The effect of DAF on speech production of post-lingual cochlear implant users

MISSING IMAGE

Material Information

Title:
The effect of DAF on speech production of post-lingual cochlear implant users
Physical Description:
viii, 143 leaves : ill. ; 29 cm.
Language:
English
Creator:
Grey, Polly Shipp, 1941-
Publication Date:

Subjects

Subjects / Keywords:
Deaf -- Rehabilitation   ( lcsh )
Cochlear implants -- Evaluation   ( lcsh )
Communication Processes and Disorders thesis Ph.D
Dissertations, Academic -- Communication Processes and Disorders -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1992
Bibliography:
Includes bibliographical references (leaves 134-142)
Statement of Responsibility:
by Polly Shipp Grey.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001753178
oclc - 26585628
notis - AJG6141
System ID:
AA00002088:00001

Full Text











THE


EFFECT


OF DAF ON SPEECH PRODUCTION
COCHLEAR IMPLANT USERS


OF POST-LINGUAL


POLLY


SHIPP


GREY


A DISSERTATION PRESENTED '
THE UNIVERSITY OF FLORIDA


rO THE GRADUATE SCHOOL OF
SIN PARTIAL FULFILLMENT















DEDICATION


This dissertation


lovingly dedicated


Edward


Crawford


Shipp


Edward Rodgers Grey


Polly Moore


Shipp


Virginia


Farrell


Grey


(1911-1963)














ACKNOWLEDGEMENTS


I wish


express


deepest


appreciation


gratitude


to Dr.


Patricia


Kricos


her


outstanding


guidance,


steadfast


encouragement


during


graduate


studies


throughout


this


project.


am al


indebted


to Dr.


Alice


Dyson


her


many


contributions


completion


of this


project.


wish


thank


other


members


committee


who


have


given


so generously


their


time,


shared


their


knowledge,


graduate


and


have


studies


, Dr.


been


so supportive


Alice


. Holmes,


me during


. Kemker,


. Dr


. Otto


von


Mering


would


Gerhardt,


also


like


Michael


express


Crary,


and


thanks


Ernest


to Dr.


Walden


Kenneth


their


assistance

problems,


with

and c


calibration,


collection


resolving


of pilot


instrumentation


data.


am grateful


to Phonic


Ear


Corporation


and


Unitron


Industries


, Incorporated


loaning


me the


necessary


equipment


and


their


understanding


when


data


collection


took


longer


than


anticipated.


am especially


grateful


Robert


Mendoza


of Phonic


Ear


Corporation


and


John


Seamans


Unitron


Industries


Incorporated


.for


their


interest


r








would


like


express


gratitude


to Mrs.


Ruby


Moore,


& R Enterprises,


her


contribution


completion


Special


this

thanks


project


to Kit


typing

Evans


final


Brinsko


manuscript.

her


friendship


her


sense


of humor,


many


hours


contributed


helping


transcribe


data.


must


also


thank


some


others


whose


friendship


been


so steadfast,


Ann


Pierson,


Darlene


Hooker,


Lori


Gonzalez,


Jamie


Barron


Schwartz.


want


support, esp

Finally


thank


ecially


give


family


Watkins


love


their


Adams

and


encouragement


Saunders,


thanks


husband,


Barry


who


fills


life


with


so much


happiness.



















TABLE OF CONTENTS


ACKNOWLEDGEMENTS..................................


ABSTRACT..........................................


CHAPTERS


Vii


INTRODUCTION..........................


REVIEW OF THE LITERATURE................


Hearing. ....... ........................
Normal Speech Production and
Perception.........................


Vowels. . . . . ..
Consonants.......................
Effect of Hearing Loss on Speech
Production and Perception...
Cochlear Implants...............
Differences in Cochlear Implants.
Delayed Auditory Feedback........
Summary..........................


000....

.......

t......
.0..000.

.......
.......
* 555
* *C

* CSCCC
* C S *


METHODOLOGY.................. ...........


Subjects..............
Test Procedures.......
Equipment.............
Conditions...........
Methods and Materials.
Analysis of Data......


"...............I.

.0".0....00...0....


"... ..... .. ... ..
....'..........Q..
* S tC* *SC ***
* C S Ct *SCC S*
* C C S SS CC ** S C S
* .S SC CS SC C* SS SC CS
* S SC CC C S *t S*SC
S S S SCSCSS C C S Ct C


DAF Effect on Reading and Counting
Duration...........................
DAF Effect on Number of Consonant
and Vowel Errors.................
Performance with Cochlear Implant


Pace


RESULTS.









DISCUSSION..... ............. ........


Reading a
Consonant
Vowel Err
Cochlear
Types of
Summary..
Conclusio:


nd Coun
Errors
ors....
Implant
Speech
....t...


ris..


ting


Vers
Error


Duration...

us..... Hearing
us Hearing


s.


* *..


* 10**

* 4S1e


* ..


Aid



* e


*400
* 0*-


APPENDIX


TOTAL
DURING


NUMBER OF
READING.


CONSONANT
".'.......


ERRORS


APPENDIX


TOTAL NUMBER
READING.....


OF VOWEL
"........


ERRORS


DURING


APPENDIX


TOTAL
DURING


NUMBER OF
COUNTING


CONSONANT
........0..


ERRORS


APPENDIX


TOTAL NUMBER
COUNTING.....


OF VOWEL
........


ERRORS
.* .. ...


DURING
t.......


APPENDIX


APPENDIX


DURATION,
READING..


DURATION,
COUNTING.


IN SECONDS,
". .... ......


IN SECONDS


DURING


DURING


)0****e* ..t.... 0*t****S


BIOGRAPHICAL


SKETCH ............ .. ..........


REFERENCES (1~1)(1~(1111(1111()))rll((~))((l








Abstract


the


Dissertation


University


Requirements


of Florida


the


Presented


Partial


Degree


Doctor


re Graduate
Fulfillment


School


of Philosophy


THE


EFFECT


OF DAF


ON SPEECH


COCHLEAR


Polly


PRODUCTION


IMPLANT


Shipp


OF POST-LINGUAL


USERS


Grey


May


1992


Chairman:


Co-Chairman:


Major


Patricia


Alice


Department


. Kricos


. Dy


Ph.D.


son,


Communication


Processes


Disorders


The


purpose


of this


study


was


to investigate


effect


of delayed


auditory


feedback


(DAF)


on the


speech


production


of adult


post-lingually


deafened


cochlear


implant


users.


Delayed


auditory


eedba


ck effects


on the


speech


production


of normal-hearing


individuals


are


well


documented


and


have


also


been


reported


in a group


of hearing


-impaired


children


using


hearing


aids.


The


eight


subjects


five


women


and


three


men,


ranged


aged


from


to 61


years.


wore


Cochlear


Corporation


Nucleus


22-Channel


cochlear


implant


had


completed


the


recommended


post


-implant


aural


rehabilitation


protocol.


The


subjects


were


tested


under


simultaneous


auditory


feedback


(SAF)


and


delayed


auditory


feedback


(DAF)


three


conditions


cochlear


implant,


hearing


aid,


no amplification


In each


condition


they


read


a passage








Alphabet and


were


analyzed


for total


number of


consonant and


vowel


errors using the


Program to


Examine


Phonetic and


Phonologic Evaluation Records


(PEPPER).


Differences between


the SAF


and DAF


conditions


the duration


of reading


and of


counting were also measured


in seconds.


The data


were subjected


to a


Friedman Two-Way


Analysis


of Variance


(ANOVA).


There were no statistically


significant differences

task or condition, sugg


performance


regardless of


testing the absence of


a DAF


effect.


However,


an apparent


effect was detected during


transcription of


recorded data.


This was


investigated


further


calculating


effect size values between SAF/DAF


and


the different experimental


tasks.


Large effect size


values under DAF were obtained


for:


reading duration


with


the cochlear


implant


(0.96)


the hearing


aid


(0.99)


counting vowel


errors


(1.02)


with


the cochlear


implant.


These


large effect


sizes


indicated


that


subjects


took longer to


read


the passage when wearing


either device


and more

cochlear


vowel cc

implant.


huntingg


errors were made


Effect size


when wearing the


value comparisons between


two devices


indicated a


large DAF


effect


size


(0.81)


vowel


reading


errors with


the cochlear


implant.


In summary


a possible DAF


effect


was evident


in some conditions but was


not supported by statistical


testing.














