009; Voss & Allen, 1994; Douglas & Levi, 1996): # #  (1.6)  Where Y0 = A/ ρc (Withnell et al., 2009; Voss & Allen, 1994; Douglas & Levi, 1996; Allen et al., 2005; Sandford & Feeney, 2008; Keefe & Simmons, 2003) and A is the cross-sectional area of the ear canal, ρ is the density of air in the ear canal, and c is the speed of sound. ρ and c are constants, while A is estimated based on the probe tip that is selected (14 A, B, or C) on the computer when the measurement is made. Thus, measurements of WBER rely on several assumptions, including the assumption that the impedance at the ear drum is similar to that at the microphone, and the cross-sectional area in all subjects who use a specific probe-tip size are approximately the same (Feeney, Grant & Marryott, 2003). At each frequency, Zs and Ps are calculated from the measurements obtained during the calibration procedure. In this procedure, a foam ear tip is placed into four cavities with a diameter of 0.74 cm and two measurements of the pressure response are made within each 16  cavity (Voss & Allen, 1994). For each measurement, the pressure-response is plotted in relation to the noise floor for each frequency (Withnell et al., 2009; Allen et al., 2005). The calibration procedure is verified by turning the cavity back to the first position where the computer determines the mean square error for each of the two measurements made in each of the four cavities (Mimosa acoustics, 2002). Another non-portable ER system called the Reflwin Interacoustics system is a prototype that is currently being distributed to researchers and is being commercialized. The calibration of this system is done by making measurements in two tubes of specified lengths. The computer can then compare the data from the two calibrations and makes a chi-square measurement to determine whether the measurements are statistically compatible (Reflwin Interacoustics, 2008). This system, as with the other system, makes measures of ambient ER. It can also, however, make three dimensional measures of WBER tympanometry, in which it is possible to view graphical representations of compliance at different pressures and frequencies. Furthermore, the Reflwin system can measure acoustic reflexes (Reflwin interacoustics, 2008). WBER measurements using the Reflwin Interacoustics system also rely on a calibration procedure followed by a measurement procedure in order to calculate all of the parameters required to measure ER. The calibration procedure for Reflwin relies on the analysis of the wave characteristics within two calibration tubes. In this method, WBER [R(f)] is calculated by using the following equation (Keefe & Simmons, 2003): { { {# #  { {{  { { "{ {  ˜{˦{  (1.7)  17  Q(f) and R0(f), which refer to the Fourier transform of the incident sound-pressure wave and the pressure at the probe respectively, can be calculated during the calibration procedure. Meanwhile, P(f), which is the recorded sound-pressure response, can be calculated during the measurement made within the subject’s ear canal, thereby allowing for the parameter of R(f) (WBER) from equation (1.7) to be solved (Keefe & Simmons, 2003). ER is related to WBER through the following equation (Keefe & Simmons, 2003; Keefe & Levi, 1996; Allen, Jeng & Levitt, 2005): È˞{˦{È$  ˗˞  (1.8)  In order to calculate the quantities of Q(f) and R0(f), the calibration procedure makes use of two plastic tubes that are closed off at the ends with a steel rod. The length of tube 1 is 295 cm, while tube 2 is 8.4 cm. The tube diameter is approximately 0.794 cm (Keefe & Simmons, 2003), which is larger than the estimation of ear canal diameter of 0.74 cm that is made for the Mimosa acoustics system (Voss & Allen, 1994). The tube lengths are chosen so that the longer tube is long enough so the waveform generated from the click stimulus can be measured without very many tube reflections. Meanwhile, the shorter tube is short enough so that its response waveform is mostly based on numerous reflections between the probe tip and the opposite end of the closed tube (Feeney & Simmons, 2003). A plastic probe tip is inserted onto the probes and is then inserted into each of the plastic calibration tubes where each of the response waveforms is measured. An analog to digital converter (ADC) picks up the response pattern from each of the tubes and graphs the waveform amplitude as a function of time as shown in Figure 1.4.  18  Figure. 1.4. Left: ADC waveforms recorded for the long tube 1 (top) and the short tube 2 (middle). The incident ADC waveform (bottom) is defined by separating the incident signal in the tube 1 waveform. Right: ADC waveform from tube 1 is redrawn by clipping the amplitude. Figures are taken from Keefe & Simmons (2003). Reprinted with permission from the American Institute of Physics.  The Fourier transform of the incident sound pressure wave (Q0F) is calculated by extracting the incident sound pressure wave from waveform 1. This is done through mathematical comparisons of the waveform responses obtained from tube 1 and tube 2 in order to extract the incident (forward-moving) waveform from the reflected (backwardmoving) waveform (Keefe & Simmons, 2003). The second parameter obtained during the calibration procedure is the pressure reflectance [R0(f)] and it can be calculated by the following equation (Keefe & Simmons, 2003). ˞I {H{  ˥-2ГL  (1.9)  Rc[K] refers to the pressure reflectance, e is a mathematical constant equal to 2.718, Г is the wave number, and L is the tube length. Based on these calculations within the two tubes, the Reflwin system calculates the SPL spectra and graphs these reponses for each of the two tubes 19  as shown in Figure 1.5. The pressure reflectance at the probe tip (R0f) can then be calculated by making comparisons between pressure reflectance measures for each of the two tubes. These comparisons involve the following basic steps, as outlined by Keefe and Simmons (2008): (1) calculating Rc[k] values for both of the two tubes using values of L and Г; (2) calculation of the source reflectance for both tubes; (3) calculating the “data” reflectance; (4) calculating the mean square error, which is the sum of the squared differences between the data reflectance and model reflectance for both tubes at each frequency. By making these comparisons between the pressure responses for both tubes, it is then possible to obtain the R0f value.  Figure. 1.5. The SPL spectra are plotted for the long tube (tube 1) as a solid gray curve, the short tube (tube 2) as a dashed black curve, and the incident signal as a solid black curve. Taken from Keefe & Simmons (2003). Reprinted with permission from the American Institute of Physics  20  1.3.4 Effects of instrumentation (Reflwin versus GSI Tympstar) on tympanometric parameters Current tympanometric systems are calibrated to make measurements at the probe-tone frequencies of 226 Hz, 678 Hz, and 1000 Hz. Comparisons between measurements made using probe-tone frequencies of 226 Hz and 1000 Hz reflect the differences in the physical characteristics that exist between the middle ears of infants and adults. In infants under the age of six months, tympanometric measurements using the probe tone of 226 Hz result in a multi-peaked admittance tympanogram, while measures made using a probe tone of 1000 Hz yield a single-peaked tympanogram (eg. Calandruccio, Fitzgerald & Prieve, 2006; Margolis, Bass-Ringdahl, Hanks et al., 2003; Shahnaz, Miranda & Polka, 2008). This finding is consistent with the fact that the middle-ear systems of infants are mass rather than stiffness-dominated. Meanwhile, adults are more likely to have a multi-peaked tympanogram at 1000 Hz and a single-peaked tympanogram at 226 Hz, which is more consistent with a stiffness-dominated system (Shahnaz et al., 2008). Interest in the 1000 Hz tympanogram has emerged because of the fact that infants with otitis media appear to have normal tympanograms when a probe tone of 226 Hz is used (eg. Hunter & Margolis, 1992; Keefe, Bulen, Arehart, et al., 1993; Paradise, Smith & Bluestone, 1976; Shahnaz, Miranda & Polka, 2008). One instrument that is routinely used in clinics to generate standard tympanograms is the Grason Stadler (GSI) Tympstar. This instrument calculates the admittance of the ear drum by measuring the volume within the ear canal as the pressure within the ear canal is changed. A probe with three ports is inserted into the ear canal. One port presents a probe-tone frequency, another changes the pressure within the ear canal, and the final port estimates the admittance of the ear drum at each frequency (Margolis & Hunter, 2000). Estimates of acoustic 21  admittance are based on calculations of the volume at each change in pressure according to the Bernouilli’s gas equation (Gaihede, Tos, Thomsen & Bale, 1997): ˜ˢ  J˞ˠ  (1.10)  Where P is pressure, V is volume, n is mole of gas, R is a universal gas constant, and T is the estimated temperature within the ear drum. It must be assumed that the moles of gas within the ear canal and the temperature within the human ear are equal to a certain average. The admittance is further calculated using the following equation (Margolis & Hunter, 2000):  I  ˢ  $$  (1.11)  Where Y is admittance, ˦J is the probe-tone frequency and V is the volume. There are certain limitations to tympanometric measures made using the GSI Tympstar. As mentioned before, the rectangular components of admittance and susceptance are not taken into account (Hunter & Margolis, 2002). As the pressure in the ear canal is changed, the ear drum is moved through pressurization, which results in a change in the middle-ear pressure (Gaihede et al., 1997). As a result, the estimates of Vea are less accurate, thereby causing inaccuracies in estimates of admittance and TW. Furthermore, Shanks & Lily (1981) tested two underlying assumptions of tympanometry: (1) Vea does not change when ear pressure is varied; and (2) an ear canal pressure of 200 daPa drives the impedance of the middle-ear transmission system to infinity so the immittance measured at 200 daPa can be attributed to Vea alone. They found that neither of these two assumptions was very accurate and could result in an error of 39% in the overestimation of the true Vea (Shanks & Lily, 1981).  22  Another alternative to using standard tympanometry and mutlifrequency tympanometry for the assessment of middle- ear disorders is the use of WBER tympanometry. As shown in Figure 1.6, the Reflwin Interacoustics system creates a three-dimensional plot of energy absorbance across a range of frequencies and pressures , effectively plotting a tympanogram. WBER tympanometry can be mathematically expressed as a plot of Ver (f,P) versus static pressure and frequency. Where Ver (f,P) refers to the reflectance equivalent volume (Ver) as a function of pressure and frequency. Ver is defined as the product of the distance between the probe tip and the ear canal and the ear canal area (Keefe & Levi, 1996).  Figure. 1.6. Three-dimensional plot of pressure, frequency, and energy absorbance as measured during WBER tympanometry. Figure taken from Reflwin interacoustics from an ear with otosclerosis. Reprinted from www.interacoustics.com with permission from Reflwin Interacoustics, Inc.  The Reflwin Interacoustics system can also extract the 226 Hz and 1000 Hz single-frequency tympanograms so that these measurements can be compared to normative data that are commonly used in clinics. The extraction of these single-frequency tympanograms are made 23  from estimates of the TPP of the WBER tympanogram rather than at ambient pressure (Reflwin interacoustics, 2008). Measurements of WBER tympanometry using the Interacoustics system rely on the use of four basic steps following the assessment of WBER. As outlined by Liu et al. (2008), these steps are: (1) artefact rejection, in which the measurement trials with energy that was too high or responses that occurred more that 10-ms after the click stimulus; (2) pressure sampling every 25 ms and interpolation of the missing pressure points; (3) fractional-octave average in which ER was averaged over 12th octave frequency bands; and (4) for every fixed frequency band, the single-frequency tympanogram was smoothed by median filtering, and resampled at every 5 daPa between -200 and 300 daPa using linear interpolation. The strategies involved in measuring single-frequency tympanograms for conventional (226 Hz) and high (1000 Hz) frequencies for the Reflwin Interacoustics system are fundamentally different from the GSI system in the sense that measures of WBER tympanometry make interpolations from measurements obtained from plots of PA versus pressure and frequency and makes use of impedance principles. Meanwhile, the GSI Tympstar plots admittance versus pressure for a single frequency and makes use of admittance principles. Admittance (Y) refers to the amount of energy that is allowed to enter the middle-ear system and is defined by the vector summation of the quantities of acoustic susceptance (Ba) and acoustic conductance (Ga). Ba refers to the energy that is exerted by the sound, while Ga refers to the absorption or dissipation of energy. These quantities are related to each other by the following equation: II  ˙I - ˪˔I  ˙I - ˪ {˔˭I - ˔JI{  (1.12)  24  Where conductance (Ga) is the real component of admittance and mass susceptance (Bma) and stiffness susceptance (Bsa) are the imaginary components of admittance. A broad, but false assumption that is made during measures of tympanometry, is that Ga is virtually zero and that Vea- can simply be subtracted from the peak admittance in order to obtain a value of Ytm (Hunter & Margolis, 1992). Measures of WBER tympanometry make use of a similar assumption because they assume that conductance is negligible.  1.4 Clinical decision analysis Once norms are established for a particular measurement, using a particular instrument, for a particular population, it is crucial to assess the efficacy with which a test can distinguish between normal and diseased conditions. The accuracy of any test can be assessed through a process known as signal detection theory, which is based on statistical decision theory. According to this theory, if the sensitivity of the test is plotted as a function of the false alarm of a test it is possible to determine how accurately the test can be used to predict what it is supposed to predict (Schwartz, Gorry, Kassirer & Essig, 1972). There are four possible outcomes that can arise when any test is measured: hits, false alarms, misses and correct rejections. The hit rate is the proportion of diseased patients correctly identified as having the disease, the false alarm rate is the proportion of non-diseased patients incorrectly identified as having the disease, the miss rate is the proportion of diseased patients incorrectly identified as not having the disease, and the correct rejection rate is the proportion of non-diseased patients correctly identified as not having the disease (Turner, Robinette & Bauch, 1999). Plotting the hit rate against the false alarm rate is the equivalent of graphing specificity versus sensitivity 25  because the sum of hits and false rejections equal one, as do the sum of the false positives and misses (Gorga, Neely & Bergman, 1992). Once the sensitivity is plotted as a function of false alarm, an ROC curve is obtained. Sensitivity and specificity are calculated by using the following formulae:  ˟˥JJ˩ˮ˩˰˩ˮ˳  ˟J˥I˩˦˩I˩ˮ˳  (1.12)  ˲ ŵŴŴ  ˲ ŵŴŴ  (1.13)  The area under the curve can be calculated to obtain a P(A) value, which is the probability that the test will accurately predict whether the ear is abnormal for that parameter. These P(A) values can range from 0.5 to 1.0, where 0.5 means the test can not distinguish between the two populations at a level better than chance, and 1.0 means the test can perfectly separate the two populations (Gorga et al., 1992). The goal of any audiological test is optimize the test’s sensitivity and specificity. The following review will outline three primary ways in which this can be done: by adjusting the criterion level, by using instrument-specific norms, by using population-specific norms. A test’s specificity and sensitivity can be improved by adjusting the criterion value (β), which is the criterion that is used to determine whether an ear is normal or abnormal (Turner et al., 1999). For instance, for measures of WBER, the criterion for the detection of otosclerosis could be ER > 0.82 for the frequency of 500 Hz. Each criterion value is associated with a particular sensitivity and specificity, which changes if the criterion is adjusted. Making the criterion more stringent (eg. increasing to ER > 0.85) will result in greater test specificity and poorer sensitivity, because fewer healthy ears will be incorrectly identified as being disordered (fewer false alarms) but 26  fewer of the disordered ears will be detected (more misses). Meanwhile, making the criterion less stringent (eg. decreasing ER to > 0.75) would result in poorer test specificity and greater test sensitivity, because more healthy ears will be incorrectly identified as being disordered (more false alarms) and more of the disordered ears will be identified as being disordered (more hits). The criterion can be adjusted to improve specificity or sensitivity depending on the population that is used and depending on the cost required to have patients be seen by a medical specialist (Turner et al., 1999). For instance, if there was a low prevalence of otosclerosis in the population, the criterion level may be more stringent because the presence of otosclerosis is less likely. If the cost of seeing a medical specialist were low, the criterion level may become less stringent because there would be fewer financial consequences to false alarms. Another way in which a test’s sensitivity and specificity may be altered is by using instrument-specific norms. If the use of instrument-specific norms does not result in improved sensitivity and specificity, then instrument-specific norms are not warranted because the test’s predictive value is not improved. One study analyzed whether tympanometric norms using a specific instrument would result in improvements in the detection of otosclerosis. Shahnaz and Bork (2008) analyzed whether the Virtual 310 System and the GSI Tympstar middle ear analyzer system would generate comparable tympanograms and whether the two instruments were comparable in the detection of otosclerosis. Standard tympanometric measures of Ytm, TPP, TW, and Vea were made. It was found that some differences existed between the two systems on measures of Vea. Differences also existed between the two systems for the multi-frequency tympanometric measures of resonant frequency and the phase angle of 45 degree. These 27  differences, however, were not as significant as the differences that exist between normal and otosclerotic ears (Shahnaz & Bork, 2008). Thus, the use of instrument-specific norms did not result in improved specificity and sensitivity for the detection of otosclerosis. It is also possible that a test’s sensitivity and specificity will be improved as a result of applying population-specific norms. As discussed earlier, Shahnaz and Davies (2006) analyzed whether ethnic-specific norms are warranted by gathering ethnic specific norms (Caucasian versus Chinese) for standard tympanometric measures of Ytm, TW, and Vea. Next, they compared these norms to data obtained from 36 individuals with otosclerosis, most of whom were Caucasian (32 of 36 were Caucasian) (Shahnaz & Davies, 2006). It was found that the use of ethnic-specific norms resulted in improvements in test sensitivity and specificity. For the probe tone of 630 Hz, the use of Chinese norms for the detection of otosclerosis offered no predictive value in the detection of otosclerosis (Shahnaz & Davies, 2006). In a subsequent study, Shahnaz et al. (2009) examined whether the parameter of ER at 500 Hz, F45 degree, or Ytm was the best predictor of otosclerosis. In that study, ethnic-specific norms were compared to 28 individuals with otosclerosis, most of whom were Caucasian. It was found that ER at 500 Hz had the best performance in detecting otosclerosis, followed by F45 degree. Ytm had the poorest overall test performance (Shahnaz et al., 2009). The previous literature illustrates that ROC analyses are invaluable tools in assessing a test performance and that improvements in test performance result from using ethnic-specific norms and from using different parameters.  