Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Speech discrimination in noise for listeners with normal hearing and listeners with noise-induced hearing… Mack, Brenda Elisabet 1978

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

Item Metadata

Download

Media
831-UBC_1978_A6_7 M34.pdf [ 3.61MB ]
Metadata
JSON: 831-1.0094281.json
JSON-LD: 831-1.0094281-ld.json
RDF/XML (Pretty): 831-1.0094281-rdf.xml
RDF/JSON: 831-1.0094281-rdf.json
Turtle: 831-1.0094281-turtle.txt
N-Triples: 831-1.0094281-rdf-ntriples.txt
Original Record: 831-1.0094281-source.json
Full Text
831-1.0094281-fulltext.txt
Citation
831-1.0094281.ris

Full Text

SPEECH DISCRIMINATION IN NOISE FOR LISTENERS WITH NORMAL HEARING AND LISTENERS WITH NOISE-INDUCED HEARING LOSS  by BRENDA ELISABET MACK B.A., U n i v e r s i t y o f B r i t i s h Columbia, 1972  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES ( DEPARTMENT OF PAEDIATRICS )  ( DIVISION OF AUDIOLOGY AND SPEECH SCIENCES )  We accept t h i s t h e s i s as conforming t o t h e r e q u i r e d standard  THE UNIVERSITY OF BRITISH COLUMBIA August, 197S (c) Brenda Elisabet Mack, 1978  In  presenting this  thesis  an advanced degree at the I  Library shall  f u r t h e r agree  for  scholarly  by h i s of  this  written  make i t  freely available  that permission  Columbia,  I agree  r e f e r e n c e and  f o r e x t e n s i v e copying o f  this  It  for financial  is understood that gain s h a l l  not  copying or  University of B r i t i s h  2075 Wesbrook P l a c e V a n c o u v e r , Canada V6T 1W5  CJLuqasl  Colximbia  <_\pteck Sciences  that  thesis or  publication  be allowed without my  of  for  study.  purposes may be granted by the Head of my Department  Aoifolo^u  Date  for  the requirements  permission.  Department  The  fulfilment of  the U n i v e r s i t y of B r i t i s h  representatives. thesis  in p a r t i a l  - i i-  ABSTRACT  The present study investigates the effect of low-pass noise on the speech discrimination performance of IB subjects with normal hearing and 18 subjects with noise-induced highfrequency hearing l o s s .  W-22  word l i s t s , low-pass f i l t e r e d  at 2000 Hz, were presented i n sound f i e l d with a pink noise masker at three stimulus levels and three signal-to-noise ratios.  Results indicated that the word discrimination  performance of both groups deteriorated with increasing l e v e l s of noise and with increasing stimulus i n t e n s i t y l e v e l s , with the hearing-impaired group performing at a lower l e v e l throughout.  While noise was shown to have a d i f f e r e n t i a l  effect on the speech discrimination of the two groups, a s a t i s f a c t o r y explanation of the effect, based on the study of Kiang and Moxon ([Tails of tuning curves of auditory nerve fibres.  Journal of the Acoustical Society of America, 1974>  55, 620-63Q~f was not supported using the present conditions.  experimental  - iii  -  TABLE OF CONTENTS Chapter  Page  ABSTRACT  i i  TABLE OF CONTENTS  .  i i i  LIST OF TABLES  . .  v  LIST OF FIGURES  vi  ACKNOWLEDGEMENT  v i i  1.  INTRODUCTION  1  2.  REVIEW OF LITERATURE  3  2.1  The e f f e c t of n o i s e on word d i s c r i m i n a t i o n 3  performance 2.2  2.3  Factors  c o n t r i b u t i n g t o reduced  speech  discrimination i n noise  16  P h y s i o l o g i c a l basis  24  3.  OBJECTIVES  28  4.  METHOD  29  4.1  Design  29  4.2  Subjects  29  4.3  Stimuli  31  4.3.1  Description  4.3.2  Preparation  31 o f Stimulus Tapes  37  4.4  Equipment  38  4.5  Procedure  40  -  iv-  5.  RESULTS  43  6.  DISCUSSION AND CONCLUSIONS  55  REFERENCES  68  -  V  LIST OF TABLES  Table  1  Page  Means and ranges of pure-tone thresholds of the Normal group and the Patient group at the test frequencies  32  Means, ranges, and standard deviations of Word Discrimination Scores f o r 18 Normal subjects  44  Means, ranges, and standard deviations of Word Discrimination Scores f o r 18 Patients with high-frequency hearing loss  45  Slopes representing the regression functions of the mean Word Discrimination Scores achieved by Normal and Patient groups versus the 3 S/N r a t i o s at the 3 stimulus l e v e l s . . .  51  5  Summary of 3-way analysis of variance of the r e s u l t s of the present study  52  6  Summary of 2-way analysis of variance of the r e s u l t s of the present study  54  2  3  4  - vi -  LIST OF FIGURES  Figure  Page  Mean pure-tone thresholds i n dB HL (ANSI 1969) of the Normal group (/_\) and the Patient group (A) at the test frequencies. Also shown i s the slope of the s k i r t of the low-pass f i l t e r used to process the speech s t i m u l i and noise ( )  33  The one-third octave band analysis of the pink noise masker measured at the e a r - l e v e l of the subject i n position i n the test booth . ••. . . . . 41 Mean WDS's of the a) NormaL group (open symbols), and b) Patient group (closed symbols) at 3 S/N r a t i o s f o r stimulus l e v e l s of 60 dB ( t r i a n g l e s ) , 70 dB ( c i r c l e s ) , and 80 dB (squares). The regression l i n e s of WDS's versus S/N r a t i o s for each stimulus l e v e l are represented by for the 60 dB group, for the 70 dB group, and for the 80 dB group 46 Mean WDS's of the Normal and Patient groups at 3 S/N r a t i o s for stimulus levels of a) 60 dB, b) 70 dB, and c) 80 dB. Symbols and regression l i n e s as defined i n Figure 3  48  Mean WDS's of the Normal group (open symbols) and Patient group (closed symbols) at the 3 stimulus levels f o r S/N ratios of 1-19 dB ( t r i a n g l e s ) , +12 dB ( c i r c l e s ) , and +5 dB (squares) . . . . . . .  60  - vii -  ACKNOWLEDGEMENT  I would l i k e to express ray thanks to a l l who had a part i n this thesis: Dr. David Chung f o r a l l his guidance during the research and w r i t i n g of the t h e s i s . Miss Noelle Lamb f o r serving on my committee. Dr. P. Gannon and the personnel of the Hearing Branch of the Workers' Compensation Board of B.C. for t h e i r help and the use of t h e i r f a c i l i t i e s f o r carrying out the necessary research. My subjects f o r t h e i r kind cooperation. Bob f o r a l l his understanding  and encouragement over the  past two years. Susan f o r her patient work on the figures. My fellow students f o r t h e i r friendship and support.  CHAPTER 1  1.  INTRODUCTION  Patients with high-frequency noise-induced hearing loss frequently report d i f f i c u l t i e s understanding speech i n a noisy environment yet experience l i t t l e d i f f i c u l t y i n a quiet listening situation.  People with normal hearing, on the  other hand, r a r e l y report any such problems i n quiet or noisy situations.  Since patients with high-frequency losses have  normal hearing through the range usually accepted as important f o r speech i n t e l l i g i b i l i t y (500 to 2000 Hz), i t seems improbable that they should react d i f f e r e n t l y to noise than do people with normal hearing.  Nonetheless, several studies i n the  laboratory s i t u a t i o n support t h e i r complaints.  These studies  have shown that word discrimination performance does d e t e r i o r ate more r a p i d l y i n noise f o r subjects with high-frequency sensorineural hearing loss than for normal l i s t e n e r s .  Few  s a t i s f a c t o r y explanations, however, have been given for t h i s phenomenon. One recent physiological study (Kiang and Moxon, 1974) does provide s i g n i f i c a n t support for the complaints of these individuals.  It  proposes that neurons with a high character-  i s t i c frequency, available to normal l i s t e n e r s but absent i n  - 2 listeners tional  h e a r i n g l o s s , may  provide  addi-  i n f o r m a t i o n on s p e e c h w h i c h a i d s d i s c r i m i n a t i o n i n  noise. for  with high-frequency  The  findings of this  study  provide a promising  basis  the a n a l y s i s o f speech d i s c r i m i n a t i o n i n n o i s e .  The  present  study,  t h e r e f o r e , was  undertaken to  determine  w h e t h e r n o i s e does d i f f e r e n t i a l l y a f f e c t  t h e word d i s c r i m i n a -  t i o n performance o f normal l i s t e n e r s  l i s t e n e r s with  induced  hearing loss,  explanation  for this  and  i f so,  difference.  and  to provide a  noise  satisfactory  - 3 CHAPTER 2  2.  REVIEW OF LITERATURE  2.1  THE EFFECT OF NOISE ON WORD DISCRIMINATION PERFORMANCE  Word recognition tests are commonly used by audiology c l i n i c s to a s s i s t i n the d i f f e r e n t i a l diagnosis of various hearing impairments.  Many individuals with high-frequency  sensorineural hearing l o s s , however, show no decrease i n performance on standard speech discrimination t e s t s . widely employed CID Auditory  The  Test W-22 (W-22), when presented  i n quiet, has been shown to be of l i m i t e d value i n separating normal-hearing patients from those with auditory pathology (Carhart, 1965;  Keith and T a l i s , 1972;  Sher and Owens, (1974).  Data published by Carhart (1965) on the word recognition scores of 170 hearing-impaired veterans tested with W-22 recordings, revealed that 60% achieved scores of 90$ or better.  In t h i s case, the speech discrimination test had  f a i l e d to i d e n t i f y either the presence, the type, or the extent of hearing loss i n 60% of the patients.  His  findings  were substantiated by Keith and T a l i s (1972) with c l i n i c a l data from 170 of t h e i r patients with sensorineural hearing losses.  Sher and Owens (1974), following subject c r i t e r i a  s i m i l a r to that used i n the present study, observed mean scores  - 4 -  of 94.6$ on the W-22 test f o r 35 l i s t e n e r s with normal hearing to 2000 Hz and high-frequency cochlear hearing l o s s e s .  In the  l i g h t of studies such as these, speech discrimination t e s t i n g i n quiet appears to be of l i t t l e diagnostic value for many people. Several studies have examined the diagnostic value of speech discrimination t e s t i n g i n the presence of noise. Simonton and Hedgecock (1953) used a mixture of white noise and two pure tones, 60 Hz and 112 Hz, with the Harvard Phonetically Balanced word l i s t s  (PB-50's), to study the  effects of noise on normal and hearing-impaired subjects. They found no difference i n the performance of l i s t e n e r s with normal hearing or with conductive hearing losses.  Subjects  with sensorineural hearing losses, however, showed increased l o s s of discrimination when tested i n noise.  Similar r e s u l t s  were reported by Palva (1955) using a continuous white noise masker and a +10 dB signal-to-noise  (S/N)  ratio.  Speech  discrimination scores for his sensorineural-impaired l i s t e n e r s suffered " d i s t i n c t l y , and i n some cases severely" (Palva, 1955 > p. 2 4 0 ) .  Palva concluded that speech discrimination t e s t i n g  i n noise might be useful i n the diagnosis of perceptive (sensorineural) deafness.  Ross et a l . (1965), as part of a  l a r g e r study, examined whether speech discrimination t e s t i n g i n white noise had any c l i n i c a l u t i l i t y .  The absolute d i s -  crimination scores achieved by t h e i r hearing-impaired subjects  - 5 were poorer than those of the normal-hearing subjects for both the quiet and noise conditions.  