CHAPTER


INTRODUCTION


The


the


most


United


prevalent


States


chronic

hearing


disability

impairment


in the

(Punch,


population

1983).


is estimated


that


million


Americans


have


some


degree


hearing


loss


approximately


these


people


are


profoundly


hearing


impaired,


i.e.


, deaf


(National


Center


Health


Statistics,


1982)


In this


group


of profoundly


hearing-impaired


people


are


some


who


no longer


receive,


have


never


received,


benefit


from


conventional


amplifica-


tion.


The


reasons


this


vary


but


include


increases


symptoms


tinnitus


or vertigo


an intolerance


vibrotactile


sensations


that


can


accompany


the


use


powerful


amplification


(Eisenberg


, Berliner,


House,


Edgerton,


1983)


as well


as the


risk


of causing


additional


damage


to the


ear


(Humes


Bess,


1981) .


Lastly,


there


some


cases


profound


hearing


loss


inability


overcome


the


hearing


loss


via


amplification.


The


critical


influence


of hearing


, and


effects


varying


degrees


hearing


well


loss,


established


acquis


(Calvert


ition


of speech


Silverman,


and


1983;


language


Gold,


1980;










production


of the


post-lingually


hearing


impaired,


individuals


who


developed


hearing


loss


after


acquiring


speech


and


language,


been


described


as to "generally


consisting]

Smith, 1989).


patterns


adult,


of anecdotal

However, a


in a normal-hearing


Zimmerman


reports"


comparison


and


Rettaliata


(Seyfried,

n study of


a post-lingually


(1981),


Hutchinson,


! articulatory


deafened


suggested


importance

maintaining


of auditory

"speech co


information


ordinative


in monitoring


structures


Boothroyd


(1988)


cited


the


significant


role


hearing


as a


feedback


system


conservation


post-lingually


some


hearing


hearing-impaired


and


restoration


impaired.

individuals


was


of speech


stated


no longer


in the


earlier,


receive,


or have


never


received,


benefit


from


conventional


amplification.


These

which


people no

Boothroyd


longer

was re


have


ferring.


access


The


the


cochle


feedback

ar implan


system t

t offers


an alternative


to conventional


amplification.


In 1957


Djourno


Eyries


reported


their


findings


following


stimulation


the


acoustic


nerve


the


direct


application


an electrode


Simmons,


1966) .


Their


patient,


a 50-year-old


deaf


man,


described


sound


sensations


experienced


as sounding


like


"crickets"


or a


"roulette


wheel


In addition


he reported


an awareness


of background


.e.,










techniques


and


design


of stimulating


devices,


resulted


interest


others


medical


community


By


1969


William


. House


and


Jack


Urban,


an electrical


engineer


, had


developed

(House, 1


the


976) .


first


cochlear


Cochlear


implant


implants


are


the United


surgically


States


implanted


devices


that


allow


profoundly


hearing


impaired


experience

electrode


sound.

(channel)


The f

but,


first

with


implants

better


used


single


technology


continuing


research


, some


of today


s devices


make


use


multiple


electrodes


(multichannel)


All


the


implants


enable


user


to experience


sound


sensation


stimulating


auditory


nerve.


Today


there


are


more


than


3000


people


who


have


rece


ived


either


a single


or multichannel


cochlear


implant


(Goldstein


Friedelwald,


1988).


In general,


cochlear


can


implant


be attributed


to the


be said


variable.


that


differences


effectiveness


Some


among


variability

people


receiving


an implant,


.e.


, differences


in audiological,


medical,


psychological,


social,


and


communication


histories,


as well


as the


type


of device


implanted.


Performance


with


an implant


can


range


from


the


individual


who


can


only


distinguish


traffic


between


versus


environmental


a doorbell


sounds


a person


(for


who


example,


able


car

engage










regardless of type of


device,


is an


improvement


speechreading ability


(Lansing,


1988).


The speech


perception


of cochlear


implant recipients


who use single or multichannel


devices has


been well


documented


(Bilger,


1983;


Eddington,


1988;


Eisenberg,


al.,


1983


Holmes,


Kemker,


& Merwin


, 1987).


There has been


less


research


on the speech


production of


these


patients.


The device does seem


evidenced by the


to offer


reported


benefit


improvement


this area as

loudness control


and vocal


quality


that has


been attributed


to the user's


ability to self-monitor his/her speech productions


(Chouard


et al.,

& Hough,


1983;

1981;


Eisenberg

Fourcin


et al.,

et al.,


1983;

1983) .


Engelmann,


Waterfall,


These studies,


while


commenting on the speech


their subjects,


were not


designed


to specifically


investigate speech production


Some post-lingual


and many prelingual hearing-impaired have


significant


problems


communicating due


their poor speech.


Cowie and


Douglas-Cowie


(1982)


reported


that


one of


their


subjects carried a


letter explaining to


people


that his


slurred speech was due


to his deafness and not


to his being


drunk.


To better address


issue of


speech


production of


cochlear


implant


users,


more


research


needed.


The


focus


of this


research


project


was


to examine


effect










to obtain


objective


evidence


their


use


their


"new"


auditory


information


to monitor


their


speech


production.


Delayed


auditory


feedback


(DAF)


refers


experimentally


induced


time


delay


a speaker


s hearing


his/her


spoken


(Yates,


voice.


utterances


1963).


Normally,


within


Under


individual


approximately


conditions,


hears


one


as an


his/her


millisecond


individual


speaks


into


a microphone,


device


to introduce


a delay


transmiss


of hi


her


speech.


The


effects


on the


speech


of normal-hearing


individuals


were


noted


over


years


(Black,


1951;


Fairbanks,


1954,


1955;


Fairbanks


Guttman,


1958


Lee,


1950).


Some


effects


included


omissions,


substitutions,


repetitions


of sounds


as well


as changes


vocal


intensity.


The


effects


of DAF


on the


speech


production


of hearing-impaired


children


using


conventional


amplification


were


studied


Maxon,


Brackett,


Riordan


and


Pfeffer


, (1987).


They


reported


that


their


subjects


experienced


some


same


DAF


effects


on speech


production


that


had


been


noted


normal-hearing


subjects,


including


increase


amount


time


needed


to read


passage


and


to recite


rote


numbers,


an increase


in vocal


intensity.


They


concluded


that


their


subj


ects


were


making


use


some


minimal


auditory


cues,


suggesting


some


degree


*










users.


A demonstrated


effect


would


seem


suggest


that


the


cochlear


implant


user


was


making


use


of auditory


cues,


as did


the


hearing-impaired


children


discussed


the


Maxon-


et al.


(1987)


study


Specifically,


a population


eight


post-lingually


deafened


cochlear


implant


subjects


, the


research


questions


to be formulated


were:


there


a difference


between


the


SAF


and


conditions


duration


of reading


and


counting


cochlear


implant


users?


there


a difference


between


SAF


and


conditions


the


total


number


of consonant


vowel


errors


made


cochlear


implant


users


during


reading


and


counting


tasks?


there


a difference


between


SAF


and


conditions


duration


of reading


and


counting


cochlear


implant


users


when


wearing


a hearing


aid


alone?


Is there


a difference


between


the


SAF


and


conditions


the


total


number


consonant


vowel


errors


cochlear


implant


users


when


wearing


a hearing


alone


and


performing


reading


and


counting


tasks?















REVIEW


CHAPTER
OF THE


LITERATURE


There


are


however,


many


one


different


which


forms


unique


of communication;


to humans


speech.


Speech


communication


can


be d


escr


ibed


using


feedback


model


which


there


is a sender


(the


speaker),


a message,


and


receiver


(the


listener)


The


feedback


a combination


auditory,


visual,


tactile,


kinesthetic


experiences.


When


there


the


a disruption


case


of profound


the


hearing


auditory


oss


feedback


, meaningful


portion,


speech


perception


is terminated


and


speech


production


may


affected


(Zimmermann


& Rettaliata,


1981


Cowie,


Dougl


as-


Cowie


Kerr,


1982


Plant,


1984)


In order


to re-establish


auditory


eedback


mechanism,


dev


ices


have


been


developed


which


directly


stimulate


auditory


nerve.


