A COMPARISON OF PURE-TONE THRESHOLDS AS M E A S U R E D BY DELAYED FEEDBACK AUDIOMETRY, ELECTRODERMAL RESPONSE AUDIOMETRY, AND VOLUNTARY RESPONSE AUDIOMETRY BRADLEY L. BILLINGS
Audiology Center ol Redlands, California THOMAS E. STOKINGER
Veterans Administration Hospital, Oklahoma City, Oklahoma One hundred unselected patients seen for medical-legal evaluation were tested for pure-tone thresholds by delayed feedback audiometry (DFA), electrodermal response audiometry (EDRA), and voluntary response audiometry (VRA). The EDRA method was successful in 73g of the patients while the DFA method was successful in 88~ of thepatients. Eighty-six percent of the DFA thresholds obtained were within 10 dB of the patients' VRA thresholds. When both DFA and EDRA were successful, 88g of the DFA thresholds were within 10 dB of the EDRA thresholds. Ninety-six percent of the EDRA thresholds obtained were within 10 dB of the patients' VRA thresholds. Although DFA is not as precise in predicting threshold as is EDRA, it is successful in a significantly greater number of patients than is EDRA and is a useful clinical tool in medical-legal evaluation for hearing loss. The effects of pure-tone delayed feedback audiometry ( D F A ) have been described in a series of experiments reported by Ruhm and Cooper (1962, 1963, 1964). The DFA testing procedure requires that the patient repeatedly tap the pattern ( . . . . . . ) on a silent electromechanical key. Determination of threshold is based on a comparison of his tapping consistency during simultaneous auditory feedback (SAF) and delayed auditory feedback ( D A F ) at different intensity levels. During SAF, each tap on the electromechanical key results in a tone burst that is presented immediately to the patient's earphone. During DAF, each tap on the key results in a tone burst that is presented to the patient's earphone after a 200-msec delay. If the tone bursts are inaudible to the patient, his tapping performance should remain unaffected under both SAF and DAF conditions. If the tone bursts are audible to the patient, his tapping performance should not be affected under the SAF condition. The DAF condition, however, will result in modification of tapping performance including time errors (greater time is generally required to tap patterns under DAF than under SAF), number errors (usually the inclusion of extra taps in 754
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
the patterns under DAF), and the use of greater tapping pressure during DAF. Ruhm and Cooper's experiments ( 1962, 1963, 1964) have shown that: 1. A delay time of 200 msec is most effective in causing tapping disruption. 2. The effects of DAF are independent of the pure-tone signal frequency from 250 to 8000 Hz. 3. There are no significant differences between DAF tapping performance of males and females. 4. Short-term practice or sophistication with DFA does not significantly affect thresholds obtained. 5. Fatigue from repeatedly tapping the patterns a reasonable number of times does not significantly affect tapping performance. 6. Normal and hard-of-hearing patients generally yield DFA thresholds that are within 5 dB of their voluntary thresholds. 7. Patients who exhibit signs of nonorganic hearing loss generally yield DFA thresholds that are within 5 dB of their electrodermaI response thresholds. The effectiveness of DFA has been reported previously on the basis of results with either normal-hearing subjects or small groups of selected subjects with a hearing loss. It was the purpose of this study to investigate the validity of DFA as a clinical tool and to compare its effectiveness with electrodermal response audiometry in an unselected series of patients seen for medical-legal evaluation. In addition, this study was designed to yield data concerning the normal variability of key-tapping performance without the effects of auditory feedback for the purpose of establishing normative criteria for use in quantitative determination of DFA thresholds. METHOD
Subiects One hundred patients of the Veterans Administration Hospital Audiology and Speech Pathology Service, Oklahoma City, Oklahoma, served as subjects for this study. Each patient had his hearing tested by delayed feedback audiometry (DFA), electrodermal response audiometry (EDRA), and voluntary response audiometrv (VRA) at either 500, 1000, or 2000 Hz in one ear only.
Testing Order One half of the patients were administered DFA prior to EDRA and the other half of the patients received the reverse order of testing. Voluntary response audiometry was administered only after DFA and EDRA testing had been completed for all patients. Although all patients received a complete BXLLINCS,STOKINCEa:Pure-Tone Thresholds 755
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
audiologic evaluation by another examiner prior to this experiment, their audiograms were not available to the experimenter until after they had completed this study.
Signal Conditiol~s. The pure-tone signals used for DFA, EDRA, and VRA were of 50-msec fullamplitude duration with 10-msec rise-decay times. The tone bursts were generated by an audio oscillator (Hewlett-Packard, Model 200ABR) and shaped by an electronic switch (Grason-Stadler, Model 829C). The electronic switch was triggered by a timing mechanism that allowed for the generation of a tone burst immediately after each tap on the electromechanical key (SAF) or for the generation of a tone burst delayed 200 msec from each tap on tlle key (DAF). In addition, the tone bursts could be initiated manually by the experimenter for the purpose of obtaining voluntary response thresholds or could be initiated through a modified psychogalvanometer (Grason-Stadler, Model E664-1) for the purpose of obtaining EDR thresholds. The experimenter monitored the generation of all tone bursts through a loudspeaker and by observation of an oscilloscope. This was especially important during the DFA trials because it enabled the experimenter to monitor auditorily the patient's tapping performance.
Acoustic Environment All patients were tested in a sound-treated examination room (Industrial Acoustics, Model 400), and all test equipment except the electromechanical key and patient's earphones was located outside the testing room. The test room conformed to criteria for background noise in audiometric examination rooms (ANSI, 1960).
Calibration Since the pure-tone signals were shorter than the critical duration for puretone thresholds (see the Results section for further discussion), the experimental apparatus was calibrated for normal hearing on the basis of mean threshold results from 10 normal-hearing young adults. These subiects reported no history of otologic abnormalities or significant noise exposure and each passed a hearing screening test at octave frequencies from 250 to 4000 Hz (using a conventional audiometer) at 15 dB HL (ANSI, 1969) in the test ear. Intensity, frequency, and timing characteristics of the experimental signals were calibrated daily. Attenuator linearity was checked before and after the experiment. The electronic switch was balanced frequently.