However, the r e l a t i v e d i s -  crimination s h i f t due to noise f a i l e d to show any s i g n i f i c a n t differences between the two groups.  Ross et a l .  (1965),  suggested that the use of d i f f e r e n t kinds and sensation l e v e l s of noise could result i n the desired d i f f e r e n t i a t i o n of the groups. Cooper and Cutts (1971) examined changes occurring i n the slope of the a r t i c u l a t i o n function with the introduction of noise.  Northwestern University Auditory Test Number 6  (NU-6) was presented monaurally with c a f e t e r i a noise at S/N r a t i o s of +5, +3,  and -+12 dB.  Consistent with the previous  studies, the mean performance of the sensorineural group was s i g n i f i c a n t l y poorer than that of the normal group. found i n Ross et a l . ' s  But, as  (1965) data, the slopes of the  a r t i c u l a t i o n functions were not s i g n i f i c a n t l y d i f f e r e n t f o r the two groups. In a study by Keith and T a l i s ( 1 9 7 2 ) , W-22 words were mixed with white noise i n an attempt to provide a more d e f i n i t i v e d i f f e r e n t i a l diagnosis of hearing impairments. Subjects with normal hearing, high-frequency cochlear losses, and f l a t cochlear losses were tested i n quiet and at three d i f f e r e n t S/N r a t i o s (-1-8, 0, -8 dB). decreased from  As the S/N r a t i o  8 to -8 dB, the discrimination scores  - 6 deteriorated and the difference i n mean scores among the groups increased.  The mean score of the normal-hearing group  deteriorated approximately 52% from the quiet condition to -8 dB S/N r a t i o , the mean score of the group with highfrequency losses deteriorated approximately 57%, and the mean score of the group with f l a t losses deteriorated approximately 67%.  These results indicated that speech discrimination  testing i n the presence of noise could help d i f f e r e n t i a t e between patients with cochlear hearing impairments and normal listeners. One s t r i k i n g f i n d i n g common to a l l of these studies of speech discrimination performance i n the presence of noise was the extreme v a r i a b i l i t y among subjects, both normal and hearing-impaired.  In a study of normal-hearing individuals,  Rupp and P h i l l i p s (1969) found marked v a r i a b i l i t y i n individual performance, both i n white noise and i n speech-spectrum noise. One normal group tested i n speech-spectrum noise at OdB S/N r a t i o a c t u a l l y demonstrated a range of 88% i n i n d i v i d u a l performance.  Simonton and Hedgecock ( 1 9 5 3 ) , Palva (1955)> and  Ross et a l . (1965) noted wider variations i n the scores of t h e i r sensorineural-impaired subjects than i n t h e i r normalhearing groups.  Discrimination scores of the impaired group  i n Cooper and Cutts' (1971) experiment increased i n v a r i a b i l i t y with lower S/N r a t i o s .  At two of the S/N r a t i o s tested,  - 7 -  they demonstrated a range of scores twice that of the normal group.  This l e d the authors to conclude that something more  than a simple masking effect was operating to reduce the performance in noise of t h i s impaired population.  Keith and  T a l i s (1972) found that the use of increasing l e v e l s of a white noise masker resulted i n increasingly wider ranges of scores within a l l three subject groups.  This finding made  the diagnosis of a p a r t i c u l a r hearing impairment based on speech discrimination scores very d i f f i c u l t .  C l e a r l y , the  v a r i a b i l i t y introduced by the use of white noise with speech limited i t s diagnostic value. In an i n t e r e s t i n g approach to t h i s problem, Olsen et a l . (1975) attempted to determine whether t h i s v a r i a b i l i t y of r e s u l t s , i t s e l f , could be diagnostically useful i n d i f f e r e n t i a t i n g kinds of hearing impairments.  They presented NU-6  word l i s t s i n quiet and i n white noise (OdB S/N ratio) to six groups of subjects including normal l i s t e n e r s and subjects with various types of auditory pathologies.  Differences of  l+Ofo or more between scores i n quiet and i n noise were observed for fewer than 1% of the normal ears tested.  Similar d i f f e r -  ences were found, however, f o r 8% of the ears with noise trauma (high-frequency loss) and Lfifo of the ears with Meniere's disease ( f l a t l o s s ) .  These findings  Keith and T a l i s (1972).  concurred with those of  Olsen et a l . (1975) concluded that a  finding of a large difference i n scores obtained under these  - aconditions i n quiet and i n noise f o r either a normal l i s t e n e r or a l i s t e n e r with a high-frequency sensorineural hearing loss could be interpreted as i n d i c a t i n g neural involvement somewhere i n the auditory system.  Therefore, r e s u l t s from speech-  t e s t i n g i n noise could be useful i n revealing abnormal auditory function " . . . but not i n suggesting a p a r t i c u l a r s i t e of involvement as being responsible f o r the dysfunction"  (Olsen  et a l . , 1975, p. 3^2). In t h e i r concluding remarks, Keith and T a l i s (1972) maintained that speech-in-noise  t e s t i n g could only become  d i a g n o s t i c a l l y f e a s i b l e i f the v a r i a b i l i t y o f r e s u l t s could be reduced.  They proposed a l t e r i n g the masking noise by  using lowpass f i l t e r e d noise as suggested by Liden  (196?).  Liden had stated that simultaneous presentation o f word l i s t s with a 500 Hz low-pass f i l t e r e d white noise at a S/N r a t i o of -3 dB, made discrimination test r e s u l t s more sensitive diagnostically.  He noted that i n such a noise, the scores  of patients with high-frequency sensorineural hearing losses and normal discrimination i n quiet could drop to as low as 10 to 20%, while a normal-hearing i n d i v i d u a l maintained 90% intelligibility. On the basis of Liden's suggestion, Cohen and Keith (1976) devised a study to determine whether normal and hearing-impaired subjects could be d i f f e r e n t i a t e d without increas-  - 9  -  ing the v a r i a b i l i t y of t h e i r scores.  A 500 Hz, low-pass  f i l t e r e d white noise was mixed with taped W-22 and presented monaurally to subjects with 1)  word l i s t s normal-hearing,  2) high-frequency hearing loss, and 3) f l a t hearing l o s s . Speech discrimination performance was then measured i n quiet and at -4 and -8 dB S/N r a t i o s .  Test results confirmed not  only Liden's (1967) findings, but also the suggestion of Ross et a l . (1965) that d i f f e r e n t types and sensation l e v e l s of noise could provide clearer d i f f e r e n t i a t i o n of speech d i s crimination performance.  While scores of the three groups  were s i m i l a r i n quiet, the more negative the S/N r a t i o , the greater the separation of group scores.  The low-pass noise  provided a greater degree of separation of the groups than did the white noise employed i n Keith and T a l i s ' s (1972) study. Type and l e v e l of continuous noise used as the speech masker, therefore, appears to be one of the c r i t i c a l factors i n the d i f f e r e n t i a t i o n of normal and hearing-impaired  listeners.  Several recent studies have investigated the effects of other types of noise on speech discrimination performance. Carhart and Tillman (1970) measured discrimination f o r monosyllables against a background of competing sentences. Four groups comprising subjects with 1) normal hearing, 2) conductive losses, and sensorineural impairments with 3) good discrimination, and with 4) f a i r discrimination, were tested  - 10 -  monaurally i n a sound f i e l d .  Northwestern University test 2  was presented i n quiet and at four S/N r a t i o s (+12, -f6, 0, and -6 dB).  Interference functions plotted from t h i s data  revealed that subjects with conductive losses performed as well as those with normal hearing when the sensation l e v e l of the  signal was the same f o r both groups.  Performances of both  sensorineural groups, i n contrast, were s i g n i f i c a n t l y d i s turbed by the competing sentences.  Their interference  functions were shifted 12 to 15 dB to the right of a reference function (plotted from previous normal data) as i f the masking e f f i c i e n c y of the sentences had increased by 12 to 15 dB compared to that exhibited f o r the normal and conductive subjects.  Carhart and Tillman (1970) suggested, therefore,  that the presence of sensorineural hearing loss reduced the subject's a b i l i t y to r e s i s t interference from the competing speech.  They noted that a s i m i l a r "overmasking" effect can  occur when the competition i s a spectrally complex steady-state noise but to a much l e s s e r degree than that found with the competing sentences.  Thus, t r a d i t i o n a l l y used maskers such  as white or speech spectrum noise might provide less s a t i s factory competition because they e l i c i t l e s s "overmasking". The effects of modulated noise on the speech i n t e l l i g i b i l i t y of hearing impaired l i s t e n e r s was examined by Shapiro et a l .  (1972).  They presented speech with continuous or  modulated white noise monaurally under headphones at four  - 11 S/N r a t i o s (-8, was  NU-6  -12,  -16, and -20 dB).  monosyllabic words.  The speech material  Subjects with sensorineural  losses performed poorly under a l l experimental ions, especially i n continuous  noise condit-  noise.  As i n previous studies,  the mean performance of t h i s group was  consistently lower than  that of the normal-hearing group.  Contrary to other findings,  however, the slope for the normal-hearing subjects was  con-  siderably steeper with increasingly negative S/N r a t i o s than that for the hearing-impaired  subjects.  no explanation f o r t h i s unusual finding.  The authors offered The difference i n  scores between t h e i r two subject groups was less than that found i n Carhart and Tillman's (1970) study, a f a c t , they attributed to the d i f f e r e n t spectra of the i n t e r f e r i n g noises. F i n a l l y , two studies by Findlay (1976), and Findlay and Denenberg (1977), compared the a b i l i t y of normal and noiseexposed subjects to discriminate speech under d i f f i c u l t l i s t e n i n g conditions.  The noise-exposed group had normal  thresholds to 2000 Hz with high-frequency hearing losses.  sensorineural  Findlay found s i g n i f i c a n t differences between  the groups on three speech discrimination tasks:  PB-50 l i s t s  presented at 40 dB sensation l e v e l (SL), and W-22  l i s t s pre-  sented at 30 dB SL i n the presence of either speech-spectrum noise or " c o c k t a i l party" noise.  The use of W-22  lists in  c o c k t a i l party noise provided the greatest and most consistent  - 12 -  d i f f e r e n t i a t i o n between the two groups, i n agreement with previous findings using competing speech as maskers. In t h e i r continuation of this study, Findlay and Denenberg (1977) compared the performances of a group with normal hearing and two groups with high-frequency hearing l o s s :  a  younger group with predominantly noise-induced hearing loss and an older group with presbycusis and some noise exposure. Again, ¥-22 words were presented with competing c o c k t a i l party noise.  