Eg‘éFLUENCE 0F AUBIOMETREC CQNFEGURAHOH 0N PURE-TONE. WARBLE-YGHE AND NARROW-BAND FéGiSE THRESHE‘ELDS FGR ADULTS WITH SENSGRWEURAL HEARENG LOSSES Thesis for the Begree of Ph. D. MICHtfiAN STATE UNIVERSITY MYRRA MANNELA STEPHENS 1973 This is to certify that the thesis entitled INFLUENCE OF AUDIOMETRIC CONFIGURATION 0N PURE-TONE, WARBLE-TONE AND NARROW-BAND NOISE THRESHOLDS FOR ADULTS WITH SENSORINEURAL HEAARLNG LOSSES presene y MYRNA MANNILA STEPHENS has been accepted towards fulfillment of the requirements for PH.D. degreein AUDIOLOGY e SPEECH SCIENCES Z/ . . 'Aii/(fl/{L/CZT " / Major professor Date 14-30-73 0-7 639 BONDING BY HnAs & SDNS' . : WK BINDERY mg I I'" sRY 8 iC/ ABSTRACT INFLUENCE OF AUDIOMETRIC CONFIGURATION 0N PURE-TONE, WARBLE-TONE AND NARROW-BAND NOISE THRESHOLDS FOR ADULTS WITH SENSORINEURAL HEARING LOSSES BY Myrna Mannila Stephens Threshold measurements for pure tones, warble tones with i 3% and i 10% frequency deviation and narrow-band noise were compared using two groups of subjects with sensorineural hearing losses. Sub- jects consisted of 16 adults whose audiograms showed a marked change in threshold as a function of frequency (sharp configuration) and 16 adults whose audiograms showed slight change in threshold across frequencies (gradual configuration). Each of the two groups was further subdivided to include 8 subjects fifty years of age or younger and 8 subjects over fifty years of age. The test stimuli were presented at, or with a center frequency of 500, 1000, 2000 and 4000 Hz. Following the initial measurements, repeat thresholds were obtained to allow analysis of test-retest reliability. Results demonstrated a small but consistent learning effect for both subject groups and all stimuli with improvement on retest of .S to 1.3 dB. Scores on the initial test showed a high (+.97) Myrna Mannila Stephens correlation with scores on retest. Comparison by age showed no significant differences between the two age groups studied. The most important findings related to differences in thresh- olds fer the four test signals, changes in the relationships between signals across frequencies and the effect configuration exerts on these relationships. For subjects whose audiograms showed a pre- cipitous drop in threshold from low to high frequencies (sharp configuration) thresholds at the lower frequencies (500 and 1000 Hz) were within i 4.4 dB for all signals while at the higher frequencies (2000 and 4000 Hz) thresholds for narrow-band noise signals were over 20 dB more sensitive than for pure tones. Thresholds for individuals exhibiting a gradually sloping audiometric configuration were similar (within 1 3.2 dB) for all signals at all frequencies. For both groups of subjects, warble-tone thresholds showed good agreement with pure-tone thresholds. However, thresholds for warble tones with i 3% frequency deviation agreed more closely to thresholds for pure tones than did thresholds for warble tones with i 10% frequency deviation (i 1.9 versus i 3.9). Based on the results obtained, it is suggested that when an alternative signal to pure tone is desired, warble tones rather than narrow-band noise be used for threshold measurements since warble- tone thresholds agree more closely to thresholds for pure tones. INFLUENCE OF AUDIOMETRIC CONFIGURATION ON PURE-TONE, WARBLE-TONE AND NARROW-BAND NOISE THRESHOLDS FOR ADULTS WITH SENSORINEURAL HEARING LOSSES By Myrna Mannila Stephens A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Audiology and Speech Sciences 1973 Accepted by the Faculty of the Department of Audiology and Speech Sciences, College of Communication Arts, Michigan State University, in partial fulfillment of the requirements for the degree of Doctor of Philosophy. Thesis Committee: // m%fl// d I Director WilliamF. Rintelmann, Phi) ./ - ' ,7”) / /, fl I: 67/a 1/ ,i '\~VJX'?‘.'.’/€’L’.(7 Daniel S. Beasley, Ph.D. éégaéixgctéf:éélfliwui /4/ May fiflhin, Ph.D. / Verling c. Trdldahl, Ph.D/ ii ACKNOWLEDGMENTS The writer wishes to thank Dr. William F. Rintelmann for his assistance and guidance as thesis adviser, and also Drs. May B. Chin, Daniel S. Beasley and Verling C. Troldahl for serving as committee members. Appreciation is also extended to Mr. Donald E. Riggs for his assistance with the instrumentation utilized in this study. Finally, and most especially, gratitude is due Peter, Sarah and John. iii TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . LIST OF FIGURES . . . . . . . . . . . Chapter I. II. III. IV. INTRODUCTION . . . . . REVIEW OF THE LITERATURE . . . . . Application of Warble Tones to Clinical Audiology . . . . Comparisons of Warble-tone and Pure-tone Stimuli . Evaluation of Warble-tone Stimulus Parameters . . . . Narrow-band Noise as a Stimulus Age . . . . . . . . . . . . Summary . . . . . . . . . . EXPERIMENTAL PROCEDURES . . . . . . . Subjects . . . . . . . . . . . Test Environment . . . . . . . . Instrumentation . . . . . . . . . Test Stimuli . . . . . . . . . . Calibration . . . . . . . . . . Procedure . . . . . . . RESULTS AND DISCUSSION . . . . . Results . . . . . . . . . Comparison of Thresholds for Pure Tones, Warble Tones and Narrow-band Noise Effect of Configuration . . . . . iv Page vi 10 13 16 16 19 19 23 23 25 26 27 28 28 29 30 Chapter Threshold as a Function of Frequency Test-Retest Comparison . . . Influence of Age . Discussion . . . Clinical Implications . V. SUMMARY, CONCLUSIONS AND RECOMMENDATIONS Summary. . Conclusions . Recommendations . REFERENCES APPENDICES Appendix A. Calibration Data B. Analysis of Variance Table . Page 38 40 41 41 49 51 51 52 53 SS 59 8S LIST OF TABLES Table Page 1. Age range, mean age, mean hearing level by frequency and mean difference between threshold at 4000 Hz and threshold at 250 Hz for subjects employed in this study . . . . . . . . . . . . 21 2. Mean threshold in dB SPL for each test signal. Thresholds for both test and retest, for all subjects and all frequencies are combined. Each mean is based on 256 thresholds . . . . . . 29 3. Mean thresholds in dB SPL by frequency for each test signal. Each mean is based on all subjects for both test and retest (64 thresholds) . . . . . 30 4. Mean threshold (in dB SPL) by frequency for subjects grouped by audiometric configuration. Data for all signals on both test and retest are combined. Each mean is based on 128 thresholds . . . . . . . . . . . . . . . 32 5. Mean threshold (in dB SPL) for each signal with subjects grouped by audiometric configuration. Each mean is based on an average of all frequencies for both test and retest (128 thresholds) . . . . . . . . . . . . . 34 6. Mean thresholds in dB SPL by frequency for each of the four test signals with subjects grouped by audiometric configuration. Data for test and retest are combined. Each mean is based on 32 thresholds . . . . . . . . . . . . . . 3S 7. Difference between mean threshold for pure tones and mean threshold for each of the other three test signals at each frequency with subjects grouped by audiometric configuration . . . . . . 35 vi Table Page 8. Comparison of thresholds for warble tones with i 3% and i 10% frequency deviation and narrow-band noise to thresholds for pure tones at 2000 and 4000 Hz for subjects with sharp audiometric configurations. Thresholds are presented for initial test only . . . . . . . . . . . . . 39 9. Mean threshold in dB SPL for four test frequencies with test and retest data for all subjects and all signals combined. Each mean is based on 256 thresholds . . . . . . . . . . . . . . 40 10. Mean threshold on initial test compared with mean threshold on repeat test for each of the factors studied . . . . . . . . . . . . . 42 11. Mean threshold (in dB SPL) by frequency for each of the four test signals with subjects grouped by age. Data for test and retest are combined. Each mean is based on 32 thresholds . . . . . . . 44 A1. Octave band and C-scale analyses of ambient noise levels in examination room in dB SPL according to standards of the American Standards Associa- tion (ASA 83.1-1960, ANSI 83.1-1971, R). . . . . . 59 A2. Audiometer earphone output data for the right earphone of the Maico MA-24 audiometer. Measure- ments in accordance with American National Standards Institute (ANSI 53.6-1969) . . . . . . 60 A3. Output data for bone conduction vibrator. Measure- ments were made in accordance with norms specified by the Hearing Aid Industry Conference (HAIC) Interim Bone-Conduction Thresholds for Audiometry (Lybarger, 1966) . . . . . . . . . 63 A4. Pre- and post-experimental linearity of Maico MA-24 audiometer attenuator, right channel. Measured at the test earphone, 1000 Hz . . . . . . 64 AS. Pre- and post-experimental rise and decay times (in milliseconds) as measured for pure tones generated by the Maico MA-24 audiometer. Measure- ments were made with the assistance of a storage oscilloscope in accordance with the American National Standards Institute (ANSI 83.6-1969) . . . 65 vii Table Page A6. Pre- and post-experimental checks of the test frequencies of the Maico MA-24 audiometer performed in compliance with the American National Standards Institute (ANSI 53.6-1969) . . . . 66 A7. Pre- and post-experimental harmonic distortion measurements of the fundamental for test fre- quencies used. Measurements were made for the right channel of the Maico MA-24 audiometer in compliance with the American National Standards Institute (ANSI 83.6-1969) . . . . . . . 67 A8. Audiometer earphone output data for the left earphone of the Maico MA-24 audiometer. Measurements in accordance with the American National Standards Institute (ANSI 83.6-1969) . . . . 68 A9. Pre- and post-experimental linearity of Maico MA-24 audiometer attenuator, left channel. Measured at the test earphone, 1000 Hz . . . . . . 69 A10. Settings required to produce and measure the warble tones with i 3% and i 10% frequency deviation. Settings include the Frequency Deviation (FD) setting of the beat—frequency oscillator; Volt Scale (VS), and Output Voltage (V) on the function generator; and the Hz/Div setting of the storage scope spectrum analyzer . . . . . . . . . . . . . 70 All. Output data for beat-frequency oscillator routed through Maico MA-24 Audiometer (left channel, right earphone). Levels were obtained by measuring the SPL output of the unmodulated warble-tone center frequencies at "0" VU reading and 70 dB Hearing Threshold Level setting through the ACCESSORY INPUW'of the Maico MA-24 Audiometer . . . . . . . . . . . 71 A12. Pre- and post-experimental harmonic distortion measurements of the fundamental of the warble- tone center frequencies used. Measurements were made for the output of the beat-frequency oscillator, routed through the left channel of the Maico MA-24 audiometer, in compliance with the American National Standards Institute (ANSI 83.6-1969). . . . . . . . . . . . . . 74 viii Table Page A13. Pre- and post-experimental acoustic spectrum data for narrow-band noise produced by Maico MA-24 audiometer . . . . . . . . 75 A14. Output data for narrow-band noise produced by Maico MA-24 Audiometer (left channel, right earphone) . . . . . . . . . . . . 76 B1. Analysis of Variance Table . . . . . . . . . . 85 ix LIST OF FIGURES Figure 1. Mean pure—tone thresholds (rounded to nearest 5 dB) for 16 subjects with sharp configuration and 16 subjects with gradual configuration . Block diagram of equipment utilized in this study . Mean thresholds in dB SPL by frequency for each Each mean is based on all subjects test signal. for both test and retest (64 thresholds) Mean thresholds in dB SPL by frequency with subjects grouped by audiometric configuration. Test and retest thresholds for all signals are Each mean is based on 128 thresholds . combined. Comparison of mean thresholds for warble tones with t 3% and i 10% frequency deviation and narrow-band noise with mean thresholds for pure tones for subjects with gradual audiometric configurations. Base line represents threshold for pure tones. Negative values indicate thresholds for pure tones were less sensitive than thresholds for comparison signals . Comparison of mean thresholds for warble tones with t 3% and 1 10% frequency deviation and narrow-band noise with mean thresholds for pure tones for subjects with sharp audiometric configurations. Base line represents thresholds for pure tones. Negative values indicate thresholds for pure tones were less sensitive than thresholds for comparison signals Mean test and retest thresholds in dB SPL at each test frequency for the two age groups Each mean threshold with all signals combined. is based on 128 thresholds . Page 22 24 31 33 36 37 43 Table Page Al. Frequency response of left (nontest) earphone. Left earphone was used in obtaining pre- experiment audiogram only . . . . . . . . . . 