Determination o[ Thresholds DFA thresholds were obtained by a modified method-of-limits procedure 756 Journal of Speech and Hearing Research
18 754-764 1975
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
using only ascending approaches to threshold with 5-dB increments. Each patient was instructed and given practice in tapping the required pattern ( . . . . . . ) prior to testing. After it was determined that the patient knew the instructions and could tap the patterns adequately, practice was terminated and testing begun. A shield was then placed in front of the patient so that he could not see either his hand or the tapping key. The patient was instructed to tap the pattern repeatedly as long as a signal light was on, and to stop tapping the pattern when the signal light was off. Eight patterns were tapped at each intensity level; the first four patterns under SAF and the last four patterns under DAF. A record of the patient's tapping performance was made with a calibrated strip-chart recorder (Techni-Rite, Model TR711) at a speed of 25 mm/sec. DFA testing was started at - 1 0 dB HL (as established on the normal-hearing group), and the intensity level was increased in 10-dB steps until the patient showed significant audible tapping disruption for the DAF condition as compared to the SAF condition (the experimenter monitored tapping performance through a loudspeaker). The intensity level was then decreased by 10 or 15 dB and another ascending approach to threshold was made in 5-dB steps. This procedure was repeated until the patient showed obvious tapping disruption twice at the same intensity level out of three ascending runs. Determination of DFA thresholds, according to normative standards as defined by this study, is discussed in the Results section of this article. EDRA thresholds were also obtained by a modified method-of-limits procedure using only ascending approaches to threshold with 5-dB intensity increments. The procedure used for EDRA was similar to that recommended by Ruhm (1961). From two to four conditioning trials were used and 100% reinforcement was maintained. To be classified as a significant response, an EDR tracing must have exceeded the baseline amplitude variation in quiet by a factor of two, must have possessed a latency of onset of from one to five seconds from the beginning of the tone burst, and must have had a slope greater than 45 degrees. EDRA (after conditioning) was started at - 1 0 dB HL (as established on the normal-hearing group), and the intensity levels were increased in 10-dB steps until the patient yielded a significant EDR. The intensity level was then decreased by 10 or 15 dB and another approach to threshold was made in 5-dB steps. This procedure was repeated until an EDR was elicited twice at the same intensity level out of three ascending runs.
VRA thresholds were also obtained by a modified method-of-limits procedure using only ascending approaches to threshold with 5-dB intensity increments. VRA was started at - 1 0 dB HL (as established on the normal-hearing group) and the intensity levels were increased in 10-dB steps until the patient indicated that he heard the signal by raising his hand. The intensity level was then decreased by 10 or 15 dB and another approach to threshold was made in 5-dB steps. This procedure was repeated until a voluntary response was elicited twice at the same intensity level out of three ascending runs. Bmt,mas, STOKINGER:eufe,-Tol'l,e, Thresholds 757
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
RESULTS
Normal Tapping Variability and Calculation of DFA Threshold It is not practical to calculate the tapping times and number errors for SAF and DAF at each intensity level during the actual testing session because of the length of time required, approximately 20 to 30 seconds for each set of eight patterns. Therefore, ascending approaches to threshold are usually terminated when the examiner is able to auditorily detect tapping performance disruption during DAF through a loudspeaker. Although this method is practical and relatively accurate, a quantitative method for determining DFA thresholds is necessary for medical-legal iustification of the obiectivity of DFA as a clinical technique. One hundred seventy-three groups of eight patterns (four SAF and four DAF) were obtained at - 1 0 dB HL from the 94 patients who were able to perform the task. Since - 1 0 dB HL was inaudible to these patients, as confirmed by voluntary thresholds and EDRA in many cases, differences in performance between SAF and DAF (that is, time and number errors) reflect normal variability of performance that is uncontaminated by the effects of audible auditory feedback, either SAF or DAF. Number Errors. Of the 173 groups of SAF and DAF patterns tapped at - 1 0 dB HL, the inclusion of extra taps during DAF compared to SAF occurred only 19 times (11~). Most number errors were limited to one extra tap with two extra taps occurring only one time. These data agree well with those presented by Cooper, Stokinger, and Billings (1971) in which 8~ of 50 normal-hearing subiects (cooperative volunteers) produced one or more extra taps during DAF compared to SAF at - 10 dB SL. Time Errors. The times required to tap the SAF and DAF pattern groups at - 1 0 dB HL were calculated from the strip-chart recorder as measured from the first tap of the first pattern to the last tap of the fourth pattern. The mean difference between the time required to tap the DAF pattern groups and the time required to tap the SAF pattern groups was 80 msec with a standard deviation of 384 msec. To compare these data with those presented by Cooper et al. (1971), the total time required to tap each pattern group was divided by four to yield an average time required to tap one pattern. The mean difference between average times required to tap one pattern under DAF and one pattern under SAF (that is, DAF minus SAF) was 20 msec with a standard deviation of .96.8 msec. These data are similar to the Cooper et al. data in which the mean difference between average times was 8.6 msec with a standard deviation of 71.4 msec. Calculation of DFA Thresholds. On the basis of the above data, disruption of tapping performance due to the audible effects of DAF was defined as the inclusion of one or more extra taps during DAF as compared to SAF, or a difference between the time required for tapping the SAF pattern group and the DAF pattern group (that is, four patterns of DAF minus four patterns of 758 Jourrurlof Speech and Hearing Research
18 754-764 1975
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
SAF) exceeding the range from - 6 8 8 msec to +848 msec (mean q-_ 2 SD). That is, variability of tapping perfolTnance that exceeds these limits is in excess of normal variability and therefore is considered to be the result of the audible effects of DAF. Each DFA threshold that was calculated on the basis of the normative criteria (N = 88) was compared to its estimated level as
determined by the examiner listening for tapping disruptions through a loudspeaker (see Method section). Sixty-one of the 88 thresholds (69.3%) were identical as determined by both methods, 17 (19.3%) were 5 dB different, and nine (10.2%) were 10 dB different. The examiner's judgment was not as
sensitive as was the application of the normative criteria to determine threshold in all but a few instances. Only one of the 88 thresholds estimated by the examiner was more than 10 dB worse than when calculated according to the normative standards.