These


devi


ces


are


called


cochlear


linniants


The


additional


auditory


information


understand


provided


speech


implant


auditorily


enabl


(Berliner,


some


Tonokawa,


users


Dye,


House,


1989;


Holmes,


Kemker


Merwin,


1987


Tye-Murray


Tyler,


1989)


and


enhances


speechreading


ability


(Ballantyne,


1985)


In addition,


researchers


have


reported










1983;


Leder,


Spitzer,


Kirchner,


Flevaris-Phillips,


Milner,


Rilchardson,


1986).


The


purpose


this


study


examine


the


re-


established f

investigating


eedback


the


system


effects


cochlear


of delayed


implant


auditory


users


feedback


the


speech


production


of cochlear


implant


users


with


post-


lingual


deafness.


This


chapter


will


review


normal


hearing,


speech


production


perception,


the


effects


delayed


auditory


feedback


on normal-hearing


and


hearing-impaired


individuals


(using


conventional


amplification),


and


speech


perception


and


production


cochlear


implant


users.


Hearing


The


ear


has


been


described


as having


several


functions


enable


us to interpret


sounds.


The


outer


ear


offers


protection


and


gives


some


enhancement


the


sound


stimulus.


The


middle


ear


acts


as an impedance


matching


network


thereby


restoring


acoustical


energy


that


would


otherwise


lost


the


sound


passes


through


the


inner


ear


or cochlea.


The


cochlea


once


was


thought


as a


very


sensitive


microphone;


however,


now


understood


that


is a series


of filters


which


pass


signals


one


frequency


and


rejects


signals


other


frequencies.


This


filtering


ability


results










Normal


Speech


Production


PercePtion


Verbal


communication


is a complex


series


of acts


dependent


sounds


upon


which,


the


when


speaker


producing


heard


a string


listener,


can


contrasting


decoded


the


message


understood.


This


process


usually


takes


place


with


ease


due


person


s knowledge


language


tern


anatomy


physiology


the


speech


hearing


structures.


Because


these


unique


structures,


human


beings


have


capacity


to learn


to produce,


recognize,


many


different


sounds.


In the


English


language


there


are


about


forty


phonemes,


classes


of sounds


that


differentiate


one


sound


from


another


(Ling


Ling,


1978;


Mackay,


1987) .


Speech


sounds


are


divid


ed into


two


large


groups,


vowels


consonants


, depending


upon


how


they


are


produced


and


their


acousti


character


stics.


In addition


ability


to produce


and


understand


these


speech


sounds


normal-hearing


listener


able


differentiate


whether


speaker


is a male,


female,


or a


child.


This


ability


result


the


vocal


folds


vibrating


at different


frequencies


depending


on the


sex


age


the


speaker.


This


vocal


fold


vibration


referred


as the


fundamental


freauencv


(FO).


The


fundamental


frequency


adult


males


typically


occurs


between


and










Vowels


Vowel


production


results


from


a combination


of exhaled


air


passing


tongue,


and


through


out


vibrating


mouth.


vocal


Changes


folds,


the


over


degree


which


shape


the


and


is open,


position


position


tongue


lips,


influence


and


acoustical


properties


the


produced


sounds


yielding


a distinct


fourteen


sounds


which


comprise


English


vowels


(Dew


& Jensen,


1979) .


The


shape


position


tongue


exert


greatest


influence


in differentiating


the


different


vowels.


Vowel


perception


result


of changes


in the


resonating


patterns


cavities


of concentrated


vocal


tract


acoustical


creating


energy


which


different


occur


different


frequencies.


These


patterns


of concentrated


energy


are


called


formants


and


are


numbered


from


lowest


highest


frequency.


Although


the


vowel


will


sound


most


natural


first


three


formants


are


heard,


the


first


two


formants


which


are


critical


differentiating


one


vowel


from


another


(MacKay,


1987).


Each


vowel


has


own


characteristic


formant


pattern


and


each


individual


his/her


own


characteristic


size


and


shaped


vocal


tract


result


that


two


different


speakers


will


pronounce


the


same


vowel


somewhat


differently.


Despite


these


differences










1987).


In order


a person


to perceive


those


critical


first


two


formants


they


need


to be


able


hear


the


range


to 1000


the


first


formant


(Fl)


and


between


to 3330


the


second


formant


(Skinner,


1978).


Diohthonas


are


phonemes


made


two


vowel


sounds


occur


as phonemes


which


glide


from


one


vowel


position


another.


The


frequency


of a diphthong


determined


first


the


formant


second


values


formant


of each


values


vowel


being


being


added


added


together


together


giving


approximation


the


frequency


range


the


particular


diphthong


(Ling


& Ling,


1978).


Consonants


Consonants


are


the


second


large


class


sounds


and


are


produced


changing


the


airflow


out


the


mouth.


The


manner


in which


the


flow


interrupted


yields


three


distinct


types


consonants


plosives,


which


result


when


airflow


completely


blocked


such


sounds


fricatives,


resulting


from


the


airflow


being


restricted


the


sounds


/f,v,e


nasals,


where


the


air


flow


directed


through


the


nose


the


sounds


In addition


the


above


types,


another


distinctive


.~~~~~ U -A.--


/Prt, k/;


,sh/;


/mn,0/.


I


-r I I


II


1










this


feature


which


allows


distinction


between


voiced


plosive


voicel


ess


cognate


During


production


of /b/


vocal


fold


vibration


the


rele


ase


the


occur


at approximately


the


same


time


(within


to 40 msec.)


in contrast,


there


a time


between


vibration


vocal


folds


and


the


release


(greater


than


to 40 msec.)


making


"voicele


speech


sound


(Dew


& Jensen,


1977).


Lisker


and


Abramson


(1964)


coined


the


term


"voice


onset


time


(VOT)"


to describe


relative


timing


the


consonant


release


and


the


onset


voicing.


Voicing


perception


dependent


on low


frequency


information,


usually


the


range


of 80


to 350


adult


speakers


and


at higher


frequencies


children


(Stevens,


1983)


The


range


of hearing


necessary


consonant


perception


varies


with


different


manner


production


and


also


within


either v

plosives


a particular


class


oiced


of consonants.


or voiceless


are


relatively


Plosives


U/P,


frequency


t, k/)

sounds


can


The

having


the


majority


1500


The


their


greatest


concentrated


energy


energy


between


found


between


4000


and


5000


and


the


energy


concentration


varies


from


1000


to 4000


Hz depending


on the


frequency


/b/


/P/


/b/


/b p/


/a, t/


/ k/










range


the


pair


approximately


3500


8000


and


the


/J,3 /


pair


around


2500


to 4500


Another


acoustic


cue


to differentiating


fricative


sounds


intensity


and


duration.


The


are


less


intense


and


of shorter


duration


than


S, 3/


the


phonemes


(Skinner,


1978)


Nasal


sounds


are


produced


the


sound


being


emitted


from


the


nose


rather


than


mouth.


They


are


voiced


are


characterized


frequency


nasal


"murmur"


the


range


to 300


In addition


to plosives,


fricatives,


and


nasals


there


are


other


called


consonant


slides


,j/)


sounds


and


which


resemble


semivowels


(/r,


vowels


1/).


and


are


Glides


are


characterized


changing


formants


and


semivowels


formant


configuration


similar


to vowels


but


more


restricted


total


frequency


region


(Skinner,


1978).


Finally,


the


consonant


equivalent


to the


diphthong


the


affricate.


The


affricate


is a combination


of a voiced


plosive

voiceless


combined

plosive


with


combined


voiced

with


fricative


(/dg/)


voiceless


or the


fricative


(/tJ/).


Effect


of Hearing


Loss on Snsseh Produnt I nit anti Dorrnn* I an
-- -- - - --- --- app -. .~ a -- --- -p~**


/s, 2/


/vlflh, e/~/


/8,2,


Lass


an SnsEch


Prnrilt~ti ah


Da rr an+ ; nh


*










their


frequency


components.


Cochlear


or inner-ear


hearing


loss


compromises


listener


receiving


that

g a d


filtering

istorted


system

signal,


resulting


any


at all


Depending


on the


severity


hearing


loss


the


signal


may


not


be recognized


, or heard,


stener


SSince


subjects


this


this


discussion


stud

will


are


post


focus


-lingually


research


deafened

reported


adults,

on their


speech


production


perception.