28  1.5 Significance of the current study 1.5.1 Clinical implications This study will examine the consistency of measurements of ER and PA obtained using the Mimosa Acoustics system and the Reflwin Interacoustics system. By observing any differences in these measurements, it may be possible to determine the validity of the calibration methods and the validity of the assumptions that are made in calculating ER for both systems. Furthermore, this study will investigate whether the differences that exist between the two systems are clinically relevant or if separate norms are warranted for the two systems. This will be done by first comparing the ER norms for static pressure obtained from each instrument to data obtained from 28 ears with surgically-confirmed otosclerosis. The sensitivity and specificity for optimal criterion values will be compared for all of the frequencies at which static ER values offer some predictive value for the detection of otosclerosis. This study will also analyze whether there are differences between WBER measurements made at static and dynamic pressures using the Reflwin Interacoustics system. ER is typically measured at static pressure because changing the pressure within the ear canal is unnecessary. In order to most accurately reflect the pressure within the ear canal for measurements made at dynamic pressures, the Reflwin system calculates ER values made from the peak of the admittance tympanogram that is generated during tympanometry. The assumption that the tympanometric peak pressure most accurately reflects the pressure within the ear canal under normal conditions is another assumption that will be tested in this study.  29  It is also within the scope of this study to test the assumptions that are inherent in measurements of tympanometry. The consistency of measurements obtained using WBER tympanometry and conventional tympanometry for low (226 Hz) and high (1000 Hz) probetone frequencies will be assessed in this study. WBER tympanometry uses impedance principles and extrapolates single-frequency tympanograms from a three-dimensional plot of frequency, ER, and pressure. Meanwhile, standard tympanometry using the GSI Tympstar measures the tympanogram directly and makes use of admittance principles. The underlying methods and assumptions will be assessed in relation to the results that are obtained in this study. Furthermore, the parameters of sweep direction and pump speed will be considered to determine whether these variables could be responsible for the observed differences that exist between the tympanometric assessments made using the Reflwin Interacoustics system and the GSI Tympstar. The results of this study will be compared to the findings of previous studies in which significant differences have been found between the tympanometric and reflectance profiles of Caucasian and Chinese young adults. Replication of differences between these groups would support the idea that different norms should be used for Caucasian and Chinese individuals when assessing middle-ear functioning. Most researchers agree that systematic replication and cross-validation of research findings is a necessary ingredient for knowledge advancement in a discipline and essential for establishing external validity. Campbell and Stanely (1963) stated that, “...The experiments we do today, if successful, will need replication and cross-validation at other times under other conditions before they can become an established part of science, before they can be theoretically interpreted with confidence” (1963, p.3). 30  If similar results are obtained between instruments, it may be possible for clinicians to purchase the middle ear analyzer that serves the practical needs of their practice. For instance, the Reflwin system may be purchased by new clinicians who do not yet own a system that measures tympanograms. In this way, new clinicians could rapidly obtain measures of ER and WBER tympanograms in the same time required to perform tympanometric assessments with the GSI Tympstar. Meanwhile, clinicians may prefer to purchase the Mimosa Interacoustics system if they require a portable system that allows for more flexibility and mobility. Flexibility in purchasing either system will depend on whether the norms from either instrument can be used to detect middle-ear pathologies.  1.5.2 Purposes The five primary purposes of this study can be summarized as follows: (1) To determine whether the Mimosa Acoustics and Reflwin Interacoustics systems yield comparable results in terms of ER and PA; (2) To compare the static versus dynamic mode in the Interacoustics system to see whether pressurization has any impact on this measure; (3) To determine whether both systems are capable of detecting differences between reflectance profiles that exist between Chinese and Caucasian young adults; (4) To determine whether the Reflwin system and the GS Tympstar are both capable of making similar assessments of the middle ear; and (5) To determine if both the Reflwin and Mimosa systems are capable of detecting the disorder of otosclerosis.  31  Chapter 2 Methods 2.1 Participants 2.1.1 Description of participants A total of 61 adults with normal hearing from two ethnic groups participated in this study. One Caucasian female was excluded from the analyses because of her age and hearing status. In the Chinese group there were 14 males with a mean age of 24.5 years (range, 18-36 yrs.) and 16 females with a mean age of 23.4 years (range, 18-32 yrs.). In the Caucasian group there were 15 males with a mean age of 27.4 years (range, 20-38 yrs.) and 15 females with a mean age of 27.5 years (range, 23-36 yrs.). There were a total of 30 participants in the Caucasian group (60 ears) and 30 in the Chinese group (60 ears). A total of seven ears were excluded from the study due to hearing status. Overall, a total of 113 normal ears were included in the study. Participants from the normal hearing group were students or affiliates of the University of British Columbia who were informed of the study through poster advertisements, email advertisements, and word of mouth. Subjects were divided into two groups. Participants in the Chinese group were immigrants or Canadian-born children of immigrants from Mainland China, Hong Kong, or Taiwan. The Caucasian group was defined based on the definition of the Caucasian by statistics Canada (2002). Based on this definition, the Caucasian group is comprised of individuals of non-Hispanic, non-Aboriginal, non-Arab/West Arab, non-black, and non-East/South/South-East Asian, with white or light skin and of European descent. Subjects were placed in the appropriate group, based on their self-reported ethnic origin. The definition 32  of ethnicity in health research has been ambiguous and at times a difficult term to define. In this paper, we are interested in the physiological and anatomical differences rather than socioeconomic and cultural differences The otosclerotic group consisted of 28 patients with surgically confirmed otosclerosis, 8 males and 20 females with a mean age of 41.6 years (range, 24-56 yrs.). Among the otosclerotic group, 20 were white, 2 Chinese, 1 Hispanic, 4 East Indian, and 1 was Filipino. Otosclerotic group participants were recruited from academic hospitals affiliated with the University of British Columbia. Only results from the candidate ear for the surgery were analyzed (total of 28 ears). Five patients were excluded as a result of tympanic membrane abnormalities or negative pressure in excess of -200 daPa. The otosclerotic participants from the current study were the same ones analyzed by Shahnaz et al., 2009.  2.1.2 Inclusion and exclusion criteria To be included in this study, subjects had to meet the following criteria: (1) present pure-tone audiometric thresholds better than 25 dB HL at octave bands between 250 and 8000 Hz; (2) report no history of head trauma or middle-ear disease; (3) present no gross eardrum abnormalities or excessive cerumen as documented by otoscopic examination; (4) pass a distortion-product otoacoustic emission (DPOAE) screening. A pass consisted of emissions that were two standard deviations above the noise floor at three out of five of the following frequency bands: 1105 Hz, 1560 Hz, 2211 Hz, 3125 Hz, and 4416 Hz. DPOAE testing was performed to further verify the condition of the cochlea and the middle ear (Koivunen, Uhari, 33  Laitakari, Alho & Luotonen, 2000); and (5) have normal tympanometric parameters for the probe-tone frequency of 226 Hz using the GSI Tympstar. These tympanometric data were compared to the tympanometric norms for Caucasian and Chinese young adults established by Shahnaz and Bork (2008) using a similar system.  2.2 Instrumentation 2.2.1 Extended high-frequency audiometry instrumentation Pure tone audiometric thresholds were obtained using the GSI 61 audiometer and HDA 200 circumaural headphones. The GSI 61 audiometer was calibrated according to the procedure outlined by Frank (2001), in which the sound pressure level (SPL) using HDA 200 circumaural headphones is compared to the dial setting displayed on the audiometer and correction values are applied to all of the pure tone audiometric thresholds in units of hearing level (HL) that were measured in order to obtain values in units of SPL. The audiometer was calibrated prior to, once during, and following data collection using a flat-plate attachment to a coupler, one half-inch microphone, preamplifier, measuring amplifier, and a Larson Davis sound level meter. The flat-plate coupler output sound pressure levels (SPLs) remained stable (± 1.5 dB) and the measured frequency was within 1 % of the test frequency.  34  2.2.2 Wideband energy reflectance instrumentation 2.2.2.1 Mimosa Acoustics (RMS-system v4.0.4.4) Measures of WBER were made on two different middle ear analyzer systems. As described in the Introduction, one is the Mimosa system that uses an ER-10C probe with foam tips for children and adults or rubber tips for infants, and relies on determining the parameters of Ps and Zs through a calibration procedure. The equipment for this system consists of a type II PCMCIA audio data acquisition card, a probe interface cable and an ER-10CP acoustical probe with two output transducers and a microphone. These components are then connected to a laptop computer and the probe tip is then placed on the acoustical probe and measurements can be made within the calibration cavities and the subjects’ ears as shown in Figure 2.1.  Figure. 2.1. Illustration of the Mimosa Acoustics Instrumentation. The probe, four-wall cavity, PIC, DSP card, and computer screen are depicted. Reprinted from www.mimosaacoustics.com with permission from Mimosa Acoustics, Inc.  35  The reflectance measurement protocol for the Mimosa system is comprised of two steps. The first step is to calibrate the ER-10C ear probe by measuring Ps and Zs with the same tip that is used for the test ears. This requires measurements of the pressure-frequency responses of the probe for four cavities of different sizes and computation of Thevenin equivalent parameters from the four pressure-frequency responses of the probe. The second step involves measuring the pressure-frequency response from the test ear and the computation of power reflectance with the use of the Thevenin equivalent parameters of the probe. The test ear’s reflectance and immittance can be computed and plotted from these measurements. In all four cases, the probes and ear tips that were used for reflectance measurements were calibrated immediately before the measurement. Please refer to section 1.3.3 for a detailed description of the calibration procedure for the Mimosa system.  2.2.2.2 Reflwin Interacoustics (Eclipse-system v.1) The Reflwin Interacoustics system, shown in Figure 2.2 consists of an AT235 probe tip with two output transducers and one input transducer and a probe interface cable that is connected to an AT235h audiometer that is capable of changing the pressure within the ear canal for making measurements of WBER tympanometry. The AT235h audiometer is also connected to a personal computer by means of an IFC69 cable that enables communication between the computer and the audiometer (Figure. 2.2).  36  Figure. 2.2. Illustration of the Reflwin Interacoustics WBER system. Note the AT235h audiometer used for changing the pressure in the ear canal and the probe tip (bottom right), which are connected to a PC. Printed with permission from Reflwin Interacoustics, Inc.  PA measurements first require a calibration phase in which the waveform responses obtained in two plastic tubes with lengths of 295 cm and 8.4 cm are obtained. The waveform characteristics obtained within these two tubes are compared in order to determine the Fourier transform of the incident sound-pressure wave [Q(f)] and the SPL spectra within the tubes are also compared in order to determine the pressure reflectance of the probe [R0(f)]. The calibration procedure for Reflwin relies on the analysis of the wave characteristics within two calibration tubes. Please refer to section 1.3.3 for a detailed description of the calibration procedure.  37  2.2.3 Tympanometric instrumentation Measures of WBER tympanometry using the Reflwin Interacoustics system were made by extrapolating from three-dimensional plots of pressure, frequency, and PA to generate 226 Hz and 1000 Hz tympanograms. This process involved four basic steps, which were described more thoroughly in the Introduction: (1) artefact rejection; (2) time alignment; (3) fractionaloctave averaging; and (4) tympanogram resampling (Liu et al., 2008). For the Reflwin system, the pressure sweep occurred from the positive (+200 daPa) to negative (-300 daPa) direction at a rate of 200 daPa/s. Single-frequency tympanograms for probe-tone frequencies of 226 Hz and 1000 Hz were also generated directly from admittance measures obtained from the GSI Tympstar. As described previously, this instrument consists of a probe tip connected to the GSI Tympstar. The probe tip consists of three acoustical tubes: one that presents the probe-tone frequency, one that changes the pressure within the ear canal, and one that serves as a microphone and measures the admittance at the probe tip location, which is assumed to be similar to the admittance at the ear canal. Measurements using this instrument do not take into account the admittance differences that exist between the sound source and the ear canal, as shown by the lack of a calibration step. For the GSI Tympstar, the sweep pressure occurred from a positive (+200 daPa) to negative (-400 daPa) direction at a rate of 200 daPa/s.  38  2.3 Procedures 2.3.1 Consent Participants first signed a consent form for a study entitled, “The Effects of Ethnicity and Diet (vegetarian versus non-vegetarian), Caucasian versus East Asian, on Immittance Audiometry, Wide Band Reflectance, real ear to coupler difference (RECD), and Laser Doppler Vibrometry Norms”, which was approved by the Clinical Research Ethics Board (H02-70609) C02-0609. A copy of the consent form is included in Appendix II.  2.3.2 Distortion-product otoacoustic emissions The first testing procedure was an otoscopic examination of the ear canal using a WelchAllyn clinical otoscope with disposable tips. The second procedure was DPOAE testing. Participants were instructed to sit in a sound booth while wearing a plastic ear tip that fit their ears. DPOAEs were measured using Intelligent Hearing Systems software in which an ear calibration was calculated, followed by the presentation of sounds into the ear canal. The tones occurred at the following f2 frequencies: 1105 Hz, 1506 Hz, 2211 Hz, 3125 Hz, 4416 Hz, 6250 Hz, and 8837 Hz. During DPOAE testing, two frequencies (f1 and f2) are presented into the ear canal in an f2/f1 ratio of approximately 1.2. The intensities of the two tones are known as primaries, where L1 refers to the intensity of f1 and L2 refers to the intensity of f2. L1 was 65 dB and L2 was 55 dB. The internal amplifier within the cochlea creates an echo that is compared to the noise floor in order to determine if the measured response is at least two standard deviations above the noise floor, and therefore statistically significant. The sampling 39  rate was two frequencies per octave, with 32 samples starting at the high frequencies and going to the low frequencies. The order of DPOAE testing was randomized, meaning for some participants the right ear was tested first and for others the left ear was tested first.  2.3.3 Tympanometry The third procedure was tympanometric testing at probe-tone frequencies of 226 Hz and 1000 Hz in both ears using the GSI Tympstar. In this procedure, a small plastic tip is placed at the opening of the ear canal in order to make a seal. Ytm and Vea were obtained from both 226 Hz and 1000 Hz probe-tone frequencies and from both positive and negative tails. TPP and TW were only obtained for the 226 Hz prone tone frequency. The order of standard tympanometric testing was randomized, meaning for some participants the right ear was tested first and for others the left ear was tested first. Some participants were first tested using the probe-tone frequency of 226 Hz while others were first tested using the probe-tone frequency of 1000 Hz.  2.3.4 Conventional and extended high-frequency audiometry The fourth testing procedure was conventional and extended high-frequency audiometric testing. The procedure that was used was the one outlined by Frank (2001). This method makes use of a down ten up two approach in which the dial on the audiometer was brought down 10 dB each time the participant pushed the response button to indicate that they heard a sound. The dial on the audiometer was then brought up in 2 dB steps until the 40  participant made another response. This same procedure was repeated six times for each frequency so that the hearing thresholds for six ascending trials could be averaged together. Testing was completed in both ears at the following frequencies: 250 Hz, 500 Hz, 750 Hz, 1000 Hz, 1500 Hz, 2000 Hz, 3000 Hz, 4000 Hz, 6000 Hz, 8000 Hz, 9000 Hz, 10000 Hz, 11200 Hz, 12500 Hz, 14000 Hz, and 16000 Hz. The order of pure tone audiometric testing was randomized, meaning for some participants the right ear was tested first and for others the left ear was tested first. Conventional audiometric frequency testing was used for the purpose of determining if participants would be included in the data analysis and if they qualified for inclusion in the study. The conventional and extended high frequency results from this study, in units of HL, were converted to SPL values through means of a calibration procedure and these SPL values were analyzed in another study. The audiometer was calibrated prior to, once during, and following data collection using a flat-plate attachment to a coupler, one half-inch microphone, preamplifier, measuring amplifier, and a Larson Davis sound level meter (Model 831).  