79 A2. Frequency response of right (test) earphone . . . . 80 A3. Pre- and post-experimental analysis of narrow-band noise with 500 Hz center frequency . . . . . . . 81 A4. Pre- and post-experimental analysis of narrow-band noise with 1000 Hz center frequency . . . . . . 82 AS. Pre- and post-experimental analysis of narrow-band noise with 2000 Hz center frequency . . . . . . 83 A6. Pre- and post—experimental analysis of narrow—band noise with 4000 Hz center frequency . . . . . . 84 xi CHAPTER I INTRODUCTION The stimuli most frequently employed as a measure of threshold sensitivity are pure tones. Such stimuli, however, are not neces- sarily appropriate in all situations. Harris (1963) lists several disadvantages of pure tones including problems associated with stand- ing waves in the earphone-eardrum coupling, irregularities at higher frequencies of the earphone frequency reSponse curve and dependence on frequency of the impedance for which the earphone is used. He suggests use of tones modulated two to five percent in frequency, or narrow bands of noise centered on the usual audiometric frequencies as substitutes for pure tones. In addition to eliminating the previously mentioned problems, frequency modulated or warble tones and noise bands are assumed to be more interesting to the subject (especially children) and easier to detect than are pure tones. Since they do not produce standing waves, they can be introduced into sound field for threshold testing with difficult to test patients, or for evaluation of hearing aid gain and frequency response. The warble tone is produced by frequency modulation while amplitude is held constant resulting in a warbling or modulating sensation. Warble tones can be altered along three basic parameters: the center or base frequency or the frequency around which modi- fication takes place; the amount of frequency deviation expressed in Hz or percent; and the modulation rate or the number of times per second the frequency varies from one extreme to the other. For example, a 500 Hz tone with a + 10% frequency deviation and an 8 per second modulation rate would vary from $00 to 550 Hz, 8 times each second. It should be noted that a i frequency deviation around a center frequency produces a greater actual change than does a + or a - change of the same amount. In other words, a + 5% change produces an actual 5% change in frequency (i.e., 500 to 525 Hz) whereas a i 5% produces an actual 10% change (i.e., 475 to 525 Hz). The literature relating to warble tones includes numerous references to its use with neonates and young children (Mendel, 1968; Beadle and Crowell, 1962; Peck, 1970; Downs, 1967; Miller and Rabinowitz, 1969; and Liden and Kankkunen, 1969) as well as to its use with adults (Allison Laboratories, Inc., Bulletin A-S; Sivian and White, 1933; Webster, 1952; Dallos and Tillman, 1966 and Young and Harbert, 1970). Recently researchers (Staab, 1971 and Rintel- mann, Orchik and Stephens, 1972) have explored stimulus parameters and established the relationship between warble-tone and pure-tone thresholds for normal-hearing adults. To date, there is little information regarding the response of hard-of—hearing individuals to the warble-tone stimulus. Since both warble tones and narrow-band noise employ a range of frequencies, one might expect responses to the two signals to be similar. Researchers comparing the responses of hearing-impaired individuals to narrow-band noise with responses obtained to pure tones have shown narrow-band noise to yield more sensitive thres- holds (Myers, 1957; Simon and Northern, 1968 and Gengel et al., 1971). Sanders and Josey (1970), however, found no significant differences between the pure-tone and narrow-band noise thresholds of hard-of-hearing children. The above authors have speculated that differences between pure-tone and narrow-band noise thresholds would probably increase for those with a more sharply sloping audiogram. However, none of the studies have considered the effect of audio- metric configuration on thresholds for narrow band noise. Another variable which may affect reSponse to auditory stimuli is age. Older individuals exhibit reduced sensitivity to pure tones and, in many cases, impaired speech discrimination. Since warble tones and narrow-band noise are thought to be more interesting and easier to detect than pure tones, one might speculate that they would hold the attention of the older subject and, as a result, would yield more sensitive and more reliable thresholds. Previous research has not considered the effects of age on reliability of audiometric thresholds or sensitivity to different stimuli. The purpose of the present study was to explore the relation- ship between pure tones, warble tones with i 3% and i 10% fre— quency deviation and narrow-band noise for adults with sensorineural hearing losses. A further purpose was to determine the effect audiometric configuration and age exert on such relationships. Finally, test-retest reliability for warble tones and narrow-band noise were determined. Specific questions asked were: 1. What is the relationship between pure-tone and warble-tone thresholds for adults with sensorineural hearing losses? How do thresholds obtained with i 3% frequency deviation compare to those obtained with t 10% frequency deviation? How do thresholds obtained with narrow-band noise compare to those obtained with pure tones and warble tones? What effect does the audiometric configuration exert on the relationships between pure-tone, warble-tone and narrow- band noise thresholds? What effect does age exert on the relationships between pure-tone, warble-tone and narrow-band noise thresholds? What is the test-retest reliability of thresholds for warble tones and narrow-band noise for subjects with sensorineural hearing losses? Does the test-retest reliability of such thresholds differ as a function of audiometric configuration? Does test-retest reliability of such thresholds differ as a function of age? CHAPTER II REVIEW OF THE LITERATURE A review of the literature pertinent to the present study concerned the two types of stimuli to be employed; warble tones and narrow-band noise. The first section, regarding warble tones, includes the application of warble tones to clinical audiology, comparisons of warble-tone and pure-tone measurements, and evaluation of warble-tone stimulus parameters. In the section on narrow-band noise only literature which relates to use of noise bands as a threshold measure is reviewed. The effects of age on response to auditory stimuli are discussed in a third section. Application of Warble Tones to Clinical Audiology A primary application of warble tones to clinical audiology has been in the evaluation of the hearing of young children. Warble tones have been used to obtain gross reSponses to sound from neonates as well as to obtain threshold reSponses from young children tested in sound-field. Several authors (Mendel, 1968; Beadle and Crowell, 1962; and Peck, 1970) have advocated use of warble tones for testing neonates. For the most part, such studies have been concerned with gross, supra-threshold responses, rather than determination of thresholds. Although a warble tone is utilized in neonatal hearing testing devices, these generally employ a large frequency deviation and a high modula- tion rate which result in the perception of a narrow band of noise, rather than a frequency modulated tone. Downs (1967) described the use of warble tones to determine the mean levels at which infants from 6 weeks to 24 months respond to tones and compared these to the level of reSponse to speech stimuli. The frequency deviation and modulation rate of the tones were not reported. Authors of pediatric audiology texts have recommended warble tones as a means of breaking up standing waves during sound-field testing (Miller and Polisar, 1964) and as a means of making the auditory signal more interesting for the child (Langenbeck, 1965). Studies reporting use of warble tones as a signal for a conditioned response when testing young children in sound-field include those of Huizing (1953), Reilly (1958), Miller and Rabinowitz (1969), and Liden and Kankkunen (1969). References to the use of warble tones for purposes other than pediatric audiology are found primarily in literature published by audiometer manufacturers. Allison Laboratories (Allison Labor- atories, Inc. Bulletin A-S, not dated) reported responses to a ques- tionnaire sent to clinics utilizing Allison equipment with a warble- tone feature. Most respondents (85%) indicated they used warble tones as a stimulus for threshold measurement either with earphones or in sound-field for children, the elderly and individuals with tinnitus. Others mentioned its use in masking, measurement of fre- quency response and gain of hearing aids, and for frequency differ- ence limen tests. Similarly, the Beltone manual (1964) suggested the use of warble tones for those with tinnitus, in sound-field pure-tone testing, and for frequency difference limen tests. Carver (1965) noted the use of warble tones for aided and unaided sound-field threshold measurements. The Madsen Company (Operating Manual for Model 0860) suggested use of warble tones only as a test of differ- ence limen for frequency while Tracor (Specification Sheet for the Allison Model 22 Audiometer) included use in masking and as a sound- field signal. The first published information regarding stimulus parameters for warble tones available on commercial audiometers appeared in 1972 with the survey by Staab and Rintelmann. Responses to a questionnaire completed by 24 audiometer manufacturers in 9 countries revealed that only 25% (6 manufacturers) provided the warble-tone stimulus. The frequency deviations provided varied from + 0.2% above the base frequency to i 10% around the base frequency. Modulation rates provided ranged from 2 to 10 per second. Warble tones were produced by modulating either around or above the base frequency with either a sinusoidal or rectangular wave form. Manufacturers who did not pro- vide warble tones on their audiometers stated there was a lack of demand for such a stimulus and a lack of research demonstrating the significance of warble tones. Comparisons of Warble—tone and Pure-tone Stimuli Several studies, designed primarily for other purposes, have involved comparisons of reSponses obtained with pure tones to those obtained with warble tones. In general, such studies have employed warble tones with one frequency deviation and modulation rate. Studies which have utilized more than one set of stimulus parameters in com- parison with each other and with pure tones are included below. Sivian and White (1933) employed a warble-tone stimulus to determine the upper portion of the monaural minimum audible field. The warble tone was used as the stimulus in order to reduce fatigue and uncertainty on the part of the listener and to eliminate standing wave patterns. Data were obtained at a 0° azimuth for 14 normal- hearing subjects with a warble tone having a modulation rate of 10 per second and a frequency deviation which was reduced in percent from approximately 1 4.6% at 1100 Hz to approximately i 0.97% at 15000 Hz. Check measurements made with the same individual with and without the warble, indicated no systematic differences in threshold values. Webster (1952) compared a group hearing test consisting of sound-field, warble-tone stimuli to a test utilizing recorded warble tones presented through earphones and a pure-tone pulse test des- cribed by Myers et a1. (1948). The warble tones had a modulation rate of five per second and an unspecified frequency deviation. Administration of the three tests to 200 students, in sound-field and under earphones, showed the warble-tone sound-field test to be as reliable as the better of the other two tests. An Audio-frequency Wobbulator utilized by Reilly (1958) allowed frequency deviations of i 0 to i 100 Hz with an unspecified modulation rate. Reilly indicated that less intensity was required to produce a response to a warble tone than to a pure tone of the same frequency. In an attempt to evaluate the changes in results attributable to standing waves, Ross and Duffy (1961) compared pulsing pure tones and pulsing warble tones. Here, warble tones were used as a standard because they do not produce standing waves. Thresholds for both types of stimuli were obtained at 8 frequencies for 25 subjects. Differences were less than 5 dB at all frequencies except 500 and 4000 Hz. The authors concluded that pure tones can be used in sound- field with validity. They advocated sound-field evaluation of hearing aid gain and frequency response in order to obtain a measure of the aid's response as it is worn by the user. Although they did not compare the level of response to pure tones and warble tones, Heron and Jacobs (1968) noted that warble tones were more consistent than were pure tones in eliciting re- sponses from neonates. As a result they designed a warble-tone audiometer with three frequency ranges: 250 to 500, 1000 to 2000, and 4000 to 8000 Hz and frequency deviations up to t 1/3 of the center frequency. Modulation rates could be altered from 1 to 10 per second. Following evaluation of 200 neonates, these authors stated that "pure tones were discarded [p. 77]." They believed the frequency modulated tone is the "ideal test tone" for use with neonates . 