Comparison of VRA, DFA, and EDRA Thresholds VRA, DFA, and EDRA thresholds were successfully obtained from 66 of the 100 patients at 500, 1000, or 2000 Hz. In view of Ruhm and Cooper's data (1963) showing that the effects of DAF are independent of frequency and because analysis of the data of this study did not reveal consistent differences with respect to frequency, the thresholds obtained at the three fre-
quencies were pooled. Figure 1 shows that 14 of the 66 patients (21.2%) who were successfully tested by VRA, DFA, and EDRA had identical thresholds by all three methods. Forty-one of the patients (62.1% or 21.2% + 40.9%) had threshold differences
among the three methods that did not exceed 5 dB, and 55 of the patients
I
0
21.21
I 4~,9%
I
~
I 27,3%
,~7,0Z
5
~ 10
21.2%
I
13.6%
J 10.6%
>
~ 20
20
3.0%
20
3.0% 5
I0
9.11
1,51
1.51 15
20
25
30
15 20 25 30 NUMBEO RFPATIENTS FICVBr 2. Difference between DFA and EDRA thresholds for individual patients 5
NUIIBEROF PATIENTS
FIOUBE1. Greatest difference among VRA, DFA, and EDRA thresholds for individual patients (N = 66).
(iv =
10
66).
BILLINGS, STOKINGEB:Pure-Tone Thresholds 759
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
(83.3% or 21.2% H- 40.9% + 21.2%) had threshold differences among the three methods that did not exceed 10 dB. Eleven of the 66 patients (16.6%) had threshold differences among the three methods that exceeded 10 dB. One of these patients had DFA and EDRA thresholds that were considerably more sensitive than his voluntary threshold, and he had exhibited signs of nonorganic hearing loss on previous audiologic evaluations.
Comparison of DFA and EDRA Thresholds The relation between thresholds for the 66 patients who were successfully tested by both DFA and EDRA is illustrated in Figure 2. Eighteen of the 66 patients (27.3%) had identical thresholds as measured by both methods. Forty-nine of the patients (74.3%) had threshold differences between the two methods that did not exceed 5 dB, and 58 of the patients (87.9%) had threshold differences between the two methods that did not exceed 10 dB. Eight of the 66 patients had threshold differences between DFA and EDRA that exceeded 10 dB. Of these eight patients, four had DFA thresholds that were more than 10 dB better than their EDRA thresholds and four had DFA thresholds that were more than 10 dB worse than their EDRA thresholds.
Comparison of VRA and DFA Thresholds VRA and DFA thresholds were successfully obtained for 88 of the 100 patients. Figure 3 shows that 37 of the 88 patients (42.0%) who were successfully tested by both VRA and DFA had identical thresholds by both methods. Sixty-five of the patients (73.8%) had threshold differences that did not exceed 5 dB between the two methods, and 76 of the patients (86.3%) had threshold
,~.
50.7%
I 42.0%
o
5
]
I 32.96
31,8%
12.3%
,~ ]5
1.46
I 8,0~
2O }20~
~,5~ :0
>20
a5
20
2~
~o
3~
0.0%
5
~0~BER oFPAUENTS Fmums 3. Difference between VRA and DFA thresholds for individual patients (N = 88).
2.7%
10
15
20
25
30
35
NUMBER OF PATIENTS
Fmtrm~ 4. Difference between VRA and EDRA thresholds for individual patients (N = 73).
760 1ournalo[ Speech and Hearing Research
18 754-764 1975
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
differences between the two methods that did not exceed 10 dB. Twelve of the patients (13.670) had DFA thresholds that were more than 10 dB different from their VRA thresholds. Of these 12 patients, five had EDRA thresholds that were within 5 dB of their VRA thresholds, indicating that the DFA thresholds are probably in error. Two patients could not be tested by EDRA but yielded DFA thresholds that were considerably better than their VRA thresholds, and they had exhibited signs of nonorganic hearing loss on previous audiologic evaluations. Twelve of the 100 patients did not yield DFA thresholds. Six of these 12 patients did not show tapping disruption at suprathreshold levels under DAF, and six of the patients either could not tap the required patterns or could not maintain the consistent tapping performance required for complete testing.
Comparison of VRA and EDRA Thresholds Seventv-three of the 100 patients were successfully tested by both VRA and EDRA. Figure 4 shows that 37 of the 73 patients (50.7g) had identical thresholds by both methods. Sixty-one of the patients (83.6~) had thresholds that did not differ by more than 5 dB between the two methods, and 70 of the patients (95.9%) had threshold differences between the two methods that did not exceed 10 dB. Three of the 73 patients (4.17o) had EDRA thresholds that were more than 10 dB different from their VRA thresholds. One of these three patients exhibited signs of nonorganic hearing loss on previous audiologic evaluations. Twentv-seven of the 100 patients could not be or were not tested by EDRA. Nineteen of these patients could not be conditioned by the procedure or did not maintain sufficient conditioning to complete testing. The remaining eight patients could not be tested by EDRA because of a history of cardiac problems, motor tremors, and so on. Twenty of the 27 patients who could not be tested by EDRA were successfully tested by DFA and only three of the 20 patients had DFA thresholds that were more than 10 dB different from their VRA thresholds.
Hearing Loss Characteristics of the Patients Thirty of the 100 patients had hearing threshold levels that exceeded 15 dB in reference to the previously established normal calibration data. The degree of hearing loss did not affect the percentage of patients who could be tested or the threshold differences as measured by VRA, DFA, or EDRA.