2.3.5 Wideband energy reflectance and WBER tympanometry The final procedures were WBER and WBER tympanometry. WBER was measured on two different systems. For the portable Mimosa system (RMS v4.0.4.4), 14A probe tips were used for all participants. As described previously, Ps and Zs were calculated during the calibration phase. During the measurement phase, proper insertion of the probe tip was assessed by determining the level of a chirp signal in relation to the noise within the 41  environment. Next, rapid chirp stimuli ranging from 250 Hz to 8000 Hz were presented into the subject’s ear and the reflectance and absorbance profiles were determined for each frequency. Two measures of ER were made for each ear and two measurements that were within 5% of each other were considered to be sufficiently replicable. The software automatically calculated a PA value, which is the reciprocal of ER. The ER and PA values were averaged together and analyzed in third-octave bands in order to simplify the analysis and to simplify comparisons to other studies that also analyzed ER and PA in third octave bands (eg. Voss et al., 1994; Keefe et al., 1993; Feeney et al., 2004). A similar approach was used when making PA measurements with the Reflwin Interacoustics system using rapid click stimuli from 250 Hz to 8000 Hz. The calibration procedure, however, measured the waveform response, as described previously. Three dimensional plots of pressure, PA, and frequency were generated for measurements made at dynamic pressures. The software also generated a graph of the PA at the TPP. Each PA measurement was performed twice within each ear to confirm the visual replicability of the two trials. The software, however, only stored one result and the researcher selected one of two PA outputs that looked similar. A single PA value was converted into third-octave bands and ER values were generated by using the equation ER = 1-PA in excel. The exact same procedure was performed for PA measurements made at ambient pressure, except it was unnecessary for the software to generate a PA graph at the TPP since measurements were made at the pressure within the ear canal. The Reflwin software automatically recorded the WBER tympanograms for 226 Hz and 1000 Hz while measuring PA at dynamic pressures. These tympanograms were generated using 42  a procedure that was described previously. The admittance values at each pressure for both tympanograms (226 Hz and 1000 Hz) were recorded for each of the following parameters: Ytm+, Ytm-, Vea+, Vea-. The parameters of TPP and TW were only recorded for the probe-tone frequency of 226 Hz. This allowed for comparisons of these parameters between the GSI Tympstar and the Reflwin systems. The Reflwin system does not calculate TW directly, thus this parameter was calculated by subtracting half of the Ytm value from the peak admittance and determining the pressures on both sides of the peak admittance at which this value was reached. The difference between these two pressures represents the TW. The order of ER and WBER tympanometric testing was randomized, meaning for some participants the right ear was tested first and for others the left ear was tested first. Furthermore, some participants were first tested using the Reflwin system, while others were first tested using the Mimosa system. Measurements made at static and dynamic pressures were also randomized for the Reflwin system, meaning some participants were tested at static pressure first, while others were first tested at dynamic pressures.  2.4 Statistical analyses The first section of the results (3.1) will outline the descriptive statistics for WBER values obtained for Caucasian and Chinese young adults using each of the two instruments. Next, the significant interactions using an ANOVA analysis in which ethnicity, gender, and ear serve as between-subject factors and instrument serves as a within-subject factor shall be presented. These interactions shall be presented graphically and the nature of the interactions of these 43  variables shall be further explored using a post-hoc analysis. The same ANOVA and post-hoc analyses will also be presented for the PA data. Finally, a third ANOVA in which ethnicity, gender, and ear serve as between-subject factors and pressurization method (static versus dynamic) using the Reflwin system serves as a within-subject factor shall be presented to determine whether there is a difference in the PA data obtained using these two pressurization methods. The second portion of the results (3.2) will begin by presenting the norms obtained from the current study for the tympanometric variables of Ytm+, Ytm-, Vea+, and Vea- (for probe tones of 226 Hz and 1000 Hz) and TPP, and TW (for probe-tone frequency of 226 Hz) for Caucasian and Chinese young adults using the Reflwin system and the GSI Tympstar. Next, the differences between the Reflwin system and the GSI Tympstar for the parameters of Ytm and Vea at positive and negative tails will be analyzed for both probe-tone frequencies. TPP and TW will only be analyzed for the probe-tone frequency of 226 Hz. Comparisons for the first two variables (Ytm and Vea) will be made using an ANOVA analysis in which gender, ethnicity, and ear serve as between-subject factors while instrument and tails will serve as within-group factors. Comparisons for the other two variables (TPP and TW) will be made using an ANOVA analysis in which gender, ethnicity, and ear serve as between-subject factors and instrument serves as a within-subject factor. All of these significant interactions will be displayed graphically. The third section of the results (3.3) will focus on the practical significance of the differences in the ER measurements obtained at static pressure using the Reflwin and Mimosa system. First, an ANOVA analysis will be performed in which ear, gender, and frequency will 44  serve as between-subject factors and patient condition (otosclerosis using the Mimosa system versus normative data using the Mimosa system) will serve as a within-subject factor. A second ANOVA will also be performed using the same factors, except the normative data for the factor of patient condition will be those obtained using the Reflwin system. Post-hoc analyses will follow in order to determine at which frequencies each system is capable of detecting significant differences between the normative and the otosclerotic data. Next, an ROC curve will be presented in order to determine each instrument’s ability to correctly identify otosclerosis for the frequencies at which both instruments detected differences between the norms and the otosclerotic ears. Pair-wise comparisons for the area under the curves will be performed in order to determine whether one system is more efficient at detecting otosclerosis. The sensitivity and specificity for each ROC curve will be further explored through means of dot plots in which all of the ER values for all of the normal ears and all of the otosclerotic ears are compared and the sensitivity and specificity for a given criterion level is plotted.  45  Chapter 3 Results The results are divided into the following three sections: (3.1) analysis of normative data for WBER and PA using the Mimosa and Reflwin systems; (3.2) analysis of normative data for tympanometric variables using the Reflwin and GSI Tympstar systems; and (3.3) analysis of the ability for the Mimosa and Reflwin system to detect otosclerosis.  3.1.1 Energy reflectance Table 3.1 shows the descriptive statistics, including the mean, standard deviation and 90% range were determined for both instruments (Mimosa and Reflwin) at fifteen frequencies (250 Hz, 315 Hz, 400 Hz, 500 Hz, 630 Hz, 800 Hz, 1000 Hz, 1250 Hz, 1600 Hz, 2000 Hz, 2500 Hz, 3150 Hz, 4000 Hz, 5000 Hz, and 6000 Hz) for both ethnic groups (Caucasian and Chinese) for the parameter of ER at static pressure.  46  Frequency  Caucasian, Mimosa  Caucasian, Reflwin  Chinese, Mimosa  Chinese, Reflwin  (Hz)  Mean  90%  Mean  90%  Mean  Mean  250  0.90  0.06  0.83 – 0.97  0.84  0.17  0.64 – 0.95  0.92  0.04  0.85 – 0.97  0.88  0.05  0.78 – 0.95  315  0.85  0.07  0.74 – 0.95  0.81  0.16  0.61 – 0.93  0.88  0.04  0.81 – 0.94  0.85  0.06  0.75 – 0.93  400  0.80  0.09  0.65 – 0.92  0.77  0.15  0.57 – 0.91  0.84  0.05  0.76 – 0.91  0.81  0.07  0.68 – 0.91  500  0.70  0.12  0.52 – 0.87  0.69  0.14  0.49 – 0.86  0.77  0.07  0.66 – 0.86  0.73  0.09  0.58 – 0.85  630  0.59  0.16  0.33 – 0.79  0.57  0.15  0.35 – 0.76  0.68  0.09  0.52 – 0.80  0.61  0.10  0.46 – 0.76  800  0.49  0.16  0.19 – 0.68  0.51  0.15  0.24 – 0.71  0.58  0.12  0.39 – 0.73  0.56  0.11  0.42 – 0.72  1000  0.42  0.13  0.18 – 0.59  0.42  0.15  0.17 – 0.58  0.48  0.13  0.26 – 0.67  0.46  0.12  0.22 – 0.60  1250  0.35  0.12  0.17 – 0.55  0.32  0.09  0.17 – 0.47  0.43  0.13  0.20 – 0.62  0.37  0.11  0.16 – 0.52  1600  0.36  0.11  0.17 – 0.52  0.37  0.09  0.23 – 0.52  0.39  0.13  0.11 – 0.58  0.39  0.10  0.20 – 0.55  2000  0.34  0.12  0.10 – 0.50  0.36  0.12  0.15 – 0.53  0.36  0.12  0.13 – 0.54  0.36  0.11  0.18 – 0.52  2500  0.30  0.12  0.08 – 0.47  0.35  0.15  0.12 – 0.62  0.30  0.12  0.11 – 0.47  0.32  0.12  0.12 – 0.50  3150  0.28  0.13  0.06 – 0.49  0.30  0.17  0.05 – 0.54  0.25  0.12  0.07 – 0.45  0.26  0.12  0.07 – 0.45  4000  0.35  0.16  0.11 – 0.64  0.36  0.19  0.04 – 0.65  0.28  0.17  0.08 – 0.60  0.30  0.16  0.07 – 0.57  5000  0.56  0.21  0.14 – 0.64  0.50  0.19  0.09 – 0.80  0.40  0.21  0.10 – 0.75  0.40  0.19  0.11 – 0.70  6000  0.70  0.20  0.37 – 0.98  0.66  0.15  0.38 – 0.87  0.51  0.20  0.18 – 0.83  0.53  0.15  0.28 – 0.78  SD  SD  SD 90%  SD  90%  Table 3.1. Descriptive statistics, including the mean, standard deviation, and 90th percentile range at each of the fifteen frequencies for measurements made at ambient pressure using the Mimosa System and the Reflwin Interacoustics system.  47  It was found that the main effects of gender [F (1,101) = 0.02, p = 0.90], ethnicity [F (1,105) = 0.18, p =0.67], and ear [F (1,105) = 1.51, p =0.22] were not significant. The withinsubject factors of instrument [F (1,101) = 15.43, p < 0.0001] and frequency [F (1,1414) = 381.26, p < 0.0001] were significant. Following a Greenhouse-Geiser correction, only the following higher order interactions remained significant: ethnicity by frequency, instrument by frequency, and instrument by frequency by gender. There was a significant interaction between ethnicity and frequency [F (1,1414) = 10.96, p < 0.0001], indicating that ER varies differently across frequencies between the Caucasian and Chinese groups. Significant instrument and frequency [F (1,1414) = 6.70, p < 0.0001] interactions indicate that ER varies across frequencies between the two instruments. Furthermore, there were significant interactions between ethnicity, frequency and instrument [F (1,1414) = 1.86, p < 0.05], indicating that ER varies differently across frequencies between the two systems and between the two ethnic groups. The interactions between ethnicity, frequency and instrument were explored further through a post-hoc Tukey HSD analysis. ANOVA and post-hoc analysis Tables are shown in Appendix I. The interaction between the factors of ethnicity and frequency was explored using a Tukey HSD test. It was found that there were significant differences between Caucasian and Chinese ethnicities for the highest frequencies of 5000 Hz and 6000 Hz. This indicates that the changes in ER patterns between these frequencies are different in the two ethnicities. A graphical representation of the data also suggests that the ER values are higher for Chinese young adults for frequencies of 1250 Hz and below, as shown in Figure 3.1. ER values were similar for both ethnicities at frequencies of 1600 Hz to 3150 Hz, and ER values were higher for 48  Caucasian young adults at frequencies of 4000 Hz to 6000 Hz (Figure. 3.1). It is also worth noting that the ER profile of the Caucasian young adults had two distinct minima (lowest ER values) at frequencies of 1250 Hz and 3150 Hz, with the larger minima (lower ER value) being at 3150 Hz. Meanwhile, the Chinese young adults only had one distinct minimum at 3150 Hz.  Figure. 3.1. Ethnicity by Frequency Interaction: ER at each of the fifteen frequencies for Caucasian and Chinese Young Adults. Data from the Reflwin Interacoustics and Mimosa middle ear analyzer systems are Pooled together.  The interaction between instrument, frequency, and gender was explored further through means of a post-hoc analysis. As shown in Figure 3.2, there were no significant differences between ER values obtained using the two systems for females (p > 0.05 for all 49  frequencies). Meanwhile, for males, the measurements made using the Mimosa system were significantly larger than those made using the Reflwin system for frequencies of 250 Hz, 315 Hz, 400 Hz, 500 Hz, and 630 Hz (p < 0.05). It is also worth noting that the overall ER values are lower at the low frequencies for the Reflwin system compared to the Mimosa system, indicating that the Reflwin system detects less energy being transmitted through the middle ear system (Figure 3.2).  Figure. 3.2. Gender, frequency, and instrument interaction: ER at each of the fifteen frequencies for females and males. Data from Caucasian and Chinese ethnicities are pooled together.  50  3.1.2 Power Absorption It was found that there was a significant main effect for the factors of instrument [F (1,104) = 109.47, p < 0.0001] and frequency [F (1,1414) = 418.78, p < 0.0001]. Following a Greenhouse-Geiser correction, the following interactions remained significant: frequency by ethnicity, and instrument by frequency. The significant interaction between frequency and ethnicity [F (1,1414) = 18.21, p < 0.0001] indicates that PA varies differently across frequencies between the Caucasian and Chinese groups. Furthermore, the instrument and frequency interaction [F (1,1414) = 31.37, p < 0.0001] indicates that PA varies differently across frequencies between the two instruments. The higher order significant interactions of frequency by ethnicity and instrument by frequency were investigated further through a posthoc Tukey HSD analysis. ANOVA and post-hoc analysis Tables are shown in Appendix I. Contrary to the ER data, there was no significant three-way interaction for the variables of gender, frequency and instrument. A post-hoc analysis of the instrument by frequency interaction revealed that the PA measurements were larger in the Mimosa system for frequencies between 250 Hz and 1600 Hz (p < 0.05). This suggests that the Mimosa system detects more energy being absorbed by the middle ear system at those frequencies because there are differences in the calibration procedures used for each instrument. This trend is also evident in Figure. 3.3.  51  Figure. 3.3 Frequency by instrument interaction: PA at each of the fifteen frequencies for the Mimosa and Reflwin systems. Data from Caucasian and Chinese ethnicities are pooled together.  A post-hoc analysis for the ethnicity and frequency interaction revealed that there were statistically significant differences between PA values for both ethnicities at 630 Hz, 800 Hz, 5000 Hz, and 6000 Hz (p < 0.05). Figure 3.4 shows that Caucasian young adults had larger PA values for the low frequencies of 1250 Hz and below, while Chinese young adults had larger PA values for frequencies of 4000 Hz to 6000 Hz. Furthermore, the maxima for the PA data using both systems occurred at the same frequencies as the minima for the ER data using both systems (comparing Figures 3.1 and 3.3).  52  Figure. 3.4. Frequency by ethnicity interaction: PA at each of the fifteen frequencies for Caucasian and Chinese. Data from both instruments are pooled together.  3.1.3 Static versus dynamic pressure measurements using the Reflwin System It was also necessary to determine whether measurements made using static (no pressure change) and dynamic (obtained during pressure change) modes for the Reflwin system were comparable. It was found that there was a significant main effect for the factors of frequency [F (1,1428) = 368.43, p < 0.0001] and pressurization method [F (1,102) = 19.9, p < 0.0001]. Following a Greenhouse-Geiser correction, the following interactions remained significant: frequency by gender, frequency by ethnicity, pressurization method by ethnicity, and pressurization method by frequency. The interaction of frequency and gender [F (1,1428) = 4.62, p < 0.0001] was significant, indicating that PA values varied differently across frequencies 53  between males and females. The frequency and ethnicity interaction [F (1,1428) = 12.08, p < 0.0001] was also significant, showing that PA values varied differently across frequencies between Caucasian and Chinese young adults. The pressurization method and ethnicity interaction [F (1,102) = 5.3, p < 0.05] showed that PA values varied differently across pressurization method between the Caucasian and Chinese groups. Furthermore, the significant pressurization method and frequency interaction [F (1,1428) = 19.9, p < 0.0001] showed that PA values varied differently across frequencies between the two pressurization methods. The interactions of frequency by ethnicity and frequency by gender were not analyzed further, as they have been discussed previously. The interaction of frequency by pressurization method was explored through a post-hoc Tukey HSD analysis. ANOVA and posthoc analysis Tables are shown in Appendix I. For the interactions between frequency and pressurization method, post-hoc analyses revealed that there were statistically significant differences between the two pressurization methods at frequencies of 250 Hz, 400 Hz, 500 Hz, 630 Hz, 800 Hz, 1000 Hz, 1250 Hz, 1600 Hz, and 2000 Hz (p < 0.05). Figure 3.5 shows that PA values were higher for the dynamic pressurization method, indicating that more energy at the low frequencies is detected as being transmitted through the middle ear using dynamic pressures. It was previously established that measurements obtained using static pressure for the Reflwin system resulted in higher PA at the lower frequencies compared to the Mimosa system (Figure. 