Speech


production


perception


studies


prelingual


well


Hudgins


population


documented


and


(Calvert


& Numbers


suggestions

Silverman,


Levitt


for

1983;


Stromberg,


remediation,


Gold,

1983;


are


1980;

Levitt,


McGarr,


research


& Geffner


available


, 1987


on the


Monsen


, 1983)


effects


There


acquired


ess


hearing


loss


the


post


-lingual


deaf


population


, according


Seyfried e

anecdotal.

production


t al. (1989)


There


much


is great


that


variability


post-lingually


deaf


reported


the


population.


generally


speech


Thi


variability


view


has


that


been


reflected


adventitious


the


deafness


literature,


does


not


including


necessarily


result


disordered


Z immermann


and


speech


(Goehl


Rettaliata


(1981)


& Kaufman,


compared


1984)

the


articulatory


patterns


an adventitiously


deafened


adult










to include


the


frequencies.


The


hearing


loss


reportedly


occurred


between


early


and


late


adolescence


and


was


unknown


origin.


The


subject


s hearing


loss


stabilized


around


age


26 and


he referred


himself


a university


speech


and

his


hearing

speech.


clinic


that


the

time


age


27 because


audiometric


test


concern


results


about


indicated


that


he had


a profound


sensorineural


hearing


loss,


however,


an informal


assessment


of his


speech


the


conclusion


that


speech


therapy


was


not


needed.


the


age


of 32


the


subject


again


sought


an evaluation


of his


speech


voice


production.


Although


speech


exhibited


some


deviations,


the


overall


interpretation


the


results


was


that


there


were


perceptual


phonemic


findings


during


the


speech


tasks"


170) .


the


time


this


study,


an informal


assessment


subject


s speech


two


trained


listeners


judged


the


subject


utterances


to be phonemically


accurate.


The


speech


task


the


subject


was


[bib],


[bab]


, [si],


[bar],


[rab]


the


carrier


phrase,


"That


s a


" (p


. 170)


Articulatory


patterns


the


subject


were


recorded


using


high


speed


cinefluorography


and


compared


with


normal


hearing


control.


The


speech


productions


of the


deaf


subj ect


revealed


systematic


timing


differences


the










coordination


the


tongue


dorsum


with


other


structures,


consistently


late


voice


termination


compared


with


normal


control.


The


authors


suggested


that


speech


a deaf


individual


acquired


slowly


because


of overlearned


motor


patterns.


Exceeding


normal


variability


ranges


these


patterns


must


occur


repeatedly,


without


the


speaker


reali


zing


the


error,


degeneration


in speech


patterns


occur


They


concluded


that


auditory


information


"plays


a critical


role


the


long


term


monitoring


maintenance


coordinative


structures


(mainly


involving


the


tongue


dorsum


musculature)


speech"


177).


Cowie


et al. (1982


studied


twelve


post


-lingually


deafened


adults


Northern


Ireland


their


subjects


were


class


ified


as having


profound,


bilateral


sensorineural


hearing


loss,


with


exception


one


subject


who


had


some


hearing


in one


ear


but


was


losing


The


purposes


their


study


were


analyze


articulation


errors


, (2)


provide


useful


measurements


subject


intelligibility


, and


identify


problems,


other


than


intelligibility,


that


subjects


reported.


Tape


recordings


hearing


-impaired


subj ects


reading


five


short


passages


conversing,


singly










listen


to each


hearing-impaired


subject' s


recording.


Speech


intelligibility


was


measured


using


a "shadow


technique.


This


technique


involved


normal


hearing


subjects


stening


the


taped


recordings


and


attempting


to repeat,


verbatim,


what


was


said


as it


was


being


said.


The


percentage


of correctly


intelligibility.


The


repeated


authors


words


listed


was


the


the


measure


following


caveat


regarding


the


level


of information


provided


this


technique:


is a relative


measure


intelligibility;


the


results


identify


cut-off


points


between


subjects

hearing;


with

and


intelligibility


measurements


problems


using


and


this


those


technique


with

e are


normal

not


directly


comparable


to studies


the


pre-lingually


deaf.


Despite

suggest


the

two


above 1

points:


.imitations,


most


the

of th


results ob

e subjects


tainted

had m


did


easurable


losses


intelligibility;


there


was


variability


among


subjects.


General


factors


affecting


subjects


pertained


style


and


content


the


reading


passage


which


the


simpler


the


style,


and


more


familiar


the


topic,


the


higher


the


intelligibility


scores;


and


age


at onset


deafness


seemed


to have


greatest


effect


intelligibility


with


worse


speakers


becoming


deaf


before


years


age


best


speakers


becoming










can


made.


The


authors


also


noted


that


their


data


did


suggest


that


factors


such


as social


background,


intelligence


or motivation


were


relevant.


Other


problems


related


to speech


deterioration


were


explored


this


study


evaluating


normal-hearing


listener


s responses


speech


of eight


original


twelve


deaf


subjects.


subject


listening


to the


deaf


speakers


reported


that


they


would


have


problems


understanding


their


speech.


but


one


the


listeners


reported


that


they


would


keep


conversation


with


the


deaf


speaker


as brief


as possible.


In addition


these


findings


illustrated


the


tendency


of people


to associate


characteristics


of speech


with


characteristics


of speakers.


example,


a speaker


with


a harsh


vocal


quality


was


viewed


"unfriendly"


108) .


The


authors


concluded


that


these


findings


suggest


that


post-lingually


deaf


individuals


experience


some


degree


speech


deterioration


but


there


wide


deafness


appeared


this


to be


population,

a relevant f


and


'actor


in the


level


deterioration.


The


significant


speech


deterioration


many


the


subjects


illustrated


the


need


more


attention


speech


conservation


this


population,


to enable


them


function


the


hearing


world.


variability


at onset










areas:


phonetic


errors


spontaneous


read


speech;


intelligibility


ratings


errors


normal-hearing


using mo

listeners


nosyllabic


words;


subject


and


s speech


production


at 2


months


30 months


following


onset


deafness.


Analysis


of phonetic


transcriptions


revealed


omissions


to be


most


prominent


error;


specifically


the


phonemes


and


/t/.


Some


these


instances


were


attributed


normal


colloquial


usage


such


as the


word


"and"


being


shortened


to "an.


Of 69 omissions,


were


alveolars


The


author


suggested


that


audition


may


very


important


maintenance


these


sounds


due


their


weak


tactile


cues.


Vowel


quality


findings


differed


from


the


read


spontaneous


speech


samples


with


spontaneous


sample


yielding

readers.


more

This


instances


was


of schwa


attributed


that

the


normal


subj ect


hearing


s exaggerated


care


and


precision


while


reading


"leading


to a highly


artificial


sample"


40).


The


suprasegmental


features


of speech


were


found


show


deviations


from


norm


also.


The


deviations


noted


were:


inability


control


pitch;


production


of all


syllables


with


equal


stress;


and


slower


than


normal


rate


/ d/










listeners.


The


listeners


gave


written


responses


what


they


heard


and


these


responses


were


recorded


on confusion


matrices


initial


consonants,


vowels,


and


final


consonants.


Initial


consonant


errors


represented


presentations


and


those


errors


affricates


were


most


commonly


misunderstood


(21%).


Additional


analysis


affricates


revealed


that


the


voiceless


affricate


(/tf/)


accounted


errors


In addition,


of those


errors,


over


involved


substituting


the


consonantal


blend


/tr/


/t/. .


The


voiced


affricate


(/dg/)


yielded


an error


rate


of 12%


presentations


and


almost


those


errors


involved


substitution


the


blend


/dr/


for


The


author


suggested


that


relatively


weak


tactile


cues


the


release


phase


the


affricates


necessitates


auditory


monitoring


to maintain


correct


production.


Analysis


vowel


and


diphthong


errors


indicated


confusion


with


adjacent


vowels,


suggesting


"overlapping"


formant


values


45) .


This


was


further


supported


the


number


errors


between


vowels


with


similar


values


which


was


explained


the


fact


that


the


are


more


distorted


and


the


subject'


knowledge,


or memory,


tongue


placement


had


been


adversely


affected


the


hearing


loss.


This


view


was


further


supported


evidence


that


more


diphthongs


were


/ d3 /










"perceived


as steady

Conversely


state

, the


rather than

vowels were


dynamic

correctly


movements"

perceived


steady


state.


The


greatest


number


consonant


errors


involved


place


articulation.


These


findings


agree


with


those


Z immermann


and


Rettaliata


(1981)


and


further


support


importance


of audition


production


of closing


gestures


speech.


Without


synchronous


and


correct


articulatory


patterns


there


may


"blurring"


of place


cues


(Plant,


1984).