Factors Affecting the Use of Delayed Feedback Audiometry During the course of this experiment, several aspects of the administration and interpretation of DFA were also investigated. DFA Testing Procedure. DFA instructions are usually given to the patient BILLINCS, STOKINCER:Pure-Tone. Thresholds 761
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
in a manner that illustrates the desired tapping pattern by telling him to tap the pattern "tap tap tap tap-tap tap" and simultaneously showing him the pattern by tapping it on the tapping key table. The suggestion of a mental task, such as "count to yourself, 1 2 3 4-1 2" was usually avoided because it was believed that such a mental activity may allow the patient to more effectively resist the effects of DAF than if mental counting was not used. The mental counting technique, however, was found to be necessary to properly train 19 of the patients in this study. Nine of the 19 patients had DFA thresholds that were identical to their VRA thresholds and 16 of the 19 patients had DFA thresholds that were within 5 dB of their VRA thresholds. Only one of the patients had a DFA threshold that was more than 10 dB different from his VRA threshold. These results suggest that the use of a mental counting task is helpful in training some patients to tap the required patterns without resulting in higher resistance to the effects of DAF than would otherwise be the case. Rate of Tapping. Observation during the initial use of DFA suggested the possibility that patients who tapped very rapidly or very slowly might be less resistant to the effects of DAF than patients who tapped at an average speed. The time required to tap the first four SAF patterns at -- 10 dB HL was analyzed for the 94 patients who could tap the required patterns satisfactorily, yielding a median time of 8.8 seconds with a range of 5.9 seconds to 14.5 seconds. All but one of the 10 patients who did not show tapping disruption under suprathreshold DAF tapped the SAF patterns faster than the median time of 8.8 seconds. It was not possible, however, to predict whether a patient would be more or less likely to show tapping disruption under DAF on the basis of his tapping rate. Many "rapid" tappers and many "slow" tappers had DFA thresholds that were equivalent to both their VRA and EDRA thresholds. Temporal Integration and DFA. Pure-tone signals less than approximately 200 msec in duration require more intensity to achieve threshold in normal ears and in ears with conductive and retrocochlear lesions than do signals greater than 200 msec in duration (Zwislocki, 1960; Olsen, Rose, and Noffsinger, 1974; Wright, 1968). The pure-tone signals used in this study were 50-msec full-amplitude duration with 10-msec rise-decay times. Therefore, before DFA results are compared with standard audiometrie results, one may wish to take into consideration the effects of temporal integration. We compared thresholds for the signals used in this study with thresholds for 500-msee signals of the same frequencies and rise-decay times. Twelve normal-hearing subjects were tested with a forced-choice up-and-down tracking procedure. The order of test frequency (500, 1000, and 2000 Hz) and duration (50 and 500 msec) were counterbalanced among subjects. The results of this subexperiment yielded mean absolute thresholds for the 500-msec signals that were within 2 dB of ANSI (1969) standards for normal hearing at the three frequencies tested. Mean temporal integration, as determined by the difference in thresholds for the 50 and 500 msec signals, was 7.0 dB at 500 Hz, 5.4 dB at 1000 Hz, and 5.2 dB at 2000 Hz. Temporal integration corrections were not made in this study because the signals used for the three methods were of the 762 1ournal o[ Speech and Hearing Research
18 754-764 1975
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
same duration, and comparisons were not made to standard audiometric results for these patients. DISCUSSION This study was performed with adult patients who were relatively unfamiliar with audiologic techniques and methods, and who had a wide range of ages and educational backgrounds. D F A thresholds were obtained in a greater number of the 100 patients than were E D R A thresholds (88 vs 73, respectively). Agreement of D F A thresholds with both VRA and E D R A thresholds was generally good, indicating that D F A is a valid measure of organic hearing level in most patients tested. In addition, a significant number of patients yielded D F A thresholds that were in good agreement with their VRA thresholds when E D R A thresholds could not be obtained. These findings indicate that D F A is a useful substitute for E D R A or can be used in conjunction with E D R A in the battery of tests for nonorganic hearing loss. Although D F A was not as precise in predicting voluntary thresholds as was E D R A in this study, its predictive value is within those limits (10 dB) that are normally accepted as providing a valid and useful clinical procedure for most patients. D F A is an easily administered and valuable tool for testing patients who are seen for medical-legal evaluation, and the technique is not noxious to the patient. D F A may yield a permanent recording of the patient's performance for future reference. In addition, data are now available for the quantitative determination of D F A thresholds based on normal variability of tapping performance without the effects of audible auditory feedback. ACKNOWLEDGMENT This project was supported by the Research and Education Committee of the Veterans Administration Hospital, Oklahoma City, Oklahoma, as a part of Proiect #M1-65. The authors thank Dr. W. A. Cooper, Jr., of Purdue University for his assistance in the formulation of this project in its initial stages of planning and review of the manuscript and data presented. Requests for reprints should be sent to Bradley L. Billings, Audiology Center of Redlands, 242 Caion Street, Redlands, California 92373. REFERENCES AMERICAN NATIONAL STANDARDSINSTITUTE, Criteria for background noise in audiometer rooms. ANSI S3.1-1960. New York: American National Standards Institute (1960). AMERICAN NATIONAL STANDARDSINSTITUTE, Specifications for audiometers. ANSI $3.6-1969. New York: American National Standards Institute (1969). COOPER, W. A., JR., STOKINGER,T. E., and BILLINGS, B. L., Pure-tone delayed auditory feedback: Intersubject variability. Paper presented at the Annual Convention of the American Speech and Hearing Association, Chicago (1971). OMEN, W. O., ROSE, D. E., and NOFFSINGER, D., Brief-tone audiometry with normal, cochlear, and eighth nerve tumor patients. Arch. Otol., 99, 185-189 (1974). RUHM, H. B., Rapid electrodermal audiometric procedure for testing adults. ]. Speech Hear•ng Dis., 26, 130-136 ( 1961 ). RUHM, H. B., and COOPER, W. A., JR., Low sensation level effects of pure tone delayed auditory feedback. ]. Speech Hearing Res., 5, 185-193 (1969.). BILLINGS, STOKINGER: Pure-Tone Thresholds
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
763
RUHM, H. B., and COOPER, W. A., Ja., Some factors that influence pure tone delayed auditory feedback. 1. Speech Hearing Res., 6, 223-237 (1963). RUHM, H. B., and CoovErt, W. A., j~., Delayed feedback audiometry. 1. Speech Hearing Dis., 29, 448-455 (1964). WRIGItr, H. N., The effect of sensori-neural hearing loss on threshold duration functions. 1. Speech Hearing Res., 11,842-852 (1968). ZWlSLOCKI, J., Theory of temporal summation. J. acoust. Soc. Am., 32, 1046-1060 (1960). Received October 15, 1974. Accepted July 18, 1975.