3.4). Since measurements made at dynamic pressures using the Reflwin system are higher than those obtained at static pressures at the low frequencies (Figure. 3.5), it is clear that the differences in PA between the  54  two systems are even more pronounced when PA values are obtained using dynamic pressures for the Reflwin system.  Figure. 3.5. Frequency by pressurization method interaction: PA at each of the fifteen frequencies using static and dynamic pressurization for the Reflwin system. Data from ethnicities are pooled together.  The interaction of the variables of pressurization method and ethinicity revealed that for both ethnicities there was an increase in the PA values when dynamic pressures were used. As shown in Figure 3.6 the differences between the measurements made using the two systems were more pronounced for the Caucasian group. 55  Figure. 3.6. Mean and 0.95 confidence intervals (vertical bars) for PA between Caucasian and Chinese young adults using static and dynamic pressurization techniques.  3.2 Tympanometric parameters Table 3.2 shows the descriptive statistics, including the mean, standard deviation and 90% range were determined for both instruments (GSI Tympstar and Reflwin), at two probetone frequencies (226 Hz and 1000 Hz) for both ethnicities (Caucasian and Chinese) for the parameters of Vea+, Vea-, Ytm+, Ytm-, TPP, and TW. The parameters of TPP and TW were only analyzed for the probe-tone frequency of 226 Hz because the probe-tone frequency of 226 Hz has often been used for these variables when making differential diagnoses of the middle ear.  56  Furthermore, since the proportion of the notched admittance tympanogram increases at 1 kHz, there is no standard with regard to the calculation of TPP and TW at this probe-tone frequency.  57  Parameter Ytm+ Mean SD 90%  Caucasian 226 Hz GSI Reflwin 0.73 0.91 0.36 0.45 0.35 – 1.48 0.48 – 1.47  Caucasian 1000 Hz GSI Reflwin 1.43 1.24 1.15 1.33 -0.74 – 2.92 -1.36 – 3.5  Chinese 226 Hz GSI Reflwin 0.44 0.59 0.18 0.38 0.24 – 0.74 0.29 – 1.06  Chinese 1000 Hz GSI 0.96 0.92 -0.5 – 2.59  Reflwin 1.12 0.82 -0.02 – 2.54  Ytm - Mean SD 90%  0.99 0.48 0.56 – 1.68  0.80 0.37 0.44 – 1.56  2.25 1.32 0.06 – 4.7  2.51 1.11 0.80 – 4.21  0.62 0.33 0.30 – 1.32  0.50 0.23 0.20 – 0.85  1.99 0.93 0.86 – 3.96  2.14 0.9 0.8 – 3.75  Vea + Mean SD 90%  1.33 0.36 0.82 – 1.88  1.29 0.29 0.89 – 1.81  5.55 1.42 3.57 – 7.59  6.09 1.13 4.53 – 8.08  1.17 0.25 0.8 – 1.57  1.25 0.26 0.89 – 1.68  5.22 0.96 3.7 – 6.95  5.79 0.9 4.61 – 7.52  Vea - Mean SD 90%  1.43 1.17 0.72 – 1.96  1.22 0.33 0.74 – 1.77  4.54 1.05 2.83 – 6.12  5.01 0.98 3.58 – 6.57  1.13 0.26 0.78 – 1.58  1.19 0.26 0.81 – 1.71  4.19 0.79 2.86 – 5.6  4.76 0.78 3.58 – 6.18  TPP  Mean SD 90%  -2.46 15.5 -20 – 15  -24.2 13.7 -47.5 - -5  N/A  N/A  -5.67 8.85 -17.5 - 5  -27.9 10.7 -50 - -9.75  N/A  N/A  TW  Mean SD 90%  73.6 25.5 40 – 129.8  85.4 29.3 55 – 149  N/A  N/A  87.8 31.9 49.8 - 150  90.9 35.9 35 – 155.5  N/A  N/A  Table 3.2. Descriptive statistics, including the mean, median, standard deviation, 5th percentile and 95th percentile for the parametes of Vea+, Vea-, Ytm+, Ytm-, TPP, and TW made using the Mimosa Acoustics system and the Reflwin Interacoustics system for both ethnicities.  58  3.2.1 Ear canal volume 3.2.1.1 226 Hz The relationships between the variables of gender, ethnicity, and instrument were established for the tympanometric parameter of Vea at positive and negative tails for the probe-tone frequency of 226 Hz. It was found that there were significant main effects for the variable of gender [F (1,109) = 6.89, p < 0.0001]. There was also a significant effect for the interaction of instrument by ethnicity [F (1,109) = 5.02, p < 0.0001], indicating that Vea varies differently across instruments between Caucasian and Chinese young adults. Furthermore, the significant tail by gender interaction [F (1,109) = 5.02, p < 0.05] shows that measures of Vea vary differently across positive and negative tails between males and females. The relationships of these main effects and interactions were explored further through graphical analyses. ANOVA Tables are shown in Appendix I. As shown in Figure 3.7, analysis of the main effect of gender revealed that when negative and positive tails were pooled together across instruments, the Vea for males were larger than those of females.  59  Figure. 3.7. Mean and 0.95 confidence intervals (vertical bars) for Vea (in mmho) in the female and male groups.  As shown in Figure 3.8, the interaction of instrument by ethnicity revealed that for the GSI Tympstar system, the Caucasian young adults had larger Veas than the Chinese young adults when positive and negative tail data were pooled together. Meanwhile, for the Reflwin system, the Veas between Caucasian and Chinese young adults were comparable (Figure. 3.8).  60  Figure. 3.8. Mean and 0.95 confidence intervals (vertical bars) for Vea (in mmho) in the Caucasian and Chinese groups.  As shown in Figure 3.9, when the results of the positive and negative tails were pooled across instruments, the volumes for males using Vea- were larger than the volumes using Vea+. Meanwhile, females had an opposite trend in which Vea+ was larger than Vea- (Figure. 3.9).  61  Figure. 3.9. Mean and 0.95 confidence intervals (vertical bars) for Vea (in mmho) between the positive and negative tails in the female and male groups.  3.2.1.2 1000 Hz It was found that there were significant effect for the main effects of gender [F (1,108) = 61.63, p < 0.0001], instrument [F (1,108) = 31.04, p < 0.0001] and tail [F (1,108) = 121.24, p < 0.0001]. There was also a significant effect for the interaction of tail and gender [F (1,108) = 1.11, p < 0.05], indicating that Vea varied differently according to tail between the two genders. ANOVA Tables are included in Appendix I. The natures of these effects were explored further through the graphical representations of these relationships. It was found that when the values of positive and 62  negative tails were pooled across instruments and ethnicity, males had an overall larger Vea than females as shown in Figure 3.10. This relationship was even more pronounced than the effect of gender for the probe-tone frequency of 226 Hz (comparison of Figures. 3.7 and 3.10).  Figure. 3.10. Mean and 0.95 confidence intervals (vertical bars) for Vea (in mmho) in the female and male groups.  As shown in Figure 3.11, when the data were pooled across ethnicity and gender, the volumes obtained using the Reflwin system were larger than those obtained using the GSI Tympstar system.  63  Figure. 3.11. Mean and 0.95 confidence intervals (vertical bars) for Vea (in mmho) using the GSI and Reflwin middle ear analyzer systems.  As shown in Figure 3.12, there was also a significant main effect of tail (positive versus negative). When data were pooled across instrument, gender, and ethnicity, it was found that the positive tail values tended to be larger than the negative tail values (Figure. 3.12).  64  Figure. 3.12. Mean and 0.95 confidence intervals (vertical bars) for Vea (in mmho) at positive (Vea+) and negative (Vea-) tails.  The interaction of the variables tail and gender revealed that the positive and negative tail values are greater in males than in females. As shown in Figure 3.13, the nature of this relationship is more pronounced for the positive tail in comparison to the negative tail.  65  Figure. 3.13. Mean and 0.95 confidence intervals (vertical bars) for Vea (in mmho) between males and females in the positive (Vea+) and negative (Vea-) tail groups.  3.2.2 Static admittance 3.2.2.1 226 Hz There were significant main effects for ethnicity [F (1,109) = 7.4, p < 0.0001] and instrument [F (1,109) = 1.21, p < 0.05]. The interaction of the variables gender and ethnicity [F (1,109) = 5.02] was found to be significant, indicating that Ytm varies differently across genders between the Caucasian and Chinese groups. Furthermore, a significant tail by gender  66  interaction [F (1,109) = 5.02, p < 0.05] showed that admittance varied differently across tails between the two genders. As shown in Figure 3.14, when all of the data for Ytm were pooled across instrument and compensation method, Ytm was larger for Caucasian young adults than it was for Chinese young adults.  Figure. 3.14. Mean and 0.95 confidence intervals (vertical bars) for Ytm for Caucasian and Chinese young adults.  67  As shown in Figure 3.15, when the admittance data were pooled across gender and ethnicity, the mean variability in these measures was greater when using the GSI Tympstar than when the Reflwin Interacoustics system was used.  Figure. 3.15. Mean and 0.95 confidence intervals (vertical bars) for Ytm using the GSI Tympstar and the Reflwin Interacoustics system.  Further investigation of the gender by ethnicity interaction revealed that for Caucasians, the Ytm was significantly larger in females than in males. Meanwhile as shown in Figure 3.16, for Chinese young adults, there were no prominent differences between the two genders when data were pooled across instruments.  68  Figure. 3.16. Mean and 0.95 confidence intervals (vertical bars) for Ytm between Caucasian and Chinese young adults in the female and male groups.  When the data were pooled across instrument and ethnicity, there was also a significant interaction between the variables of tail and gender. It was found that females had significantly larger Ytm values when measurements were made using negative compensation. Meanwhile, as shown in Figure 3.17, the Ytm values calculated from the positive tail were similar in both genders. It is also worth noting that there was much more variability (larger 0.95 confidence intervals) for measurements made using negative compensation in comparison to measurements made using positive compensation (Figure. 3.17). 69  Figure. 3.17. Mean and 0.95 confidence intervals (vertical bars) for Ytm between negative (Ytm-) and positive (Ytm+) tails in the female and male groups.  3.2.2.2 1000 Hz It was found that there was a significant main effect for the factor of compensation method [F (1, 108) = 517.11, p < 0.0001] and there was also a significant effect for the interaction of gender and ethnicity [F (1,108) = 6.72, p < 0.05], indicating that Ytm varies differently for the two ethnicities between males and females. There were no other significant effects using this mixed-model ANOVA. ANOVA Tables are shown in Appendix I. When gender and ethnicity were pooled across instruments, it was found that Ytmmeasures were larger than Ytm+ measures (Figure. 3.18). 70  Figure. 3.18. Mean and 0.95 confidence intervals (vertical bars) for Ytm calculated using negative and positive tails.  The significant interaction of the variables gender and ethnicity was also investigated, as shown in Figure 3.19. In young adults, the differences in Ytm values between Caucasian females and Chinese females were much greater than the differences between Caucasian males and Chinese males (Figure. 3.19). Furthermore, the Ytm for females was larger in the Caucasian group compared to the Chinese group. Meanwhile, for males, the opposite was true and Ytm of Chinese males was larger than the Ytm of Caucasian males (Figure. 3.19)  71  Figure. 3.19. Mean and 0.95 confidence intervals (vertical bars) for Ytm between males and females for the Caucasian and Chinese groups.  3.2.3 Tympanometric peak pressure 3.2.3.1 226 Hz It was found that there was a main effect for instrument [F (1,109) = 324.8, p < 0.0001]. The interaction between instrument, gender and ethnicity was also found to be significant [F (1,109) = 6.12, p < 0.05], indicating that TPP varies differently across the two instruments between the two genders and between the two ethnic groups. There were no other significant main effects or interactions. ANOVA Tables are shown in Appendix I. As shown in Figure 3.20, it was found that TPP values obtained using the GSI Tympstar were more positive than those obtained using the Reflwin Interacoustics system when the data  72  were pooled across gender and ethnicity. Furthermore, the variability (0.95 confidence interval) was similar in both instruments (Figure. 3.20).  Figure. 3.20. Mean and 0.95 confidence intervals (vertical bars) for TPP measurements made using the GSI Tympstar and the Reflwin Interacoustics system.  As shown in Figure 3.21, there was a significant three-way interaction for the variables of ethnicity, gender and instrument. For the Caucasian group, TPP measurements made using the GSI system detected more negative values for females compared to males. For the Reflwin system, however, TPP measurements were more negative for Caucasian females than they were for Caucasian males. In the Chinese group, there were no significant differences between TPP in males and females using either of the two systems (Figure. 3.21).  73  Figure. 3.21. Mean and 0.95 confidence intervals (vertical bars) for TPP measurements made using the GSI and Reflwin systems for Caucasian and Chinese young adults between females and males.  3.2.4 Tympanometric width 3.2.4.1 226 Hz It was found that the main effects of gender [F (1,109) = 6.37, p < 0.05] and instrument [F (1,109) = 6.16, p < 0.05] were both significant. Furthermore, the interaction of gender and ethnicity [F (1,109) = 6.62, p < 0.05] was also significant, indicating that TW varies differently across genders between the two ethnicities. The three-way interaction of the variables ethnicity, gender and instrument was also found to be significant [F (1,109) = 6.16, p < 0.05], 74  indicating that TW varies differently across instruments between the two ethnicities and between the two genders. ANOVA Tables are shown in Appendix I. When all of data were pooled across ethnicity and instrument, it was found that TW was wider in males than it was in females (Figure. 3.22).  Figure. 3.22. Mean and 0.95 confidence intervals (vertical bars) for TW for female and male groups. The significant main effect of ethnicity revealed that Chinese young adults had wider TWs than Caucasian young adults, when data were pooled across gender and instrument (Figure. 3.23).  75  Figure. 3.23. Mean and 0.95 confidence intervals (vertical bars) for TW for Caucasian and Chinese groups.  The interaction of gender by ethnicity was investigated further, as shown in Figure 3.24. It was found that there were no differences in measures of TW between females and males for the Chinese group (Figure. 3.24). Meanwhile, in the Caucasian group, it was found that males had wider TWs than females (Figure. 3.24)  76  Figure. 3.24. Mean and 0.95 confidence intervals (vertical bars) for TW between Caucasian and Chinese young adults for the female and male groups.  There was also a significant interaction of the variables ethnicity, gender and instrument, which was investigated further as shown in Figure 3.25. For the GSI system, males from both ethnicities had wide TWs compared to females, though this relationship was much more prominent in the Caucasian group (Figure. 3.25). For measurements made using the Reflwin system, however, Chinese females had TWs that were slightly wider than those of Chinese males. Caucasian males still had wider TWs than Chinese males, and this relationship  77  was much more prominent for measurements made using the Reflwin system in comparison to measurements made using the GSI system (Figure. 3.25).  Figure. 3.25. Mean and 0.95 confidence intervals (vertical bars) for TW measurements made using the GSI and Reflwin systems for Caucasian and Chinese young adults between female and male genders.  3.3 Efficacy of Mimosa and Relwin in Identifying Otosclerosis 3.3.1 ANOVA and Post-hoc Analyses 3.3.1.1 Mimosa system In order to determine the frequencies at which both instruments could offer some predictive value in detecting the disorder of otosclerosis, a mixed-model ANOVA was 78  performed. In this model, patient condition (otosclerosis versus normal) served as a betweensubject factor, while frequency (15 frequencies) served as a within-subject factor. It was found that there was a significant interaction between the variables of frequency and patient condition [F (14, 1932) = 5.27 , p < 0.0001], thereby suggesting that ER values varied differently across frequencies between normal and otosclerotic ears. A post-hoc HSD analysis revealed that there were significant differences between normal and otosclerotic ears for ER measurements at 500 Hz, 630 Hz, and 800 Hz. Figure 3.26 illustrates that the ER for otosclerotic and normal subjects differed at the lower frequencies. ANOVA and post-hoc analysis Tables are shown in Appendix I.  Figure. 3.26. Condition by Frequency Interaction (Mimosa): ER at each of the fifteen frequencies for Mimosa normal ears (N = 60) and Mimosa otosclerotic ears (N = 28).  79  3.3.1.2 Reflwin system A second similar ANOVA was performed in which patient condition (otosclerosis versus normal) served as a between-subject factor, while frequency (15 frequencies) served as a within-subject factor. The normal hearing norms used for this analysis were the ones obtained using the Reflwin system. It was found that there was a significant interaction between the variables of frequency and patient condition [F (14, 1932) = 8.64, p < 0.0001], thereby suggesting that ER values varied differently across frequencies between normal and otosclerotic ears. A post-hoc Tukey’s HSD analysis revealed that there were significant differences between normal and otosclerotic ears for the frequencies of 315 Hz, 400 Hz, 500 Hz, 630 Hz, and 800 Hz. Figure 3.27 illustrates that ER varied differently for otosclerotic and normal ears at the lower frequencies. ANOVA and post-hoc analysis Tables are shown in Appendix I.  80  Figure. 3.27. Condition by Frequency Interaction (Reflwin): ER at each of the fifteen frequencies for Reflwin normal ears (N = 60) and Mimosa otosclerotic ears (N = 28).  3.3.2 ROC analyses ER at 500 Hz, ER at 630 Hz, and ER at 800 Hz using both systems were statistically compared with measures of ER at 500 Hz, ER at 630 Hz, and ER at 800 Hz from otosclerotic data obtained using the Mimosa system. As shown in Figure 3.28, an ROC analysis was used to determine the optimum measure of detection of ears afflicted with otosclerosis from normal ears. AUROC plots and corresponding 95% confidence intervals along with pair-wise comparison of AUROC between Mimosa ER measures and Reflwin ER measures are summarized 81  in Table 3.3. The AUROCs for all parameters with their corresponding 95% confidence intervals were above a level attributable to chance (Table 3.3a), indicating that they are able to distinguish between otosclerotic and normal ears. In other words, using any ER value at 500 Hz, 630 Hz, and 800 Hz to diagnose otosclerosis is better than chance. The pair-wise comparisons of AUROCs between the Mimosa and Reflwin systems at 500 Hz, 630 Hz, and 800 Hz did not reveal any statistically significant differences (Table 3.3b). This indicates that there is no advantage of using one instrument over the other at any of the three frequencies for distinguishing otosclerotic ears from normal ears. a, AUROC plots and 95% CIs  AUROC  Standard error  95% CIs  ER at 500 Hz Mimosa  0.875  0.0455  0.785 to 0.938  ER at 500 Hz Reflwin  0.873  0.0459  0.781 to 0.936  ER at 630 Hz Mimosa  0.813  0.0539  0.712 to 0.890  ER at 630 Hz Reflwin  0.838  0.509  0.740 to 0.909  ER at 800 Hz Mimosa  0.775  0.0578  0.670 to 0.860  ER at 800 Hz Reflwin  0.774  0.0579  0.669 to 0.859  b, Pair-wise comparison of AUROC  Difference between AUROCs  SE  95% CI  Z Statistic  P  ER at 500 Hz Mimosa vs. ER at 500 Hz Reflwin  0.0026  0.0428  -0.0812 to 0.0864  0.607  0.952  ER at 630 Hz Mimosa vs. ER at 630 Hz Reflwin  0.0247  0.048  -0.0693 to 0.119  0.514  0.607  ER at 800 Hz Mimosa vs. ER at 800 Hz Reflwin  0.0013  0.0537  -0.104 to 0.107  0.0242  0.981  Table. 