10 In another study of neonates, Peck (1970) employed tones with a 5% frequency modulation and recorded environmental sounds in a sound-field. Normative threshold data obtained with normal- hearing adults showed that the reference level for the warble tones approached that of the ISO reference levels. Evaluation of Warble-tone Stimulus Parameters Research cited above has included indirect comparisons of pure-tone and warble-tone thresholds and studies which utilized only one modulation rate and frequency deviation. Such studies were not designed to evaluate the effects of varying the frequency deviation and modulation rate. Specific information regarding these stimulus parameters has been provided by Dallas and Tillman (1966), Young and Harbert (1970), Staab (1971) and Rintelmann, Orchik and Stephens (1972). Dallos and Tillman (1966) presented information obtained from one normal-hearing subject and one patient with an acoustic neurinoma. A tone with a center frequency of 500 Hz modulated at the rate of l, 2, 5, 10 and 25 times per second, with frequency deviations of 10, 63, and 250 Hz (approximately i 1, i 6 and i 25%) served as the stimulus. For the normal—hearing subject, sensitivity improved slightly with increasing frequency deviation. Slower repetition rates also resulted in more sensitive thresholds. For the patient with the acoustic neurinoma, stable thresholds were obtained only with a combination of slow repetition rate and wide frequency deviation. It was suggested that for both normal and II pathological ears, the ratio of frequency deviation to modulation rate, or the modulation index, is the critical variable. For a given modu- lation index, employing a smaller frequency deviation will result in a more sensitive threshold. Thus, a modulation index of 20 achieved with a 100 Hz frequency deviation and a repetition rate of 5 would yield a lower threshold than would the same modulation index of 20 achieved through a 200 Hz frequency deviation and a modulation rate of 10. A similar study using frequency modulated tones with normal and abnormally adapting ears was conducted by Young and Harbert (1970). Thresholds were obtained for a tone with 1000 Hz center frequency, modulation rates of l, 4, 10 and 25 per second and frequency devia- tions of i 10, i 63 and i 250 Hz for four normal listeners and five subjects with sensorineural hearing loss and abnormal adaptation. There was a trend for the normal-hearing subjects to show improved thresholds with increased deviation and for this to be more pro- nounced with decreasing modulation rate. With the abnormally adapting ears, for a given modulation rate, threshold decreased as frequency deviation increased. For a given frequency deviation, threshold decreased as modulation rate decreased. The authors agreed with Dallos and Tillman that for normal-hearing subjects, for a given modulation index, better thresholds are obtained with smaller fre- quency deviations; and for those with abnormal adaptation, slow modulation rate and large frequency deviation are necessary for stable thresholds. 12 A subsequent study by Staab (1971) compared thresholds ob- tained with various combinations of modulation rate and frequency deviation. Three normal-hearing subjects were presented warble tones with frequency deviations of i l, 3, 6, 10 and 50% and modulation rates of l, 2, 4, 8, l6 and 32 per second in random combinations. Repeated thresholds were obtained for the octave frequencies from 250 through 8000 Hz. Warble-tone combinations up to and including frequency deviations of i 10% and modulation rates up to 32 per second were within i 5 dB of pure-tone thresholds. Changes in fre- quency deviation exerted a greater influence on threshold than did changes in modulation rate. In general, for a given frequency deviation, a lower modulation rate yielded a better threshold. For the combinations utilized by Staab, modulation index had no syste- matic effect on dB difference scores. Rintelmann, Orchik and Stephens (1972) conducted two experi- ments comparing pure-tone and warble-tone thresholds. The first study concerned thresholds obtained with warble tones with i 3% and i 10% frequency deviations under earphones, warble tones with i 3% and i 10% frequency deviations presented in sound-field, and pure tones under earphones. Agreement of pure-tone and warble-tone thresholds obtained under earphones was within i 5 dB for both amounts of frequency deviation. Sound-field thresholds for the i 3% and i 10% warble tones agreed within 1 3 dB. Comparison of earphone and sound-field thresholds revealed that sound-field thresholds were from S to 14 dB more sensitive with the least differences in the mid-frequency (500 to 2000 Hz) range. 13 The second study was an investigation of the effects of azi- muth and occlusion of the nontest ear. Here, thresholds were obtained with warble tones with t 3% and i 10% frequency deviation under the fbllowing conditions: 0° azimuth, nontest ear unoccluded; 90° azi- muth, nontest ear unoccluded; and 90° azimuth, nontest ear occluded. Thresholds were also obtained for pure tones under earphones. Results indicated that occlusion or nonocclusion of the nontest ear and use of 90° or 0° azimuth have little effect on threshold (less than 1 5 dB). In addition, test-retest reliability of warble tones was shown to be comparable to that of pure tones. Narrow-Band Noise as a Stimulus Like warble tones, narrow-band noise signals have been used as a substitute for pure tones. Such signals also eliminate the problem of standing waves, reduce irregularities in the earphone response curve, and appear to be easier stimuli to detect. Research relating to threshold measurements with narrow-band noise signals has included comparison of pure-tone and narrow-band noise thresholds for hard-of-hearing adults (Myers, 1957 and Simon and Northern, 1966) and children (Sanders and Josey, 1970 and Gengel et al., 1971). The earliest study comparing noise bands to pure tones was that of Myers (1957) who hypothesized that acuity for noise repre- sented an average of thresholds for the frequency region included in the band, and as such, would provide a better picture of the indi- vidual's hearing than would discrete frequency pure-tone thresholds. Myers obtained thresholds for noise bands of 400 to 630, 800 to 14 1250, 1500 to 2500 and 3200 to 5000 Hz and for pure tones of 500, 1000, 2000 and 4000 Hz for 12 subjects with hearing losses. He did not indicate whether the losses were conductive or sensorineural, degree of loss, or audiometric configuration. Correlations between pure tones and noise bands ranged from .94 at 1000 Hz to .75 at 4000 Hz. Differences were greater at the higher frequencies. Test- retest reliability for noise bands was reported as equivalent to that for pure tones. Myers concluded that noise bands and pure tones were interchangeable for lower frequencies, but that for higher fre- quencies the two measurements differed. He felt there was no ad- vantage to be gained from the use of noise bands. In a later study, Simon and Northern (1966) examined thresh- old measurements with narrow-band noise and pure-tone signals for normal-hearing and sensorineural hearing loss subjects. Subjects with sensorineural hearing losses were divided into two groups accord- ing to whether their Bekesy tracing patterns were classified as Type I or Type II. For the sensorineural group, data for only two frequencies, 2000 and 4000 Hz was presented. While normal-hearing subjects showed no significant differences in sensitivity between narrow-band noise and pure tones from the center of the band, those with sensorineural hearing losses had more sensitive thresholds when tested with noise. Simon and Northern ascribed this to the fact that noise bands sample a wider area of the basilar membrane than do pure tones and represent either an average sensitivity to fre- quencies within the band, or sensitivity at the most acute frequency. 15 Two studies, Sanders and Josey (1970) and Gengel et a1. (1971) have applied narrow-band noise to the evaluation of hard-of—hearing children. Sanders and Josey determined validity and reliability of narrow bands of noise as a test stimulus for hard-to-test (mentally retarded) patients. Validity of narrow-band noise signals was demon- strated through comparison of narrow-band noise thresholds with pure-tone thresholds for a group of hard-of-hearing preschool children. Differences between thresholds obtained with the two types of stimuli were not significant and correlations were high at all frequencies. Reliability of the noise stimulus was determined by comparison of test-retest data for a group of mentally retarded youngsters. Again, no significant differences were found for the two stimuli. The authors noted, however, that all subjects had relatively flat audiograms, and they questioned whether greater differences would exist for an individual with a precipitous drOp in sensitivity. In such a case, one would expect the noise bands to produce a lower threshold than would a pure-tone signal. Sanders and Josey concluded that narrow-band noise audiometry is a valid and reliable method of assessing auditory sensitivity and suggested that it be used with hard-to-test patients. In developing a technique to evaluate the frequency response of hearing aids through use of narrow bands of noise, Gengel et a1. (1971) compared unaided pure-tone and narrow-band noise thresholds for a group of 62 severely hard-of—hearing children. The audio- metric configuration was not considered. Differences between thresh- olds for narrow-band noise with a bandwidth equal to i 10% of the l6 center frequency, and pure tones ranged from 3.4 dB at 250 Hz to 6.0 dB at 2000 Hz with thresholds for noise more sensitive at all frequencies. Ag: Presbycusis, hearing loss which occurs as a result of ad- vancing age, is known to affect both auditory sensitivity and Speech discrimination. A bilaterally symmetrical reduction in sensitivity to pure tones, greatest for high frequencies, has been described by a number of authors (Beasley, 1938; Bunch, 1929 and 1931; Corso, 1963; and Glorig and Roberts, 1965). Phonemic regression, poorer speech discrimination than would be expected from the amount of pure— tone loss, was first studied by Gaeth (1948) and has since been confirmed by Pestalozzi and Shore (1955); Goetzinger et a1. (1961); Olsen (1965); Harbert, Young and Menduke (1966) and Rintelmann and Schumaier (1973). Other studies of the elderly have investigated the effects of age on pitch discrimination (Konig, 1957), $181 scores (Jerger, Shedd and Harford, 1959), Bekesy tracings (Jerger, 1960), and recruitment (Pestalozzi and Shore, 1955 and Goetzinger et al., 1961). Although investigators have reported variability of responses within the group of individuals, none have discussed individual variability or test-retest reliability of scores obtained from these older individuals. Summary Discussed in the review of the literature were the applica- tion of warble tones to clinical audiology, comparisons of pure-tone 17 and warble-tone stimuli, exploration of warble-tone stimulus para- meters, the use of narrow-band noise as a stimulus, and the effects of age on reSponse to auditory stimuli. The primary application of warble tones has been the evalua— tion of neonates and young children. There has been some mention of the use of warble tones for frequency difference limen tests and for threshold measurement of adults with tinnitus. Comparison of