764 Journal of Speech and Hearing Research
18 754-764
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
1975
Audiology Center ol Redlands, California THOMAS E. STOKINGER
Veterans Administration Hospital, Oklahoma City, Oklahoma One hundred unselected patients seen for medical-legal evaluation were tested for pure-tone thresholds by delayed feedback audiometry (DFA), electrodermal response audiometry (EDRA), and voluntary response audiometry (VRA). The EDRA method was successful in 73g of the patients while the DFA method was successful in 88~ of thepatients. Eighty-six percent of the DFA thresholds obtained were within 10 dB of the patients' VRA thresholds. When both DFA and EDRA were successful, 88g of the DFA thresholds were within 10 dB of the EDRA thresholds. Ninety-six percent of the EDRA thresholds obtained were within 10 dB of the patients' VRA thresholds. Although DFA is not as precise in predicting threshold as is EDRA, it is successful in a significantly greater number of patients than is EDRA and is a useful clinical tool in medical-legal evaluation for hearing loss. The effects of pure-tone delayed feedback audiometry ( D F A ) have been described in a series of experiments reported by Ruhm and Cooper (1962, 1963, 1964). The DFA testing procedure requires that the patient repeatedly tap the pattern ( . . . . . . ) on a silent electromechanical key. Determination of threshold is based on a comparison of his tapping consistency during simultaneous auditory feedback (SAF) and delayed auditory feedback ( D A F ) at different intensity levels. During SAF, each tap on the electromechanical key results in a tone burst that is presented immediately to the patient's earphone. During DAF, each tap on the key results in a tone burst that is presented to the patient's earphone after a 200-msec delay. If the tone bursts are inaudible to the patient, his tapping performance should remain unaffected under both SAF and DAF conditions. If the tone bursts are audible to the patient, his tapping performance should not be affected under the SAF condition. The DAF condition, however, will result in modification of tapping performance including time errors (greater time is generally required to tap patterns under DAF than under SAF), number errors (usually the inclusion of extra taps in 754
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
the patterns under DAF), and the use of greater tapping pressure during DAF. Ruhm and Cooper's experiments ( 1962, 1963, 1964) have shown that: 1. A delay time of 200 msec is most effective in causing tapping disruption. 2. The effects of DAF are independent of the pure-tone signal frequency from 250 to 8000 Hz. 3. There are no significant differences between DAF tapping performance of males and females. 4. Short-term practice or sophistication with DFA does not significantly affect thresholds obtained. 5. Fatigue from repeatedly tapping the patterns a reasonable number of times does not significantly affect tapping performance. 6. Normal and hard-of-hearing patients generally yield DFA thresholds that are within 5 dB of their voluntary thresholds. 7. Patients who exhibit signs of nonorganic hearing loss generally yield DFA thresholds that are within 5 dB of their electrodermaI response thresholds. The effectiveness of DFA has been reported previously on the basis of results with either normal-hearing subjects or small groups of selected subjects with a hearing loss. It was the purpose of this study to investigate the validity of DFA as a clinical tool and to compare its effectiveness with electrodermal response audiometry in an unselected series of patients seen for medical-legal evaluation. In addition, this study was designed to yield data concerning the normal variability of key-tapping performance without the effects of auditory feedback for the purpose of establishing normative criteria for use in quantitative determination of DFA thresholds. METHOD
Subiects One hundred patients of the Veterans Administration Hospital Audiology and Speech Pathology Service, Oklahoma City, Oklahoma, served as subjects for this study. Each patient had his hearing tested by delayed feedback audiometry (DFA), electrodermal response audiometry (EDRA), and voluntary response audiometrv (VRA) at either 500, 1000, or 2000 Hz in one ear only.
Testing Order One half of the patients were administered DFA prior to EDRA and the other half of the patients received the reverse order of testing. Voluntary response audiometry was administered only after DFA and EDRA testing had been completed for all patients. Although all patients received a complete BXLLINCS,STOKINCEa:Pure-Tone Thresholds 755
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
audiologic evaluation by another examiner prior to this experiment, their audiograms were not available to the experimenter until after they had completed this study.
Signal Conditiol~s. The pure-tone signals used for DFA, EDRA, and VRA were of 50-msec fullamplitude duration with 10-msec rise-decay times. The tone bursts were generated by an audio oscillator (Hewlett-Packard, Model 200ABR) and shaped by an electronic switch (Grason-Stadler, Model 829C). The electronic switch was triggered by a timing mechanism that allowed for the generation of a tone burst immediately after each tap on the electromechanical key (SAF) or for the generation of a tone burst delayed 200 msec from each tap on tlle key (DAF). In addition, the tone bursts could be initiated manually by the experimenter for the purpose of obtaining voluntary response thresholds or could be initiated through a modified psychogalvanometer (Grason-Stadler, Model E664-1) for the purpose of obtaining EDR thresholds. The experimenter monitored the generation of all tone bursts through a loudspeaker and by observation of an oscilloscope. This was especially important during the DFA trials because it enabled the experimenter to monitor auditorily the patient's tapping performance.
Acoustic Environment All patients were tested in a sound-treated examination room (Industrial Acoustics, Model 400), and all test equipment except the electromechanical key and patient's earphones was located outside the testing room. The test room conformed to criteria for background noise in audiometric examination rooms (ANSI, 1960).
Calibration Since the pure-tone signals were shorter than the critical duration for puretone thresholds (see the Results section for further discussion), the experimental apparatus was calibrated for normal hearing on the basis of mean threshold results from 10 normal-hearing young adults. These subiects reported no history of otologic abnormalities or significant noise exposure and each passed a hearing screening test at octave frequencies from 250 to 4000 Hz (using a conventional audiometer) at 15 dB HL (ANSI, 1969) in the test ear. Intensity, frequency, and timing characteristics of the experimental signals were calibrated daily. Attenuator linearity was checked before and after the experiment. The electronic switch was balanced frequently.