Two test conditions were applied:  one with  words and noise u n f i l t e r e d , the other with both words and noise low-pass f i l t e r e d at 1800 Hz.  This condition was to  rule out any effects on speech discrimination performance r e s u l t i n g from differences i n the high frequency s e n s i t i v i t y of the normals and the hypacusics. employed.  A -4 dB S/N r a t i o was  In the u n f i l t e r e d condition, the discrimination  performance of the normal l i s t e n e r s was s i g n i f i c a n t l y better than either hearing-impaired group.  In the low-pass f i l t e r e d  condition, however, the younger, noise-exposed subjects achieved a s i g n i f i c a n t l y higher l e v e l of performance than d i d the normal-hearing group.  Findlay and Denenberg (1977)  ventured a possible explanation for t h i s t o t a l l y unexpected result:  perhaps the normal l i s t e n e r s r e l i e d heavily on high-  frequency cues to discriminate speech in noise and therefore encountered considerable d i f f i c u l t y when presented with only  - 13 -  mid-frequency information.  Listeners with l i m i t e d high-  frequency hearing, on the other hand, might have learned to make better use of t h e i r mid-frequency hearing and hence, adjusted more r e a d i l y to the f i l t e r e d condition. Bilger et a l . (1974, 1976) also found that sensorineural hearing loss does not always cause l i s t e n e r s to perform poorl y in noise.  Ten subjects with normal hearing and eighteen  with moderate sensorineural hearing losses were tested i n quiet and i n noise for a consonant recognition task.  This  consisted of sixteen consonants paired with three d i f f e r e n t vowels i n a consonant-vowel (GV) context.  Ten l i s t s of 96  s y l l a b l e s were then presented monaurally v i a headphones at 100 dB SPL with competing broad-band noise at 95 dB SPL.  Upon  introduction of the noise, the mean per cent recognition scores of the normal group dropped from 75$ to 50%.  Of the eighteen  subjects with sensorineural hearing l o s s , however, nine demonstrated no adverse effect i n noise and only four showed the drop seen i n the normal group.  Those with low performanc-  es i n quiet tended to perform as well i n noise as i n quiet. In t h e i r discussion of these findings Bilger et a l . (1976) suggested that the a b i l i t y to t o l e r a t e noise or to l i s t e n i n noise i s distributed independently of sensorineural hearing l o s s .  It  i s only that people with hearing losses  complain about i t , whereas normal l i s t e n e r s do not.  This  - 14 -  remark c e r t a i n l y merits serious consideration.  However, t h e i r  observation that sensorineural hearing loss does not always r e s u l t i n decreased performance was based on results from subjects described simply as having "appreciable sensorineural l o s s " (p. 393).  Neither the configuration nor the extent of  t h e i r losses were d e t a i l e d .  They did note, however, that the  subjects selected had discrimination scores not exceeding 76% on t e s t i n g with ¥ - 2 2 ' s , and had speech reception thresholds of 30 dB or better.  Their subjects, therefore, probably  represented a v a r i e t y of degrees of sensorineural hearing loss including f l a t losses, and not s o l e l y the high-frequency configuration usually cited as demonstrating the greatest d i f f i c u l t i e s hearing speech in noise. The study by Cohen and Keith (1976) discussed previously, concerning the effects of low-pass noise on speech discriminat i o n t e s t i n g , reported s i m i l a r findings for subjects with f l a t hearing losses.  These subjects gave consistently higher  performances i n the presence of noise than those with highfrequency hearing losses, achieving scores of 8 5 . 4 % (at - 4 dB S/N r a t i o ) and 80.5% (at -12 dB S/N ratio) 70.5% and 5 5 . 4 % scores of the l a t t e r group.  compared with the Both levels of  performance were lower than that of the normal-hearing group. Cohen and Keith (1976) then queried what effect the d i f f e r e n t o v e r a l l sound pressure l e v e l s of speech and noise had on the  - 15  -  discrimination scores of the groups.  Since speech was present-  ed at 40 dB SL, the words were presented at average l e v e l s of 64.4 dB SPL for the normals, 75.5 dB SPL for those with highfrequency losses, and 96.2 dB SPL for those with f l a t  losses.  To examine t h i s question, a second experiment was devised i n which f i v e normal-hearing subjects were tested at l e v e l s equal to the l e v e l s presented to the f l a t - l o s s group i n the f i r s t experiment.  Word-recognition scores were obtained monaurally  at -4 and -12 dB S/N r a t i o s , with words presented at 96 dB SPL and noise at 100 and 108 dB SPL.  Results showed that  when speech and noise were presented at equal sound pressure l e v e l s , normal-hearing subjects had s i g n i f i c a n t l y poorer wordrecognition scores (70.4$ at -4 dB S/N r a t i o and 26.4$ at -12 dB S/N ratio) than subjects with f l a t cochlear hearing losses. The similar findings i n these two studies suggest that the r e s u l t s of Bilger at a l . (1976) may have been due to the presence of persons with f l a t hearing losses i n t h e i r subject group.  However, as i l l u s t r a t e d by Cohen and K e i t h ' s (1976)  o r i g i n a l experiment, subjects with high-frequency sensorineura l hearing losses consistently showed the largest deterioration i n performance i n noise of the three groups.  This group, un-  l i k e the group with unspecified sensorineural losses of the study of Bilger et a l . (1976), did perform poorly i n noise.  - 16 -  The results of these studies provide evidence that people with normal hearing and those with high-frequency sensorineura l hearing loss do perform d i f f e r e n t l y under speech d i s c r i m i nation tests i n noise.  This d i s t i n c t i o n i s not a clear-cut  one i n view of the v a r i a b i l i t y of results and the often c o n f l i c t i n g findings documented.  One fact that does emerge  from these studies is that i n d i v i d u a l performance i n noise cannot be predicted accurately on the basis of speech d i s crimination scores measured i n quiet. 2.2. FACTORS CONTRIBUTING TO REDUCED SPEECH DISCRIMINATION IN NOISE Reduced speech discrimination performance with the introduction of noise has been attributed to a number of factors.  The two most obvious are the degree of sensorineur-  a l involvement of the l i s t e n e r and the masking e f f e c t of the noise which s h i f t s the a r t i c u l a t i o n function to the r i g h t of i t s position i n quiet.  Most theories proposed in the studies  reviewed i n the previous section revolve around some aspect of the interaction of these two variables. D i f f i c u l t i e s in speech discrimination among l i s t e n e r s with high-frequency sensorineural hearing losses are most commonly ascribed to reduced high-frequency s e n s i t i v i t y .  As  Bess and Townsend (1977) pointed out, the hearing loss e f f e c t i v e l y reduces or eliminates the a u d i b i l i t y of the high-frequency  - 17 -  consonant sounds necessary for word i n t e l l i g i b i l i t y .  The  frequencies required for the optimum understanding of speech, therefore, merit consideration. French and Steinberg (1947) used various high- and lowpass f i l t e r conditions i n a s y l l a b l e discrimination task to determine the r e l a t i v e importance of d i f f e r e n t frequencies to discrimination.  With each successive c u t - o f f of the high-  frequency portion of the spectrum, they found a progressive deterioration of s y l l a b l e discrimination.  Similar  results  were obtained upon r e j e c t i o n of the low frequencies.  The  range of frequencies either 1900 Hz and above or 1950 Hz and below each gave about a 69$ correct score.  In other words,  the elimination of a l l frequencies, for example, above 1900 Hz would s t i l l leave approximately 70$ of the s y l l a b l e s i n t e l l i g i b l e to a normal-hearing l i s t e n e r .  Decreased d i s -  crimination under low-pass f i l t e r i n g conditions was also noted by Giolas and Epstein (1963).  Their study demonstrated that  the influence of frequency on discrimination was further modified by such variables as the f a m i l i a r i t y and type of speech materials employed, the a r t i c u l a t i o n  characteristics  of the speaker, and the quality of the recording. Researchers could not agree upon the importance of frequencies above 2000 Hz in the understanding of speech. Several studies on noise-induced hearing loss attempted to  - 18 -  r e l a t e auditory acuity at selected frequencies to speech discrimination a b i l i t y .  Quiggle et a l . (1957), and l a t e r ,  G l o r i g et a l . (1961) reported that thresholds at 5 0 0 , 1000, and 2000 Hz were adequate for predicting the hearing and understanding of everyday speech.  Although speech t h e o r e t i c -  a l l y contains frequencies from 300 to 4000 Hz, i t s redundancy, they f e l t , made the contribution of the higher frequency information unnecessary.  The same three frequencies were  recommended for evaluation of impairment for purposes of compensation by the Committee on the Conservation of Hearing of the American Academy of Opthalmology and Otolaryngology (AA0O) ( L i e r l e , 1959). On the other hand, Mullins and Bangs (1957) found that speech discrimination scores correlated most highly with auditory thresholds obtained at 2000 and 3000 Hz.  This l e d  them to conclude "that these two frequencies are r e l a t i v e l y more important for speech discrimination than are the other frequencies" (Mullins and Bangs, 1957, p. 1 5 4 ) .  In an  excellent treatment of t h i s issue, Kryter et a l . (1962) suggested that test measures, such as those employed by Quiggle et a l . (1957), and Glorig et a l . (1961), tended to underestimate the importance of acuity at frequencies above 2000 Hz for understanding speech.  Their use of "thresholds  of i n t e l l i g i b i l i t y " rather than word i n t e l l i g i b i l i t y scores, spondees rather than phonetically balanced monosyllabic words  - 19 -  (PB's), and quiet, d i s t o r t i o n - f r e e test conditions did not adequately assess the a b i l i t y to understand speech under actual.day-to-day l i s t e n i n g conditions.  Hence Kryter et a l .  (1962) designed a study to determine the frequencies necessary for the understanding of speech under various conditions of noise and frequency d i s t o r t i o n more closely resembling r e a l everyday l i s t e n i n g conditions.  Seven groups of subjects with  normal hearing and d i f f e r e n t degrees of noise-induced hearing loss took part i n the experiment.  Recorded phonetically  balanced words and Harvard sentences were presented monaurally v i a headphones, both i n quiet and i n speech-spectrum noise at selected S/N r a t i o s . at  7 00 0  The speech materials were low-pass f i l t e r e d  Hz for some tests and at 2000 Hz for others.  