3.3. Summary of AUROC plots and 95% CI along with pair-wise comparison of AUROC between Mimosa and Reflwin for ER at 500 Hz, ER at 630 Hz, and ER at 800 Hz. 82  83  Figure. 3.28. ROCs showing plots of the sensitivity versus specificity for the detection of otosclerosis using the ER norms for Reflwin and Mimosa systems at 500 Hz, 630 Hz, and 800 Hz.  3.3.3 Dot Plot Analysis Next, the data were presented in the form of dot plots, as shown in Figures 3.29, 3.30, 3.31, 3.32, 3.33, and 3.34in order to compare the sensitivity and specificity of the two instruments for a specific criterion value at a specific frequency. For ER values obtained at 500 Hz, it was found that when the criterion of ER > 0.80 was used, the sensitivity and specificity for the Mimosa system were 82.1% and 81.8% respectively (Figure. 3.29). For the Reflwin system, the sensitivity and specificity were 82.1% and 75% (Figure. 3.30). This suggests that both systems have the same sensitivity, but that test specificity is better for the Mimosa system for the same criterion. 84  A similar comparison of the dot plots was made for ER values obtained at 630 Hz. The dot plot analysis revealed that when the criterion levels were the same (ER at 630 Hz > 0.81), the Mimosa system had a sensitivity and specificity of 64.3% and 94.5% respectively (Figure. 3.31). The sensitivity and specificity were 64.3% and 98.2% for the Reflwin system (Figure. 3.32). Both systems were good at detecting normal ears (specificity), but were not very good at detecting ears with otoslerosis (sensitivity). For ER values obtained at 800 Hz, it was found that when the criterion was set at > 0.66, the sensitivity and specificity for the Mimosa system were 64.3% and 92.7% respectively (Figure. 3.33). For the Reflwin system, the sensitivity and specificity were 64.3% and 85.7% (Figure. 3.34). Both systems had similar sensitivities, but the Reflwin system was somewhat poorer at detecting normal ears (specificity).  Figure. 3.29. Dot plot showing the sensitivity and specificity for the Mimosa norm at detecting otosclerosis for the criterion ER > 0.80 at 500 Hz. 85  Figure. 3.30. Dot plot showing the sensitivity and specificity for the Reflwin norm at detecting otosclerosis for the criterion ER > 0.80 at 500 Hz.  Figure. 3.31. Dot plot showing the sensitivity and specificity for the Mimosa norm at detecting otosclerosis for the criterion ER > 0.79 at 630 Hz. 86  Figure. 3.32. Dot plot showing the sensitivity and specificity for the Reflwin norm at detecting otosclerosis for the criterion ER > 0.79 at 630 Hz.  Figure. 3.33. Dot plot showing the sensitivity and specificity for the Mimosa norm at detecting otosclerosis for the criterion ER > 0.66 at 800 Hz.  87  Figure. 3.34. Dot plot showing the sensitivity and specificity for the Reflwin norm at detecting otosclerosis for the criterion ER > 0.66 at 800 Hz.  88  Chapter 4 Discussion  This study evaluates the effects of gender, ethnicity and instrument on measures of WBER and single-frequency tympanometry at 226 Hz and 1000 Hz. The parameters of ER and PA are assessed, as well as the tympanometric parameters of Vea+, Vea-, Ytm+, Ytm-, TPP and TW. All these measurements were made in 60 normal-hearing subjects who were balanced according to ethnicity and gender. Furthermore, 28 participants with surgically-confirmed otosclerosis who were tested by Shahnaz and colleagues using the Mimosa system (Shahnaz et al., 2009), served as a diseased group for the purpose of assessing whether the Reflwin and Mimosa systems could be used accurately to detect otosclerosis. In order to effectively use ER in clinical practice, all sources of variability arising from instrument, gender, and ethnicity must be taken into account in order to determine whether the use of more specific norms results in significant improvements in the detection of otosclerosis. It has been shown that ethnic-specific norms result in improvements in the detection of otosclerosis (Shahnaz & Bork, 2006). The following study examines whether instrument-specific norms also result in improvements in the detection of otosclerosis. Furthermore, the sources of differences between instruments for WBER and tympanometric parameters were analyzed. The discussion is divided into three sections: (4.1) sources of differences for WBER measurements; (4.2) clinical significance of WBER variability between instruments, and (4.3) sources of differences for Tympanometric parameters. The first part (4.1) is divided into four different sub-sections in which the Mimosa system is compared to previous studies, the Reflwin 89  system is compared to previous studies, measurements from the two instruments for the current study are compared to each other, and ethnicity as a source of variability is discussed. The second part (4.2) discusses the clinical significance of the ER findings for each system and compares test sensitivity and specificity to previous studies. The final section (4.3) is divided into four subsections in which the results of the current study are discussed for each of the four tympanometric parameters: Vea, Ytm, TPP, and TW. For each parameter, the results are first discussed in relation to previous studies and then the sources of differences between the two instruments are addressed.  4.1 Sources of differences for WBER measurements The overall pattern of ER throughout all of the comparisons outlined in the results is similar to the general trend that has been obtained in a number of other studies (eg. Keefe et al., 1993; Margolis, Saly & Keefe, 1999; Voss & Allen, 1994). That is, the ER is highest at the lowest frequencies, reaches the minimum around 3000 Hz, and increases to moderate levels at the higher frequencies. Furthermore, the variability (standard deviation) increases as frequency increases.  4.1.1 Comparison of the Mimosa results to previous studies In order to determine the effects of instrumentation, Caucasian and Chinese data were pooled together before making comparisons to previous studies because most studies did not report the ethnic origins of their participants. First, the measures made at ambient pressure 90  using the Mimosa system were compared to the overall data obtained by Shahnaz and Bork (2006) who measured WBER using similar equipment and calibration technique. The hardware and software version for the current study (MEPA) was a newer version of the hardware and software used by Shahnaz and Bork (2006). Both studies were also balanced with regards to gender and ethnicity. As shown in Figure 4.1, results between the two systems were comparable, with some of the ER means being slightly lower at some of the lower frequencies. These slight differences could be explained by poorer test-retest reliability for the lower frequencies that arise due to sensitivity to environmental interference, as reported in previous studies (eg. Vander Werff, Prieve & Georgantas, 2007). Although the test-retest reliability is poorer in the lower frequencies, the variability between subjects is actually larger at the higher frequencies. This is clear in both the current study and in the results obtained by Shahnaz and Bork (2006), as shown by larger error bars at the higher frequencies (Figure. 4.1). Greater variability in ER values above 1000 Hz is a common finding in a number of studies (eg. Keefe & Ellen, 1996; Margolis, Saly & Keefe, 1999).  91  Figure. 4.1. Comparison of pooled ER data for frequencies obtained at ambient pressure using the Mimosa system for the current study in dots (N=60) and the data obtained by Shahnaz & Bork (2006) using the same instrument in squares (N=126). Both studies were balanced by gender and ethnicity.  The data obtained at ambient pressure for the Mimosa system in the current study were compared to previous studies by calculating the average percent difference in ER values across all of the fifteen frequencies, using the following formula: É  É  ˲ ŵŴŴ  (4.1)  Where A was the ER value from the current study, and B was the ER value from the comparison study. The average percent differences were also calculated across two major frequency bands (low frequency band < 1250 Hz, high frequency band > 1250 Hz) to determine whether there were substantial differences in the variability between studies at the low versus high frequencies. For the remainder of the discussion, the reported average percent differences will 92  be for the entire frequency range unless the difference in the variability between studies at the low versus high frequencies is greater than 5%. It was found that the results from the Mimosa system for the current study were comparable to the values obtained by Shahnaz and Bork (2006) and Voss et al. (1996), as the average percent differences for the current study and those studies were 3% and 11% respectively. The differences between the Mimosa system norms for the current study and those by Feeney et al. (2004), Keefe et al. (1993), and Sanford and Feeney (2008) were 34%, 16%, and 11% respectively. Given the fact that Shahnaz and Bork (2006) used a similar system to the Mimosa system from the current study, Voss et al. (1996) used an early version of the Mimosa system, and that Feeney et al. (2004) and Keefe et al. (1993) both used early versions of the Reflwin system; it is clear that instrumentation is a substantial source of variability between measurements. It is also worth noting that the deviation between the ER data from the current study and the data obtained by Feeney et al. (2004) and Sanford and Feeney (2008) were mostly due to the differences at the frequencies above 1250 Hz (Tables A24 & A25). Although both instruments rely on determining Ps and Zs, the ways in which they make these calculations are fundamentally different because the Mimosa system calculates the sound pressure within calibration cavities, while the Reflwin system calculates the reflection profiles within calibration tubes. Furthermore, the calculation of ER depends on the crosssectional area of the ear canal. Huang et al. (2000) showed that accurate WBR measurements require that the Thevenin equivalents of the acoustic measurement system be determined with loads that have diameters within 10-15% of the actual ear canal diameter. The Mimosa Acoustics System used in this study estimates the cross-sectional area of the ear canal based on 93  the probe tip diameter and the calibration cavity used during calibrations. With this system, the cross-sectional area is either estimated to be 4.5 mm (rubber-tip cavity) or 7.5 mm (foam-tip cavity). Numerical simulations that explore the effects of variations in ear canal cross-sectional area show that increasing the cross-sectional area increases ER at most frequencies (Huang et al., 2000).  Figure. 4.2. Comparison of ER data between the current study (N = 60; 18 – 38 yr.), Shahnaz & Bork (2006) (N = 128; 18 – 32 yr.), and Voss et al. (1994) (N = 10; 18 – 24). The current study and Shahnaz & Bork (2006) use the same version of the Mimosa system, while the study by Voss et al. (1994) uses an older version of the Mimosa system.  A graphical representation of the relationship between the ER results obtained from the current study and those obtained from the study by Shahnaz and Bork (2006) and Voss et al. (1994) reveals another important difference. The raw data for these comparisons are found in 94  Appendix I. As shown in Figure 4.2, the study by Voss et al. (1994) reveals a double minima pattern (two lowest ER values) rather than a single minimum (one lowest ER value). The double minima pattern may have arisen in the study by Voss et al. (1996) if all participants were Caucasian, though ethnicities were not reported for that study. Both the current study and the results obtained by Shahnaz and Bork (2006) suggest that Caucasian young adults tend to have double minima patterns, but that a single minimum pattern emerges once the norms are pooled with the Chinese young adults. This finding suggests that the mechano-acoustical properties of the Chinese middle ear are different from the Caucasian middle ear, which results in a different level of energy absorption at different frequencies and may also result in different ER patterns.  4.1.2 Comparison of the Reflwin results to previous studies For the Reflwin system, results were obtained at static and dynamic pressures and both of these results are compared to the results from previous studies. The results obtained from the Reflwin system for the current study are compared to those obtained by Sanford and Feeney (2008), as both studies used the same software (Eclipse – v.1) and calibration technique.  95  Figure. 4.3. Comparison of pooled ER data for frequencies obtained at ambient pressure using the Reflwin system for the current study in dots (N=60) and the data obtained by Sandford and Feeney (2008) using the a similar instrument in squares (N=20).  The consistency between the data obtained in the current study and the data obtained by Sandford and Feeney (2008) was not as great as the consistency between the Mimosa system for the current study and Shahnaz and Bork (2006) (comparing Figures 4.2 and 4.3). The variability between the Reflwin data for the current study and the data from Sanford and Feeney (2008) at the frequencies above 1250 Hz (22%) was substantially larger than the variability for the frequencies of 1250 Hz and below (9%). One source of difference between the results of the two studies could be the fact that the sample size in the study by Sanford and Feeney (2008) was substantially smaller (N = 60 vs. N = 20). Furthermore, the norms from the current study were balanced by gender and ethnicity, while the adults in the study by Sanford and Feeney (2008) were not balanced by gender and ethnicities were not reported. 96  Reflwin ambient data were also compared to other studies. The raw data for these studies are included Appendix I. The average percent difference for ER values obtained using the Reflwin system at static pressure for the current study was compared to the results from Voss et al. (1996) and Shahnaz and Bork (2006), which both used a calibration technique that was similar to the Mimosa system. The average percent differences were 11% and 8% respectively. The average percent difference between the Reflwin static results for the current study and the studies by Keefe et al. (1993) and Feeney et al. (2004) were 18% and 35% respectively. Most of the variability between the Reflwin system for the current study and the data from Keefe et al. (1993) was due to differences at the frequencies at 1250 Hz and below (Tables A24 and A25). Figure 4.4 illustrates that the studies that used the same calibration technique as the Reflwin system (or an earlier version of the Reflwin system) did not have very similar ER patterns (Figure. 4.4).  97  Figure. 4.4. Comparison of ER data between the current study (N = 60; 18 – 38 yr.), Sanford & Feeney (2008) (N = 20; 22 – 30 yr.), Feeney et al. (2004) (N = 40; 18 – 28 yr.), and Keefe et al. (1993) (N = 10; 20 – 50 yr.). The current study and Sanford & Feeney (2008) used the same version of the Reflwin system, while the studies by Feeney et al. (2004) and Keefe et al. (2004) used an older version of the Reflwin system.  These findings are different from what one would expect because the static ER results using Reflwin from the current study are more similar to the results from studies that used the calibration technique that is employed by the Mimosa system. This illustrates that other sources of differences may be more important than the calibration technique used in each system. For instance, the ER results from the current study may be dissimilar to those obtained by Keefe et al. (1993) because they only tested 10 participants ranging from 20 to 50 years in age and previous studies have shown that ER is affected by age (eg. Sanford & Feeney, 2008; Keefe & Bulen, 1993; and Feeney et al., 2004). Although the calibration technique used by Feeney et al. (2004) relied on similar principles as the one in the current study, they used six 98  calibration tubes while the current study only used two. Furthermore, Feeney et al. (2004) had a smaller sample size and ethnicities were not reported.  4.1.3 Comparison between the Mimosa and Reflwin system in the current study ER measurements made using the Reflwin system for ambient pressures were significantly lower at frequencies below 2000 Hz than those obtained using the Mimosa system (Figure. 3.3). These differences arise because the calibration procedures used for the Reflwin and Mimosa systems are fundamentally different in that they rely on measures of the waveform characteristics and pressure-response data respectively. During the calibration procedure for the Reflwin system, the waveform characteristics at each frequency are measured for a small tube and a large tube so that the waveform characteristics of both tubes can be compared for the purpose of separating the waveform of the incident sound from the reflected sound (Keefe & Simmons, 2003). This procedure may be less accurate for lower frequencies because there are multiple reflections within the tube due to greater wavelengths for lower pitches. This possibility is supported by the fact that the differences between ER measurements between the two systems become less pronounced as frequency increases (Table 3.1). Another source of difference between the Reflwin and the Mimosa system for the current study is the way in which both systems estimate the cross-sectional area of the ear canal. The Reflwin system uses a rubber probe tip and makes an estimation that is similar to the one used by Feeney et al. (2008), in which the cross-sectional area of the ear canal is 99  estimated using a set value based on the calibration tube, which is an approximation of the diameter of the human ear canal. Meanwhile, the Mimosa system makes an acoustic estimate of the Vea that is similar to the one outlined by Voss and Allen (1994). The Reflwin system will have more errors if the subject’s ear canal differs substantially from the norm. Meanwhile, the Mimosa system will have more errors if the corrections for the assumptions made during the acoustic estimate are inaccurate for the subject’s ear canal. Sanford and Feeney (2008) compared ER results in 12 week and 14 week-old infants and adults to previous studies and they also reported that the most significant differences between ER results occur at frequencies below 2000 Hz and that these differences are partially due to estimates in ear canal crosssectional area. For frequencies of 2000 Hz and below, ER measurements were substantially lower for measurements made at dynamic pressures versus static pressure when using the Reflwin system. ER measurements made at dynamic pressures may be lower than those obtained at static pressure because static ER measurements assume that the maximum compliance occurs at 0 daPa. In reality, however, the average TPP for the current study was negative regardless of whether measurements were made using the Reflwin system or the GSI Tympstar (Figure. 