Determination o[ Thresholds DFA thresholds were obtained by a modified method-of-limits procedure 756 Journal of Speech and Hearing Research
18 754-764 1975
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
using only ascending approaches to threshold with 5-dB increments. Each patient was instructed and given practice in tapping the required pattern ( . . . . . . ) prior to testing. After it was determined that the patient knew the instructions and could tap the patterns adequately, practice was terminated and testing begun. A shield was then placed in front of the patient so that he could not see either his hand or the tapping key. The patient was instructed to tap the pattern repeatedly as long as a signal light was on, and to stop tapping the pattern when the signal light was off. Eight patterns were tapped at each intensity level; the first four patterns under SAF and the last four patterns under DAF. A record of the patient's tapping performance was made with a calibrated strip-chart recorder (Techni-Rite, Model TR711) at a speed of 25 mm/sec. DFA testing was started at - 1 0 dB HL (as established on the normal-hearing group), and the intensity level was increased in 10-dB steps until the patient showed significant audible tapping disruption for the DAF condition as compared to the SAF condition (the experimenter monitored tapping performance through a loudspeaker). The intensity level was then decreased by 10 or 15 dB and another ascending approach to threshold was made in 5-dB steps. This procedure was repeated until the patient showed obvious tapping disruption twice at the same intensity level out of three ascending runs. Determination of DFA thresholds, according to normative standards as defined by this study, is discussed in the Results section of this article. EDRA thresholds were also obtained by a modified method-of-limits procedure using only ascending approaches to threshold with 5-dB intensity increments. The procedure used for EDRA was similar to that recommended by Ruhm (1961). From two to four conditioning trials were used and 100% reinforcement was maintained. To be classified as a significant response, an EDR tracing must have exceeded the baseline amplitude variation in quiet by a factor of two, must have possessed a latency of onset of from one to five seconds from the beginning of the tone burst, and must have had a slope greater than 45 degrees. EDRA (after conditioning) was started at - 1 0 dB HL (as established on the normal-hearing group), and the intensity levels were increased in 10-dB steps until the patient yielded a significant EDR. The intensity level was then decreased by 10 or 15 dB and another approach to threshold was made in 5-dB steps. This procedure was repeated until an EDR was elicited twice at the same intensity level out of three ascending runs.
VRA thresholds were also obtained by a modified method-of-limits procedure using only ascending approaches to threshold with 5-dB intensity increments. VRA was started at - 1 0 dB HL (as established on the normal-hearing group) and the intensity levels were increased in 10-dB steps until the patient indicated that he heard the signal by raising his hand. The intensity level was then decreased by 10 or 15 dB and another approach to threshold was made in 5-dB steps. This procedure was repeated until a voluntary response was elicited twice at the same intensity level out of three ascending runs. Bmt,mas, STOKINGER:eufe,-Tol'l,e, Thresholds 757
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
RESULTS
Normal Tapping Variability and Calculation of DFA Threshold It is not practical to calculate the tapping times and number errors for SAF and DAF at each intensity level during the actual testing session because of the length of time required, approximately 20 to 30 seconds for each set of eight patterns. Therefore, ascending approaches to threshold are usually terminated when the examiner is able to auditorily detect tapping performance disruption during DAF through a loudspeaker. Although this method is practical and relatively accurate, a quantitative method for determining DFA thresholds is necessary for medical-legal iustification of the obiectivity of DFA as a clinical technique. One hundred seventy-three groups of eight patterns (four SAF and four DAF) were obtained at - 1 0 dB HL from the 94 patients who were able to perform the task. Since - 1 0 dB HL was inaudible to these patients, as confirmed by voluntary thresholds and EDRA in many cases, differences in performance between SAF and DAF (that is, time and number errors) reflect normal variability of performance that is uncontaminated by the effects of audible auditory feedback, either SAF or DAF. Number Errors. Of the 173 groups of SAF and DAF patterns tapped at - 1 0 dB HL, the inclusion of extra taps during DAF compared to SAF occurred only 19 times (11~). Most number errors were limited to one extra tap with two extra taps occurring only one time. These data agree well with those presented by Cooper, Stokinger, and Billings (1971) in which 8~ of 50 normal-hearing subiects (cooperative volunteers) produced one or more extra taps during DAF compared to SAF at - 10 dB SL. Time Errors. The times required to tap the SAF and DAF pattern groups at - 1 0 dB HL were calculated from the strip-chart recorder as measured from the first tap of the first pattern to the last tap of the fourth pattern. The mean difference between the time required to tap the DAF pattern groups and the time required to tap the SAF pattern groups was 80 msec with a standard deviation of 384 msec. To compare these data with those presented by Cooper et al. (1971), the total time required to tap each pattern group was divided by four to yield an average time required to tap one pattern. The mean difference between average times required to tap one pattern under DAF and one pattern under SAF (that is, DAF minus SAF) was 20 msec with a standard deviation of .96.8 msec. These data are similar to the Cooper et al. data in which the mean difference between average times was 8.6 msec with a standard deviation of 71.4 msec. Calculation of DFA Thresholds. On the basis of the above data, disruption of tapping performance due to the audible effects of DAF was defined as the inclusion of one or more extra taps during DAF as compared to SAF, or a difference between the time required for tapping the SAF pattern group and the DAF pattern group (that is, four patterns of DAF minus four patterns of 758 Jourrurlof Speech and Hearing Research
18 754-764 1975
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
SAF) exceeding the range from - 6 8 8 msec to +848 msec (mean q-_ 2 SD). That is, variability of tapping perfolTnance that exceeds these limits is in excess of normal variability and therefore is considered to be the result of the audible effects of DAF. Each DFA threshold that was calculated on the basis of the normative criteria (N = 88) was compared to its estimated level as
determined by the examiner listening for tapping disruptions through a loudspeaker (see Method section). Sixty-one of the 88 thresholds (69.3%) were identical as determined by both methods, 17 (19.3%) were 5 dB different, and nine (10.2%) were 10 dB different. The examiner's judgment was not as
sensitive as was the application of the normative criteria to determine threshold in all but a few instances. Only one of the 88 thresholds estimated by the examiner was more than 10 dB worse than when calculated according to the normative standards.