Test  results showed that information i n the speech frequencies above 2000 Hz made a s i g n i f i c a n t contribution to the i n t e l l i g i b i l i t y of the sentences i n the presence of noise even for subjects with very large losses in those frequencies.  On  the basis of t h e i r data, Kryter et a l . (1962) found 2000, 3000, and 4000 Hz to be the most important frequencies f o r predicting the i n t e l l i g i b i l i t y of speech.  In l i g h t of the AA00 recommend-  ations and various other studies, however, they concluded that an average of the losses at 1000, 2000, and 3000 Hz for pred i c t i n g the a b i l i t y of hearing-impaired persons to understand everyday speech, would be a reasonably v a l i d compromise.  - 20 -  Harris  (1965), i n his exploration of the effects of  hearing loss upon the discrimination of mildly and severely distorted speech, also concluded that 1000, 2000, and 3000 Hz were the most important frequencies for understanding speech. Like Kryter et a l . (1962), he argued that one r a r e l y has the opportunity under normal l i s t e n i n g conditions, to hear c l e a r l y a r t i c u l a t e d speech i n quiet.  He observed that  . . . o r d i n a r i l y . . . there i s considerable masking noise, often the acoustics of the l i s t e n i n g space are poor, the peaks of conversation occur during meal time, often the t a l k e r i s smoking, chewing gum, or at least t a l k i n g with slovenly vocal gesture . . . (Harris, 1965, p. 830). In an e a r l i e r paper, Harris  (I960) demonstrated how mild  sources of d i s t o r t i o n , taken i n d i v i d u a l l y , may reduce i n t e l l i g i b i l i t y only s l i g h t l y , but, when combined may r e s u l t i n a drastic deterioration of i n t e l l i g i b i l i t y .  In the 1965 study,  he reasoned that i f a p a t i e n t ' s audiogram could be considered a type of d i s t o r t i o n , then an audiometric defect, irrelevant to the understanding of normal speech, may exert a pronounced e f f e c t on even mildly distorted speech.  Accordingly, he  conducted discrimination tests on 52 subjects with sensorineura l hearing l o s s , using sentences distorted by 1) speakers wearing nose clamps, 2) speed, 3) interruptions, and 4) reverberations (Harris, 1965).  The mean score from these  four conditions was averaged with the score for undistorted  - 21  -  speech to create "50$ distorted speech," a condition assumed to approximate everyday l i s t e n i n g .  As previously stated, the  three frequencies found to correlate most highly with t h i s condition were 1000,  2000, and  3000  Hz.  That frequencies above 2000 Hz do play an important i n the p e r c e p t i b i l i t y of speech was a paper by Sher and Owens ( 1 9 7 4 ) .  role  given further support i n They discovered that  individuals with normal hearing to 2000 Hz accompanied by a high-frequency  l o s s do have d i f f i c u l t y i d e n t i f y i n g a substant-  i a l number of phonemes as compared to normal l i s t e n e r s .  Con-  sideration of these findings along with the reports of d i f f i c u l t i e s experienced 1 9 6 5 ) , and  i n hearing distorted speech (Harris,  speech i n noise (Kryter et a l . , 1 9 6 2 ) indicates,  i n the words of Sher and Owens ( 1 9 7 4 ) , "that the problems these people often complain of are r e a l " (p. 6 7 8 ) . Findlay ( 1 9 7 6 ) and Findlay and Denenberg ( 1 9 7 7 ) advanced the theory that decreased speech discrimination among subjects with noise-induced hearing loss above 2000 Hz may due to undetected  be p a r t l y  midfrequency auditory dysfunction.  Using  fixed-frequency Bekesy audiometry, Findlay found that t h i s group consistently demonstrated separation of the  continuous  and pulsed-tone tracings at 2000 Hz, i n d i c a t i n g the presence of cochlear dysfunction. normal group.  No such separation was  found i n the  Findlay and Denenberg ( 1 9 7 7 ) then evaluated  - 22 -  speech discrimination performance  i n noise with the signal  low-pass f i l t e r e d at 200G Hz, to determine i f auditorydysfunction at frequencies less than 2000 Hz was contributing,  i n fact, to the discrimination d i f f i c u l t i e s .  The r e s u l t s ,  reviewed i n the previous section, revealed that the noiseexposed l i s t e n e r s performed at a higher l e v e l than d i d the normal l i s t e n e r s .  There was no evidence that subtle mid-  frequency cochlear dysfunction was hampering the d i s crimination performance  of t h i s group.  Findlay and Denenberg  (1977) concluded that: . . . the complaint of noise-exposed l i s t e n e r s that they experience undue d i f f i c u l t y discriminating speech i n the presence of competing noise appears to be wholly attributable to the loss of high-frequency s e n s i t i v i t y (p. 2 5 7 ) . Some researchers believe that a possible secondary f a c t o r i n the decreased discrimination performance  of l i s t e n e r s with  high-frequency sensorineural hearing loss, i s the occurrence of an upward spread of masking.  Bess and Townsend (1977)  suggest that " . . . the better hearing i n the lower frequencies causes a masking effect on important high-frequency cues" (p. 2 3 2 ) . Danaher and Pickett (1975) noted that i f a lowfrequency sound i s presented at a high i n t e n s i t y l e v e l , i t w i l l produce masking that reduces the a u d i b i l i t y of sounds i n the higher frequency regions.  Interest i n t h i s concept  - 23 -  increased when research by Jerger, Tillman, and Peterson (I960), and Rittmanic (1962) suggested that l i s t e n e r s with sensorineural impairment exhibit a greater spread of masking than do normal l i s t e n e r s .  Perhaps then, this could help  explain the d i f f e r e n t i a l discrimination performance of normal and hearing-impaired l i s t e n e r s i n noise. A recent study by Leshowitz (1977) attempted to r e l a t e tonal masking to this same problem.  He found that masked  speech i n t e l l i g i b i l i t y thresholds for subjects with noiseinduced hearing loss or presbycusis, were approximately 10 dB higher than those found for normal l i s t e n e r s . masking patterns were then measured.  Pure-tone  As much as 30 dB more  upward spread of masking was revealed for the group with high-frequency hearing loss than for the normal group.  In  l i g h t of the strong positive relationship between the masked speech i n t e l l i g i b i l i t y threshold and the upward spread of masking, Leshowitz suggested that the masked threshold could be used to predict speech perception handicap i n noise. Results of a study by Martin and Pickett (1970), however, f a i l e d to support the idea of increased upward spread of masking i n sensorineural l i s t e n e r s suggested by previous studies.  Instead they found similar amounts of masking spread  i n both t h e i r normal-hearing subjects and i n those with various degrees of hearing impairment.  Within the sensorineural group  - 24 -  they noted marked differences in masking spread, leading them to conclude: . . . Sensorineural subjects, as a group, cannot be described as c h a r a c t e r i s t i c a l l y showing either greater than normal or less than normal upward spread of masking (Martin and Pickett, 1970, p. 436). Consideration of upward spread of masking, therefore, as a possible contributing factor to decreased discrimination i n noise, should be viewed with caution. These factors can supply at best a p a r t i a l explanation of why l i s t e n e r s with high-frequency sensorineural hearing l o s s experience such d i f f i c u l t i e s understanding speech i n noise. 2.3. PHYSIOLOGICAL BASIS One important physiological study (Kiang and Moxon, 1974) does support the complaints of people with high-frequency hearing l o s s .  In d e t a i l e d studies of cats, they discovered  that neurons with high c h a r a c t e r i s t i c frequencies (CF) also carry considerable information concerning s t i m u l i in the speech-frequency region.  Previous research (Bekesy, I960)  had suggested that the entire cochlear p a r t i t i o n including the high-frequency basal region, may respond to low-frequency s t i m u l i i f presented at s u f f i c i e n t l y high i n t e n s i t y .  Support  f o r Bekesy's suggestion may be found by examination of the  \  - 25  -  tuning curve of an auditory nerve f i b r e .  The t y p i c a l tuning  curve of a neuron with a high CF consists of a sharp t i p i n the low threshold region at the CF and a long broad t a i l i n the high threshold region extending  into the low frequencies.  This neuron w i l l be most sensitive to stimuli whose frequencies f a l l near i t s CF.  However, presentation of a low-frequency  stimulus at a suitably high i n t e n s i t y l e v e l may the neuron v i a i t s low-frequency t a i l .  also activate  Kiang and Moxon (1974)  hypothesized that as these high CF neurons are broadly tuned throughout the speech frequencies, they could make a valuable contribution to the understanding  of speech.  In a quiet s i t u a t i o n , most of the information i n speech i s carried by neurons with CF. i n the speech region. information i s expressed  This  i n terms of the discharge rates  associated with the phonetic elements of the speech stimulus, and i n terms of the synchrony of f i r i n g of these with the acoustic waveform of the stimulus.  discharges  If a listener  has normal hearing i n the speech frequencies, he should have no d i f f i c u l t i e s hearing speech c l e a r l y i n quiet.  With the  introduction of noise, however, the information on discharge rates carried by the speech frequency neurons, i s eliminated. This leaves only the information regarding the synchronous f i r i n g of these neurons.  In individuals with normal hearing,  extra cues are s t i l l available i n information carried by the  - 26 -  high CF neurons.  Provided the speech s i g n a l i s sufficiently-  intense to activate these neurons, they w i l l carry the needed information on the rate of discharge and make the speech signal more i n t e l l i g i b l e .  This a d d i t i o n a l information i s not  available to l i s t e n e r s with high-frequency  hearing losses.  Hence, while speech i n t e l l i g i b i l i t y i n noise remains high f o r normal-hearing individuals, i t deteriorates for those with high-frequency  losses due to the absence of t h i s extra cue.  The results of Kiang and Moxon's (1974) study have implications for the examination of noise effects on the speech discrimination performance of normal-hearing and highfrequency impaired subjects.  They suggest that as the  i n t e n s i t y of a speech stimulus i s increased beyond that l e v e l at which responses from the high CF neurons are e l i c i t e d , the difference between the performance of the two groups should become more marked.  This would be noted p a r t i c u l a r l y i f the  discharge rate information of the low CF neurons was abolished through the introduction of noise.  Kiang and Moxon (1974)  also demonstrated that with presentation of either wide-band noise or narrow-band noise centered around the CF, the entire tuning curve of a high CF neuron i s elevated i n threshold. This interference with the threshold of e x c i t a t i o n of the neuron would r e s u l t , therefore, i n less d i f f e r e n t i a t i o n i n the performance of the two groups.  