3.20). Negative middle-ear pressure results in a stiffer tympanic membrane (Beers, Shahanaz, Kozak & Westerberg, in press) which results in greater reflectance at the low frequencies where stiffness makes a significant contribution to impedance (Allen et al., 2005). As a result, measurements made using the Reflwin system in the dynamic mode should be representative of the true middle-ear pressure, resulting in lower ER and higher PA at low frequencies. Another factor that must be considered is the effect of changing the pressure within the ear 100  canal. Gaihede (1996) showed that changing the pressure within the ear canal decreases the stiffness of the tympanic membrane and increases the absorption of energy. Thus, the inaccurate estimation of TPP and the effects of changing the pressure within the ear canal could account for the observed differences between static and dynamic conditions. Another finding in the current study was the significant three-way interaction of the variables gender, instrument, and frequency. It was found that ER measurements at static pressure were the same between the two systems for females but they were lower for males at the low frequencies using the Reflwin system (Figure. 3.2). The Reflwin system may detect lower ER values for males because the plastic tip used for the calibration procedure is often different from the one used during the measurement of ER. In contrast, the Mimosa system makes use of foam tips so that the tip used during the calibration can also be used for the measurement of ER.  4.1.4 Ethnicity as a source of variation in ER The effects of ethnicity (Caucasian versus Chinese) on WBER have previously been analyzed by Shahnaz and Bork (2006). Similar trends exist between that study and the current study. For instance, ANOVA analyses in both studies were statistically significant for ethnicity by frequency interactions (Figure. 3.4; Shahnaz & Bork, 2006). Furthermore, both studies found that the ER profiles for Chinese young adults were larger at the low frequencies and that Caucasian young adults had larger ER profiles at the high frequencies (Figure. 3.1; Shahnaz & Bork, 2006). In both studies, it was found that the Caucasian young adults had ER minima at 101  around 1250 Hz and 3150 Hz, while Chinese young adults had a single minimum at 3150 Hz (Figure. 3.1; Shahnaz & Bork, 2006). Another trend was also noted in the interaction between pressurization method (static versus dynamic) and ethnicity (Caucasian versus Chinese) (Figure. 3.6). It was found that while measurements made using dynamic pressures resulted in greater PA, these differences were more distinct for the Caucasian group (Figure. 3.6). Since Caucasians tend to have body sizes that are larger and more variable (Bell et al., 2002), their WBER measurements are also more likely to be affected by any sources of error inherent in the assumptions made during ER measurements. In particular, estimations of the cross-sectional area of the ear canal are more likely to be a source of variability in ER measurements because body size is more variable in Caucasians. Furthermore, since TPP is more variable in the Caucasian group (Figure. 3.21), it follows that ER measurements using the dynamic pressurization technique could potentially result in bigger differences in the Caucasian group because static measurements are always at ambient pressure regardless of the middle-ear pressure.  4.2 Clinical significance of WBER variability between instruments As outlined in the Introduction, previous studies have shown that test sensitivity and specificity may be altered by using ethnic-specific norms, instrument-specific norms, or the use of different parameters (Shahnaz & Davies, 2006; Shahnaz et al., 2009; Shahnaz & Bork, 2006; Shahnaz & Bork, 2006). In order to determine whether one of the two instruments was better at detecting otosclerosis, an area under the ROC was statistically compared between the two 102  systems. In this analysis, the areas under each curve were compared between instruments for each frequency that was shown to be statistically different between normal and otosclerostic ears. It was found that there were no significant differences between the AUROCs for each of these comparisons, suggesting that both instruments perform equally well at detecting otosclerosis (Table 3.3). Thus, the results of the current study suggest that test sensitivity and specificity are not improved as a result of using instrument-specific norms. This relationship was explored further by comparing the sensitivity and specificity between the two instruments for each of the three frequencies. It was found that setting the criterion for ER at 500 Hz resulted in high sensitivity and specificity (Figures. 3.30 and 3.31). When the criteria levels were set for ER at 630 Hz and ER at 800 Hz, the test specificity improved at the expense of sensitivity. This suggests that if the cost of misses is greater than the cost of false alarms, it may be best to use the criterion of 500 Hz. If the cost of false alarms is greater than the cost of misses, then it may be best to use the criteria of ER at 630 Hz and 800 Hz. Thus, sensitivity and specificity can be adjusted by using different parameters (ER at 500 Hz vs. ER at 630 Hz and ER at 800 Hz) for the detection of otosclerosis. Dot plots depicting test sensitivity and specificity for pooled norms revealed that test sensitivity was consistently high (above 85%) and the test specificity was consistently low (between 60% and 70%) when pooled norms were used. These plots were not included in the Results because Shahnaz and Bork (2006) already found that ethnic-specific norms are warranted when using ER to detect otosclerosis. The findings from the current study lend further support to the use of ethnic-specific norms when using ER to detect otosclerosis.  103  4.3 Tympanometric parameters The descriptive statistics of the current study were compared to those obtained from previous studies that examined the effects of ethnicity and gender on the same tympanometric parameters as shown in Table 4.1. Comparisons for measures of Ytm and Vea were only included for the positive tail because this was the approach used in previous studies due to less variability in measurements obtained at the positive tail. There are clear differences between the results of the current study and those of previous studies. For instance, Ytm is substantially larger for females compared to males in the current study, while the opposite is true in previous studies (Table 4.1). Furthermore, overall Ytm values are noticeably smaller for males for the current study in comparison to other studies, and Ytm values for females are noticeably larger in the current study compared to other studies (Table 4.1). Measurements of TPP using the GSI system are comparable to previous studies but TPP measurements using the Reflwin system are substantially lower (Table 4.1). Given the large variability in results for TW, there are no obvious differences between the results of the current study and those of previous studies (Table 4.1). The rest of this discussion, however, shall focus on the statistical analyses that were presented in the Results and will attempt to explain these results in relation to previous studies and the underlying methods and assumptions made using both instruments.  104  GSI Current study  M (n=30) F (n=30) Overall (n=60)  Reflwin Current study  M (n=30) F (n=30) Overall (n=60)  GSI Shahnaz & Davies (2006)  M (n=34) F (n=42) Overall (n=76)  GSI Shahnaz & Bork (2006)  M (n=27) F (n=26) Overall (n=53)  Mean SD 90% Range Mean SD 90% Range Mean SD 90% Range Mean SD 90% Range Mean SD 90% Range Mean SD 90% Range Mean SD 90% Range Mean SD 90% Range Mean SD 90% Range  Ytm (mmho) C 0.56 0.35 0.30 – 0.74 0.98 0.58 0.46 – 2.19 0.90 0.44 0.48 – 1.44 0.41 0.14 0.24 – 0.66 0.80 0.44 0.34 – 1.51 0.73 0.36 0.35 – 1.48 0.79 0.33 0.30 – 1.38 0.75 0.41 0.30 – 1.70 0.77 0.37 0.30 – 1.60  Mean SD 90% Range Mean SD 90% Range Mean SD 90% Range  0.80 0.28 0.30 – 1.30 0.66 0.24 0.30 – 1.20 0.73 0.27 0.30 – 1.20  0.63 0.40 0.29 – 1.29 0.82 0.19 0.53 – 1.09 0.59 0.38 0.29 – 1.06 0.47 0.22 0.25 – 0.94 0.66 0.23 0.4 – 1 0.44 0.18 0.24 – 0.74 0.58 0.35 0.20 – 1.31 0.40 0.24 0.17 – 1.10 0.47 0.29 0.20 – 1.10  TW (daPa) C 84.4 29.4 55 – 150 66.9 24.8 40 – 116.8 73.6 25.5 40 – 129.8 92.66 42.06 32.8 – 164 71.6 20.2 45.5 – 109.5 85.4 29.3 55 – 149 93.1 22.3 54.5 – 131.6 97.1 30.3 51 – 149 95.3 26.9 51.7 – 136.3  0.67 0.29 0.30 – 1.20 0.37 0.20 0.20 – 0.70 0.51 0.29 0.20 – 1.10  79 18 55 – 110 92 27 60 – 135 85 24 60 – 115  A  91.61 34.72 47 – 140 80.83 24.57 47 – 130 87.8 31.9 50 – 150 88.7 27.62 47 – 134 100.2 30.6 65 – 161 90.9 35.9 35 – 156 104.1 28.5 51 – 160 112.3 29.2 66 – 157 109.2 29.1 66 – 158  TPP (daPa) C -4.5 9.1 -17.5 – 7 -3.9 16.8 -33 – 12 -2.46 15.49 -20 – 15 -27.5 9.88 -46 - -14 -20.7 12.91 -36 - -2.8 -24.2 13.74 -47.5 – 5 -3.37 11.5 -22 – 14 -6.5 17.9 -39 – 14 -5.1 15.4 -23 – 14  107 72 40 – 290 128 70 70 – 225 118 61 50 – 265  0.63 5.58 -10 – 5 0.65 9.21 -25 – 5 0.64 7.49 -10 – 5  A  -6.96 8.54 -17.5 – 0 -0.92 14.1 -18 – 18 -5.7 8.9 -17.5 – 5 -28.43 11.85 -47 - -8 -27.92 13.84 -50 - -6.13 -5.7 8.9 -17.5 – 5 0.73 15.3 -38 – 28 -1.41 12.7 -18 – 14 -0.6 13.8 -18 – 19  ECV (mmho) C 1.41 0.39 0.83 – 1.87 1.25 0.32 0.82 – 1.85 1.33 0.36 0.82 – 1.88 1.34 0.31 0.87 – 1.78 1.25 0.26 0.91 – 1.73 1.29 0.29 0.89 – 1.81 1.30 0.37 0.72 – 2.06 1.05 0.27 0.67 – 1.60 1.16 0.34 0.70 – 1.80  1.25 0.25 0.99 – 1.58 1.09 0.23 0.80 – 1.44 1.17 0.25 0.80 – 1.57 1.39 0.27 1.05 – 1.83 1.13 0.19 0.88 – 1.45 1.25 0.26 0.89 – 1.68 1.10 0.33 0.60 – 1.80 0.98 0.29 0.60 – 1.50 1.02 0.32 0.60 – 1.60  -5 13.8 -35 – 5 -4.04 7.5 -15 – 5 -4.5 10.9 -20 – 5  1.06 0.25 0.70 – 1.60 1.28 0.22 1.0 – 1.7 1.37 0.32 1.0 – 1.9  1.32 0.25 1.0- 1.7 1.06 0.25 0.7 – 1.6 1.18 0.28 0.7 – 1.6  A  A  Table 4.1. Descriptive statistics for Ytm, TW, TPP and Vea obtained using both GSI and Reflwin systems. Ytm and Vea were calculated using positive compensation. Some other published normative data studies are also included for comparison. C= Caucasian; A=Chinese; M=male; F=female.  105  GSI Wan & Wong (2002)  M (n=50) F (n=50) Overall (n=100)  GSI Roup et al. (1998)  M (n=51) F (n=51) Overall (n=102)  Mean SD 90% Range Mean SD 90% Range Mean SD 90% Range Mean SD 90% Range Mean SD 90% Range Mean SD 90% Range  N/A  N/A  N/A  0.87 0.46 0.30 – 1.80 0.58 0.27 0.30 – 1.12 0.72 0.40 0.30 – 1.19  0.58 0.29 0.30 – 1.10 0.52 0.28 0.20 – 1.10 0.55 0.28 0.20 – 1.10 N/A  N/A  N/A  N/A  N/A  N/A  59.8 17.3 35 – 87 73.9 17.2 45 – 107 66.9 18.6 32.8 – 95  88.3 34.1 45 – 175 94.2 29.2 45 – 145 91.2 31.8 45 – 159 N/A  N/A  N/A  N/A  N/A  N/A  -26.2 31.66 -110 – 9 -27.8 23.5 -80 – 3 -29.96 27.76 -104 – 4  4.8 20.73 -24.5 – 29.7 3.10 15.8 -19.8 – 24.7 3.95 18.41 -19.8 – 25.0 N/A  N/A  N/A  N/A  N/A  N/A  1.40 0.32 1.00 – 2.01 1.18 0.22 0.80 – 1.60 1.29 0.29 0.90 – 1.80  1.22 0.25 0.81 – 1.70 1.13 0.31 0.70 – 1.60 1.17 0.28 0.80 – 1.60 N/A  N/A  N/A  Table 4.1. Descriptive statistics for Ytm, TW, TPP and Vea obtained using both GSI and Reflwin systems. Ytm and Vea were calculated using positive compensation. Some other published normative data studies are also included for comparison. C= Caucasian; A=Chinese; M=male; F=female.  106  4.3.1 Ear canal volume 4.3.1.1 226 Hz The results for measurements of Vea revealed some findings that were consistent with previous studies and other findings that differed from previous studies. In the current study, it was found that Vea was larger in males than it was in females (Figure. 3.7). This finding is consistent with previous studies (eg. Davies & Shahnaz, 2006; Wan & Wong, 2002; Roup, et al., 1998) in which it was hypothesized that larger Vea values are found in males who have a larger body size and that a correlation exists between Vea and body size. Analysis of the instrument by ethnicity interaction, however, revealed an interesting trend. It was found that there was a significant difference in Veas between Caucasian and Chinese for measurements made using the GSI Tympstar but not for measurements made using the Reflwin system (Figure. 3.8). It is also worth noting that the results obtained using the GSI system were more variable than those obtained using the Reflwin sytem (Figure. 3.8). The larger variability for measurements obtained using the GSI system s can be explained by the fact that the GSI system derives tympanograms from admittance and estimates of Vea are shown to be affected by the distance of the probe tip from the tympanic membrane (Shanks & Lily, 1981). Meanwhile, the Reflwin system derives tympanograms from measurements of PA, which have been shown to be unaffected by the depth of the probe tip insertion (Stinson et al., 1992; Voss & Allen, 1994). The variability in Veas between ethnicities for the GSI system is difficult to explain; however this finding might suggest that the GSI system gives a more  107  accurate estimate of the actual Vea since Caucasian ear canals are likely to be larger than Chinese ear canals. The gender by tail interaction revealed that the variability was much larger for measurements made at the negative tail in comparison to those made at the positive tail (Figure. 3.9). This is consistent with previous researchers who have shown that measures made at the positive tail are much less variable than those made at the negative tail (Margolis & Goycoolea, 1993). One unusual finding in this study, however, was the fact that measurements made at the negative tail were significantly larger than those obtained at the positive tail for males when the two instruments were pooled together (Figure. 3.9). This is inconsistent with a number of studies in which Vea- is consistently smaller than Vea+ (Margolis & Goycoolea, 1993; Margolis & Shanks, 1991; Shanks & Lily, 1981). This difference may have arisen, however, because measurements made using the Reflwin system go from dynamic pressures of +200 daPa to -300 daPa, while the GSI system measures at dynamic pressures of +200 daPa to -400 daPa. As a result, the negative tail values for the current study differ from previous studies in which the admittance obtained at –400 daPa was taken to be the negative tail.  4.3.1.2 1000 Hz For measurements obtained at 1000 Hz, males also had significantly larger Veas compared to females (Figure. 3.10) when the two instruments were pooled together. At 1000 Hz, however, there were also significant main effects for instrument (Figure. 3.11) and tails (Figure. 3.12), which did not exist for the probe-tone frequency of 226 Hz. The Vea estimate was larger for the Reflwin system and Vea- was smaller than Vea+. 108  The Vea estimate using the Reflwin system may have been larger than the estimate using the GSI Tympstar because the Reflwin system extracts single-frequency impedance data from plots of PA, frequency, and pressure before making conversions to admittance data. Meanwhile, the GSI Tympstar makes a direct measure of admittance. As such, the conversion of PA data into impedance data may result in larger estimates of Vea because PA does not change as a function of depth insertion (Keefe & Levi, 1996), while tympanometry does change as a result of depth of insertion (Shanks & Lilly, 1981). The significant main effect for tails was consistent with previous studies in which Veawas smaller than positive tail Vea+ (Margolis & Goycoolea, 1993; Margolis & Shanks, 1991; Shanks & Lily, 1981). It is also logical for this effect to be present at 1000 Hz and absent at 226 Hz because the differences between positive and negative tail are more pronounced at 1000 Hz (Calandruccio et al., 2006). Thus, Vea- would not need to be measured at lower pressures of 400 daPa for the negative tail to have reached a significantly negative value. The interaction of the effects of tail and gender revealed that while Vea+ and Vea- are both greater in males than in females, the nature of this relationship is more pronounced for the positive tail in comparison to the negative tail (Figure. 3.13). Once again, this may be a reflection of the fact that a negative tail pressure of –300 daPa was used. Furthermore, the larger variability for measurements obtained at the negative tail could also cause the differences between genders to be less pronounced.  109  4.3.2  Static admittance  4.3.2.1 226 Hz The current study showed that measures of Ytm were significantly larger for Caucasian young adults compared to Chinese young adults for probe-tone frequencies of 226 Hz and 1000 Hz (Figures. 3.14 & 3.18), which is consistent with previous studies (eg. Shahnaz & Bork, 2006; Shahnaz & Davies, 2008). Admittance values were found to be larger and more variable using the GSI Tympstar system in comparison to measurements made using the Reflwin system (Figure. 3.15). This finding is consistent with the fact that measures of Vea were also found to be less variable using the Reflwin system (Figure. 3.8), thereby resulting in less variability for measures of Ytm, which depend on Vea. Further, analysis of the gender by ethnicity interaction (Figure. 3.16) revealed that there were no substantial differences between Chinese males and Chinese females but that Ytm for Caucasian females was substantially larger than for Caucasian males. As a result, body mass indices obtained by Bell et al. (2002) were examined to determine whether the variability in female body sizes could account for these observed differences. It was found that the standard deviations for body mass indices for Caucasian females, Caucasian males, Chinese females and Chinese males were 6.4, 4.8, 3.2, and 2.8 respectively (Bell et al., 2002). This suggests that the overall variability in Caucasian female body sizes could account for the fact that Ytm for females was substantially larger than it was for males in the current study. Meanwhile, previous studies showed an opposite trend in which Caucasian females had larger Ytms than Caucasian males (e.g. Shahnaz & Davies, 2006; Shahnaz & Bork, 2006; Roup et al., 1998). It is possible that the males in the current study had unusually small body sizes, while the females had unusually large body sizes, as there were no controls on body size indices for the current study. 110  The gender by tail interaction revealed that there were no gender differences for measures of Ytm at the positive tail but that there were differences between genders for measurements made at the negative tail (Figure. 