The


three


final


area


spontaneous


investigation


speech


samples,


involved


recorded


presenting


deaf


subject


at approximately


months


and


months


after


onset


deafness,


twelve


normal-hearing


listeners


listeners

pitch, in

month int


were


tonatio

erval,


asked


to rate


rate.


post-onset,


the

Th

was


samples on voice

recording made


rated


near


normal,


quality,


whereas


recordings


at 30 months,


post-onset,


was


rated


abnormal.


The


parameters


rated


most


deviant


were


pitch


and


intonation.


The


above


research


delineated


errors


in speech


production


which


authors


attributed


directly


adventitious


profound


hearing


loss.


Goehl


and


Kaufman










years


0 months)


the


time


hearing


was


lost.


They


emphasized


that


onset


of deafness


resulted


in an


interruption


the


maturation


child


s phonologic


system.


learn


They


speech,


speculated


auditorily,


that


from


this


inability


others


would


continue


account


speech


deterioration


would


important


as loss


auditory


feedback.


They


further


questioned


the


influence


speech


training


techniques


were


techniques


used


described)


with

and f


the


inall


child

v the


(these

possible


continuing


residual


effects


from


the


etiology


of the


hearing


loss


(meningitis).


These


authors


also


raise


questions


regarding


Z immermann


and


Rettaliata


(1981)


interpretation


their


data.


They


noted


that


these


authors


pointed


out


inability


generalize


their


findings


due


sample


size


= 2)


considerable


variability


in speech


production


across


speakers


regardless


their


ability


hear


or not.


addition,


they


questioned


whether


the


differences


between


the


hearing


and


deaf


subjects


were


real


differences


because


the


limitations


of describing


the


utterances


being


judged


as phonemicallyy


accurate"


170)


two


trained


listeners,


i.e.,


no formal


clinical


assessment


was


made


subjects'


speech


prior


their


experiment.













certified


speech-language


pathologists


as judges


determine


there


were


identifiable


changes


the


speech


of normal

addressed


hearing


the


and

study


deaf

were


subj ects.

whether


The

those


specific

expert 1


questions


isteners


would


able


"(1)


judge


articulation


as within


normal


limits,


and


identify


speakers


as being


adventitiously


deafened"


The


subjects


were


divided


into


two


groups,


deaf


hearing


were


matched


sex


and


were


similar


(range


years


old)


The


deaf


group


was


selected


from


a pool


twelve


who


were


attending


a speechreading


group


results


the


they


time


were


study


expected


Based


to develop


on their


"speech


audiometric


deviations"


due


their


degree


hearing


oss


(Calvert,


1982)


The


control


group


, based


on audiometri


evaluation,


was


determined


to have


normal


(for


level)


hearing.


Each


subject


was


taped


reading


the


Grandfather


Passage.


The


passages


were


randomized


the


order


presentation


was


reversed


every


other


judge.


The


second


question


identification


of deaf


speakers)


was


always


presented


after


the


first


, identification


normal


articula-


tion)


order


to prevent


listener


bias.


was


also


done


to give


the


listener


unlimited


trials


order


to make


.e.










The


results


indicated


that


subjects


were


judged


to have


normal


articulation,


but


that


deafened


speakers


could


be identified


at better


than


chance


level.


In addition


neither


number


years


since


onset


deafness


nor


the


degree


of impairment


seemed


related


to the


listener's

reported t

describing


labeling


hat

the


speaker


a review of

speech of


the

those


as deaf.


comments

subj ects


The


the


authors

judges


identified


as "deaf"


included


differences


rhythm


, rate,


voice


quality,


and


"slightly


inaccurate


articulation"


particularly


regard


sibilants


63) .


Goehl


Kaufman


suggested


that


these


findings


may


be supported


Z immermann


and


Rettaliata


(1981);


judges


however,


their


they


study


also


pointed


misidentified


that


some


some


normal


subjects.


They


also


point


out


the


possible


influence


the


expectations


which


judges


brought


their


task


once


they


knew


that


the


subjects


were


deaf.


These


expectations


were


described


as a willingness


on the


part


the


judges


tolerate


mild


changes


speech


that


are


associated


with


elderly


population


(rate,


voice,


articulation)


and


still


conclude


that


speech


production


was


within


normal


limits.


Goehl


and


Kaufman


concluded


that


clinically


significant


speech


production


deterioration,


as a result


of adventitious










maintaining


speech


based


on a servo-mechanism


theory


speech


control


was


also


supported


their


study.


Acquired


hearing


loss


produces


variable


effects


on the


speech


of post-lingually


deafened


adults.


The


loss


auditory


monitoring


capabilities


has


been


the


attributing


cause


of speech


deterioration


some


this


population


(Cowie


et al.,


1982


Plant,


1984;


Zimmermann


& Rettaliata,


1981)


A dissenting


view


was


reported


Goehl


and


Kaufman


(1984).


Finally,


need


more


research


this


area


apparent


due


to the


lack


of conclusive


empirical


studies


(Seyfried


et al.,


1989).


Cochlear


Implants


Cochlear


implants


are


surgically


implanted


devices


that


directly


stimulate


inner


ear


and


produce


a sensation


sound


The


Committee


on Cochlear


Implants


(Hopkins,


1986)


listed


eleven


different


types


implants


which


are


being


manufactured


United


States


or Europe;


however,


Gibson


(1987)


wrote


that


there


are


reportedly


over


different


cochlear


implant


devices


being


investigated


worldwide


Despite


number


of implants


being


investigated,


their


basic


composition


remains


the


same:


microphone;


oDeech


processor: (3


transmitter


ccii r


r~rni varP


... J


,










Differences


in Cochlear


Implants


The


differences


in cochlear


implants


can


be described


examining


electrodes


following


electrode


features


position;


number


speech


processing


strategy.


The


employed

(House,


earliest


today

1976) .


simple


had a single

Some example


cochlear


electrode


of multi


implant


"single


-electrode,


still


Lannel")

or "multi-


channel"


implants


cited


Gibson


(1987)


and


the


Ad Hoc


Committee


(Hopkins


, 1986)


included


the


Symbion


channels


Chorimac


channel


and


Nucleus


channel


devices.


The


configuration


electrodes


multichannel


devi


ces


differ,


but


subjects


study


will


wearing


Nucleus


device,


a description


electrode


configuration


will


be given


in more


detail


later


the


chapter.


The


second


difference


cochlear


implants


the


electrode


sitioning.


An extracochlear


electrode


may


placed


on the


membrane


the


round


window;


type


being


investigated


Hochmair


(#M/Vienna


implant)


Fraser


(Finetech/RNID)


as cit


ed by


Gibson


(1987)


Example


intracochlear


devi


ces


include


Symbion,


the


Chorimac,


Nucleus


(Gib


son,


1987)










filtered between


2700


(Boothroyd,


1989).


This


filtered signal


used


to vary the amplitude of


a 16 KHz


sinusoidal


carrier


signal


which


just below


audibility.


There


no other processing


after the


amplitude modulation.


The clinician


is able


to adjust


carrier


level


and


limit


the maximum output


through


manipulation of


processor


controls


each


patient.


Limiting


of the maximum output


by peak clipping.


Moore


(as cited by

processing as


Kessler,


1989),


referred to this


"modulated-carrier,


analog


type of

approach,


defining


it as a


strategy


in which


the waveform itself


used to drive


the electrode and


in which an attempt


is made


to squeeze the entire speech signal


into a single channel"


188).


Another type of


speech


processing,


and


the one used by


the Nucleus device,


selectively


extracts components of


speech


from the acoustic signal.


The directional


microphone


picks up sounds and sends the


information


the speech


processor.

extracts f


The speech


* aan


processor


information


filters


from the


the


sound and


speech signal.


That


information passes


to a


"map"


which


the digital


memory


the speech processor


electrode,


here


current amplitude,


that


and stimulus


the appropriate


rate are










signal,


allowing


pairs


the


electrodes


to be stimulated


to produce


speech


-like


sounds.


The


Nucleus


device


extracts


fundamental


frequency


(FO)


from


speech


person


and


sends


that


information


speaker


intensity


rate


information


of stimulation


is conveyed


Signal


amplitude


ses.


First


(Fl)


and


second


formant


information


provided


from


position


the


electrodes.


Initially,


the


Wearable


Speech


Processor


(WSPII)


provided


only


FO and


information.