On the other hand, with  - 27 -  presentation of a low-frequency noise, only the t i p of the tuning curve i s elevated, leaving the threshold of the broad t a i l unaffected.  These findings are consistent with the  r e s u l t s of Keith and T a l i s ( 1 9 7 2 ) , Gohen and Keith ( 1 9 7 6 ) , and Liden (1967) which found that the use of low-pass noise improves the diagnostic effectiveness of speech discrimination measures i n separating normal l i s t e n e r s from those with highfrequency sensorineural hearing l o s s .  - 26" -  CHAPTER 3  3.  OBJECTIVES  The masking e f f e c t of noise on the i n t e l l i g i b i l i t y of speech i n individuals with high-frequency sensorineural hearing loss remains a controversial issue.  Information  gained from past studies has been incomplete, sometimes contradictory, and d i f f i c u l t to integrate into a clear picture of the problem.  Moreover, no adequate explanation has yet  evolved for the d i f f e r e n t i a l effect of noise on speech i n t e l l i g i b i l i t y i f indeed such an e f f e c t e x i s t s .  The findings  of Kiang and Moxon (1974) provide a promising physiological framework on which to bfise a systematic study of the issue. Their data, while compatible with c l i n i c a l observations, were obtained from cats.  The present study attempts to determine  whether t h e i r findings hold for human subjects as w e l l . The objectives of t h i s experiment are as follows: 1)  To determine whether masking noise does d i f f e r e n t i a l l y affect the speech i n t e l l i g i b i l i t y of subjects with normal hearing and subjects with high-frequency sensorineural hearing losses.  2)  I f i t does, what i s the explanation?  3)  To test whether Kiang and Moxon's findings can be v e r i f i e d with human data.  - 29 -  CHAPTER 4  4.  METHOD  4.1. DESIGN The study has a two by three by three f a c t o r i a l design. The independent variables consist  of:  1)  two subject groups: Normals and Patients with high-frequency, noise-induced hearing loss,  2)  three stimulus l e v e l s : and  3)  three S/N r a t i o s :  60, 70, and 80 dB SPL,  +-5, +12, and+-19 dB.  The dependent variable is the word discrimination score (WDS). 4.2. SUBJECTS Eighteen men with normal hearing and eighteen men with high-frequency sensorineural hearing losses served as subjects. The groups were designated N and P respectively. The N group consisted of twelve employees from the Workers' Compensation Board of B r i t i s h Columbia (W.C.B.), and s i x students from the University of B r i t i s h Columbia.  All  were i n good health with no known histories of either ear pathology or prolonged noise exposure.  Auditory thresholds  - 30 -  for t h i s group were better than 25 dB HL (ANSI 1969) at the octave frequencies from 250 to 8000 Hz i n the better ear. Their ages ranged from 22 to 37 years (mean age — 2 8 . 6 years). The P group was composed of eighteen patients from the Hearing Branch C l i n i c of the W.C.B.  As i n the N group,  subjects were healthy with no known h i s t o r y of middle ear pathology.  A l l P group subjects, however, had extensive  h i s t o r i e s of prolonged  noise exposure o f i n d u s t r i a l o r i g i n .  Case h i s t o r i e s and audiological evaluations conducted by W.C.B. audiologists confirmed frequency  that each patient had a high-  cochlear hearing l o s s with no evidence o f r e t r o -  cochlear involvement.  The o r i g i n a l s e l e c t i o n c r i t e r i a c a l l e d  f o r auditory thresholds no greater than 25 dB HL from 250 to 2000 Hz, and no less than 60 dB HL at 4000 Hz i n the better ear.  Time r e s t r i c t i o n s and the shortage of suitable subjects,  however, necessitated a modification of the c r i t e r i o n regarding 2000 Hz to allow thresholds up to 35 dB HL.  A t o t a l of  eight subjects were included on the basis of the modified criterion.  The effect of t h i s change on the r e s u l t s of the  study w i l l be considered i n a l a t e r section.  Patients were  also to be further categorized on the basis of the s e v e r i t y of t h e i r hearing impairments at 4000 Hz, into mild, moderate, and severe c l a s s i f i c a t i o n s .  As a l l but one subject (moderate)  f e l l into the mild class (60 to 75 dB HL threshold at 4000 Hz),  - 31 -  t h i s further breakdown scheme was abandoned.  Ages of the P  group subjects ranged from 37 to 84 years (mean age — 57.2 years). The mean pure-tone thresholds of the two subject groups are presented i n Table 1, and diagramatically i n Figure 1. 4.3. STIMULI 4.3.1.  Description  Taped recordings of CID Auditory Test ¥ - 2 2 were presented simultaneously with pink noise to compare the WDS's of the two groups.  Subjects were assigned to one of three stimulus l e v e l  groups, each consisting of s i x N's  and s i x P's.  For each  subject, the speech s t i m u l i then were presented at a constant i n t e n s i t y l e v e l while the noise l e v e l was varied to produce the required S/N ratios of +19, +12, and +5 dB.  A preliminary  sample of subjects had l e d to the establishment of t h i s range of S/N r a t i o s to allow a broad spread of performance. Inasmuch as 500 Hz is important i n speech and i s less affected by noise exposure and presbycusis than other speech frequencies, the pure-tone threshold at 500 Hz was selected as the reference for the stimulus l e v e l .  Within each of the  stimulus-level conditions a certain amount of allowance was given to the subjects depending on t h e i r 500 Hz thresholds.  Table 1.  Means and ranges of pure-tone thresholds of the Normal group and the Patient group at the test frequencies.  HEARING SUBJECT GROUP  25O Hz  THRESHOLD  500 Hz  1000 Hz  LEVELS  (ANSI 1969)  2000 Hz  4000 Hz  8000 Hz  -I.67  5.56  Normals: Mean  3.05  2.78  1.39  .28  Range  -5 - 10  -10 - 10  -10 - 10  -10 - 10  -10 - 20  -10 - 25  Mean  10.27  13.33  16.11  22.5  65  62.77  Patients:  Range  0-20  0 - 30  5-30  5-35  50-85  35 - 100  - 33 -  Figure 1. Mean pure-tone thresholds i n dB HL (ANSI 1969) of the Normal group (A) and the Patient group (J_) at the t e s t frequencies. Also shown i s the slope of the s k i r t of the low-pass f i l t e r used to process the speech s t i m u l i and noise ( . .) .  HEARING  THRESHOLD  LEVEL  IN  dB  3)  m o c m z o -<  N  •F-  - 35 -  Subjects with a threshold of 10 dB HL or more were given 10 dB more i n stimulus l e v e l ;  e.g., a subject i n the 60 dB  group with a threshold of 15 dB at 500 Hz was actually presented a stimulus l e v e l of 70 dB SPL.  Subjects with a threshold  of 5 dB HL at 500 Hz were given a 5 dB allowance.  Subjects  with 0 dB HL or l e s s at 500 Hz were given no allowance.  While  e f f o r t s were made, therefore, to accommodate i n d i v i d u a l variations i n threshold, i t was necessary to establish a c e i l i n g l e v e l to avoid overlap with the succeeding stimulus l e v e l group.  Any subject with a 500 Hz threshold exceeding  t h i s l i m i t was assigned a compromise threshold l e v e l of 10 dB HL f o r the purposes of the experiment. The three stimulus l e v e l s , 60, 70, and 80 dB SPL were chosen i n an attempt to traverse  the approximate threshold  region of excitation of the high CF neurons.  In t h i s region,  a marked d i f f e r e n t i a t i o n i n the performance of the N and P groups should appear.  Data from Kiang and Moxon's (1974)  study, obtained from cats, indicated that a c t i v a t i o n of these high CF neurons by low frequency stimuli occurred at i n t e n s i t y l e v e l s of 50 to 80 dB SPL.  In the present study, the lowest  l e v e l , 60 dB SPL, also corresponds to the region of quiet conversational speech.  The highest l e v e l , 80 dB SPL,  corresponds to very loud speech such as that found when conversing in a noisy environment.  If the excitation of the  - 36 -  high CF neurons does play an important role i n a s s i s t i n g speech i n t e l l i g i b i l i t y i n noise, i t l o g i c a l l y should occur within this stimulus i n t e n s i t y range. A d i f f e r e n t 50-word l i s t was presented at each S/N r a t i o . The same three l i s t s ,  2 - E , 3-A, and 4-A were used for a l l  subjects, however, the p a r t i c u l a r combination of l i s t and S/N r a t i o was varied systematically from subject to subject, to avoid bias due to any possible variations of the t e s t s .  i n the d i f f i c u l t y  Fifty-word l i s t s rather than the more commonly  employed h a l f - l i s t s  of 25-words were used on the basis of  findings by C h a i k l i n (1968), and Keith and T a l i s According to Keith and T a l i s  (1972).  (1972), unpublished data reported  by Chaiklin at the 1968 A.S.H.A. convention, indicated that h a l f - l i s t scores, although r e l i a b l e for normal l i s t e n e r s , were unreliable with hearing-impaired l i s t e n e r s due to the v a r i a b i l i t y of t h e i r responses.  Keith and T a l i s  (1972) also  found poor c o r r e l a t i o n between h a l f - l i s t scores f o r t h e i r patients with sensorineural hearing losses.  Moreover, as the  noise l e v e l of the masker was increased, t h i s c o r r e l a t i o n grew even poorer.  They concluded, therefore, that use of h a l f -  l i s t s with a sensorineural population might r e s u l t i n a spurious score that could not be reproduced i n a retest situation.  - 37 -  The word-lists were passed through a low-pass f i l t e r with a cut-off at 2000 Hz.  This was to ensure that any r e a l  differences noted i n the speech discrimination performances of the two groups were not due to t h e i r inequality i n thresholds i n the high frequencies.  Figure 1 shows the  s i m i l a r i t y between the slopes of the f i l t e r  dB/octave)  and the P subjects' audiograms between 2000 and 4000 Hz. Similar use of f i l t e r e d speech tests to compare the performance of normal hearers and subjects with sensorineural hearing loss were reported by Sher and Owens (1974), Cohen and Keith (1976), and Findlay and Denenberg (1977). The pink noise used as the masker i n the study was also low-pass f i l t e r e d at 2000 Hz.  As noted e a r l i e r , Kiang and  Moxon (1974) found that low-pass noise kept the excitation threshold of the t a i l of high CF neurons low so as to give a greater d i f f e r e n t i a t i o n of scores between the two groups. 4.3.2.  Preparation of Stimulus Tapes  One master tape was prepared with the f i l t e r e d word-lists on track 1, and f i l t e r e d pink noise on track 2.  From this  master, three cassette tapes were made, each with a d i f f e r e n t ordering of the l i s t s ,  for use in the actual t e s t s i t u a t i o n .  Records containing l i s t s  2-E, 3-A, and 4-A of CID  Auditory Test W-22 were played on a BSR 710 turntable.  From  - 38 -  there the signal was passed through a Marantz 2215 a m p l i f i e r , and then through a Rockland Programmable Dual Hi/Lo F i l t e r with a low-pass cut-off frequency of 2000 Hz.  