3.17). This finding suggests that the discrepancy between gender and Ytm can be explained by measurements obtained from the negative tail, which was taken to be -300 daPa rather than -400 daPa. Thus, it could be that males require measurements that are made at more negative pressures before the minimum tail value can be reached.  4.3.2.2 1000 Hz Caucasian females were found to have larger overall Ytms compared to Chinese females when the data were pooled across instruments (Figure. 3.19). This finding is consistent with previous studies in which Chinese young adults were found to have lower Ytms compared to Caucasian young adults (Shahnaz & Bork, 2006; Shahnaz & Davies, 2008). For males, however, Ytm was found to be slightly higher for the Chinese group than for the Caucasian group (Figure. 3.19). This finding is somewhat difficult to explain; however, this trend could have arisen if measurements in Caucasian males require more negative pressures to obtain their most negative impedance values.  4.3.3  Tympanometric peak pressure TPP values obtained using the GSI system were more positive than those obtained using  the Reflwin system. Neither system had high variability (Figure 3.20). A three-way interaction 111  of the variables ethnicity, gender, and instrument also revealed that there were no significant differences between Chinese males and females for either instrument, but that females in the Caucasian group had more positive TPPs compared to males for measurements made using the Reflwin system (Figure. 3.21). This finding is consistent with the idea that more variability may exist within the Caucasian group because of more variability in Caucasian body sizes. The results from the current study are consistent with those conducted by Shahnaz and Bork (2006) because there were no main effects of gender on TPP values in that study, nor in the current study. Meanwhile, contrary to the current study, Shahnaz and Davies (2006) found that TPP varied differently according to ethnicity and that Chinese young adults had more positive TPPs. These inconsistencies and the lack of consensus between studies suggest there are no true effects of ethnicity on measures of TPP.  4.3.4 Tympanometric Width Measurements of TW were perhaps the most consistent with those from previous studies. In the current study, males had larger TWs than females and Chinese young adults had larger TWs than Caucasian young adults. Both of these findings are consistent with Shahnaz and Davies (2006) and Shahnaz and Bork (2008). Analysis of a gender by ethnicity interaction revealed that TWs were substantially wider for Caucasian males compared to females, but that there were no significant differences between genders for the Chinese group (Figure. 3.24). This finding suggests that there is less variability in TWsfor the Chinese group. This may be  112  attributed to the variability in the body sizes of Caucasian females as discussed previously (Bell et al., 2002). It is also worth noting that TW is derived from Ytm, which in turn is derived from Veaand Vea+. As such, all the sources of variability that affect these measurements could potentially effect measurements of TW. Thus, it may be surprising to note that there were no significant differences between instruments for measures of TW but that the main effect of instrument was significant for the measure of Ytm and Vea at 1000 Hz. This is most likely a result of the fact that measures of TW are so variable to begin with, thereby effectively masking any significant main effects of instrument.  113  Chapter 5 Conclusions 5.1 Summary The current study was designed to answer five primary questions: (1) To determine whether the Mimosa Acoustics and Reflwin Interacoustics systems yield comparable results in terms of ER and PA ; (2) To compare the static versus dynamic mode in the Interacoustics system; (3) To determine whether both systems are capable of detecting differences between reflectance profiles that exist between Chinese and Caucasian young adults; (4) To determine whether the Reflwin system and the GSI Tympstar are both capable of making similar assessments of the middle ear; and (5) To determine if the Reflwin and Mimosa systems are capable of detecting the disorder of otosclerosis. In order to address purpose 1, the current study compared ER and PA from the Mimosa and Reflwin systems. There were no significant differences in ER values between the two instruments for frequencies above 630 Hz. For frequencies below 630 Hz, however, the ER values were significantly lower (PA significantly higher) for the Reflwin Interacoustics system (Figure 3.2). These differences were attributed to the calibration procedures used for the two instruments and the way in which each instrument estimates the cross-sectional area of the ear canal. When using the Reflwin Interacoustics system, it is also possible to obtain ER data and tympanometric data simultaneously by making measurements for the Reflwin system at dynamic pressures. The current study addressed purpose 2 by comparing ER measurements made at static and dynamic pressures to determine whether ER values obtained during WBER 114  tympanometry could be reliably compared to those obtained using static pressure. It was found that ER values were even lower for the Reflwin system when dynamic pressures were used and that these differences were significant for frequencies of 1600 Hz and lower (Figure 3.5). These differences are attributed to the fact that static ER always measures ER at ambient pressure and some subjects may have their maximum compliance at more negative pressures. A number of studies have shown that physical differences exist between the middle ears of Chinese and Caucasian young adults (Dreisbach et al., 2007; Shahnaz & Davies, 2006; Wan & Wong, 2002; Roup et al., 1998; Shahnaz & Bork, 2006). The ability for both instruments to assess impedance was assessed by determining whether both instruments could detect differences that exist between the ER profiles of Caucasian and Chinese young adults. It was found that the trends in ER were similar to those from previous studies in that Chinese young adults had greater reflectance at the lower frequencies but that Caucasian young adults had greater reflectance at the higher frequencies (Figure 3.4). Furthermore, in the current study there was a single minimum ER pattern for Chinese and a double minima pattern for Caucasians, which is also consistent with previous research (Figure 3.4; Shahnaz & Bork, 2006). These findings indicate that both instruments are capable of detecting the differences in impedance that exist between the two ethnicities. This is further supported by the fact that none of the ANOVA analyses revealed any differences between the two instruments. Another goal of this study was to determine whether the Reflwin and GSI system both yield comparable tympanometric results. A number of differences were found to exist between the two systems for the parameters of Ytm, TPP and Vea. Ytm was larger and more variable for the GSI Tympstar (Figure 3.15), TPP was more negative for the Reflwin system compared to the 115  GSI Tympstar (Figure 3.20), and Vea was larger for the Reflwin system for the probe-tone frequency of 1000 Hz (Figure 3.11). Other differences also existed between the two systems as a function of gender and ethnicity. No significant differences between instruments were found for the parameter of TW, which could be due to the fact that measures of TW are highly variable. The differences between the two instruments can be attributed to the fact that the GSI Tympstar makes a direct measure of admittance. Meanwhile, the Reflwin Interacoustics system measures PA across different frequencies, which must then be extracted for the generation of a single-frequency tympanogram. Once the PA data at a single frequency are extracted, the PA values must be converted to admittance values for each pressurization point. In order to assess the clinical significance of the differences between the Reflwin and Mimosa system, the ER norms obtained for the current study were compared to those obtained from otosclerotic ears for both systems. It was found that ER at 500 Hz, 630 Hz, and 800 Hz all provided some statistically significant predictive value in the detection of otosclerosis for both systems. The two instruments are very comparable in their ability to detect otosclerosis (Table 3.3) and system-specific norms are not warranted for the detection of otosclerosis. The test sensitivity was poor for both instruments for the frequencies of 630 Hz and 800 Hz. At 500 Hz, however, test sensitivity improved. Since audiologists use a test battery approach, it is more desirable to have better sensitivity because it is unlikely for those who are falsely labelled as having otosclerosis to have a conductive hearing loss.  116  5.2 Limitations of the current study and areas for future research It was found that the ER values for females did not differ between the two instruments but that they did differ for males (Figure. 3.2). Since otosclerosis is twice as common in females as males (Ginsberg & White, 1985), the current study may not have detected any differences in each system’s ability to detect otosclerosis because the differences between the two instruments occur for males rather than females. In order to confirm that instrument-specific norms are not warranted, future studies should compare the ER results to norms from pathologies that are equally prominent in males and females. Furthermore, it is important to consider the instrument on which the otosclerotic data were obtained in order to ensure that instrument-specific norms are not warranted. Thus, future studies could analyze test performance making the following comparisons: Mimosa norms compared to otosclerotic norms obtained using the Mimosa system, Mimosa norms compared to otosclerotic norms obtained using the Reflwin system, Reflwin norms compared to otosclerotic norms obtained using the Reflwin system, and Reflwin norms compared to otosclerotic norms obtained using the Mimosa system. The use of ethnic-specific norms would be useful in improving ER’s sensitivity in detecting otosclerosis and a more rigorous analysis of instruments could help to confirm whether norms from either instrument can be used interchangeably. In the current study it was also found that greater differences exist between the Reflwin and Mimosa system when dynamic pressurization is used for the Reflwin system. It may be worth investigating whether the dynamic pressurization technique results in better test performance than the static technique for measures of ER using the Reflwin system. If test 117  performance was better or the same, it may be possible for clinicians to make a single measurement of WBER tympanometry to obtain standard tympanograms and ER values simultaneously. Another limitation of the current study was the fact that the clinical significance of the differences between the tympanometric norms obtained using the Reflwin system and the GSI Tympstar were not assessed. Future studies could perform a similar ROC analysis to determine whether instrument-specific norms are warranted for the detection of middle-ear pathologies.  5.3 General conclusions Overall, the findings from the current study suggest that for measures of ER, there are significant differences between the Reflwin and Mimosa systems at low frequencies. Furthermore, the tympanometric parameters of Vea, TPP, and Ytm also differ between the Reflwin system and the GSI Tympstar. The differences in ER between the Reflwin and Mimosa systems are not clinically significant and the use of instrument-specific norms does not result in improved test performance for the detection of otosclerosis. The clinical significance of the differences between the two instruments for tympanometric parameters using the Reflwin system and the GSI Tympstar requires further investigation.  118  References Allen, J., Jeng, P., and Levitt, H. (2005). Evaluation of human middle ear function via an acoustic power assessment. Journal of Rehabiliation Research and Development, 42 (4), 63-78. Bell, A., Adair, L. and Popkin, B. (2002). Ethnic differences between body mass index and hypertension. American Journal of Epidemiology, 155, 346-353. Browning, G., Swan, I., and Gatehouse, S. (1985). The doubtful value of tympanometry in the diagnosis of otosclerosis. The Journal of Laryngology and Otology, 99: 545-547 Calandruccio, L., Fitzgerald, T., and Prieve, A. (2006) Normative multifrequency tympanometry in infants and toddlers. Journal of the American Academy of Audiology, 17:470–480. Cambell, D., and Stanley, J. (1963). Experimental and Quasi-Experimental Designs for Research. Chicago: Rand McNally. Fowler, C. and Shanks, J. (2002). Tympanometry. In J. Katz (ed.), Handbook of Clinical Audiology (pp. 175-205). Baltimore, Maryland: Williams and Wilkins. Feeney, M. and Sanford, C. (2004). Age effects in the human middle ear: wideband acoustical measures. Journal of the Acoustical Society of America, 116, 3546-3458. Feeney, G., Grant, I. and Marryott, L. (2003). WBER measurements in adults with middle ear disorders. Journal of Speech, Language, and Hearing Research, 46, 901-911. Frank, T. (2001). High-Frequency (8 to 16 kHz) Reference Thresholds and Intrasubject Threshold Variability Relative to Ototoxicity Criteria Using a Sennheiser HDA 200 Earphone. Ear and Hearing, 22 (2): 161-168. Gaihede, M. (1996). Tympanometric preconditioning of the tympanic membrane. Hearing Research, 102, 28-34. Gaihede, M., Tos, M., Thomsen, J. and, Bale, V. (1997). Negative middle ear pressures explained by tympanometric artefacts in secretory otitis media. Otitis Media Today, 101-106. Ginsberg, I. and White, T. (1985). Otologic considerations in audiology. In: Katz, J. Handbook of Clinical Audiology (3rd ed.). Baltimore: Williams and Wilkins. Gorga, M., Neely, S. and Bergman, B. (1992). Otoacoustic emissions from normal-hearing and hearingimpaired subjects: Distortion product responses. Journal of the Acoustic Society of America, 93, 20502060. Greenhouse, S. and Geisser, S. (1959). On methods in the analysis of profile data. Psychometrika, 24, 95-112. 119  Holte, L., Cavanaugh, R., Margolis, R. (1990). Ear canal wall mobility and tympanometric shape in young infants. Journal of Pediatrics, 117, 77-80. Huang, G., Rosowski, J., Puria, S., and Peake, W. (2000). A non-invasive method for estimating acoustic admittance at the tympanic membrane. Journal of the Acoustical Society of America, 108, 1128-1146. Hunter, L., Bagger-Sjoback, D., and Lundberg, M. (2008). WBER associated with otitis media in infants and children with cleft palate. International Journal of Audiology, 47, 57-61. Hunter, L., and Margolis, R. (1992). Multifrequency tympanometry: current clinical applications. American Journal of Audiology, July 1992, 33-43. Jeng, P., Levitt, H., Lee, W., Gravel, J. (2002). Reflectance measures for detecting otitis media with effusion in children: Preliminary findings, Proceedings of the 7th International Symposium on Recent Advances in Otitis Media with Effusion (pp. 194-199). Fort Lauderdale, FL: BC Decker Inc. Keefe, D., and Bulen, J. (1993). Ear canal impedance and reflection coefficient in human infants and adults. Journal of the Acoustical Society of America, 94, 2617-2638. Keefe, D., Bulen J., Arehart, K., and Burns, E. (1993). Ear canal impedance and reflection coefficient in human infants and adults. Journal of the Acoustical Society of America, 94, 2617-2638. Keefe, D., and Levi, E. (1996). Maturation of the middle and external ears: acoustic power-based responses and reflectance tympanometry. Ear and Hearing, 17 (5), 361-373. Keefe, D. H., and Simmons, J.L. (2003) Energy transmittance predicts conductive hearing loss in older children and adults. Journal of the Acoustical Society of America, 114, 3217-3238. Kei, J., Mazlan, R., Hickson, L. Et al. (2007) Measuring middle ear admittance in newborns using 1000 Hz tympanometry: a comparison of methodologies. Journal of the Acoustical Society of America. 18:739-48. Kemp, D.T. (1979). Evidence of mechanical nonlinearity and frequency selective wave amplification in the cochlea. Archive of Otorhinolaryngology, 224, 37-45. Koivunen, P., Uhari, M., Laitakari, K., Alho, O. and Luotonen, J. (2000). Otoacoustic emissions and tympanometry in children with otitis media. Ear and Hearing, 21, 212-217. Liu, Y., Cohn, E. Ellison, J., Fitzpatrick, D., Gorga, M., Gortemaker, M., Sanford, C., and Keefe, D. (2008). Wideband acoustic ear canal reflectance, including wideband tympanometry and acoustic reflex thresholds: System development and results on children with middle ear fluid and adults. Association of Research in Otolaryngology. Meeting, Phoenix, AZ (February). Lynch, T., Nedzelnitsky, V.,Peake, W. (1982). Input impedance of the cochlea of in the cat. Journal of the Acoustical Society of America, 72, 108-130. Margolis, R., Bass-Ringdahl, S., Hanks, W., Holte, L., Zapala, D. (2003). Tympanometry in newborn infants: 1 kHz norms. Journal of the American Academy of Audiology, 14, 383-392. 120  Margolis, R., Goycoolea, H. (1993). Multifrequency tympanometry in normal adults. Ear Hear. 1993 14: 408–413 Margolis, R., and Hunter, L. (2000). Immitance measurements. Audiology Diagnosis, 381-422. Margolis, R., Saly, G., and Keefe, D. (1999). WBER tympanometry in normal adults. Journal of the Acoustical Society of America, 106 (1), 265-280. Margolis RH and Shanks JE (1991) Tympanometry: Principles and procedures In W. F. Rintelmann (Ed.), Hearing assessment (pp. 179-246). Texas: Pro-Ed Margolis, R. H., Paul, S., Saly, G.I., Schachern, P.A. and Keefe, D.H. 2001. WBER tympanometry in chinchillas and human. Journal of the Acoustical Society of America, 110, 1453-1464. Mimosa Acoustics. (2002). Power Flow Measurement System User’s Manual. Mimosa Acoustics, November 6, 2002. Retrieved July 2009 from http://www.mimosaacoustics.com/. Paradise, J., Smith, C., and Bluestone, C. (1976). Tympanometric detection of middle ear effusion in infants and young children. Pediatrics, 58, 198-210. Piskorski, P., Keefe, D., Simmons, J., and Gorga, M. (1999). Prediction of conductive hearing loss based on acoustic ear canal response using a multivariate clinical decision theory. Journal of the Acoustical Society of America, 105, 1749-1764. Puria, S., and Allen, B. (1998). Measurement and model of the cat middle ear: evidence of tympanic membrane acoustic delay. Journal of the Acoustical Society of America, 104, 3463-3481. Reflwin Interacoustics. (2008). Metlab Utilities for Reflwin: Summary of Functions. Reflwin Interacoustics, May 28, 2008. Roup, C., Wiley, T., Safady, S. and Stoppenbach, D. (1998). Tympanometric screening norms for adults. American Journal of Audiology, 7, 55-60. Sanford, C., and Feeney, M. (2008). Effects of maturation on tympanometric wideband acoustic transfer functions in human infants. Journal of the Acoustical Society of America, 124, 2106-2122. Sang, Y., Park, S., Park, Y., and Do, B. (2003). Prognostic value of otoacoustic emissions in children with middle ear effusion. Head and Neck Surgery, 129, 136-140. Schwartz, W., Gorry, A., Kassirer, J., and Essig, K. (1972). Decision analysis and clinical judgement. The American Journal of Medicine, 55 (3), 459-472. Shahnaz, N., and Bork, K. (2008). Comparison of standard and multi-frequency tympanometric measures obtained with Virtual 310 system and Grason-Stadler Tympstar. Canadian Journal of SpeechLanguage Pathology and Audiology, 32 (4), 146-157. Shahnaz, N., Bork, K., Polka, L., Longridge, N., Bell, D., and Westerberg, B. (2009). ER and tympanometry in normal and otosclerotic ears. Ear and Hearing, 30, 219-233. 