thresholds for normal-hearing subjects obtained with warble tones and with conventional pure tones has shown close agreement when both signals are presented under earphones. Greater differences have been found when the warble-tone thresholds are obtained in a sound field. To date, only limited information is available concerning the response of pathological ears to warble tones. Such data has been obtained with abnormally adapting ears and has been primarily concerned with varying the frequency deviation and modulation rate in order to obtain a stable threshold. None of the studies have investigated the relationship between pure-tone and warble-tone thresholds for subjects with sensorineural hearing loss. Research with hard-of—hearing subjects using narrow-band noise as a stimulus has shown noise bands to yield more sensitive thresholds than pure tones. Although one might expect this same relationship to hold for warble tones, it has not been demonstrated. Similarly, previous research has not evaluated the possible effects of age on sensi- tivity for these stimuli. Before warble tones can be meaningfully employed as a threshold measure, the pure-tone and warble-tone 18 relationship must be established for pathological ears. In addition, it should be determined whether audiometric configuration or age influence possible differences between pure-tone and warble-tone thresholds. The present study has been designed to provide such information. CHAPTER III EXPERIMENTAL PROCEDURES Thresholds for pure tones, warble tones with i 3% frequency deviation, warble tones with t 10% frequency deviation and narrow- band noise were obtained for adults with sensorineural hearing losses. Repeat thresholds were also obtained for each of the stimuli so that test-retest reliability could be determined. Sub- jects were divided into two groups based on audiometric configura- tion, and each of these main groups was further subdivided to include two different age groups. Subjects Thirty-two adults (18-69 years of age) with sensorineural hearing losses participated in this study. All subjects had pre- viously been tested at the Audiology Clinic of Michigan State University or at Rehabilitation Medical Center, Sparrow Hospital, Lansing. All subjects had an average loss of at least 25 dB for the frequencies 500, 1000 and 2000 Hz with interweaving air and bone conduction thresholds. Only the better ear for each subject was used in testing. The thirty-two subjects were divided into two categories, 16 per category, based on audiometric configuration. 19 20 These two categories were further divided to contain 8 individuals aged 50 years or less and 8 individuals aged over 50 years. The categories were defined as follows: Gradual Configuration. Individuals in this category had a flat or gradually sloping audiogram with the 1055 beginning in the low frequencies. The threshold at 500 Hz was 25 or greater; changes by octave were 10 dB or less with the exception that one octave could change by 15 dB and the difference between the highest and lowest threshold for the octave frequencies 250 through 4000 Hz was no greater than 35 dB. In addition, no threshold exceeded 90 dB. Sharp Configuration. Individuals in this category had normal or near normal hearing in the low frequencies, with a sharp drOp in threshold between 500 and 1000 Hz or between 1000 and 2000 Hz. The threshold at 500 Hz was 30 dB or better; there was at least a 20 dB drop between the thresholds for 500 and 1000 Hz, or between thresholds for 1000 and 2000 Hz; and the difference between the highest and lowest threshold for the octave frequencies from 250 to 4000 Hz was 40 dB or more. No threshold was greater than 90 dB. The mean age, age range, and audiometric data for each of these categories is displayed in Table 1. Figure 1 shows the audiometric configuration for each group. 21 00.00 50.00 00.50 00.00 00.50 50.00 0: 000 00 000000000 0:0 00 0000 00 000gmounu coozuon mucouommfiv :00: 00.05 00.00 00.00 00 00 00.50 00.00 N: 0000 00.05 00.05 00.05 50.00 00.50 50.00 0: 0000 00.00 00.00 00.00 00.00 50.00 50.00 0: 0000 50.00 50.00 50.00 00.00 50.00 00.00 0: 0000 50.00 50.00 50.00 00.50 00.00 05.00 0: 000 00.00 50.00 00.00 00.00 05.00 00.00 00 000 Ho>og weaken: ado: 00 0 0 00 0 0 00000000 00 000002 50.00 00.00 00.00 05.00 00.00 00.00 000 000: 00-00 00-00 00-00 00-00 00-00 00-00 00000 00< @3000 0000» 0006» @5000 0006» 0000» 00000 00 A 00.w 00000 00 A 00.w coflumusmfimcou mhmnm cofipmusmflmcou 0036000 .xvspm 00Eu 00 00000960 muoonnsm pom um omm um 000cmoucu 0:0 N: 0000 um vfionmounu coozuon coconommfiv name new socoscoum x9 Ho>oH mcfiumo; cams .oma :moa .omcwu om cowuoasm umog zoom qoxkzou 25 Beat-frequency Oscillator. A Bruel and Kjaer (BGK) Beat- frequcncy Oscillator model 1013 was used as the source of the warble tones. The oscillator, tOgethcr with the Function Generator, allowed for modulation of the output signal. Function Generator. The frequency of the sine wave genera- tion produced by the Hewlitt-Packard model 3310A Function Generator determined the modulation rates and its output voltage was instru- mental in obtaining the desired frequency deviation. Voltmeter. A Bruel and Kjaer type 2409 Electronic Voltmeter allowed find adjustments of the output voltage of the Function Generator. Frequency Counter. The Bekman Eput and Timer, model 6148, was used initially to determine the accuracy of frequency of modula- tion rates and during the experiment to monitor the base or center frequencies of the warble-tone stimuli. Test Stimuli Stimuli consisted of pure-tone (PT), warble-tone (WT) and narrow-band noise (NBN) signals at, or with a center frequency of, 500, 1000, 2000 and 4000 Hz. For the warble-tone signals, fre- quency deviations of i 3% and i 10% and a constant modulation rate of 8 modulations per second were employed. The frequency deviations were accomplished using a sinusoidal waveform. Settings required to produce the correct frequency deviations are shown in Appendix A. 26 The narrow-band noise was produced by the Maico MA-24 audiometer. The bandwidth and rejection rates for each test frequency are in- cluded in Appendix A. Calibration Calibration of all test equipment took place at the beginning and end of the experiment. The Maico MA-24 audiometer was cali- brated, or checked, at each test frequency for fundamental frequency, harmonic distortion, rise and decay time, and sound pressure level (SPL) output. Attenuator linearity was checked at 1000 Hz and a frequency response curve was obtained for each earphone. Calibra- tion was consistent with the ANSI 1969 specifications. Measurements of the center or base frequency of the warble tones (output of the beat-frequency oscillator without introduction of frequency modulation) included base frequency, SPL output and distortion. Such measurements were made at the earphone after routing the signal through the accessory input of the Maico MA-24 and calibrating to 0 "VU" in a manner recommended by ANSI. The narrow-band noise, as produced by the Maico MA-24 audio- meter, was analyzed pre- and post-experimentally. The acoustic spectrum of the noise is displayed in Appendix A. The SPL output of the Maico MA-24 for pure tones and narrow- band noise, and the output of the beat-frequency oscillator (source of the warble tone) were determined prior to each day's testing. Specific calibration data is presented in Appendix A. No systematic changes were feund in the calibration of the equipment during the course of the investigation. 27 Procedure The procedure for each subject was as follows: 1. Pure—tone air conduction thresholds were obtained for both ears at octave frequencies 250 through 8000 Hz. Bone conduc- tion thresholds were obtained for the better (test) ear. These measurements determined whether the subject met the criteria for inclusion in the study. 2. The right earphone was placed on the test ear and thresholds were obtained for pure tones, warble tones with i % and i 10% frequency deviation and for narrow-band noise. For each of the 3 types of stimuli test frequencies, or center frequencies, were 500, 1000, 2000 and 4000 Hz. 3. After the above measurements the subject was given a ten- minute rest period. Repeat thresholds were then obtained for pure tones, warble tones and narrow-band noise. All thresholds were obtained using a 5 dB step of attenuation and the modified, ascending, Hughson-Westlake technique as described by Carhart and Jerger (1959). The order of presentation of frequencies and the order of presentation of stimuli (pure tones, warble tones with i 3% frequency deviation, warble tones with i 10% frequency deviation and narrow-band noise) were randomly determined. However, the order of presentation of frequencies and stimuli during the repeat test session was the same as that employed during the initial session. CHAPTER IV RESULTS AND DISCUSSION This chapter presents the results of the comparison of reSponses of adults with sensorineural hearing losses to the four test signals. Also included is a discussion of the findings and the implications for clinical audiology. Results The data were subjected to a multifactor analysis of variance. Significant effects were associated with the four signals, test- retest scores, the two audiometric configurations, and the four test frequencies. There were no significant effects associated with the age factor. A complete analysis of variance table can be found in Appendix B. With the exception of the sections of this chapter relating Specifically to test-retest comparisons, data for both test and retest have been combined in presenting results. Justification for combining test and retest scores will be presented later. All thresholds were converted to SPL re 0.0002 dynes/cm2 to allow a direct comparison of signals. 28 29 Comparison of Thresholds for Pure Tones, Warble Tones and Narrow-band Noise A significant main effect was associated with the signal factor (F = 65.87; df = 3, 84; p < 0.0005). Also there were signi- ficant interactions between signal and frequency (F = 26.11; df = 9, 252; p < 0.0005), configuration and signal (P = 43.54; df = 3, 84; p < 0.0005) and configuration, signal and frequency (F = 19.94; df = 9, 252; p < 0.0005). When scores for all subjects were averaged across all fre- quencies, there was a significant difference between the mean scores for the four test signals. The mean score for each signal is shown in Table 2. It can be seen that differences, with all frequencies combined, were slight. The mean thresholds for i 3% (52.2 dB) and i 10% (50.8 dB) warble tones were both within 2 dB of the mean threshold for pure tones (52.7) and the mean threshold for narrow-band noise (46.1) was within 6.6 dB of the threshold for pure tones. TABLE 2.--Mean threshold in dB SPL for each test signal. Thresholds for both test and retest, for all subjects and all fre- quencies are combined. Each mean is based on 256 thresholds. Signal Mean Threshold Pure tone 52.7 Warble tone i 3% 52.2 Warble tone 1 10% 50.8 Narrow-band noise 46.1 30 The signal by frequency interaction is shown in Table 3 and Figure 3. As shown, the thresholds at 500 and 1000 Hz were similar for all signals. At 2000 and 4000 Hz thresholds for narrow-band noise differed markedly from those for the other three signals. At 2000 Hz the mean threshold for narrow-band noise was 11.8 dB more sensitive and at 4000 Hz, 12.3 dB more sensitive than the mean threshold for pure tones. Other signals were within t 5 dB of thresholds for pure tones. TABLE 3.--Mean thresholds in dB SPL by frequency for each test signal. Each mean is based on all subjects for both test and retest (64 thresholds). Signal Frequency Pure Tones WT i 3% WT i 10% NBN 500 Hz 37.4 37.4 37.5 37.0 1000 Hz 45.1 44.4 43.3 43.2 2000 Hz 60.9 59.8 58.1 49.1 4000 Hz 67.6 67.0 64.2 55.3 Interactions between configuration and signal and between configuration, signal and frequency are discussed in the following section. Effect of Configuration Significant interactions occurred between configuration and frequency (F = 59.2; df = 3, 84; p < 0.0005) and configuration and signal (P = 43.54; df = 3, 84; p < 0.0005). In addition, the 31 O—O NBN 30 *- ' ' WT i” 3°’o o__--. WT i 10% NE 40 '- U \ U) 0 E. "U N 8 <2 50 - O 0 H 6.1 Oz U) 93 60- 70. 1 n L l 500 1000 2000 4000 Frequency in Hz Figure 3.--Mean thresholds in dB SPL by frequency for each test signal. Each mean is based on all subjects fbr both test and retest (64 thresholds). 