Two f i l t e r s  were connected i n cascade to give a f i l t e r slope, above the c u t - o f f frequency, of 48 dB per octave.  The f i l t e r e d output  was then taped by means of a Revox A 77 tape recorder, d i r e c t l y onto track 1 of the master tape.  The test words were recorded  so that t h e i r i n t e n s i t y peaked at 0 dB on the VU meter. Pink noise,' generated by a random noise generator (General Radio Corporation, Model 1382), was s i m i l a r l y processed through the low-pass f i l t e r i n g system and recorded on track 2 of the master tape. Three cassette tapes were produced from the o r i g i n a l master.  A Sony Stereo Cassette-Corder (model TC-158 SD)  incorporating a Dolby B noise reduction system, was used to record the taped speech s t i m u l i and noise onto Memorex ATC cassettes.  A 30-second s i l e n t i n t e r v a l was inserted between  each l i s t on the cassettes. 4.4. EQUIPMENT A l l experimental t e s t i n g was conducted i n a soundtreated test suite  (Tracoustics, Model RS Z52) at the  Hearing Branch of the W.C.B. i n Richmond, B.C.  The physical  arrangement of the test room positioned the subject's head  - 39  -  equidistant between two side speakers (Madsen FF72) placed i n opposite corners of the room. at a distance of 1 meter, was  D i r e c t l y i n front of the subject, a centre loudspeaker  (Madsen  FF74). The test tape, with i t s separate tracks for word l i s t s and noise, was  introduced into a Madsen audiometer, Model  0B70, by means of an Akai dual-channel tape recorder, Model GXC-740D.  The word l i s t s were routed through channel 1 of  the audiometer and presented to the subject v i a the centre loudspeaker.  S i m i l a r l y , the noise was routed through channel  2 of the audiometer and presented v i a the two side speakers. Equipment c a l i b r a t i o n was  checked p e r i o d i c a l l y with a  Bruel and Kjaer (B and K) sound l e v e l meter, Model 2204.  A  half-inch microphone equipped with nose cone (B and K Model 4165) on a 1 meter tripod, was placed at the position to be occupied by the subject's head, but with the subject absent from the f i e l d .  It was  connected to the sound l e v e l meter,  located outside the test room. calibrated separately.  Speech and noise signals were  The i n t e n s i t y l e v e l of the incoming  taped signal was adjusted to peak at zero on the VU meter of the audiometer.  Then the output s i g n a l , either speech or  noise from t h e i r respective speakers, was picked up by the microphone and measured by the sound l e v e l meter.  The  audiometer settings required to produce the desired S/N r a t i o s were then recorded.  - 40 -  The acoustic spectrum of the pink noise was measured by a t h i r d octave spectrometer (B and K, Model 2114), i n conjunction with a graphic l e v e l recorder (B and K, Model 2307). The t h i r d octave spectrum of t h i s noise i s presented i n Figure 2. 4.5. PROCEDURE Subjects were seated on a chair maintained i n a fixed position i n the sound-treated chamber.  Both the noise and  the word l i s t s were presented at the required i n t e n s i t y levels to produce S/N r a t i o s of 4-19, +12, and +5 dB i n that order.  The subjects were f a m i l i a r i z e d with the task and  instructed to repeat the word heard.  They were encouraged  to guess i f they were not sure of the correct response. Responses were scored by the examiner i n the standard manner: the entire word had to be accurately i d e n t i f i e d to be marked as correct.  The administration of the entire test took  approximately t h i r t e e n minutes.  - 41 -  Figure 2. The one-third octave band analysis of the pink noise masker measured at the ear-level of the subject i n position i n the t e s t booth.  1/3  OCTAVE  BAND CD O  <J1  O  o  O  3J  m o c m z o -<  X  N  -  tn  LEVEL  -  - 43 -  CHAPTER 5  5.  RESULTS  The results of the experiment are summarized i n Tables 2 and 3 f o r the N and P groups r e s p e c t i v e l y .  These tables  report the means, ranges, and standard deviations of scores obtained by the two groups under the d i f f e r e n t experimental conditions.  In addition, Figures 3 and 4 show the WDS's  plotted as a function of the experimental conditions.  Figure  3a shows the WDS's obtained by the N group, and Figure 3b shows the WDS's obtained by the P group, under the nine possible combinations of stimulus l e v e l and S/N r a t i o . Figure 4 presents another view of the same data, with the WDS's obtained by both groups at the three S/N r a t i o s at stimulus l e v e l s of 60 dB ( 4 a ) , 70 dB ( 4 b ) , and 80 dB SPL (4c).  In these f i g u r e s , the data points represent the mean  WDS's under each condition while the l i n e s indicate the regression functions for each stimulus l e v e l . Examination of the data for both subject groups reveals a decrease i n performance with increasing noise.  The highest  mean scores of the N group were obtained at a stimulus l e v e l of 60 dB, with successive decreases in mean scores at the 70 and 80 dB stimulus l e v e l s .  The P group, on the other hand,  - 44 -  Table 2.  Means, ranges, and standard deviations of Word Discrimination Scores f o r 18 Normal subjects.  STIMULUS LEVEL  60 dB  70 dB  80 dB  S/N RATIO  MEAN  RANGE  STANDARDDEVIATION  + 5  57.67  42 - 76  12.98  *H2  79.00  72 - 86  6.29  +19  83.33  74 - 88  5.47  + 5  55.00  50 - 66  6.42  +12  67.67  48 - 82  12.48  +19  79.67  74 - 90  5.99  + 5  48.00  44 - 52  2.83  +12  64.00  46 - 76  12.07  +19  75.00  66 - 84  6.90  - 45 -  Table 3.  Means, ranges, and standard deviations of Word Discrimination Scores for 18 Patients with high-frequency hearing l o s s .  STIMULUS LEVEL  60  70  dB  dl  80 dB  S/N RATIO  MEAN  RANGE  STANDARD DEVIATION  + 5  41.67  34 - 46  4.27  +12  62.67  50 -  74  9.69  +19  76.00  64 - 90  11.38  +  45.00  28 - 64  14.35  +12  58.67  44  - 82  13.59  +19  61.67  76 - 90  6.12  + 5  37.67  30 - 50  7.42  +12  52.67  23 - 62  12.75  +19  67.67  5a - 76  7.94  5  -  46  -  Figure 3. Mean WDS's of the a) Normal group (open symbols), and b) Patient group (closed symbols) at 3 S/N ratios f o r stimulus l e v e l s of 60 dB (triangles), 70 dB (circles} and 80 dB (squares). The regression l i n e s of WDS's versus S/N r a t i o s for each stimulus l e v e l are represented by for the 60 dB group, f o r the 70 dB group, and for the 80 dB group.  W O R D  DISCRIMINATION  S C O R E  - 48 -  Figure 4. Mean WDS's of the Normal and Patient groups at 3 S/N r a t i o s for stimulus l e v e l s of a) 60 dB, b) 70 dB, and c) 80 dB. Symbols and regression l i n e s as defined i n Figure 3.  WORD  DISCRIMINATION  S C O R E  - 50 -  obtained t h e i r highest mean scores at the 60 and 70 dB stimulus l e v e l s with a decrease i n performance at the 80 dB stimulus l e v e l .  The o v e r a l l mean WDS's of the P group were  consistently i n f e r i o r to those of the N group with one notable exception at a stimulus l e v e l of 70 dB with a +19 dB S/N ratio.  In this condition, the P group achieved a mean WDS of  81.67$, 2$  higher than the 79.67$ score achieved by the N  group. The slopes of the mean regression functions plotted i n Figures 3 and 4 are summarized i n Table 4.  These slopes  represent the regression of the mean WDS's achieved by the N and P groups at the three S/N r a t i o s for stimulus l e v e l s of 60, 70, and 80 dB SPL.  Again, the slopes of the P group were  consistently steeper than those of the N group at a l l three stimulus  levels.  The raw WDS's of the two groups were subjected to an analysis of variance appropriate f o r a three-factor experiment (subject group, S/N r a t i o , and stimulus l e v e l ) . t h i s three-way analysis  i s presented i n Table 5«  A summary of Results  indicated that s i g n i f i c a n t differences existed between the discrimination performances of the N and P groups  (p z. .01).  Significant differences were also noted among WDS's achieved at the three S/N r a t i o s l e v e l s (p  (p ^ - . 0 1 ) , and at the three stimulus  . 0 1 ) . A borderline case of i n t e r a c t i o n was noted  - 51 -  Table 4«  Slopes representing the regression functions of the mean Word Discrimination Scores achieved byNormal and Patient groups versus the 3 S/N r a t i o s at the 3 stimulus l e v e l s .  STIMULUS LEVELS SUBJECT GROUP  60 dB  70 dB  80 dB  Normals  1.79  1.79  1.93  Patients  2.43  2.64  2.14  Table 5.  Summary of 3-way analysis of variance of the results of the present study.  SUMMARY SOURCE OF VARIANCE  SUM OF SQUARES  A  Subjects  2,446•26  B  S/N Ratios  C  Stimulus Levels  OF  ANALYSIS  VARIANCE  MEAN SQUARES  F  pZ.01  1  2,446.23  26.94  Yes  15,974.39  2  7,937.45  37.96  Yes  1,630.39  2  340.45  9.26  Yes  AB  373.74  2  139.37  2.09  No  AC  257.13  2  123.59  1.42  No  BC  295.55  4  73.39  .31  No  ABC  69.49  4  17.37  .19  No  3,172.67  90  29,275.67  107  Within C e l l TOTAL  DEGREES OF FREEDOM  OF  90.31  -  53  -  between the factors of subject group and S/N  ratio.  This  was  investigated further by a second s t a t i s t i c a l treatment of the data. A two-way analysis of variance was  performed using  the  slope of the regression function of the raw WDS's versus the three S/N  r a t i o s as the dependent variable.  The two  independ-  ent variables were the subject groups and the stimulus l e v e l s . The r e s u l t s of t h i s analysis, summarized i n Table 6, a s i g n i f i c a n t difference between the slopes of the functions of the two  groups (p £ - . 0 5 ) .  the speech discrimination of the two affected by noise l e v e l .  revealed  regression  This suggests that  groups was  differentially  Table 6.  Summary of 2-way analysis of variance of the results of the present study.  SUMMARY  OF  ANALYSIS  OF  VARIANCE  MEAN SQUARES  SUM OF SQUARES:  DEGREES OF FREEDOM  2.861  1  .147  2  .0735  .16  No  .637  2  .3185  .68  No  Within C e l l  14.139  30  .4713  TOTAL  17.784  35  SOURCE OF VARIANCE  A  Subjects  B  Stimulus Levels  AB  2.861  P .05  6.07  Yes  VJ1  - 55 -  CHAPTER 6  6.  DISCUSSION AND CONCLUSIONS  The results of the experiment support the findings o f previous studies that noise does d i f f e r e n t i a l l y affect the WDS's of people with normal hearing and those with highfrequency, sensorineural hearing l o s s .  The mean WDS's of  the P group i n noise consistently f e l l below those of the N group with the exception o f a high P score at 70 dB with a S/N r a t i o of -KL9 dB.  The steeper slopes of the regression  functions of the P group were shown to be s i g n i f i c a n t l y different from those of the N group.  Noise, therefore, did  have a more devastating effect on the discrimination performance of the hearing-impaired group. The d i f f e r e n t i a l performance o f the two subject groups was not, however, as marked as that found i n e a r l i e r studies, notably that of Cohen and Keith (1976).  They achieved a  separation of mean WDS's between t h e i r normal-hearing and high-frequency loss groups of 24.3$ at -4 dB S/N r a t i o and 37$ at -12 dB S/N r a t i o .  The maximum separation achieved i n  the present case was approximately 16$ at S/N ratios o f both +12  and +5 dB at the 60 dB stimulus l e v e l .  