121  Shahnaz, N., and Bork, K. (2006). WBER norms for Caucasian and Chinese Young Adults. Ear and Hearing, 27 (6), 774-788. Shahnaz, N., and Davies, D. (2006). Standard and multifrequency tympanometric norms for Caucasian and Chinese young adults. Ear and Hearing, 27, 75-90. Shahnaz, N. and Polka, L. (1997). Standard and multi-frequency tympanometry in normal and otosclerotic ears. Ear and Hearing, 18, 326-341. Shanks, J., and Lilly, D. (1981). An evaluation of tympanometric estimates of Vea. Journal of Speech and Hearing Research, 24, 557-566. Shanks, J., and Shelton, C. (1991). Basic principles and clinical applications of tympanometry. Otolaryngology Clinics of North America, 24, 299-328. Statistics Canada. (2004). Statistics Canada Web site. Available at http://www.statcan.ca/english/concepts/definitions/ethnicity01.htm. Accessed January 21, 2004. Stinson, M. R. (1990). Revision of estimates of acoustic ER at the human eardrum. Journal of the Acoustical Society of America, 88, 1773-1778. Stinson, J.A., Shaw, E.A. and Lawson, B. W. (1982). Estimation of acoustical ER at the eardrum from measurements of pressure distribution in the human ear canal. Journal of the Acoustical Society of America, 72, 766-773. Tonndorf, J. and Khanna, S. (1970). The role of the tympanic membrane in middle ear transmission. Annals of Otorhinolaryngology, 79, 743-753. Turner, R., Robinette, M. and Bauch, C. (1999). Clinical Decisions in Musiek, F. and Rintelmann, W., editors, Contemporary Perspectives in Hearing Assessment (pp 437 – 463). Boston: Allyn and Bacon. Wan, I., and Wong, L. (2002). Tympanometric norms for Chinese young adults. Ear and Hearing, 23, 416-421. Withnell, R., Jeng, P., Waldvogel, K., Morgenstein, K. and Allen, J. (2009). An in situ calibration for hearing thresholds. Journal of the Acoustical Society of America, 125, 1605-1611. Wiley, T., Cruickshanks, K., Nondahl, D., Tweed, T., Klein, R., & Klein, B. (1996). Tympanometric measures in older adults. Journal of the American Academy of Audiology, 7, 257-266. Vander Werff, K., Prieve, B., and Georgantas, L. (2007). Test-retest reliability of WBER measures in infants under screening and diagnostic conditions. Ear and Hearing, 28 (5), 669-681.  122  Voss, S., and Allen, B. (1994). Measurement of acoustic impedance and reflectance in the human ear canal. Journal of the Acoustical Society of America, 95, 372-384.  123  Appendix I Statistical Analysis Tables Note: The sections in this Appendix refer to the corresponding sections of the Thesis. 3.1.1 Wideband energy reflectance  Table. A1. ANOVA Table for ER with ethnicity, gender, and ear as within-subject factors and instrument and frequency as between-subject factors. 124  Table. A2. Greenhouse-Geiser Correction for the ANOVA of ER using the Mimosa versus Reflwin system.  125  Table. A3. Tukey’s HSD analysis for the ethnicity by frequency interaction from the ANOVA of ER using the Mimosa versus Reflwin system.  126  Table. A4. Tukey’s HSD analysis for the gender, instrument, and frequency interaction from the ANOVA of ER using the Mimosa versus Reflwin system.  127  3.1.2 Power absorption  Table. A5. ANOVA Table for PA with ethnicity, gender, and ear as within-subject factors and instrument and frequency as between-subject factors.  128  Table. A6. Greenhouse-Geiser Correction for the ANOVA of ER using the Mimosa versus Reflwin system.  129  Table. A7. Tukey’s HSD analysis for the ethnicity by frequency interaction from the ANOVA of PA using the Mimosa versus Reflwin system.  130  Table. A8. Tukey’s HSD analysis for the instrument by frequency interaction from the ANOVA of PA using the Mimosa versus Reflwin system.  131  3.1.3 Static versus dynamic pressure measurements using the Reflwin System  Table. A9. ANOVA Table for PA with ethnicity, gender, and ear as within-subject factors and pressurization method and frequency as between-subject factors.  132  Table. A10. Greenhouse-Geiser Correction for the ANOVA of PA using the dynamic versus static pressurization technique for Reflwin.  133  Table. A11. Tukey’s HSD analysis for the pressurization technique by frequency interaction from the ANOVA of PA using the dynamic versus static pressurization technique for Reflwin.  134  3.2.1.1 Vea at 226 Hz  Table. A12. ANOVA Table for Vea at 226 Hz with ethnicity, gender, and ear as within-subject factors and instrument and tail as between-subject factors.  3.2.1.2 Vea at 1000 Hz  Table. A13. ANOVA Table for Vea at 226 Hz with ethnicity, gender, and ear as within-subject factors and instrument and tail as between-subject factors. 135  3.2.2.1 Ytm at 226 Hz  Table. A14. ANOVA Table for Ytm at 226 Hz with ethnicity, gender, and ear as within-subject factors and instrument and tail as between-subject factors.  Table. A15. ANOVA Table for Ytm at 1000 Hz with ethnicity, gender, and ear as within-subject factors and instrument and tail as between-subject factors. 136  3.2.3.1 TPP at 226 Hz  Table. A16. ANOVA Table for TPP at 226 Hz with ethnicity, gender, and ear as within-subject factors and instrument as a between-subject factor.  3.2.4.1 TW at 226 Hz  Table. A17. ANOVA Table for TW at 226 Hz with ethnicity, gender, and ear as within-subject factors and instrument as a between-subject factor.  137  3.3 Efficacy of Mimosa and Reflwin in identifying otosclerosis - ANOVA and Post-hoc Tables  Table. A18. ANOVA Table for the comparison between otosclerotic and normal ears assessed on the Mimosa system. Condition (normal versus otosclerosis) served as a between-subject factor and frequency served as a within-subject factor.  Table. A19. Greenhouse-Geiser Correction for the ANOVA of the comparison between otosclerotic and normal ears assessed on the Mimosa system.  138  Table. A20. Tukey’s HSD analysis for the condition by frequency interaction from the ANOVA of the comparison between otosclerotic and normal ears assessed on the Mimosa system.  Table. A21. ANOVA Table for the comparison between otosclerotic and normal ears assessed on the Reflwin system. Condition (normal versus otosclerosis) served as a between-subject factor and frequency served as a within-subject factor.  139  Table. A22. Greenhouse-Geiser Correction for the ANOVA of the comparison between otosclerotic and normal ears assessed on the Reflwin system.  Table. A23. Tukey’s HSD analysis for the condition by frequency interaction from the ANOVA of the comparison between otosclerotic and normal ears assessed on the Reflwin system.  140  4.1.1 Comparison of the Mimosa system to previous studies  Frequency (Hz)  Current Study (N =60) Mimosa Mean 0.91 0.87 0.82 0.74 0.64 0.54 0.45 0.39 0.37 0.35 0.30 0.27 0.32 0.48 0.60  250 315 400 500 630 800 1000 1250 1600 2000 2500 3150 4000 5000 6000  SD 0.05 0.06 0.07 0.10 0.14 0.14 0.13 0.13 0.12 0.12 0.12 0.13 0.17 0.22 0.22  Shahnaz & Bork 2006 (N = 128) Mimosa  Voss et al., 1996 (N = 10) Early Mimosa  Mean 0.91 0.88 0.84 0.76 0.68 0.58 0.47 0.40 0.37 0.34 0.29 0.26 0.30 0.47 0.62  Mean 0.92 0.90 0.87 0.82 0.75 0.64 0.48 0.37 0.39 0.41 0.36 0.30 0.31 0.38 0.54  SD 0.08 0.08 0.09 0.12 0.14 0.16 0.16 0.15 0.13 0.12 0.13 0.14 0.17 0.22 0.19  SD 0.05 0.05 0.06 0.07 0.08 0.11 0.13 0.16 0.13 0.08 0.10 0.12 0.16 0.19 0.19  Table. A24. Comparison of ER data between the current study (N = 60; 18 – 38 yr.), Shahnaz & Bork (2006) (N = 128; 18 – 32 yr.), and Voss et al. (1994) (N = 10; 18 – 24). The current study and Shahnaz & Bork (2006) use the same version of the Mimosa system, while the study by Voss et al. (1994) uses an older version of the Mimosa system.  Frequency (Hz)  250 315 400 500 630 800 1000 1250 1600 2000 2500 3150 4000 5000 6300  Current study N = 60 Reflwin static Mean 0.86 0.83 0.79 0.71 0.59 0.54 0.44 0.35 0.38 0.36 0.33 0.28 0.33 0.45 0.59  SD  0.13 0.12 0.12 0.12 0.13 0.14 0.14 0.10 0.09 0.11 0.14 0.14 0.18 0.20 0.16  Current study N = 60 Reflwin dynamic Mean SD  Sanford & Feeney (2008); N = 20 Reflwin static Mean SD  Keefe et al., 1990; N = 10 Early Reflwin Mean SD  0.79 0.80 0.74 0.63 0.51 0.43 0.33 0.28 0.32 0.33 0.30 0.27 0.34 0.47 0.60  0.87 0.85 0.83 0.81 0.62 0.58 0.50 0.42 0.41 0.39 0.32 0.28 0.27 0.48 0.67  0.96 0.95 0.93 0.88 0.82 0.76 0.69 0.63 0.60 0.58 0.49 0.35 0.24 0.33 0.65  0.08 0.07 0.10 0.12 0.14 0.15 0.11 0.08 0.08 0.10 0.12 0.13 0.17 0.18 0.15  unavailable  0.04 0.04 0.05 0.07 0.08 0.11 0.14 0.13 0.16 0.17 0.18 0.20 0.28 0.27 0.19  Feeney et al., 2004 N = 40 Early Reflwin Mean SD 0.92 0.03 0.90 0.04 0.87 0.05 0.82 0.07 0.75 0.10 0.64 0.13 0.48 0.15 0.37 0.17 0.39 0.18 0.41 0.17 0.36 0.16 0.30 0.16 0.31 0.15 0.38 0.20 0.54 0.20  Table. A25. Comparison of ER data between the current study (N = 60; 18 – 38 yr.), Sanford & Feeney (2008) (N = 20; 22 – 30 yr.), Feeney et al. (2004) (N = 40; 18 – 28 yr.), and Keefe et al. (1993) (N = 10; 20 – 50 yr.). The current study and Sanford & Feeney (2008) used the same version of the Reflwin system, while the studies by Feeney et al. (2004) and Keefe et al. (2004) used an older version of the Reflwin system. 141  Appendix II Consent Form for Normal Hearing Subjects  THE UNIVERSITY OF BRITISH COLUMBIA Appendix II Consent form for normal hearing subjects School of Audiology & Speech Sciences Faculty of Medicine 5804 FairviewAvenue  Consent Form Vancouver, BC, Canada V6T 1Z3  Project Title The Effects Of Race, Caucasian Versus East Asian, Diet (vegetarian vs. non-vegetarian) On Immittance Audiometry, Wide Band Reflectance, real ear to coupler difference (RECD) and Laser Doppler Vibrometry Norms  Principal Investigator Dr. Navid Shahnaz Assistant Professor School of Audiology & Speech Sciences  Co-Investigator Jefferey Shaw Graduate Student School of Audiology & Speech Sciences  Phone: 604-822-5953 Email: nshahnaz@ audiospeech.ubc.ca  Phone: 604-822-9474  Co-Investigator Dr. Susan Small Assistant Professor School of Audiology & Speech Sciences  Co-Investigator Ning Hu Graduate Student School of Audiology & Speech Sciences  Phone: 604-822-5696 Email: ssmall@ audiospeech.ubc.ca  Phone: 604-822-9474  Purpose This project will evaluate the effectiveness of newmiddle ear analysis techniques for assessing middle ear function in different ethnic groups and different diet regimes. These new procedures are a variation of a test procedure that is used frequently to detect middle ear problems in clinics which is called tympanometry. Multi-frequency tympanometry is a modification of this technique that assesses the function of the middle ear across much wider frequency range, therefore, providing a more detailed picture of your child’s middle ear. Wide band reflectance is a newmiddle 142  ear analysis technique and poses no risks or danger to your ear or hearing. The project will also incorporate a newand safe way to assess the middle ear function. Laser Doppler vibrometry (LDV) is a well established technique used throughout the hearing research community for measurements of the middle ear function. This method is also a safe procedure and poses no risks or danger to your ear or hearing. A lowintensity light will be projected on your ear drum and you will hear a sound. The changes in the light as a result of your ear drum movements can be detected and analysed. Your height, weight, and skull dimensions will also be measured. The purpose of this project is 1) to investigate the effects of race, Caucasian versus Chinese, on hearing sensitivity, immittance audiometry, wide band reflectance, and laser Doppler vibrometry results, and 2) to establish race dependent guidelines and protocols for multifrequency tympanometry, wide band reflectance and laser Doppler vibrometry. Results from this study will help us determine if there are differences in the middle ears of subjects from different racial backgrounds. If differences are found, this will lead to a better diagnosis of middle ear problems in individuals from different races. You are invited to participate in this research.  Study Procedures As a subject for this study, you will attend one session 1.5 hour in duration. All appointments will be scheduled at your convenience. Testing will be carried out in Room B-28 in the basement of the Woodward Instructional Resources Centre or School of Audiology and Speech Sciences located on UBC campus. In order to be included in this study you should be a Caucasian or East Asian young adult between the ages of 18-34, with a normal hearing system as assessed by the following tests. You should also be free of any history of head trauma and middle ear infection. You have been invited to participate in this study because you meet the above criteria Before the test procedures are conducted, your hearing will be tested using conventional and extended high frequencies and your eardrum and ear canal will be inspected. Five tests will be done in this study: 1) Transient otoacoustic emissions, 2) Wide band reflectance, 3) Multifrequency tympanometry; 4) Real Ear to Coupler Difference, and 5) laser Doppler vibrometry. In the first three tests, a small earphone will be placed into the entrance of your ear canal using a soft and delicate plastic or sponge tip. It is designed not to cause any allergic reactions. The presence of the earphone may be a bit uncomforTable to some subject, but not painful. The earphone and the attached tip pose no risks or danger to your ear or hearing and have been used extensively in testing newborns, children, and adults. The first test, transient-evoked otoacoustic emissions, is commonly used in clinics to detect hearing loss. A click sound will be presented through the small earphone placed at the entrance of your ear canal. The level of sound is the same as normal talking. Echoes to the sound come back out of your ear. These echoes are measured by a computer. This test will give us information about your middle and inner ear. It will take 2-3 minutes in each ear. The second test, multi-frequency tympanometry, presents a pure tone while the air pressure in the ear canal is changed. We will test at the frequencies of speech. You will be tested on two 143  commercially available systems to also check for test-retest-reliability of the systems. It will take about 8 minutes in each ear. The third test involves a test called the Real Ear to Coupler Difference. In this test, the same soft tip as in tympanometry remains in your ear, while a small, narrowtube is inserted alongside the tip in your ear canal. The tube is soft and will not cause you any pain. Some adults report a tickling sensation when the tube is in their ear. You will then hear a series of tones that increase in pitch. The tube will help measure the amount of sound close to your eardrum. The test time varies between 2-4 minutes per each ear. The fourth test, wide band reflectance, presents chirping sounds to the ear. This helps us see howwell the middle ear reacts to sounds that span the human speech range. This test will take about 2-3 seconds for each ear. In the fifth test, you will lay on your back on a comforTable examination bed with head turned to raise the ear to be measured. Mounted on a microscope is a laser-vibrometer. A standard microscope will be used to observe the ear drum and focus the laser beam on your ear drum. An ear speculum, with associated sound source will be placed in the ear canal to allowvisualization of the ear drum. You will hear a chirping sound very much similar to the wide band reflectance. Measurements for both ears are done within 15 minutes, without pain or discomfort. You may withdrawfrom this study at any time and without providing any reasons for your decision. Should you decide to withdrawfrom this study all the data collected before your withdrawal will be discarded permanently from our database. Withdrawal will in no way jeopardize your present or future clinical care. Risks There are no known or anticipated side effects of the above procedures. All tests completed in this study are routine, non-invasive, clinical procedures that are used on newborns, children, and adults. However, some people may experience slight discomfort during the test procedures. Adequate breaks will be provided if this occurs. Confidentiality Your identity will be coded using a code known only to the researchers, and all information that is collected from you will remain confidential. Only group results or coded individual results will be given in any reports about the study. Coded results only (no personal information) will be kept in computer files on a password protected hard drive. Your confidentiality will be respected. No information that discloses your identity will be released or published without your specific consent to the disclosure. However, research records and medical records identifying you may be inspected in the presence of the Investigator or his designate by representatives of Health Canada, and the UBC Research Ethics Board for the purpose of monitoring the research. However, no records that identify you by name or initials will be allowed to leave the Investigators’ offices. 144  Reimbursement You will be paid an honorarium of $10 for your participation in this research. Your participation in this study is entirely voluntary; should you withdrawfrom the study at any point you will still receive the $10 honorarium.  Compensation for Injury Signing this consent form in no way limits your legal rights against the sponsor, investigators, or anyone else. Consent: I, _________________________, have read the above study consent form and I consent to participate in this study undertaken by Dr. Navid Shahnaz at the School of Audiology & Speech Sciences at UBC. The researcher assures me that my participation in this experiment is completely voluntary and that I may withdrawfrom this research at any time without consequences. If I have any question or desire further information with respect to this study, I may contact Dr. Navid Shahnaz at 604-822-5953. If I have any concerns about my treatment or rights as a research subject, I may contact the Research Subject Information Line at the University of British Columbia Office of Research Services, at 604-822-8598. I have received a signed and dated copy of this consent form for my records.  _________________________ Subject name (please print)  __________________________ Witness name (please print)  Subject signature  ______________________________ Witness signature  __________________________ ______________________________ Name of principal/co-investigator (please print) Co/Investigator signature  Date  Date  Date  145  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0070857/manifest

Comment

Related Items