32 three-way interaction between configuration, signal and frequency was significant (F = 19.94; df = 9, 252; p < 0.0005). The mean threshold for all signals combined at each frequency for the gradual and sharp groups are shown in Table 4 and Figure 4. Thresholds for subjects in the gradual group showed a small (5.3.1 dB per octave) drop across frequencies, while thresholds for those in the sharp group showed a large ( > 10 dB per octave) drop across frequencies. Further, thresholds for the sharp group were more sensitive than thresholds for the gradual group at the lower fre- quencies (500 and 1000 Hz) and less sensitive at the higher fre- quencies (2000 and 4000 Hz). TABLE 4.--Mean threshold (in dB SPL) by frequency for subjects grouped by audiometric configuration. Data for all signals on both test and retest are combined. Each mean is based on 128 thresholds. Configuration Frequency Gradual Sharp 500 Hz 47.4 27.2 1000 Hz 50.5 37.5 2000 Hz 53.5 60.4 4000 Hz 54.0 73.1 Table 5 displays the mean thresholds for each signal averaged across frequency for the two groups of subjects. Mean thresholds for all four signals for the gradual group were within a 1.3 dB range (52.0 to 50.7 dB). For the sharp group, the mean thresholds 33 20 t 0———————o Gradual o \. 30 p \\ O———O Sharp \. \ ‘\ \o \ 40 ~ \ \ \ U1 0 l O / / dB SPL re 0.0002 dynes/cm2 O‘ O I o/ 70. \ J l l l 500 1000 2000 4000 Frequency in Hz Figure 4.--Mean thresholds in dB SPL by frequency with subjects grouped by audiometric configuration. Test and retest thresholds for all signals are combined. Each mean is based on 128 thresholds. 34 for pure tones and for the two warble-tone signals were within a 2.7 dB range (50.7 to 53.4 dB) while the mean threshold for the most sensitive measure (41.6 dB) narrow-band noise, differed from that for the least sensitive measure (53.4 dB) pure tones, by 11.8 dB. TABLE S.--Mean threshold (in dB SPL) for each signal with subjects grouped by audiometric configuration. Each mean is based on an average of all frequencies for both test and retest (128 thresholds). Configuration Signal Gradual Sharp Pure tones 52 .o 53.4 Warble tones i 3% 51.8 52.5 Warble tones i 10% 50.9 50.7 Narrow-band noise 50.7 41.6 The interaction between the factors of configuration, signal and frequency is shown in Tables 6 and 7, and Figures 5 and 6. The mean thresholds for the gradual group, at any given frequency, were within 3.2 dB for all signals. For the sharp group, thresholds at 500 and 1000 Hz were within 4.4 dB for all signals; however, at 2000 and 4000 Hz, warble-tone and pure-tone thresholds agreed within 3.8 dB while the mean threshold for narrow-band noise differed from the mean threshold for pure tones by 20.7 dB at 2000 Hz and by 21.2 dB at 4000 Hz. Figures 5 and 6 display the differences between pure tones and each of the other three signals at each frequency for the two groups of subjects. 35 TABLE 6.--Mean thresholds in dB SPL by frequency for each of the four test signals with subjects grouped by audiometric configura- tion. Data for test and retest are combined. Each mean is based on 32 thresholds. Signal Configuration Frequency Pure Tone WT i 3% WT i 10% NBN Gradual 500 Hz 47.7 47.2 47.8 46.9 1000 Hz 50.1 50.7 50.4 50.6 2000 Hz 54.7 54.3 52.8 52.1 4000 Hz 55.6 55.1 52.4 52.9 Sharp 500 Hz 27.0 27.6 27.1 27.0 1000 Hz 40.1 38.2 36.2 35.7 2000 Hz 66.8 65.3 63.4 46.1 4000 Hz 79.8 78.8 76.0 57.6 TABLE 7.--Difference between mean threshold for pure tones and mean threshold for each of the other three test signals at each frequency with subjects grouped by audiometric configuration. Signal Configuration Frequency WT i 3% WT i 10% NBN Gradual 500 Hz - .5* .1 - .8 1000 Hz .6 .3 .S 2000 Hz - .4 -1.9 - 2.6 4000 Hz - .S -3.2 - 2.7 Sharp 500 Hz .6 .1 0.0 1000 Hz -l.9 -3.9 - 4.4 2000 Hz -1.5 -3.4 -20.7 4000 Hz -1.0 -3.8 -22.2 *Negative values indicate thresholds for pure tones were less sensitive than thresholds for comparison signals. 36 Threshold re threshold for pure tones c> 0.0 l I I l / // O O 1 l 1 l 500 1000 2000 4000 Frequency in Hz Figure S.--Comparison of mean thresholds for warble tones with i 3% and i 10% frequency deviation and narrow—band noise with mean thresholds for pure tones for subjects with gradual audio- metric configurations. Base line represents threshold for pure tones. Negative values indicate thresholds for pure tones were less sensitive than thresholds for comparison signals. Threshold re threshold for pure tones 37 + 5 -15 ~ -20 r- -25 J l J l 500 1000 2000 4000 Frequency in Hz Figure 6.--Comparison of mean thresholds for warble tones with i 3% and i 10% frequency deviation and narrow-band noise with mean thresholds for pure tones for subjects with sharp audiometric configurations. Base line represents thresholds for pure tones. Negative values indicate thresholds for pure tones were less sensitive than thresholds for comparison signals. 38 Inspection of the reSponses of individual subjects at 2000 and 4000 Hz revealed that for subjects in the gradual group, only 8 of the 64 thresholds (both test and retest) for the narrow-band noise differed from thresholds for pure tones by more than 5 dB. However, in the sharp group only 3 of the 64 thresholds were within 5 dB and only 8 were within 10 dB of thresholds for pure tones. The majority of the narrow-band noise thresholds (53) were more sensitive than pure-tone thresholds by 10 to 49 dB. For both groups of subjects, all thresholds for warble tones were within 10.3 dB of thresholds for pure tones. Table 8 displays the differences between thresholds for pure tones and thresholds for the other three signals for the subjects in the sharp group. A similar comparison is not shown for subjects in the gradual group since differences between their pure- tone, warble-tone and narrow-band noise thresholds were negligible as shown in Figure 5. As stated above, subjects in the gradual group showed a similar response to all signals, while subjects in the sharp group demonstrated more sensitive thresholds for narrow-band noise. Further, the configuration-signal-frequency interaction demonstrates that such differences occur only at the higher frequencies (2000 and 4000 Hz). Threshold as a Function of Frequency Significant effects involving frequency were the main effect (F = 104.93; df = 3, 84; p < 0.0005), the interaction TABLE 8.--Comparison of thresholds for warble tones with i 3% and i 10% frequency deviation and narrow-band noise to thresholds for pure tones at 2000 and 4000 Hz for subjects with sharp audiometric configurations. presented for initial test only. Thresholds are Drop in hearing 2000 Hz 4000 Hz level from 500 to 4000 Hz WT i 3% WT t 10% NBN WT i 3% WT i 10% NBN 75 - 5.0* -10.0 -34.5 - 4.5 —4.5 -49.0 70 - .l - 5.1 -28.3 0.0 -5.0 -32.S 70 + 4.5 - .S -14.2 + 4.8 - .2 -23.2 60 - 4.7 - 4.7 -38.2 + .8 -9.2 -37.5 60 0.0 0.0 -14.0 - 4.2 -9.2 -22.2 60 + .5 + .5 -28.7 + .7 -4.3 -32.8 55 0.0 + .3 —23.2 - 3.8 -3.8 -l7.5 55 - 4.3 - 4.7 -18.7 + 1.0 -4.0 -27.6 55 0.0 0.0 -23.3 + .3 -4.7 -l7.3 55 -10.0 - 5.0 -23.5 + .6 +5.6 - 7.2 50 - 4.7 - 9.7 -23.2 +10.3 +5.3 - 2.5 50 - 4.5 - 9.5 -l3.7 - 4.5 -4.5 -17.5 50 + .5 - 4.5 -23.7 - 4.4 + .6 -l7.5 45 + .5 - 4.5 -13.7 + .5 -4.4 -12.5 45 + .5 — 4.5 -l3.S 0.0 -S.0 -l7.S 40 + .3 + .3 + .l + .3 + .3 + 1.8 *Negative values indicate thresholds for pure tones were less sensitive than thresholds for comparison signals. 40 between signal and frequency (F = 26.11; df = 9, 252; p < 0.0005), and the interaction between configuration, signal and frequency (F = 19.94; df = 9, 252; p < 0.0005). A comparison of the mean thresholds for all subjects (gradual and sharp configurations) with all signals combined, showed a significant difference by frequency as seen in Table 9. Inspection of this table shows that threshold sensitivity decreased with in- creasing frequency. Examination of the configuration by frequency interaction shows this to be due primarily to the differences across frequencies for the sharp group. The interactions between signal and frequency and between configuration, signal and frequency which have been reported above indicate that differences occur for narrow- band noise at the higher frequencies for the sharp group. TABLE 9.——Mean threshold in dB SPL for four test frequencies with test and retest data for all subjects and all signals combined. Each mean is based on 256 thresholds. Frequency Mean Threshold 500 Hz 37.3 1000 Hz 44,0 2000 Hz 56.9 4000 Hz 63.5 Test-retest Comparison When thresholds for all subjects were averaged across signals and frequencies, the mean score on initial test was 50.9 and that on repeat test, 50.0 dB. This difference was significant at greater 41 than the 0.0005 level (F = 19.15; df = 1, 28). The lack of inter- action between test-retest and other factors indicates that the slight improvement on the repeat test was consistent for both configurations and both age groups, as well as for all signals and all frequencies. A comparison of test-retest data is presented in Table 10. The table shows that all test-retest differences were slight (.5 to 1.3 dB). To examine the test-retest data more closely, a rank order correlation was computed for the mean score for each subject on test and retest. Scores on initial test showed a +.97 correlation with scores on repeat test. Influence of Age Age was not significant as a main effect, nor was the age factor involved in any significant interaction. Table 11 displays the mean thresholds for the two age groups for each of the four signals as a function of frequency. The small differences in thresh- old for the two age groups did not reach significance. As can be seen in Figure 7, both age groups showed slight improvement on retest and this improvement was consistent for all test signals and all frequencies. Discussion The most important findings in the present study relate to the differences between thresholds for the four test signals, the changes in the relationship between these signals across frequencies and the effect audiometric configuration exerts on these relationships. 42 TABLE 10.--Mean threshold on initial test compared with mean threshold on repeat test for each of the factors studied. Improvement * Factor Category Test Retest on Retest N Configuration Gradual 51.8 50.9 .9 256 Sharp 50.0 49.1 .9 256 Age 5, so years 51.9 51.1 .8 256 > 50 years 49.9 48.8 1 1 256 Signal Pure tones 53.1 52.3 .8 128 WT r % 52.6 51.7 .9 128 WT t 10% 51.2 50.3 .9 128 NBN 46.6 45.6 1.0 128 Frequency 500 Hz 37.9 36.6 1.3 128 1000 Hz 44.6 43.4 1.2 128 2000 Hz 57.3 56.6 .7 128 4000 Hz 63.8 63.3 .5 128 Composite ------- 50.9 50.0 9 512 *N indicates the number of thresholds on which each mean is based. 43 30 r . o :_50 years Test 0—-0 < 50 years 8\\ " Retest > 50 years Test A O T > 50 years Retest tn 0 I dB SPL re 0.0002 dynes/cn2 O‘ O I 70 . 500 1000 2000 4000 Frequency in Hz Figure 7.--Mean test and retest thresholds in dB SPL at each test frequency for the two age groups with all signals combined. Each mean threshold is based on 128 thresholds. 44 TABLE 11.--Mean threshold (in dB SPL) by frequency for each of the four test signals with subjects grouped by age. Data for test and retest are combined. Each mean is based on 32 thresholds. Signal Age Frequency 0 0 Pure Tone WT + 36 WT + 106 NBN 5.50 years 500 Hz 38.6 37.6 38.7 38.1 1000 Hz 47.4 45.8 45.1 45.8 2000 Hz 63.0 62.1 60.8 50.2 4000 Hz 66.3 66.5 63.7 54.8 > 50 years 500 Hz 36.1 37.3 36.3 35.8 1000 Hz 42.8 43.0 41.5 40.6 2000 Hz 58.5 57.6 55.4 48.0 4000 Hz 69.0 67.4 64.8 55.8 The fact that subjects in the two groups represented different audiometric configurations is demonstrated by the significant inter- action between configuration and frequency. The main effect for frequency and the interaction between configuration and frequency both reflect the basic audiometric characteristics of the subjects and the criteria for subject selection. Differences by frequency reflect the fact that subjects showed increasingly greater hearing loss with increasing frequency. As would be expected according to the criteria for subject selection, those in the sharp group showed greater differences by frequency than did those in the gradual group. The interaction between configuration and frequency, then, demon- strates that thresholds for the two groups differed by frequency with the sharp group showing less hearing loss at low frequencies and more at high frequencies. 