In 1985,


another


coding


strategy


(WSPIII)


providing


Seligman,


basal


FOF1F2


Blamey


electrode


information


& Clark,


was


became


available


1987)


stimulated


followed


coding

rapid


(Dowell,


strategy


succession


stimulation


a more


apical


electrode


yielding


and


cues,


respectively


The


addition


of F1


the


processing


strategy


perception


would


expected


as those


formants


to further


are


enhance


known


vowel


critic


(MacKay,


1987)


The


most


recent


development


coding


strategy


became


available


1989


with


the


Mini


Speech


Processor


(MSP)


(Skinner,


Holden


, Dowell,


Seligman,


Brimacombe,


Beiter,


1991


Clark


1991)


The


MSP


employs


"multi-peak"


strategy


which


measures


dominant


spectral


peaks


frequency


ranges


containing


270-730


and


(840-










addition three other frequency bands can be activated A3


(2000-2800 Hz),


(2800-4000


Hz),


and A5


(4000-7000 Hz)


(Patrick &


Clark,


1991).


and A2


are always


activated and


activation of


the higher


frequency bands


is dependent upon


the particular acoustic signal


being


sampled.


In the case


of unvoiced stimuli


there


is a


small


amount of


energy


in the


frequency range


so the


electrode


not


stimulated.


band


In its


is activated


place


thereby


fixed electrode


extracting high


for the A5


frequency


information


(Clark,


1991) .


The opposite


true


for voiced


stimuli.


Since


there


is only


a small


amount of high


frequency energy


7000


in voiced stimuli


not activated.


the A5 electrode


The ability


(4000-


of the speech


processor to provide


combination with FOF1F2


consonant


frequency bands A3,


information provides more vowel


information.


In addition


to extracting


information


important


understanding


designed to


speech.


The electrode configuration


take advantage of the physiological


response of


the cochlea.


This


is accomplished by


stimulating discrete


areas of the basilar membrane.


The configuration


described as


"bipolar.


Bipolar electrodes


are


small


are


located at equal


distances


from the target nerve and










electrodes


stimulate


large


groups


nerve


cells


and


are


effective


providing


discrete


areas


of stimulation


(Fravel,


1986).


The


design


Nucleus


implant,


using


bipolar


electrode


configuration,


takes


advantage


place

the h


pitch


igh


theory


frequencies


f hearing.

necessary


For

for


example,


stimulation


perception


can


accomplished


because


number


of electrodes


and


fact


that


high


frequency


information


coded


on the


basis


of place


on the


basilar


membrane.


In contrast,


the


frequency

electrode


information

at a rate D


found


roportional


used

the v


to stimulate


oicing


each


frequency


(Clark,


1986).


This


is consistent


with


what


known


about


frequency


hearing


which


coded


based


on the


"place


maximal,


discharge


periodicity


the


discharge


pattern....


" (Yost,


1985,


92-93).


Patient


Selection


Criteria


There


has


been


a wide


range


selection


criteria


and


large


variations


in patients'


responses


as more


people


have


received


devi


ces.


the


present


time


there


are


standardized


implant


criteria


candidate


accepting


(Goldstein


or rejecting


Friedelwald,


1988)


a cochlear

Maddox


and


Porter


(1983)


have


suggested


that


rather


than


asking,


"Who


is a candidate


a cochlear


implant?"


would


r.


u


G


LV.










basis and should not be generalized


the profoundly


deafened population as a

The criteria suggest


whole"


250).


ted by the National


Institutes of


Health


Consensus


Development


Conference Statement


(1988)


include


following:


Audiolocical


Criteria:


Bilateral,


profound sensorineural


hearing


loss


Bilateral,


aided


thresholds greater than


dB HL


0 percent correct on


open-set speech


recognition


Lack


substantial


increase


lipreading


with appropriate amplification


Electrophvsioloqical


Criteria:


Measurement of


early,


middle,


and late


latency


evoked potentials


Absence of neural


responses may


or may not


prove


to be a


contraindication


Medical Surqical


Criteria:


Usual


candidate


is a


postlingual


onset,


healthy


adult


Possible complicating factors


include:










Psychouhvsical


Criteria:


None


psychophysical


data


available,


example


gap


detection,


are


considered


critical


issue


of candidacy


this


time


Psychophy


sical


data


have


been


shown


good


predi


ctors


of speech


recognition


performance


Psvcholog ical


and


Linguistic


Criteria:


Most


psychological


exclusionary


purposes


testing


such


done


as mental


retardation


or psychiatric


disorders


Risks


and


Benefits


Cochlear


Implants


As with


any


surgical


procedure


and


use


general


anesthesia,


risks


there


cited


are


the


inherent


Hoc


risks


Committee


the


Report


patient.


(Hopkins,


The


1986)


include:


The


reduction


or total


loss


on any


presurgical


residual


hearing


Complications


due


to anesthesia


surgery


Post-surgical


sks:


Bioincompatibility


internal


components










Facial


nerve


damage/paralysis


New


bone


growth


within


cochlea


Intracochlear


scarring


and


fibrous


tissue


growth


Degeneration


Unknown


of surviving


reactions


nerve


to prolonged


fibers

electrical


stimulation


Required


replacement


of failed


internal


components


Unrealistic

patients and


expectations


family


leading


of cochlear


implant


to psychological


disturbances


Possible

Possible


vestibular c

exacerbation


complications


tinnitus


Possible


benefits


cited


committee


include:


Increased


awareness


of sound


improved


ability


to monitor


speech


production


Recognition

Improved aw


some


areness


everyday


sounds


of surprasegmental


features


speech


Supplement


to lipreading


Perception

leading to


some


limited


segmental


word


aspects


recognition


of speech

and










improved


social


interaction


and


employment


potential


Significant


word


recognition


sentences


sound-alone


Rehabilitation


conditions


Issues


Aural


rehabilitation


cochlear


implant


patient


begins


with


evaluation


procedures


employed


during


selection


process.


Lansing


(1988)


has


divided


the


rehabilitation


process


into


three


areas:


development


realistic


expectations


implant;


training


the


interpretation


and


use


new


auditory


information


with


without


visual


cues;


development


coping


strategies


to enhance


communication.


Some


variability


responses


from


patients,


no doubt,


due


their


individual


differences


however,


some


may


be due


differences


protocol


among


the


centers.


While


most


centers

direct


offer c

training


counseling


and


measures


, ongoing


adjustment


of performance,


others


device,

only


provide


counseling


and


adjustment


the


implant


(Hopkins,


1986).


Others,


like


House


Ear


Institute


and


centers


implanting


Nucleus


devices


have


specified


extensive


protocols


adiustina


implant


and t


rainina


~ral


. ..


L.


. .


1IC










Soeech


Perception


Studies


Cochlear


Irnolant


Users


Speech


perception


studies


with


the


implant,


including


both


single


multi-channel


have


been


an area


of wide


interest


to the


resea


community


(Bilger,


1983


Eddington,


1988


Eisenberg,


et al


1983


Holmes


, Kemker,


& Merwin,


1987)


The


perception


of speech


using


auditory


cues


only


been


seen


some


patients


and


recently


been


reported


House


(1989)


children


some


children.


evaluated


over


Berliner,


50 profoundly


years.


Tonokawa,


hearing


All


Dye,


-impaired


the children


were


implanted


with


the


3M/House


single-channel


devi


ce.


Their


ages


the


time


study


ranged


from


to 15.9


years,


with


average


age


the


onset


of deafness


of 2


years.


Age


the


time


receiving


implant


ranged


from


to 15


years


, with


a mean


of 6.2


years.


Test


materials


were


administered


level


using


Portions


live-voi


at a normal


Glendonald


Auditory


conversation


Screening


Procedure


(GASP)


(Erber,


1982


were


administered


assess


open


-set


speech


recognition.


The


authors


noted


that


GASP


designed


as a closed


-set


test


, picture


pointing


task;


however,


study


was


administered


auditorily

sentence s


only


timuli


The children


using


were


an auditory


tested


mode


with


only


word


They


and

reported










deafness


was


noted


to be


shorter


children


able


score


on the


open-set


discrimination


task


than


those


who


were


unable


to do


SO.


Berliner,


et al.


(1989)


also


cited


a study


Geers


and


Moog


press


which


they


assessed


of 12


high-performing


children


using


3M/House


device,


they


also


reported


that


children


study


achieved


some


open-set,


auditory-only


recognition


simple


sentences.