Comparison of VRA, DFA, and EDRA Thresholds VRA, DFA, and EDRA thresholds were successfully obtained from 66 of the 100 patients at 500, 1000, or 2000 Hz. In view of Ruhm and Cooper's data (1963) showing that the effects of DAF are independent of frequency and because analysis of the data of this study did not reveal consistent differences with respect to frequency, the thresholds obtained at the three fre-
quencies were pooled. Figure 1 shows that 14 of the 66 patients (21.2%) who were successfully tested by VRA, DFA, and EDRA had identical thresholds by all three methods. Forty-one of the patients (62.1% or 21.2% + 40.9%) had threshold differences
among the three methods that did not exceed 5 dB, and 55 of the patients
I
0
21.21
I 4~,9%
I
~
I 27,3%
,~7,0Z
5
~ 10
21.2%
I
13.6%
J 10.6%
>
~ 20
20
3.0%
20
3.0% 5
I0
9.11
1,51
1.51 15
20
25
30
15 20 25 30 NUMBEO RFPATIENTS FICVBr 2. Difference between DFA and EDRA thresholds for individual patients 5
NUIIBEROF PATIENTS
FIOUBE1. Greatest difference among VRA, DFA, and EDRA thresholds for individual patients (N = 66).
(iv =
10
66).
BILLINGS, STOKINGEB:Pure-Tone Thresholds 759
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
(83.3% or 21.2% H- 40.9% + 21.2%) had threshold differences among the three methods that did not exceed 10 dB. Eleven of the 66 patients (16.6%) had threshold differences among the three methods that exceeded 10 dB. One of these patients had DFA and EDRA thresholds that were considerably more sensitive than his voluntary threshold, and he had exhibited signs of nonorganic hearing loss on previous audiologic evaluations.
Comparison of DFA and EDRA Thresholds The relation between thresholds for the 66 patients who were successfully tested by both DFA and EDRA is illustrated in Figure 2. Eighteen of the 66 patients (27.3%) had identical thresholds as measured by both methods. Forty-nine of the patients (74.3%) had threshold differences between the two methods that did not exceed 5 dB, and 58 of the patients (87.9%) had threshold differences between the two methods that did not exceed 10 dB. Eight of the 66 patients had threshold differences between DFA and EDRA that exceeded 10 dB. Of these eight patients, four had DFA thresholds that were more than 10 dB better than their EDRA thresholds and four had DFA thresholds that were more than 10 dB worse than their EDRA thresholds.
Comparison of VRA and DFA Thresholds VRA and DFA thresholds were successfully obtained for 88 of the 100 patients. Figure 3 shows that 37 of the 88 patients (42.0%) who were successfully tested by both VRA and DFA had identical thresholds by both methods. Sixty-five of the patients (73.8%) had threshold differences that did not exceed 5 dB between the two methods, and 76 of the patients (86.3%) had threshold
,~.
50.7%
I 42.0%
o
5
]
I 32.96
31,8%
12.3%
,~ ]5
1.46
I 8,0~
2O }20~
~,5~ :0
>20
a5
20
2~
~o
3~
0.0%
5
~0~BER oFPAUENTS Fmums 3. Difference between VRA and DFA thresholds for individual patients (N = 88).
2.7%
10
15
20
25
30
35
NUMBER OF PATIENTS
Fmtrm~ 4. Difference between VRA and EDRA thresholds for individual patients (N = 73).
760 1ournalo[ Speech and Hearing Research
18 754-764 1975
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
differences between the two methods that did not exceed 10 dB. Twelve of the patients (13.670) had DFA thresholds that were more than 10 dB different from their VRA thresholds. Of these 12 patients, five had EDRA thresholds that were within 5 dB of their VRA thresholds, indicating that the DFA thresholds are probably in error. Two patients could not be tested by EDRA but yielded DFA thresholds that were considerably better than their VRA thresholds, and they had exhibited signs of nonorganic hearing loss on previous audiologic evaluations. Twelve of the 100 patients did not yield DFA thresholds. Six of these 12 patients did not show tapping disruption at suprathreshold levels under DAF, and six of the patients either could not tap the required patterns or could not maintain the consistent tapping performance required for complete testing.
Comparison of VRA and EDRA Thresholds Seventv-three of the 100 patients were successfully tested by both VRA and EDRA. Figure 4 shows that 37 of the 73 patients (50.7g) had identical thresholds by both methods. Sixty-one of the patients (83.6~) had thresholds that did not differ by more than 5 dB between the two methods, and 70 of the patients (95.9%) had threshold differences between the two methods that did not exceed 10 dB. Three of the 73 patients (4.17o) had EDRA thresholds that were more than 10 dB different from their VRA thresholds. One of these three patients exhibited signs of nonorganic hearing loss on previous audiologic evaluations. Twentv-seven of the 100 patients could not be or were not tested by EDRA. Nineteen of these patients could not be conditioned by the procedure or did not maintain sufficient conditioning to complete testing. The remaining eight patients could not be tested by EDRA because of a history of cardiac problems, motor tremors, and so on. Twenty of the 27 patients who could not be tested by EDRA were successfully tested by DFA and only three of the 20 patients had DFA thresholds that were more than 10 dB different from their VRA thresholds.
Hearing Loss Characteristics of the Patients Thirty of the 100 patients had hearing threshold levels that exceeded 15 dB in reference to the previously established normal calibration data. The degree of hearing loss did not affect the percentage of patients who could be tested or the threshold differences as measured by VRA, DFA, or EDRA.