45 The main effect for signal shows that thresholds differed as a function of the type of signal utilized. It is evident from the results that the major source of variation was narrow-band noise. A review of the configuration-signal interaction, as displayed in Table 5, indicates that this difference was seen for the sharp group only, and analysis of the three-way interaction between configura- tion, signal and frequency as shown in Table 6 and Figures 5 and 6 shows that marked differences were seen between narrow-band noise and the other three signals at 2000 and 4000 Hz for the sharp group. Table 7 reveals that differences between all test signals were 3.2 dB or less for those in the gradual group. For the sharp group, thresholds for pure tones and both warble-tone signals were within 3.9 dB at all frequencies while those for narrow-band noise were within 4.4 dB of pure-tone thresholds at 500 and 1000 Hz only. At 2000 and 4000 Hz narrow-band noise differed from pure tones by over 20 dB. The close agreement between warble tones and pure tones for subjects with sensorineural hearing loss compares favorably with the results obtained for subjects with normal hearing by Rintelmann, Orchik and Stephens (1972). The relationship between pure-tone and warble-tone thresholds is not influenced by audiometric con- figuration. However, the relationship between pure tones and narrow- band noise is affected by configuration. The data for the gradual group in the present study also compares favorably with results of studies utilizing hearing-impaired subjects described as having relatively "flat" audiometric 46 configurations. Both Gengel (1971) and Sanders and Josey (1970) reported close agreement between pure-tone and narrow-band noise thresholds for children with "flat" sensorineural hearing losses. The present results for the sharp group agree with results reported by Simon and Northern (1966). In their study of individuals with sensorineural hearing losses and type I and II Bekesy tracings, they found more sensitive thresholds for narrow-band noise than for pure tones at 2000 and 4000 Hz. The present results also agree with those of Myers (1957) who concluded that noise bands were inter- changeable with pure tones for the lower frequencies, but that for the higher frequency region the two measurements differ. For subjects in the sharp group, the greatest difference between pure tones and narrow-band noise was seen for the subject with the most precipitous change in threshold from 500 to 4000 Hz. The least difference between pure tones and narrow-band noise was seen for the individual with the smallest change in thresholds from 500 to 4000 Hz. Although there was a tendency for those with the most change in threshold across frequencies to show the most difference between pure tones and narrow-band noise at 2000 and 4000 Hz, sufficient variation in this trend prevents use of amount of change in pure-tone threshold as a means for predicting the pure-tone versus narrow-band noise differences. It has been hypothesized (Simon and Northern, 1966 and Myers, 1957) that thresholds for narrow-band noise represent either average acuity for the band or thresholds at the most sensitive frequency within the band of the noise. Judging from the large 47 differences between thresholds for pure tones found among subjects in this study with a sharp audiometric configuration, it would appear that the response to narrow-band noise may occur to fre- quencies contained within the noise, but outside the bandwidth proper. Table 13 and Figures 3 through 6 of Appendix A display the pre- and post—experimental analysis of the narrow-band noise. It can be seen that the rejection rate was 20 dB for the first octave. Given this spectrum, those subjects with sharp audiometric con- figurations may respond to frequencies outside the bandwidth of the noise at levels well below the pure-tone threshold for the test frequency. To explain, an individual with a large difference in threshold across frequencies may respond to low frequency energy which is present in the noise band (outside the bandwidth) at an intensity above his threshold for the lower frequencies. Thus, a response to narrow-band noise may occur even when the intensity of the noise is not sufficient to reach threshold for the center, or test, frequency. The results of this study also demonstrate that there is a slight learning effect from test to retest. Inspection of Table 10 shows that an improvement on retest of from .S to 1.3 dB was seen for all factors investigated. It has been demonstrated that variables other than physiological change in threshold, such as placement of the earphone, may lead to threshold variability of up to 5 dB (Steinberg and Munson, 1936; Munson and Wiener, 1950; and Myers and Harris, 1949). Studies of repeat threshold measurements on clinical populations (Bunch, 1930; Witting and Hughson, 1940; 48 Carhart and Hayes, 1949; Carhart and Jerger, 1959 and Jerger, 1962) have demonstrated that test-retest agreement of pure tones is within one step of the audiometer attenuator dial, namely i 5 dB. Thus, clinical audiologists generally accept agreement within 1 5 dB as good clinical reliability. Rintelmann, Orchik and Stephens (1972) in a study of normal— hearing adults found comparable test-retest reliability ( i 5 dB) for warble tones with i 3% and t 10% frequency deviation. The present study indicates that this agreement holds for individuals with sensorineural hearing loss. The absence of interaction between test-retest and the factors of age and configuration shows that both age groups and both configuration groups exhibited a slight improvement on repeat measurement. Similarly, the absence of interaction between signal and test-retest shows that no signal was superior to the others in terms of test—retest reliability. The high correlation (+.97) between scores on initial and repeat tests indicates a high level of test-retest consistency. The lack of significant effects or interactions for the age factor suggests that at least for the age range included in this study, age does not affect response to the four test stimuli, nor does it affect the test-retest performance. Since all subjects in this study were required to come to the Audiology Clinic to participate, only older individuals in good health were included. It is possible that older adults, beyond the age included in the present study, or those in poor health may show differences in 49 their responses to the four test signals which were not detected in the present investigation. Clinical Implications Based on the relationship between pure tones, warble tones and narrow-band noise, one must conclude that warble tones are a better alternative to pure tones than are narrow bands of noise. The audiologist with both warble tones and narrow-band noise available to him would be wise to utilize warble tones, rather than narrow-band noise, since warble-tone thresholds agree more closely to thresholds for pure tones. This is particularly true when the individual is likely to have a marked drop in threshold from low to high frequencies. An example of a situation in which use of narrow-band noise could lead to a false picture of the hearing loss is in the evalua- tion of a child whose hearing loss is related to Rh incompatibility. The audiograms obtained by Matkin and Carhart (1968) in their study of 22 children with hearing loss due to Rh incompatibility, showed near normal hearing for low frequencies, a drop of approximately 20 dB per octave from 250 to 1000 Hz and a loss of approximately 80 dB at the higher frequencies. Since these audiogram character- istics resemble those of subjects in the sharp group of the present study, one could expect that thresholds obtained by narrow-band noise would be more sensitive than those obtained by pure tones. As a result, a child evaluated with narrow-band noise would appear to have less hearing loss at high frequencies, and a flatter 50 audiometric configuration than if he had been evaluated using either pure tones or warble tones. Another area in which use of narrow-band noise could yield misleading results is in the evaluation of hearing aid gain through comparison of aided and unaided thresholds. Here, thresholds at any frequency may be influenced by changes in thresholds at adjacent frequencies giving a false measure of the hearing aid gain. As is true in obtaining basic audiometric thresholds, warble tones again would be preferable to narrow—band noise for evaluating hearing aid gain. Finally, although thresholds obtained with either i 3% or i 10% warble tones show good agreement to thresholds obtained with pure tones, the closest agreement to pure-tone thresholds would be obtained with warble tones utilizing : 3% frequency deviation. CHAPTER V SUMMARY, CONCLUSIONS AND RECOMMENDATIONS Summary Threshold measurements for pure tones, warble tones with i 3% and i 10% frequency deviation, and narrow-band noise were compared using two groups of subjects with sensorineural hearing losses. Sub- jects consisted of 16 adults whose audiograms showed a marked change in threshold across the range of frequencies (sharp configuration) and 16 adults whose audiograms showed slight change in threshold across frequencies (gradual configuration). Each of the two groups contained 8 subjects fifty years of age or younger and 8 subjects over fifty years of age. The test stimuli were presented at, or with a center frequency of, 500, 1000, 2000 and 4000 Hz. Following the initial measurements, repeat thresholds were obtained to allow analysis of test-retest reliability. Results demonstrated a small but consistent learning effect for both subject groups and all stimuli, with a mean improvement on retest of 0.9 dB. Comparison by age showed no significant differ— ences between the two age groups studied. Thresholds for those with a gradual configuration were within 1 3.2 dB for all stimuli. Thresholds for those with a sharp configuration were within i 3.9 dB 51 52 for pure tones and warble tones. Thresholds for the sharp group for the narrow-band noise agreed with thresholds for pure tones within i 4.4 dB at 500 and 1000 Hz, while thresholds for narrow-band noise at 2000 and 4000 Hz were over 20 dB more sensitive than thresholds for pure tones. Conclusions The following conclusions were drawn: 1. Warble-tone and narrow-band noise thresholds were equal to pure-tone thresholds in terms of test-retest reliability. 2. Neither audiometric configuration or age influenced test- retest reliability for the four test signals. 3. For subjects whose audiograms showed relatively little change in threshold across the range of frequencies (gradual configuration), both warble tones and narrow—band noise yielded thresholds which were in good agreement (3.2 dB) with thresholds for pure tones. 4. For subjects whose audiograms showed marked changes in threshold across the range of frequencies (sharp configura- tion), thresholds for warble tones agreed closely (i 3.9 dB) with those for pure tones. Thresholds for narrow-band noise agreed (i 4.4 dB) with those for pure tones only at the lower frequencies (500 and 1000 Hz), while narrow-band noise produced markedly more sensitive (> 20 dB) thresholds than pure tones at the higher frequencies (2000 and 4000 Hz). 53 5. Differences between narrow-band noise thresholds and pure- tone thresholds were sufficient to preclude use of narrow— band noise as a threshold measurement stimulus for individuals with sharp audiometric configurations. 6. Thresholds obtained via warble tones with i 3% frequency deviation agreed more closely to pure-tone thresholds than did thresholds obtained with warble tones using 1 10% frequency deviation. 