Dorman


et al.


1989)


investigated


the


ability


patients


using


Symbion


multi


-channel


implant


to identify


synthetic


vowels.


at least


Eight


years


subj ects


old.


The


participated


length


deafness


the


study,


ranged


from


year


to 28


years


with


a mean


of 6


years.


The


results


this


study


indicated


that


eight


subjects


scored


above


70 percent


correct


on a test


of spondee


word


identification.


Twelve


synthetic


vowels


the


"bvt"


format


comprised


stimuli.


Vowels


characterized


F1 and


formants


were


well


identified


whereas


those


with


high


formants


were


well


identified.


The


authors


concluded


that


their


results


were


consistent


with


the


model


Fl being


specified


a rate


code


and


extreme


values


specified


rate/place


code.


Waltzman


and


Hochberg


(1990)


investigated










age


Speech


from


Pattern


26 to 85,


with


Contrast


a mean


(SPAC)


test


of 51


years.


(Boothroyd,


The


1987)


was


used


to evaluate


perception


of speech


pattern


contrasts


the


subjects.


This


test


evaluates


the


perception


of four


suprasegmental


eight


segmental


speech


contrasts.


a forced


-choice


format


using


real


words,


phrases,


sentences.


The


results


of this


investigation


indicated


that


both


processing


strategies


provided


suffi


cient


information


about


speech


contrasts;


especially,


fundamental


frequency


temporal,


and


intensity


cues.


Lastly,


Tye-Murray


Tyler


(1989)


investigated


auditory


consonant


and


word


recognition


skill


subjects

implants


seven


wore

n wor


3M/House

e Nucleus


implants,

implants


three

and


wore


wor


3M/Vienna

e Symbion


implants


The


speech


material


consisted


of 14


consonants


in an /


context


Sentence


Test


Without


Version


(Tyler,


Lansing


, & Preece,


1984)


Their


conclusions


were


follows


Most


subjects


recognized


consonants


in an


audition-only


condition


at above


chance


level


Subjects


perceive


the


envelope


feature


relatively


well

The


and

voic


the


ing,


ace


feature


nasality,


relatively


duration,


and


poorly.

envelope











Most


multi-channel


users


achieved


some


open-set


word


recognition


an auditron-only


mode,


there


was


wide


range


of performance.


None


3M/House


or 3M/Vienna


users


achieve


open-set


word


recognition.


Subjects


who


are


able


to utilize


middle


high


frequency


information


are


more


likely


score


better


on open-set


sentence


tests.


Speech


Production


Cochlear


Imolant


Patients


The


speech


production


of cochlear


implant


patients


been


investigated


as thoroughly


as speech


perception.


As stated


introduction,


the


device


seems


to offer


benefit


as evidenced


reported


improvement


loudness


control


and


vocal


quality.


Those


improvements


have


been


attributed


the


re-establishment


the


auditory


feedback


channel


(Chouard,


et al.,


1983;


Engelmann


et al.,


1981


Fourcin


et al.,


1983).


One


study


designed


specifically


to investigate


the


effect


single


channel


cochlear


implant


on voice


parameters


was


reported


Kirk


and


Edgerton


(1983).


The


subjects


were


two


men


two


women,


ages


to 65


years.


Speech


samples


subjects


reading


the


Rainbow


Passaae


vti~h Iba


i4mnrl nn+


al.*$ A f lat'


A-- -rA~~ Y


CYIAI~


A--_ _


hR dldrd


--~-


~~A


i


nT~ 31111










These


changes


were


direction


the


normal-hearing


controls.


The


female


subjects


produced


voices


that


were


higher


than


when


implant


was


turned


off.


Again,


this


finding


was


direction


the


normal-hearing


control


speakers.


subj ects


continued


demonstrate


prolongation


sentences


and


pauses


consistent


with


the


speech


deaf


even


when


the


implant


turned


In conclusion,


authors


cited


need


further


study


this


area.


Speech


production


output


three,


post-lingual,


Nucleus,


22-channel


implant


users


was


reported


Medwetsky,


Hanin,


Boothroyd


(1987)


Their


subjects


had


been


j udged


as successful


implant


users


based


on their


speech


recognition

excellent s


ability,


sentence


with


per


only


ception


the

using


implant,


and


implant


their

with


lipreading.


Their


subj ects


had


become


deaf


the


ages


, 18,


and


years,


respectively.


They


had


received


their


implants


ages


, 38,


respectively.


The


purpose


the


study


was


to evaluate


their


pre-


and


post-implant


speech


production


as defined


intelligibility.


Boothroyd'


Speech


Perception


Contrast


(SPAC)


test


was


used


measure


intelligibility


of phonetic


contrasts.


While


the


test


was


originally


designed


measure


speech










three


subjects


twice.


A pre-implant


recording


was


initially


made


and


months


after


implantation


a second


recording


was


obtained.


The


subjects


did


receive


any


speech


therapy


during


interval


between


receiving


implant


and


making


the


second


recording.


Twelve


normal-hearing


listeners


were


presented


the


recordings


of each


implant


user


and


they


gave


written


responses


to what


they


heard.


The


results


obtained


indicated


that


there


were


significant


improvements


production


Speech


Pattern


Contrasts


three


subjects.


These


improvements


were


mostly


voicing,


stress,


and


intonation.


was


suggested


that


these im

provided


provements


were


implant.


due


the


Two


auditory


the


subject


feedback

s also


demonstrated


significant


improvement


phoneme


and


word


production.


Oster


ten


3M/Vien


(1988)


re


post-lingual

na extracochl


ported

Swedish


ear


changes


patients


the

who


single-channel


speech

received


device.


production

the


The


research


was


conducted


after


18-24 months


implant


use.


The


patients


had


varying


causes


hearing


loss


and


their


ages


ranged


from


25-55


years


old.


Recordings


were


made


subjects


' reading


"standard


passage


words"


15) .


The


subjects


read


prior


to receiving


the


implant


one,










normal


limits


appropriate


pitch


prior


to receiving


implant.


However,


any


changes


noted


post-implant


were


reportedly

as evidence


towards

that s


normal


subjects


are


ranges.


able


This


to make


was

use


interpreted

of low


frequency


information


received


through


auditory


feedback


order


to control


monitor


their


pitch.


Those


subj ects


whose


pre-implant


voices


had


been


described


as tense


harsh


were


reported


to exhibit


improved


voice


quality


following


implantation.


Subjects


relatives


responded


a subjective


questionnaire


assess


the


perceived


benefit


of the


implant.


the


subjects


reported


feeling


more


confident


their


communication


following


receiving


their


implant.


Lastly,


was


used


at durations


of 200-600


milliseconds


to attempt


to determine


the feedback


provided


with


implant.


Responses


reported


the


literature


normal-hearing


individuals


.e.,


slow


speech,


repetitions


or words


or syllable


higher


intensity,


and


increase


pitch")


were


not


found


these


subjects.


Two


explanations


were


offered


this


lack


response:


years


total


deafness


had


resulted


patients


learning


to rely


tactile


feedback,


FO and


durational


auditory


cues


are


perceived


as background


noise


and


do not


influence


articulation.










hearing


their


own


voice.


Normally,


as people


speak,


they


hear


their


own


voices


within


approximately


one


millisecond


(Yates,


1963).


Lee


(1950)


first


described


a condition


DAF


which


produced


changes


speech


production


some


individual


He reported


that


his


subjects


read


"moderately


difficult


text


at a comfortable


speed....


While


under


different


conditions


of DAF.


The


changes


noted


their


speech


included


a slower


rate


and


stuttering


characterized


repetition


syllables


or fricativeq.


suggested


finding


a model


describing


speech


production


a closed-loop


to explain


system


which


this

h the


length


of each


loop


was


approximately


proportional


the


time


necessary


particular


function.


He defined


four


loops,


arranged


a hierarchy


, made


articulatory


loop


(phonemes),


a voice


loop


(syllables),


a word


look,


and


a thought


loop.


He proposed


that


the


auditory


system


was


series


with


the


voice


loop


auditory


monitoring


and


that


the


articulatory


and


voice


loops


were


most


affected


DAF.


stressed


that


loop


the


the


model,


auditory


down


monitor


syllable


dissatisfied


level,


with


may


the


be repeated


previous


performance.