Aside from  differences i n S/N r a t i o , one of the reasons for t h i s  reduced  -  56 -  separation could be the d i f f e r e n t s t i m u l i used i n the two studies.  In Cohen and Keith's study, the s t i m u l i were  u n f i l t e r e d CID ¥ - 2 2 word l i s t s , whereas i n the present study, f i l t e r e d W-22 word l i s t s were employed.  The normal-hearing  subjects of the former study had, therefore, a l l the highfrequency consonantal cues which contribute s i g n i f i c a n t l y to word i n t e l l i g i b i l i t y .  As the hearing-impaired subjects  lacked these cues due to t h e i r high-frequency hearing losses, the d i f f e r e n t i a t i o n of speech discrimination performance between the groups with the introduction of noise was exaggerated.  F i l t e r i n g of the s t i m u l i i n the present study  e f f e c t i v e l y eliminated these high-frequency cues for the normal subjects thereby reducing the difference between the mean ¥ D S ' s of the two subject groups. A similar reduction i n the discrimination a b i l i t y of normal l i s t e n e r s with f i l t e r e d stimuli was reported by Sher and Owens (1974).  They presented a phoneme i d e n t i f i c a t i o n  task, i n quiet, to two groups of l i s t e n e r s .  One group had  normal hearing to 2000 Hz with a high-frequency hearing l o s s , similar to the P group of the present study.  The other group  had normal hearing and received the speech s t i m u l i through a low-pass f i l t e r with a cut-off frequency of 2040 Hz.  Whereas  the normal subjects generally scored 100% on this test i n the u n f i l t e r e d condition, with f i l t e r e d s t i m u l i , there was no  - 57 -  s i g n i f i c a n t difference between the mean scores (approximately 7 5 to 7 6 $ mean scores) of the two groups.  When the f i l t e r  s k i r t and the slope of the hearing loss were c l o s e l y matched, the test behavior of the two groups was v i r t u a l l y the same. Due to the f i l t e r e d speech s t i m u l i , higher S/N r a t i o s were employed i n the present study than i n previous  experiments.  As expected, the range of '•19 to +5 dB S/N r a t i o resulted i n a broad range of scores.  The combination of d i s t o r t i o n s  introduced by f i l t e r i n g of the s t i m u l i and the addition of noise made these S/N r a t i o s s u f f i c i e n t l y d i f f i c u l t f o r the purposes of the study.  Lower S/N r a t i o s might be considered  for future research i n t h i s area to assess the course of the observed trends under increasingly d i f f i c u l t  listening  situations. As noted e a r l i e r , an exception to the generally poorer speech discrimination performance of the P group was found at the highest S/N r a t i o at the 7 0 dB stimulus l e v e l (Figure 4b).  Here no s i g n i f i c a n t difference between the mean scores  of the N and P groups was found. Examination of the data presented i n Figure 3a f o r the N group reveals decreasing speech discrimination performances with increasing stimulus l e v e l s .  Examination of the data  presented i n Figure 3b f o r the P group, however, reveals  -  58 -  l i t t l e change i n the performances at the 60 and 70 dB stimulus l e v e l s although a decrease i n performance does occur at the 80 dB l e v e l .  A l i k e l y explanation f o r these findings rests  i n the pure-tone thresholds of the two subject groups.  As  discussed i n the "Subjects" section ( 4 » 2 ) , the o r i g i n a l subject c r i t e r i a f o r both groups had called for normal hearing (thresholds no greater than 25 dB HL) up to 2000 Hz.  Lack  of suitable subjects, however, necessitated a modification of these c r i t e r i a to allow the inclusion of several P subjects with thresholds of up to 35 dB HL at 2000 Hz. Whereas, i d e a l l y the mean pure-tone thresholds of the two groups would have been matched i n t h i s frequency range, the actual mean thresholds d i f f e r e d by from 7 dB at 250 Hz to 22 dB at 2000 Hz (see Table 1 ) .  E f f o r t s were made to accommodate  i n d i v i d u a l variations i n auditory function up to a l i m i t of 10 dB when assigning stimulus presentation l e v e l s .  However,  due to t h i s inequality of thresholds, some P subjects were presented s t i m u l i at the 60 dB l e v e l , which were, i n a c t u a l i t y , 10 to 15 dB softer than those stimuli presented to t h e i r N counterparts.  This p a r t i a l l y accounts f o r the large  d i f f e r e n t i a t i o n of scores seen at the 60 dB l e v e l for the N and P groups (Figure 4 a ) .  The addition of a further 10 dB  of stimulus i n t e n s i t y at the 70 dB l e v e l then probably brought the P group up to a more optimum l e v e l f o r word discrimination performance, already reached by the N group at the 60 dB l e v e l .  -  59 -  Increasing the stimulus l e v e l from 60 dB to 70 dB SPL f o r the P group therefore helped counteract the difference i n thresholds between the two groups.  However, the decrease  i n discrimination performance with increasing stimulus l e v e l s noted for the N group, was a trend i n the opposite d i r e c t i o n . Perhaps the opposition of these two effects in the P group therefore resulted i n the s i m i l a r i t y of performance noted at the 60 and 70 dB l e v e l s . On the basis of Kiang and Moxon's (1974) study, the results of the present experiment were expected to demonstrate a marked difference i n the discrimination performance of the two groups when the high-frequency neurons of the N group were c a l l e d into play.  It was anticipated that this would  be revealed by a change i n the slopes of the regression functions of the groups r e l a t i v e to each other when stimulus intensity was increased. to support t h i s reasons.  The present study, however, f a i l e d  (Figure 5 ) .  There are a number of possible  The f i r s t i s the confoundment of results  arising  from the inequality of thresholds of the two groups. application of more stringent subject c r i t e r i a ,  The  ideally  specifying matched group thresholds no greater than 15 dB HL through to 2000 Hz, would c l a r i f y the actual affects of stimulus and noise levels on the WDS's of the two groups.  - 60 -  Figure 5 . Mean WDS's of the Normal group (open symbols) and Patient group (closed symbols) at the 3 stimulus l e v e l s f o r S/N r a t i o s of +19 dB ( t r i a n g l e s ) , +12 dB ( c i r c l e s ) , and -/-5 dB (squares).  - 61 -  STIMULUS  PRESENTATION IN  dB  SPL  LEVEL  - 62 -  A second, more fundamental reason i s inherent i n the experimental design.  In an attempt to simulate more closely-  actual everyday l i s t e n i n g conditions, the experiment was conducted i n a sound f i e l d s i t u a t i o n , using loudspeakers, rather than under headphones.  This resulted i n a number of  unforeseen complications, the most notable of which was the aforementioned drop i n the l e v e l of discrimination performance with increasing stimulus l e v e l .  Ordinarily, for normal  l i s t e n e r s under headphones i n quiet, a r t i c u l a t i o n curves for PB W-22 word l i s t s reach t h e i r maximum (PB max) at approximately 50 dB SPL and remain at a constant plateau to stimulus l e v e l s of 90 dB SPL or more (Davis and Silverman, 1970, p. 212).  With a f i l t e r e d word l i s t , the maximum height  of the plateau might be reduced (French and Steinberg, 1947), but the PB max should s t i l l be maintained at a constant l e v e l . In the case of the present study, however, the sound f i e l d WDS did not plateau but demonstrated a " r o l l - o v e r " e f f e c t , where a speech i n t e n s i t y of 80 dB SPL was less i n t e l l i g i b l e than speech of 70 or 60 dB SPL.  (This can be seen in Figure  5 which plots the mean WDS's of the two groups as a function of stimulus l e v e l with S/N r a t i o as the parameter.)  One  possible reason for t h i s " r o l l - o v e r " of WDS's with increasing stimulus l e v e l s could be the high o v e r a l l noise l e v e l s used (Pollack and Pickett, 1958).  Or, i t may have resulted from  the use of the sound f i e l d test condition rather than the  - 63 headphones.  The sound-treated t e s t suite used i n the  experiment was a semi-reverberant room, not an anechoic chamber.  Reverberation effects could have caused the  observed deterioration of scores. Pollack and Pickett (1958) examined the deterioration of word i n t e l l i g i b i l i t y at high noise l e v e l s under headphones. With S/N r a t i o held constant, they observed a decrease i n the i n t e l l i g i b i l i t y of monosyllabic words with increasing overall sound l e v e l s .  The bend-down of the a r t i c u l a t i o n curves  occurred at l e v e l s of 80 dB SPL and higher and was accentuated by decreasing the S/N r a t i o .  A stringent series of controls  was carried out to ensure that the effect was not a r e s u l t of equipment d i s t o r t i o n .  Pollack and Pickett (1958), therefore,  concluded that the decrease i n i n t e l l i g i b i l i t y was most l i k e l y a r e s u l t of "overloading"  (p. 130) of the auditory system of  the l i s t e n e r . Support for the possible deletrious effects a r i s i n g from the sound f i e l d test condition may be found i n several studies, two of which deal with the effect of hearing protection on speech i n t e l l i g i b i l i t y .  Kryter (1946) assessed the effect  of wearing earplugs on a r t i c u l a t i o n scores obtained i n reverberant and anechoic environments.  In the reverberant  room, he discovered that for normal l i s t e n e r s i n the presence of 80 dB or more of noise, earplugs improved speech i n t e l -  - 64  -  l i g i b i l i t y f o r a l l speech i n t e n s i t i e s , while i n less noise, they resulted i n decreased i n t e l l i g i b i l i t y .  Earplugs  produced  t h i s improved a r t i c u l a t i o n by e f f e c t i v e l y reducing the o v e r a l l i n t e n s i t y l e v e l of speech and noise from high to medium levels while keeping S/N r a t i o constant.  S i m i l a r l y i n the present  experiment, WDS's generally improved with decreasing o v e r a l l l e v e l s of speech and noise.  Kryter found that f o r subjects  not wearing earplugs, increasing the speech l e v e l with  S/N  r a t i o held constant resulted i n a decrease of a r t i c u l a t i o n scores, the same r o l l - o v e r effect demonstrated i n the present study.  As the S/N r a t i o was decreased, the r o l l - o v e r of the  a r t i c u l a t i o n curves was  accentuated and occurred at success-  i v e l y lower speech i n t e n s i t y l e v e l s . Tests i n the anechoic chamber, on the other hand, produced maximum a r t i c u l a t i o n scores f o r both groups which did not decrease with higher speech l e v e l s but stayed at a constant plateau.  Speech at 80 dB or more above threshold was  heard  equally w e l l with plugs of without. Kryter (1946) attributed  the divergence of a r t i c u l a t i o n  performance (approximately 10%)  for the two test rooms at high  speech l e v e l s , to reverberation effects which are present i n the reverberant room and absent i n the anechoic one.  