7. Age of subjects did not affect the response to the four test stimuli. Recommendations The present study has demonstrated that audiometric config- uration does not significantly affect the relationship between pure tones and warble tones. However, it does affect the relationship between pure tones and narrow-band noise with the differences most apparent at 2000 and 4000 Hz. Subjects in the present study were classified into two broad groups based on overall change in threshold and change per octave. A systematic investigation should be made of responses of subjects with varying drOps in threshold for the octaves from 500 to 1000 and 1000 to 2000 Hz. Evaluation should be made of the effect of different rejection rates for the narrow-band noise to determine the most suitable noise spectrum for threshold measurement. Since thresholds for mid-octave frequencies could affect response to narrow-band noise, it is suggested that in future studies S4 thresholds be measured utilizing Bekesy audiometry to sample more frequencies within the area of audition. In this study age did not emerge as an influence on the relationship between pure tones and the other test signals. It is felt this may have been due, in part, to the requirement that subjects be able to come to the Clinic to participate. In addition, all subjects were under 70 years of age. Study of difficult-to-test geriatric subjects, such as hOSpital patients, aphasics, or residents of nursing homes, may show that one signal is superior to others in terms of maintaining interest and obtaining more sensitive, reliable thresholds. Although there is no reason to expect that children with sensorineural hearing losses should respond to pure tones, warble tones and narrow-band noise differently from adults, this has not been studied. A study, similar to the present one, should be carried out with hard-of—hearing children. REFERENCES REFERENCES Allison Laboratories, Inc. The use of warble tone in clinical audiometry. Bulletin A-S. LaHabra, Calif.: Allison Laboratories, Inc. (not dated). Beadle, C., and Crowell, D. H. Neonatal EKG responses to sound. J. Speech Hearing Res., 5, 112-123 (1962). Beasley, W. Generalized age and sex trends in hearing loss. Hearing Study Series Bulletin No. 7. National Health Survey. Washington, D.C.: Public Health Service (1938). Beltone Electronics Corp. Qperating_lnstructions. Beltone WT-401 Warble Tone Adapter. Chicago: Beltone Electronics Corp. (1964). Bunch, C. Age variations in auditory acuity. A.M.A. Arch. Otolaryngx, 9, 625-636 (1929). Bunch, C. Further observations on age variations in auditory acuity. A.M.A. Arch. Otolaryng., 13, 170-180 (1931). Carhart, R., and Jerger, J. F. Preferred method for clinical determination of pure-tone thresholds. J. Speech Hearing_ Dis., 24, 330-345 (1959). Carhart, R., and Hayes, C. Clinical reliability of bone conduction audiometry. Laryngpscope, 59, 1084-1101 (1949). Carver, W. F. Major Audiometric Measurements. Chicago: Beltone Electronics Corp. (1965). Corso, J. Age and sex differences in pure-tone thresholds. A.M.A. Arch. Otolaryngs, 77, 385-405 (1963). Dallos, P., and Tillman, T. The effects of parameter variations in Bekesy audiometry in a patient with acoustic neurinoma. J. Speech Hearigg Res., 9, 557-572 (1966). 55 S6 Downs, M. P., Testing hearing in infancy and early childhood, in F. McConnell and P. Ward (Eds.), Deafness in Childhood. Nashville, Tennessee: Vanderbilt UnIVersity Press, 25-33 (1967). Gaeth, J., A study of phonemic regression associated with hearing loss. Unpublished doctoral dissertation, Northwestern University (1948). Gengel, R. W., Pascoe, D. and Shore, I., A frequency reSponse procedure for evaluating and selecting hearing aids for severely hearing impaired children. J. Speech Hearing Dis., 36, 341-353 (1971). Glorig, A. and Roberts, J., Hearing levels of adults by age and sex. Public Health Services Publication, n.1000-Series, ll-No. 11. Washington, D.C.: U.S. Pfiblic Health Service (1965). Goetzinger, C., Proud, G., Dirks, D. and Embrey, A., A study of hearing in advanced age. A.M.A. Arch. Otolaryngr, 73, 662-674 (1961). Harbert, F., Young, 1., and Menduke, H., Audiologic findings in presbycusis. J. Aud. Res., 6, 297-312 (1966). Harris, J. D., Research frontiers in audiology, in J. Jerger (Ed.), Modern Developments in Audiology: New York: Academic Press, 414-415 (1963). Heron, T. G., and Jacobs, R., A physiological response of the neo- nate to auditory stimulation. Int. Audiol., 7, 41-47 (1968). Heron, T. G., and Jacobs, R., Respiratory curve responses of the neonate to auditory stimulation. Int. Audiol., London Congress, 8, 77-84 (1969). Jerger, J., Bekesy audiometry in analysis of auditory disorders. . Speech Hearing_Dis., 3, 275-287 (1960). C... Jerger, J., Comparative evaluation of some auditory measures. . Speech Hearing Res., 5, 3—17 (1962). L4 Jerger, J., Shedd, J. and Harford, E., On the detection of extremely small changes in sound intensity. A.M.A. Arch. Otolaryng., 69, zoo-211 (1959). S7 Konig, E., Pitch discrimination and age. ACTA Oto-Laryng,, 48, 475-489 (1957). Langenbeck, B., Textbook of Practical Audiometry. Baltimore: The Williams and Wilkins Co., 23-25 (1965). Liden, G., and Kankkunen, A., Visual reinforcement audiometry in the management of young deaf children. Int. Audiol., London Congress, 8, 99-106 (1969). Madsen Electronics, Clinical Audiometer Model DB 60. Copenhagen, Denmark: Madsen Electronicsfi(not dated). Matkin, and Carhart, R., Hearing acuity and Rh incompatibility. N., A.M.A. Arch. Otolaryngs, 87, 383-388 (1968). Mendel, M. I., Infant responses to recorded sounds. J. Speech Hearing Res., 11, 811-816 (1968). Meyers, C. J., Harris, J. D. and Fowler, E. P., The feasibility of group audiometry. Indust. Medicine, 17, 245-252 (1956). Miller, M. H., and Polisar, I. A., Audiological Evaluation of the Pediatric Patient. Springfield: Charles C. Thomas, 29—30, 747(1964). Miller, M. H., and Rabinowitz, M., Audiological problems associated with pre-natal rubella. Int. Audiol., London Congress, 8, 90-98 (1969). Munson, W. A., and Wiener, F. M., Sound measurements for psycho- physical tests. J. Acoust. Soc. Amer., 22, 382-386 (1950). Myers, C. K., Noise bands vs pure tones as stimuli for audiometry. J. Speech HearingDis., 22, 757-760 (1957). Myers, C. K. and Harris, J. D., The Inherent Stability of the Auditory Threshold. New London, Conn.: Naval Médical Research Lab., Report No. 3 (1949). Olson, I., Discrimination of auditory information as related to aging. J. Geront.,20, 394-397 (1965). Peck, J. K., The use of bottle-feeding during infant hearing testing. J. Speech Hearing Dis., 35, 364-368 (1970). Pestalozza, G., and Shore, I., Clinical evaluation of presbycusis on the basis of different tests of auditory function. Laryngoscope, 65, 1136-1163 (1955). 58 Reilly, R. N., Frequency and amplitude modulation audiometry. A.M.A. Arch. Otolaryngx, 60, 363-366 (1958). Rintelmann, W. F., Orchik, D., and Stephens, M., A comparison of pure-tone and warble-tone thresholds. Laboratory Report SHSLR 172, Michigan State University (1972). Rintelmann, W. F., and Schumaier, D., Factors affecting Speech discrimination scores in a clinical setting: List equiva- lency, hearing loss and phonemic regression. Laboratory Report SHSLR 173, Michigan State University (1973). Ross, M., and Duffy, J. K., Clinical uses of pure-tone sound-field audiometry. Paper presented at the Annual Convention of ASHA (1961). Sanders, J. W., and Josey, A. F., Narrow-band noise audiometry for hard-to-test patients. J. Speech Hearing_Res., 13, 74-81 (1970). Simon, G. R., and Northern, J. L., Automatic noiseband audiometry. J. Aud. Res., 6, 403-407 (1966). Sivian, L. J., and White, S. D., On minimum audible sound fields. J. Acoust. Soc. Amer., 4, 288-321 (1933). Staab, W., Comparison of pure-tone and warble-tone thresholds. Unpublished doctoral dissertation, Michigan State University (1971). Staab, W., and Rintelmann, W. F., Status of warble-tone in audiometers. Aud.: J. of Aud. Communication, 11, 244-255 (1972). Steinberg, J. C., and Munson, W. A., Deviations in the loudness judgements of 100 people. J. Acoust. Soc. Amer., 8, 71-80 (1936). Tracor, Inc., Allison Model 22 Clinical and Research Audiometer. Austin, Texas: Tracor, Inc. Webster, J. C., A recorded warble tone audiometer test suitable for group administration over loudspeakers. J. Speech Hearing Dis., 17, 213-223 (1952). Witting, E. G., and Hughson, W., Inherent accuracy of a series of repeated clinical audiograms. Laryngoscope, 50, 259-269 (1940). Young, I. M., and Harbert, F., Frequency modulated tone thresholds in normal and abnormally adapting ears. Ann. Otol. Rhinol. Laryng., 79, 138-144 (1970). APPENDICES APPENDIX A CALIBRATION DATA TABLE A1.--Octave band and C-scale analyses of ambient noise levels in examination room in dB SPL according to standards of the American Standards Association (ASA 83.1-1960). Test Room - IAC 1200 ACT Microphone - BGK 4145 Sound Level Meter - BGK 2204 Octave Band Filter Set - 38K 1613 Center Frequency in Hz dB SPL 31.5 37 63 27 125 12 250 10 500 10 1000 10 2000 10 4000 10 8000 10 C-Scale 45 dB SPL 59 60 TABLE A2.-—Audiometer earphone output data for the right earphone of the Maico MA-24 audiometer. Measurements in accordance with American National Standards Institute (ANSI S3.6-1969). Audiometer - Maico MA-24 Earphone Type - TDH 39/102 Cushion Type - MX-41/AR Artificial Ear - BGK 4152 Microphone - BGK 4144 Sound Level Meter - BGK 2204 Octave Band Filter Set — BGK 1613 Frequency Output Reference Interval in Hz (70 dB Input) Threshold Pre-experiment 250 95.40 25.40 Subjects 1 and 2 500 81.40 11.40 1000 76.80 6.80 2000 77.60 7.60 4000 78.15 8.15 8000 85.50 15.50 Subjects 3 and 4 250 95.70 25.70 500 81.80 11.80 1000 77.20 7.20 2000 77.90 7.90 4000 78.55 8.55 8000 85.90 15.90 Subjects 5 and 6 250 96.80 26.80 500 82.90 12.90 1000 78.20 8.20 2000 79.00 9.00 4000 79.25 9.25 8000 86.50 16.50 Subjects 7 and 8 250 96.70 26.70 500 82.80 12.80 1000 78.20 8.20 2000 78.80 8.80 4000 79.25 9.25 8000 86.80 16.80 Subject 9 250 96.80 26.80 500 82.80 12.80 1000 78.20 8.20 2000 79.00 9.00 4000 79.05 9.05 8000 86.70 16.70 61 TABLE A2.--Continued Frequency Output Reference Interval in Hz (70 dB Input) Threshold Subjects 10 and 11 250 96.50 26.50 500 82.70 12.70 1000 77.80 7.80 2000 78.50 8.50 4000 78.75 8.75 8000 86.50 16.50 Subjects 12, l3, 14 250 96.70 26.70 500 83.00 13.00 1000 78.20 8.20 2000 78.80 8.80 4000 78.95 8.95 8000 86.80 16.80 Subjects 15 and 16 250 96.50 26.50 500 83.00 13.00 1000 78.00 8.00 2000 78.80 8.80 4000 78.95 8.95 8000 86.50 16.50 Subjects 17, 18, 19 250 96.50 26.50 500 83.00 13.00 1000 78.20 8.20 2000 78.50 8.50 4000 79.25 9.25 8000 86.70 16.70 Subject 20 250 97.00 27.00 500 83.00 13.00 1000 78.50 8.50 2000 79.10 9.10 4000 79.25 9.25 8000 86.80 16.80 Subject 21 250 96.50 26.50 500 83.00 13.00 1000 78.20 8.20 2000 78.50 8.50 4000 79.25 9.25 8000 86.50 16.50 62 TABLE A2.--Continued ‘n3: Frequency Output Reference Interval in Hz (70 dB Input) Threshold Subjects 22, 23, 250 96.50 26.50 24, 25 500 83.00 13.00 1000 78.20 8.20 2000 78.80 8.80 4000 79.25 9.25 8000 86.80 16.80 Subjects 26, 27, 28 250 96.70 26.70 500 83.00 13.00 1000 78.30 8.30 2000 79.00 9.00 4000 79.45 9.45 8000 87.00 17.00 Subject 29 250 96.30 26.30 500 82.70 12.70 1000 78.00 8.00 2000 78.60 8.60 4000 79.25 9.25 8000 86.80 16.80 Subject 30 250 96.70 26.70 500 83.00 13.00 1000 78.20 8.20 2000 78.50 8.50 4000 79.25 9.25 8000 87.00 17.00 Subjects 31 and 32 250 96.30 26.30 500 82.50 12.50 1000 77.80 7.80 2000 78.50 8.50 4000 78.55 8.55 8000 86.20 16.20 Post-experiment 250 96.50 26.50 500 82.50 12.50 1000 78.00 8.00 2000 78.50 8.50 4000 79.25 9.25 8000 86.80 16.80 63 TABLE A3.--Output data for bone conduction vibrator. Measurements were made in accordance with norms specified by the Hearing Aid Industry Conference (HAIC) Interim Bone- Conduction Thresholds for Audiometry (Lybarger, 1966*). Audiometer - Maico MA-24 Artificial Mastoid - Beltone MSA Channel - Right Mastoid Amplifier - Beltone MSA MicrOphone Amplifier - BGK 2603 Frequency dB level re 1 millivolt Difference Actual minus Expected Actual Expected Pre-experiment 250 27.1 28.4 -1.3 500 40.8 45.1 -4.3 1000 25.8 26.7 - .9 2000 20.0 20.2 - .2 4000 18.9 15.8 +3.1 Post-experiment 250 , 26.8 28.4 -1.6 500 41.6 45.1 —3.5 1000 26.8 26.7 + .l 2000 21.9 20.2 +1.7 4000 20.0 15.8 +4.2 *Lybarger, S. F., Interim bone conduction thresholds for audiometry, J. Speech Hearing Res., 9, 483-487 (1966). 64 TABLE A4.--Pre- and post-experimental linearity of Maico MA-24 audiometer attenuator, right channel. Measured at the test earphone, 1000 Hz. Audiometer - Maico MA-24 Artificial Ear - BGK 4152 Audiometer Channel - Right Sound Level Meter - BGK 2204 Earphone - Right TDH-39/IOZ Microphone - B8K 4144 Earphone Cushion - MX-41/AR Octave Band Filter Set - BGK 1613 Post-experiment Post-experiment dB HTL dB SPL Difference dB SPL Difference in dB in dB 100 106.4 107.5 10.0 9.9 90 96.4 97.6 9.9 10.0 80 86.5 87.6 9.9 9.9 70 76.6 77.7 10.0 10.0 60 66.6 67.7 9.8 9.7 50 56.8 58.0 10.0 10.0 40 46.8 48.0 9.9 10.2 30 36.9 37.8 10.1 10.2 20 26.8 27.6 65 TABLE A5.--Pre- and post-experimental rise and decay times (in milli- seconds) as measured for pure tones generated by the Maico MA-24 audiometer. Measurements were made with the assistance of a storage oscilloscope in accordance with the American National Standards Institute (ANSI S3.6-1969). Storage Oscilloscope Audiometer - Maico MA-24 Tektronix Type 564B Ch Frequency pre-experiment Post-experiment annel in Hz , . R158 Decay Rise Decay Right 250 42 7o 36 66 500 45 70 40 70 1000 40 72 40 70 2000 40 80 36 76 4000 30 78 32 76 8000 33 76 40 70 Left 250 32 68 34 70 500 36 70 30 80 1000 36 70 4O 72 2000 30 76 36 74 4000 36 72 30 70 8000 28 74 28 76 Notes: Rise time--Time for SPL to rise from -20 to -1 dB re its final steady value. Decay time--Time for SPL to decay by 20 dB. 66 TABLE A6.