Factors Affecting the Use of Delayed Feedback Audiometry During the course of this experiment, several aspects of the administration and interpretation of DFA were also investigated. DFA Testing Procedure. DFA instructions are usually given to the patient BILLINCS, STOKINCER:Pure-Tone. Thresholds 761
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
in a manner that illustrates the desired tapping pattern by telling him to tap the pattern "tap tap tap tap-tap tap" and simultaneously showing him the pattern by tapping it on the tapping key table. The suggestion of a mental task, such as "count to yourself, 1 2 3 4-1 2" was usually avoided because it was believed that such a mental activity may allow the patient to more effectively resist the effects of DAF than if mental counting was not used. The mental counting technique, however, was found to be necessary to properly train 19 of the patients in this study. Nine of the 19 patients had DFA thresholds that were identical to their VRA thresholds and 16 of the 19 patients had DFA thresholds that were within 5 dB of their VRA thresholds. Only one of the patients had a DFA threshold that was more than 10 dB different from his VRA threshold. These results suggest that the use of a mental counting task is helpful in training some patients to tap the required patterns without resulting in higher resistance to the effects of DAF than would otherwise be the case. Rate of Tapping. Observation during the initial use of DFA suggested the possibility that patients who tapped very rapidly or very slowly might be less resistant to the effects of DAF than patients who tapped at an average speed. The time required to tap the first four SAF patterns at -- 10 dB HL was analyzed for the 94 patients who could tap the required patterns satisfactorily, yielding a median time of 8.8 seconds with a range of 5.9 seconds to 14.5 seconds. All but one of the 10 patients who did not show tapping disruption under suprathreshold DAF tapped the SAF patterns faster than the median time of 8.8 seconds. It was not possible, however, to predict whether a patient would be more or less likely to show tapping disruption under DAF on the basis of his tapping rate. Many "rapid" tappers and many "slow" tappers had DFA thresholds that were equivalent to both their VRA and EDRA thresholds. Temporal Integration and DFA. Pure-tone signals less than approximately 200 msec in duration require more intensity to achieve threshold in normal ears and in ears with conductive and retrocochlear lesions than do signals greater than 200 msec in duration (Zwislocki, 1960; Olsen, Rose, and Noffsinger, 1974; Wright, 1968). The pure-tone signals used in this study were 50-msec full-amplitude duration with 10-msec rise-decay times. Therefore, before DFA results are compared with standard audiometrie results, one may wish to take into consideration the effects of temporal integration. We compared thresholds for the signals used in this study with thresholds for 500-msee signals of the same frequencies and rise-decay times. Twelve normal-hearing subjects were tested with a forced-choice up-and-down tracking procedure. The order of test frequency (500, 1000, and 2000 Hz) and duration (50 and 500 msec) were counterbalanced among subjects. The results of this subexperiment yielded mean absolute thresholds for the 500-msec signals that were within 2 dB of ANSI (1969) standards for normal hearing at the three frequencies tested. Mean temporal integration, as determined by the difference in thresholds for the 50 and 500 msec signals, was 7.0 dB at 500 Hz, 5.4 dB at 1000 Hz, and 5.2 dB at 2000 Hz. Temporal integration corrections were not made in this study because the signals used for the three methods were of the 762 1ournal o[ Speech and Hearing Research
18 754-764 1975
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
same duration, and comparisons were not made to standard audiometric results for these patients. DISCUSSION This study was performed with adult patients who were relatively unfamiliar with audiologic techniques and methods, and who had a wide range of ages and educational backgrounds. D F A thresholds were obtained in a greater number of the 100 patients than were E D R A thresholds (88 vs 73, respectively). Agreement of D F A thresholds with both VRA and E D R A thresholds was generally good, indicating that D F A is a valid measure of organic hearing level in most patients tested. In addition, a significant number of patients yielded D F A thresholds that were in good agreement with their VRA thresholds when E D R A thresholds could not be obtained. These findings indicate that D F A is a useful substitute for E D R A or can be used in conjunction with E D R A in the battery of tests for nonorganic hearing loss. Although D F A was not as precise in predicting voluntary thresholds as was E D R A in this study, its predictive value is within those limits (10 dB) that are normally accepted as providing a valid and useful clinical procedure for most patients. D F A is an easily administered and valuable tool for testing patients who are seen for medical-legal evaluation, and the technique is not noxious to the patient. D F A may yield a permanent recording of the patient's performance for future reference. In addition, data are now available for the quantitative determination of D F A thresholds based on normal variability of tapping performance without the effects of audible auditory feedback. ACKNOWLEDGMENT This project was supported by the Research and Education Committee of the Veterans Administration Hospital, Oklahoma City, Oklahoma, as a part of Proiect #M1-65. The authors thank Dr. W. A. Cooper, Jr., of Purdue University for his assistance in the formulation of this project in its initial stages of planning and review of the manuscript and data presented. Requests for reprints should be sent to Bradley L. Billings, Audiology Center of Redlands, 242 Caion Street, Redlands, California 92373. REFERENCES AMERICAN NATIONAL STANDARDSINSTITUTE, Criteria for background noise in audiometer rooms. ANSI S3.1-1960. New York: American National Standards Institute (1960). AMERICAN NATIONAL STANDARDSINSTITUTE, Specifications for audiometers. ANSI $3.6-1969. New York: American National Standards Institute (1969). COOPER, W. A., JR., STOKINGER,T. E., and BILLINGS, B. L., Pure-tone delayed auditory feedback: Intersubject variability. Paper presented at the Annual Convention of the American Speech and Hearing Association, Chicago (1971). OMEN, W. O., ROSE, D. E., and NOFFSINGER, D., Brief-tone audiometry with normal, cochlear, and eighth nerve tumor patients. Arch. Otol., 99, 185-189 (1974). RUHM, H. B., Rapid electrodermal audiometric procedure for testing adults. ]. Speech Hear•ng Dis., 26, 130-136 ( 1961 ). RUHM, H. B., and COOPER, W. A., JR., Low sensation level effects of pure tone delayed auditory feedback. ]. Speech Hearing Res., 5, 185-193 (1969.). BILLINGS, STOKINGER: Pure-Tone Thresholds
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
763
RUHM, H. B., and COOPER, W. A., Ja., Some factors that influence pure tone delayed auditory feedback. 1. Speech Hearing Res., 6, 223-237 (1963). RUHM, H. B., and CoovErt, W. A., j~., Delayed feedback audiometry. 1. Speech Hearing Dis., 29, 448-455 (1964). WRIGItr, H. N., The effect of sensori-neural hearing loss on threshold duration functions. 1. Speech Hearing Res., 11,842-852 (1968). ZWlSLOCKI, J., Theory of temporal summation. J. acoust. Soc. Am., 32, 1046-1060 (1960). Received October 15, 1974. Accepted July 18, 1975.
764 Journal of Speech and Hearing Research
18 754-764
Downloaded From: https://jslhr.pubs.asha.org/ by a University Library Utrecht User on 04/17/2018 Terms of Use: https://pubs.asha.org/ss/rights_and_permissions.aspx
1975