Earplugs  attenuated these effects below the l i s t e n e r ' s threshold. l i s t e n e r s without earplugs, however, as the speech  For  intensity  - 65 -  w a  s  raised, the masking effects of reverberation on i n t e l -  l i g i b i l i t y increased.  This resulted i n the r o l l - o v e r of  a r t i c u l a t i o n scores noted i n the reverberant room. (1946) concluded that these reverberation effects unintelligible  Kryter "constitute  ' n o i s e ' that interferes with speech reception"  (p. 416), and that the masking e f f e c t of this reverberant speech increased \\rith higher o v e r a l l l e v e l s . Similar findings were reported i n a recent study by Martin et a l . (1976) investigating the influence of earplugs and earmuffs on communication i n noise.  In agreement with  Kryter ( 1 9 4 6 ) , they found that speech discrimination improved with the wearing of ear protection i n high noise levels  (above  85 dBA) but was degraded i f protectors were worn i n noise l e v e l s less than 65 dBA.  Examination of t h e i r data obtained  i n a semi-reverberant room again reveals the r o l l - o v e r of discrimination scores with high speech i n t e n s i t i e s unoccluded ears.  for  They, however, did not attribute t h i s  decreased speech discrimination to reverberation e f f e c t s . They referred to the possible occurrence of d i s t o r t i o n i n the cochlea with increasing noise l e v e l s above 65 dB.  This would  appear to be akin to the auditory "overloading" of Pollack and Pickett ( 1 9 5 $ ) .  As the wearing of ear protectors i n high  noise levels would reduce this d i s t o r t i o n without a f f e c t i n g the S/N r a t i o , improved a r t i c u l a t i o n could r e s u l t .  - 66  j  n  -  a study cited e a r l i e r , Kryter et a l . (1962)  presented  monosyllabic word l i s t s i n noise monaurally to subjects with normal hearing and various degrees of sensorineural impairment. The speech s t i m u l i , low-pass f i l t e r e d at 7000 Hz, were presented at overall l e v e l s of either 65 dB or 95 dB SPL.  For the  majority of t h e i r subjects, Kryter et a l . , found the speech presented at 65 dB was more i n t e l l i g i b l e than that presented at 95 dB.  As i n the present study, i n t e l l i g i b i l i t y  decreased  with higher stimulus i n t e n s i t y . A t h i r d possible reason why the anticipated change i n the slopes of the regression functions of the N and P groups f a i l e d to occur with increasing stimulus intensity, l i e s i n the p a r t i c u l a r stimulus l e v e l s used. and 80 dB SPL may  These l e v e l s , 60,  70,  have been so high i n i n t e n s i t y that most of  the high CF neurons were already being activated at the 60 stimulus l e v e l .  dB  In such a case, l i t t l e change i n the slopes  would occur as a l l three l e v e l s would be representing the s i m i l a r condition of high CF neuron excitation.  To validate  Kiang and Moxon's (1974) findings, the range of stimulus l e v e l s employed must span the threshold region of e x c i t a t i o n of these neurons i n order to demonstrate any change t h e i r e x c i t a t i o n may  bring about i n word discrimination a b i l i t y .  Stimulus l e v e l s , therefore, lower than 60 dB SPL may  be  necessary to produce the anticipated r e s u l t s i n future studies i n t h i s area.  - 67 -  In summary, the r e s u l t s of the present study support the complaints of individuals with high-frequency sensorineural hearing losses that they experience d i f f i c u l t i e s hearing speech i n noise.  Presentation of a f i l t e r e d word discrimination  task i n the presence of masking noise demonstrated that noise does have a s i g n i f i c a n t d i f f e r e n t i a l effect on the WDS's of the subject groups, N and P.  A s a t i s f a c t o r y explanation of  these r e s u l t s based on the physiological findings of Kiang and Moxon (1974) was not accomplished, however, due to at l e a s t three possible complicating f a c t o r s .  These included  the f a i l u r e to adequately match the pure-tone thresholds of the two groups up to 2000 Hz, the use of the sound f i e l d test condition rather than headphones, and the employment of too high stimulus l e v e l s .  Kiang and Moxon s (1974) study s t i l l f  provides a promising framework for the investigation of speech discrimination i n noise.  Future use of a s i m i l a r  f i l t e r e d speech task under headphones (to eliminate reverberation e f f e c t s ) , using c a r e f u l l y matched subjects and a lower range of stimulus l e v e l s , may provide a better experimental design for t e s t i n g the application of Kiang and Moxon's data to humans.  The more r e a l - l i f e s i t u a t i o n tested  i n this study demonstrates that the deteriorating effect of noise on hard-of-hearing people i s a complex problem that involves the i n t e r a c t i o n of many f a c t o r s .  - 68 -  REFERENCES  Bekesy, G. Von. Experiments i n Hearing. McGraw-Hill Book Company, I960.  New York:  Bess, F. H., and Townsend, T. H. Word discrimination for l i s t e n e r s with f l a t sensorineural hearing losses. Journal of Speech and Hearing Disorders, 1977, 4 2 , 232-237. B i l g e r , R . C , S t i e g e l , M. S., and Rose, N. R. Consonant recognition i n noise: sensorineural l o s s . Journal of the Acoustical Society of America, 1974, 55» S87. (Abstract). B i l g e r , R. C , S t i e g e l , M. S., and Stenson, N. E f f e c t s of sensorineural hearing loss on hearing speech i n noise. Transactions of the American Academy of Ophthalmology and Otolaryngology., 1976.' 82, 363-365. Carhart, R. Problems i n the measurement of speech discrimination. Archives of Otolaryngology, 1965, 82, 253-260. Carhart, R., and Tillman, T. W. Interaction of competing speech signals with hearing losses. Archives of Otolaryngology, 1970, 91, 273-279. Cohen, R. L., and Keith, R. W. Use of low-pass noise i n word-recognition t e s t i n g . Journal of Speech and Hearing Research, 1976, l g , 48-54. Cooper, J . C , and Cutts, B; P. Speech discrimination i n noise, journal of Speech and Hearing Research, 1971,  lit.  Danaher, E. M., and P i c k e t t , J . M. Some masking e f f e c t s produced by low frequency vowel formants i n persons with sensorineural hearing l o s s . Journal of Speech and Hearing Research, 1975, 18> 261-271. Davis, H., and Silverman, S. R. Hearing and Deafness. New York: Holt, Rinehart, and Winston, 1970.  - 69 -  Findlay, R. C. Auditory dysfunction accompanying noiseinduced hearing l o s s . Journal o f Speech and Hearing Disorders, 1976, _1, 374-384. Findlay, R. C , and Denenberg, L. J . Effects of subtle mid-frequency auditory dysfunction upon speech discrimination i n noise. Audiology, 1977, 16, 252-259. French, N. R., and Steinberg, J . C. Factors governing the i n t e l l i g i b i l i t y of speech sounds. Journal of the Acoustical Society of America, 1947, 12, 90-119. Giolas, T. G., and Epstein A. Comparative i n t e l l i g i b i l i t y of word l i s t s and continuous discourse. Journal of Speech and Hearing Research, 1963, 6, 349-358. G l o r i g , A., Ward, W. D., and Nixon, J . Damage r i s k and noise-induced hearing l o s s . Archives of Otolaryngology, 1961, 7_, 413-423.  criteria  Harris, J . D. Combinations of d i s t o r t i o n i n speech. Archives of Otolaryngology, I960, J7_2, 227-232. Harris, J . D. Pure-tone acuity and the i n t e l l i g i b i l i t y of everyday speech. Journal of the Acoustical Society of America, 1965, 32, 824-830. Jerger, J . F., Tillman, T. W., and Peterson, J . L. Masking by octave bands of noise in normal and impaired ears. Journal of the Acoustical Society of America, I960, 32, 385-390.  1  —  Keith, R. W., and T a l i s , H. P. The effects of white noise on PB scores of normal and hearing-impaired l i s t e n e r s . Audiology, 1972, 11, 177-186. Kiang, N. Y. S., and Moxon, E. C. T a i l s of tuning curves of auditory nerve f i b r e s . Journal of the Acoustical Society of America, 1974, J_, 620-630. Kryter, K. D. Effects of ear protective devices on the i n t e l l i g i b i l i t y of speech i n noise. Journal of the Acoustical Society of America, 1946, 18, 413-4T7T Kryter, K. D., Williams, C , and Green, D. M. Auditory acuity and the perception of speech. Journal of the Acoustical Society of America, 1962, ^4., 1217-1223.  - 70 -  Leshowitz, B. Relationship of tonal masking to speech i n t e l l i g i b i l i t y i n noise f o r l i s t e n e r s with sensorineural hearing damage. Journal of the Acoustical Society of America, 1977, 62, S93T (Abstract). :  Linden, G. Undistorted speech audiometry. In B. Graham (Ed.), Sensorineural Hearing Processes and Disorders. Boston: L i t t l e , Brown, 1967. L i e r l e , D. M. Guide for the evaluation of hearing impairment: a report of the committee on conservation of hearing. Transactions of the American Academy of Ophthalmology ancf Otolaryngology, 1959. 63, 236-238. Martin, A. M., Howell, K., and Lower, M. C. Hearing protection and communication i n noise. In S. D. G. Stephens (Ed.), Disorders of Auditory Function (Vol. 2). London: Academic Press, 1976. Martin, E. S., and Pickett, J . M. Sensorineural hearing loss and upward spread of masking. Journal of Speech and Hearing Research, 1970, 13, 426-437. Mullins, C. J . , and Bangs, J . L. Relationships between speech discrimination and other audiometric data. Acta Otolaryngologica, 1957, JtZ, 149-157. Olsen, W. 0., Noffsinger, D., and Kurdziel, S. Speech discrimination i n quiet and i n white noise by patients with peripheral and central l e s i o n s . Acta Otolaryngologica, 1975, 80, 375-382. Palva, T. Studies of hearing for pure-tones and speech i n noise. Acta Otolaryngologica, 1955, lt£, 231-243. Pollack, I., and P i c k e t t , J . M. Masking of speech by noise at high sound l e v e l s . Journal of the Acoustical Society  of America, 195^, 20, 127-130.  Quiggle, R. R., G l o r i g , A., Delk, J . H., and Summerfield, A. Predicting hearing loss for speech from pure-tone audiograms. Laryngoscope, 1957, 62,, 1-15. Rittmanic, P. A. Pure-tone masking by narrow-noise bands i n normal and impaired ears. Journal of Auditory Research, 1962, 2, 287-304.  - 71 -  Ross, M., Huntington, D. A., Newby, H. A., and Dixon, R. F. Speech discrimination of hearing-impaired individuals i n noise. Journal of Auditory Research. 1965, j>, 47-72. Rupp, R. R., and P h i l l i p s , D. The effect of noise background on speech discrimination function i n normal hearing i n d i v i d u a l s . Journal of Auditory Research, 1969, 9, 60-63. Shapiro, M. T . , Melnick, W., and VerMeulen V. Effects of modulated noise on speech i n t e l l i g i b i l i t y of people with sensorineural hearing l o s s . Annals of Otology, Rhinology, and Laryngology, 1972, 81, 241-248. Sher, A. E., and Owens, E. Consonant confusions associated with hearing loss above 2000 Hz. . Journal of Speech and Hearing Research, 1974, 12, 669-681. Simonton, K. M., and Hedgecock, L. D. A laboratory assessment of hearing acuity for voice signals against a background of noise. Annals of Otology, Rhinology, and Laryngology, 1953, 62, 735-747.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0094281/manifest

Comment

Related Items