-~Pre- and post-experimental checks of the test frequencies of the Maico MA-24 audiometer performed in compliance with the American National Standards Institute (ANSI S3.6-1969).* Audiometer - Maico MA-24 Channel - Right Frequency Analyzer - BGK 2107 Frequency Counter — Bekman 6148 Measured Test Frequency Frequency Difference Difference in Hz in Hz in Hz in percent Pre-experiment 250 247 - 3 1.20 500 500 00 0.00 1000 991 - 9 .90 2000 2005 + 5 .25 4000 4031 +31 .77 8000 7968 -32 .40 Post-experiment 250 247 - 3 1.20 500 500 00 0.00 1000 991 - 9 .90 2000 2005 + 5 .25 4000 4032 +32 .80 8000 7972 -28 .35 *The unmodulated warble-tone center frequencies were observed during all testing and manually varied to be within 3 percent of the indicated frequency. 67 TABLE A7.--Pre- and post-experimental harmonic distortion measurements of the fundamental for test frequencies used. Measure- ments were made for the right channel of the Maico MA-24 audiometer in compliance with the American National Standards Institute (ANSI 83.6-1969). Audiometer - Maico MA-24 Channel _ Right Frequency Analyzer - BGK 2107 Frequency SPL of Fuistmziigl Difference in Hz Fundamental Rejected* in dB Pre-experiment 250 104.0 72.0 32.0 500 111.0 75.0 36.0 1000 104.0 73.0 31.0 2000 109.0 79.0 30.0 4000 104.0 70.0 34.0 8000 108.0 77.0 31.0 Post-experiment 250 102.0 68.0 34.0 500 109.5 72.0 37.5 1000 105.0 69.5 35.5 2000 109.2 75.0 34.2 4000 103.5 68.0 35.5 8000 106.0 62.5 43.5 *These values represent the total SPL remaining after the fundamental has been rejected. 68 TABLE A8.--Audiometer earphone output data for the left earphone of the Maico MA-24 audiometer. with the American National Standards Institute (ANSI 83.6-1969 ). Measurements in accordance Audiometer - Maico MA-24 Earphone Type - TDH-39/IOZ Cushion Type - MX-41/AR Artificial Ear - BGK 4152 Microphone - BGK 4144 Sound Level Meter - 86K 2204 Octave Band Filter Set - BGK 1613 Frequency Output Reference Interval in Hz (70 dB Input) Threshold Pre-experiment 250 94.40 24.40 500 81.00 11.00 1000 76.30 6.30 2000 77.10 7.10 4000 78.05 8.05 8000 85.00 15.00 After Subject 4 250 95.00 25.00 500 81.50 11.50 1000 77.80 7.80 2000 78.00 8.00 4000 78.25 8.25 8000 86.00 16.00 After Subject 16 250 96.00 26.00 500 82.00 12.00 1000 78.60 8.60 2000 78.10 8.10 4000 78.85 8.85 8000 86.20 16.20 Post-experiment 250 95.30 25.30 500 81.80 11.80 1000 77.50 7.50 2000 77.80 7.80 4000 .78.75 8.75 8000 86.00 16.00 69 TABLE A9.--Pre- and post-experimental linearity of Maico MA-24 audiometer attenuator, left channel. Measured at the test earphone, 1000 Hz. Audiometer - Maico MA-24 Artificial Ear - BGK 4152 Audiometer Channel - Left Sound Level Meter - BGK 2204 Earphone - Right TDH-39/IOZ Microphone - BaK 4144 Earphone Cushion - MX-41/AR Octave Band Filter Set - BfiK 1613 Pre-experiment Post-experiment dB HTL Difference Difference dB SPL in dB dB SPL in dB 100 106.6 107.2 9.1 10.0 90 97.5 97.2 10.5 9.9 80 87.0 87.3 10.7 9.8 70 76.3 77.5 10.0 10.0 60 66.3 67.5 9.3 9.8 50 57.0 57.7 9.7 10.0 40 47.3 47.7 9.8 9.7 30 37.5 38.0 9 5 9.5 20 28.0 28.5 70 TABLE A10.--Settings required to produce and measure the warble tones with i 3% and i 10% frequency deviation. Settings include the Frequency Deviation (FD) setting of the beat-frequency oscillator; Volt Scale (VS), and Output Voltage (V) on the function generator; and the Hz/Div setting of the storage scope Spectrum analyzer. Base Frequency Deviation Frequency FD VS V Hz/Div in Hz in % in Hz Pre-experiment 500 : 3 30 100 3 1.2 20 :10 100 100 10 3.4 20 1000 : 3 60 100 3 1.9 10 :10 200 160 10 4.4 50 2000 : 3 120 100 10 3.6 50 :10 400 400 10 3.2 100 4000 : 3 240 250 10 3.6 100 :10 800 630 10 3.4 200 Post-experiment 500 : 3 30 100 3 1.2 20 :10 100 100 10 3.2 20 1000 : 3 60 100 3 1.9 20 :10 200 160 10 4.2 50 2000 : 3 120 100 10 3.4 50 :10 400 400 10 3.0 100 4000 : 3 240 250 10 3.6 100 :10 800 630 10 3.2 200 71 TABLE A11.--Output data for beat-frequency oscillator routed through Maico MA-24 Audiometer (left channel, right earphone). Levels were obtained by measuring the SPL output of the unmodulated warble-tone center frequencies at "O" VU reading and 70 dB Hearing Threshold Level setting through the ACCESSORY INPUT of the Maico MA-24 Audiometer. Audiometer - Maico MA-24 Earphone Type - TDH-39/1OZ Cushion Type - MX-41/AR Beat-frequency oscillator - Microphone - BGK 4144 Artificial Ear - 88K 4152 Sound Level Meter - BGK 2204 Octave Band Filter Set — BGK 1613 86K 1013 Frequency Output Reference Interval in Hz (70 dB Input) Threshold Pre-experiment Subjects 1 and 2 500 92.00 22.00 1000 91.80 21.80 2000 88.30 18.30 4000 94.05 24.05 Subjects 3 and 4 500 91.50 21.50 1000 91.70 21.70 2000 88.20 18.20 4000 94.05 24.05 Subjects 5 and 6 500 93.00 23.00 1000 92.70 22.70 2000 89.30 19.30 4000 94.55 24.55 Subjects 7 and 8 500 92.30 22.30 1000 92.50 22.50 2000 88.80 18.80 4000 94.75 24.75 Subject 9 500 93.00 23.00 1000 92.60 22.60 2000 89.50 19.50 4000 94.25 24.25 Subjects 10 and 11 500 93.00 23.00 1000 92.50 22.50 2000 88.80 18.80 4000 94.55 24.55 72 TABLE All.--Continued Interval Frequency Output Reference in Hz (70 dB Input) Threshold Subjects 12, 13, 14 500 93.00 23.00 1000 92.50 22.50 2000 88.80 18.80 4000 94.55 24.55 Subjects 15 and 16 500 92.50 22.50 1000 92.50 22.50 2000 88.80 18.80 4000 94.75 24.75 Subjects 17, 18, 19 500 92.50 22.50 1000 92.70 22.70 2000 89.00 19.00 4000 94.75 24.75 Subject 20 500 93.00 23.00 1000 92.80 22.80 2000 89.10 19.10 4000 94.55 24.55 Subject 21 500 93.00 23.00 1000 92.50 22.50 2000 88.80 18.80 4000 94.55 24.55 Subjects 22, 23, 500 92.30 22.30 24, 25 1000 92.50 22.50 2000 89.30 19.30 4000 94.25 24.25 Subjects 26, 27, 28 500 92.30 22.30 1000 92.20 22.20 2000 88.50 18.50 4000 94.25 24.25 Subject 29 500 92.50 22.50 1000 92.50 22.50 2000 88.50 18.50 4000 94.25 24.25 73 TABLE All.--Continued Frequency Output Reference Interval in Hz (70 dB Input) Threshold Subject 30 500 92.50 22.50 1000 92.30 22.30 2000 88.50 18.50 4000 94.25 24.25 Subjects 31 and 32 500 92.20 22.20 1000 92.00 22.00 2000 88.50 18.50 4000 94.25 24.25 Post-experiment 500 92.30 22.30 1000 92.20 22.20 2000 88.50 18.50 4000 94.25 24.25 74 TABLE A12.--Pre- and post-experimental harmonic distortion measure- ments of the fundamental of the warble-tone center frequencies used. Measurements were made for the output of the beat-frequency oscillator, routed throngh the left channel of the Maico MA-24 audiometer, in compliance with the American National Standards Institute (ANSI S3.6-1969). Audiometer - Maico MA-24 Beat-frequency Oscillator - BfiK 1013 Channel - Left Frequency Analyzer - BGK 2107 Frequency SPL of SPL "1th Difference . Fundamental . in Hz Fundamental . * in dB Rejected Pre-experiment 500 103.0 73.0 30.0 1000 102.0 70.0 32.0 2000 103.0 72.0 31.0 4000 103.5 73.0 30.5 Post-experiment 500 107.0 69.0 38.0 1000 107.0 65.0 42.0 2000 107.0 70.0 37.0 4000 107.0 68.0 39.0 *These values represent the total SPL remaining after the fundamental has been rejected. 75 TABLE A13.--Pre- and post-experimental acoustic spectrum data for narrow-band noise produced by Maico MA-24 audiometer. Audiometer - Maico MA-24 Channel - Left Frequency Analyzer - BGK 2107 Power Level Recorder - 88K 2305 Rejection Rate Center . Band Width in dB/Octave Frequency NOise Band in Hz Lower Higher Pre-experiment 500 475 S60 85 25 20 1000 950 1100 150 25 21 2000 1900 2250 350 22 21 4000 3800 4600 800 19 21 Post-experiment 500 490 575 85 25 20 1000 950 1125 175 25 20 2000 1900 2250 350 20 20 4000 3750 4500 750 18 22 76 TABLE A14.--Output data for narrow—band noise produced by Maico MA-24 Audiometer (left channel, right earphone). Audiometer - Maico MA-24 Microphone - BGK 4144 Earphone Type - TDH-39/lOZ Artificial Ear - BGK 4152 Cushion Type - MX-41/AR Sound Level Meter - 88K 2204 Octave Band Filter Set - BGK 1613 Center Interval Frequency Output Reference . (70 dB Input) Threshold in Hz Pre-experiment 500 76.70 6.70 Subjects 1 and 2 1000 74.00 4.00 2000 68.90 -1.10 4000 70.45 .45 Subjects 3 and 4 500 77.00 7.00 1000 74.20 4.20 2000 69.10 - .90 4000 70.75 .75 Subjects 5 and 6 500 78.00 8.00 1000 75.00 5.00 2000 70.00 0.00 4000 71.05 1.05 Subjects 7 and 8 500 71.00 1.00 1000 71.00 1.00 2000 69.30 - .70 4000 70.25 .25 Subject 9 500 72.00 2.00 1000 71.50 1.50 2000 70.30 .30 4000 71.25 1.25 Subjects 10 and 11 500 71.50 1.50 1000 71.50 1.50 2000 70.30 .30 4000 71.25 1.25 Subjects 12, 13, 14 500 70.50 .50 1000 71.00 1.00 2000 70.30 .30 4000 71.75 1.75 77 TABLE Al4.--Continued Center Interval Frequency OUtPUt Reference . (70 dB Input) Threshold in Hz Subjects 15 and 16 500 71.00 1.00 1000 70.50 .50 2000 69.80 - .20 4000 71.75 1.75 Subjects 17, 18, 19 500 71.00 1.00 1000 70.50 .50 2000 69.80 - .20 4000 71.75 1.75 Subject 20 500 72.50 2.50 1000 71.50 1.50 2000 70.80 .80 4000 71.95 1.95 Subject 21 500 72.50 2.50 1000 71.50 1.50 2000 70.30 .30 4000 71.75 1.75 Subjects 22, 23, 500 72.50 2.50 24, 25 1000 71.50 1.50 2000 70.30 .30 4000 71.75 1.75 Subjects 26, 27, 28 500 71.50 1.50 1000 71.00 1.00 2000 69.80 - .20 4000 71.25 1.25 Subject 29 500 72.50 2.50 1000 71.50 1.50 2000 70.30 .30 4000 71.75 1.75 Subject 30 500 71.50 1.50 1000 71.00 1.00 2000 69.80 - .20 4000 71.25 1.25 78 TABLE A14.--Continued Interval nggfigicy OUtPUt Reference . (70 dB Input) Threshold in Hz Subjects 31 and 32 500 72.00 2.00 1000 71.00 1.00 2000 69.80 - .20 4000 71.25 1.25 Post-experiment 500 71.00 1.00 1000 71.00 1.00 2000 69.80 — .20 4000 71.25 1.25 79 .xaco Enumofinsm ucoefluomxoumua mcflcflmuno a“ new: mm: ecocmhmo umoq .oconmhme flamencocg umofl mo omcoamon socoscoum--.a< enamfim N: :H socosconm oooom OOOOH ooom ooom oooH com com ooH om .u>> - a — d _ q — . 1 O O N H 1 O N) Kitsuaiul aAtietau 80 .ococmumo npmouv unmfln mo omcommoa socoscoauuu.m< ounmwm .%OOON OOOOH ooom ooom oooH a: :a xucoscohm com com ooH om - u q k (ll); OH ON on ow Airsueiul antueteu Relative Intensity 81 Pre—experiment r >\ 0» p -H m a 0 4.1 .5 w -10 - > H ‘3 H 32 -20 t 250 500 1000 Frequency in Hz Post-experiment oh -10 . ~20 “ 250 300 A I 1000 Frequency in Hz Figure A3.--Pre- and post-experimental analysis of narrow-band noise with 500 Hz center frequency. 82 >‘ Pre-experiment p ‘51 a 0 ' 0 p c H 0 .3 .u -10 _ 2 d.) :z -20 . J 1 l 500 1000 200} Frequency in Hz Post-experiment 3: 0 * -H U) a 0) p E a -10 > «4 {J 2 0 CI: 5 -20 500 1000 2000 Frequency in Hz Figure A4.--Pre- and post-experimental analysis of narrow-band noise with 1000 Hz center frequency. 83 Pre-experiment Relative Intensity L. O O I N O t 1000 2000 4000 Frequency in Hz Post-experiment 5: 0 «4 m r: o p t: H o -10 > w-O U d H 6‘2 -20 L l 1000 2000 4000 Frequency in Hz Figure A5.--Pre- and post-experimental analysis of narrow-band noise with 2000 Hz center frequency. Relative Intensity 84 Pre-experiment 0 r -10' -20 . 2000 4000 8000 Frequency in Hz Post-experiment >. O r .0.) «4 U) a d) 4.) a H o -10 ’ > m1 9 a F4 1) :z -20 _ 2000 4000 8000 Frequency in Hz Figure A6.--Pre- and post-experimental analysis of narrow-band noise with 4000 Hz center frequency. APPENDIX B ANALYSIS OF VARIANCE TABLE TABLE B1.--Analysis of Variance Table. Probability source 55 dF MS Ratio of Statistic Configuration (A) 789.26 1 789.26 0.25 0.621 Age (B) 1173.49 1 1173.49 0.37 0.547 AB 101.63 1 101.63 0.03 0.859 Error 88252.46 28 3151.87 Test-Retest (C) 211.16 1 211.16 19.15 <0.0005 AC 0.22 l 0.22 0.02 0.889 BC 7.06 l 7.06 0.64 0.430 ABC 0.22 l 0.22 0.02 0.889 Error 308.69 28 11.02 Signal (D) 6868.42 3 2289.47 65.87 <0.0005 AD 4540.51 3 1513.51 43.54 <0.0005 80 27.13 3 9.04 0.26 0.854 ABD 40.02 3 13.34 0.38 0.765 Error 2919.78 84 34.76 CD 1.83 3 0.61 0.07 0.974 ACD 17.85 3 5.94 0.72 0.544 BCD 14.91 3 4.97 0.60 0.616 ABCD 1.83 3 0.61 0.07 0.974 Error 695.60 84 8.28 Frequency (E) 110818.92 3 36939.64 104.93 <0.0005 AE 62524.65 3 20841.55 59.20 <0.0005 BE 1276.92 3 425.64 1.21 0.312 ABE 2241.77 3 747.26 2.12 0.103 Error 29570.80 84 352.03 CE 28.78 3 9.59 0.97 0.410 ACE 8.08 3 2.69 0.27 0.845 BCE 17.65 3 5.88 0.59 0.620 ABCE 28.78 3 9.59 0.97 0.410 Error 329.98 84 9.88 DE 4875.42 9 541.71 26.11 <0.0005 ADE 3723.58 9 413.73 19.94 <0.0005 BDE 197.13 9 21.90 1.06 0.396 ABDE 221.36 9 24.59 1.18 0.304 Error 5228.21 252 20.75 CDE 38.30 9 4.26 0.85 0.571 ACDE 27.37 9 3.04 0.61 0.791 BCDE 48.27 9 5.36 1.07 0.384 ABCDE 39.87 9 4.43 0.88 0.540 Remaining Error 1261.03 252 5.00 85