Further,


there


was


a common


junction


the


loops


which


was


represented


a cortical


speech


*










Fairbanks


(1954)


proposed


a more


elaborate


closed-loop


model,


or servosystem,


speech


production


based


auditory


utilizing


monitoring


DAF.


illustrated


model


his


included


model


sensor


with


unit"


studies


which


auditory,


tactile,


and


proprioceptive


information


back


the


"comparator


unit.


This


information


was


compared


output


and


comparator


identified


any


discrepancies


between


the


two


signals.


He noted


that


the


delayed


auditory


feedback


condition


sin forms"


system


about


success


effecting


and


ordering


intended


output


units,


thus


impairing


basic


produce"


334).


In 1955


, Fairbanks


reported


on the


articulation


errors


and


changes


the


rate


of speech


his


subjects


under


DAF


condition.


Subjects


this


study


were


sixteen


male


college


students


who


had


never


experienced


DAF.


They


read


the


Rainbow


Passage


under


five


different


stimulus


conditions


with


speech


output


amplified


a constant


amount


and


the


delay


intervals


varied


from


to 800


milliseconds.


Results


indicated


that


DAF


resulted


a variety


speech


disturbances


divided


into


groups:


"direct


effects"


and


"indirect


effects.


Direct


effects


included


an increase


articulatory


errors


and


longer


duration


and


indirect


effects


included


an increase


in sound


pressure,


.e.,


increased










The


data


were


from


subjects


the


1955


study


of sixteen


male


college


students


and


the


various


types


errors


found


the


original


study


were


investigated.


Their


analysis


indicated


that


there


was


agreement


with


original


study,


the


general


effect


of DAF


to a reduction


number


correct


words


read


an increase


the


total


time


reading.


They


also


noted


that


the


maximum


effect


was


seen


using


a 200


milli


second


delay.


articulatory


errors


reported


included


substitutions,


omissions,


were


and


additions


additions.


and


The


about


greatest


those


number


were


errors


repetitive.


The


authors


noted


that


repetitions


were


not


primarily


corrective


but


rather


a purposeless


response.


The


amount


delay


which


is most


disruptive


has


been


shown


vary


as a function


with


the


delay


necessary


to disrupt


speech


ecreas


with


age


(MacKay,


1968


Siegel,


Fenst,


1951;


Garber,


Fairbanks,


Harrington,


& Pick


1955


have


1988)


, 1980).


MacKay,


reported


Many


1968


that


researchers


Buxton,


between


(Black,


1969


and


milliseconds


produces


most


disruption


the


speech


adults,


with


exception


of older


adults


years


old)


who


experience


maximal


disruption


their


speech


milliseconds


(Buxton,


1969).


.e.










Timmons


& Boudreau,


1972),


however,


Buxton


(1969)


reported


no differences.


The


demonstrated


effect


has


been


interpreted


many


researchers


as evidence


that


speech


acts


as a


servomechanism


and


auditory


feedback


the


primary


control


channel


(Borden


& Harris,


1984).


However,


Borden


and


Harris


(1984)


also


noted


that


view


been


challenged


those


who


point


out


that


DAF


effect


can,


example,


overridden


speaker


attends


the


reading


and


ignores


the


acoustic


information


being


received.


In addition,


measures


other


than


will


interfere


with


speech


production,


example,


amplifying


air


conducted


sound


will


cause


speakers


to decrease


their


vocal


intensity


(Siegel


& Pick,


1974),


attenuating


the


air


conducted


sound


will


produce


to hear


their


the


opposite


speech


at all


effect

they


and

will


listeners


Increase


are


their


unable

vocal


intensity


and


prolong


voicing


(Lane


& Tranel,


1971).


Finally,


Borden


and


Harris


(1984)


suggested


that


audition


a feedback


mechanism


monitoring


ongoing


fluent


speech


production


does


provide


a complete


explanation


"because


many


transient


sounds


provides


information


speaker


too


late;


he has


already


spoken


and


can


only


make


corrections


after


fact"


136) .


However,


there


was










with


auditory


feedback


they


will


try


overcome


attempting


to correct


their


speech


output.


Harrington


(1988


offered


another


model


to explain


effect


including


influence


of rhythmic


structure


the


production


literature


addressing


fluent

the a


speech.


acoustic


He noted


and


that


articulatory


the

aspects


of coarticulation


suggested


that


consonants


and


vowel


are


produced

overlap


using


time


"prespeci fies


"separate


the


articulatory


He stated

intervals


that r

between


strategies


hythmic

vowels


that


structure


stressed


syllables"


and


that


same


structure


allows


the


speaker


predict


when


vowels


of stressed


syllables


will


be auditorily


perceived.


illustrated


his


model


describing


production


two


syllables,


Sl and


, by


a fluent


speaker


under


DAF.


The


time


of production


the


vowel


would


some


arbitrary


unit


of time


after


vowel


of S1 and


amount


of DAF


was


milliseconds.


He suggested


that


the


effect


produces


a mismatch


between


the


"actual


time


production


of the


vowel


of S2


and


the


expected


time


perception"


39) .


The


speaker'


attempt


to correct


the


mismatch


results


an effort


onset


to study


of dysfluent


auditory


reliance


speech.


, Maxon,


Brackett,


Riordan,


Pfeffer


(1987)


studied


the


effect











oral/aural


educational


programs.


The


authors


reported


that


their


subj ects


displayed


disturbances


in their


speech


production


under


which


were


similar


to those


reported


literature


normal


hearing


individuals,


i.e.,


increase


the


time


to read


passage


or reciting


note


numbers


an increase


in vocal


intensity.


Based


on their


results

impaired


they


interpreted


children


can


make


data


use


to suggest


of minimal


that


auditory


hearing-

cues


during


self-monitoring.


A review


of the


literature


indicates


that


while


there


may


be various


explanations


to explain


the


effect,


the


presence


an effect


is not


disputed.


The


literature


also


suggests


that


audition


influences


an individual


s fluent


speech


production.


Snmmarv


The


importance


auditory


feedback


to the


normal


acquisition


speech


and


language


is unquestioned.


However,


there


are


differing


views


as to its


degree


importance


the


maintenance


adventitiously


of normal


profoundly


speech


production


hearing-impaired


(Cowie


Douglas-Cowie,


Rettaliata,


1983;


1981).


Goehl


Many


Kaufman,


this


1984


population


Zimmerman

experience


ah-rn4 fi~ran* rwhn n na n-Aiin4~C 4 nnn.


r n eraaan


o i nn C f; ~P11C


rrh ~nna


mk,,,


CL *I~ 1)~1


i










amplification.


A cochlear


implant


the


only


type


device


available


which


has


capability


of providing


auditory


feedback


them.


Some


studies


of speech


perception


cochlear


implant


users


have


reported


improvement


loudness


control


and


voice


quality


and


have


attributed


improvement


to the


reestablished


auditory


feedback


(Chouard,


et al.,


1923


Engelmann


et al.,


1981


Fourcin


et al.,


1983).


The


area


of speech


production


adventitiously


deafened


cochlear


implant


user


has


been


extensively


studied.


an area


needing


further


study


document


effectiveness


reestablished


auditory


feedback


system


and


to begin


to evaluate


effectiveness.


Delayed


measures


auditory


the


feedback


cochlear


offers


implant


a tool


users


making


speech


objective


production


with


and


without


the


implant.















CHAPTER


METHODOLOGY


The

whether


procedures


there


this


is a delayed


study w

auditory


ere


designed


feedback


to ascertain


(DAF)


effect


speech


users.


production


Specifically,


of post-lingual


parameters


cochlear


measured


implant


evidence


of a DAF


effect


were:


reading


rate


as measured


total


time


to read


the


Grandfather


Passage,


speech


duration


measured


total


time


to count


backwards


from


and


(c3)


total


number


of consonant


and


vowel


errors


reading


and


counting


backwards.


The


cochlear


implant


users'


speech


was


recorded


presence


and


absence


of DAF


with


their


cochlear


implant


speech


processor


off


and


on.


addition,


using


a third


conventional


experimental c

amplification


condition


the


was


investigated


unimplanted


ear.


Subjects


Eight


subjects


participated


this


study--five


women


and


three


men.


All


had


completed


the


recommended


number


post-implant


aural


rehabilitation


sess


ions


and


were


wearing


the


Melbourne


(Cochlear


Corporation)


22 multichannel










subjects


ranged


age


from


years


with


a mean


of 47


years.


Reported


hearing


loss


etiologies


included


meningitis,


noise,


Meninere


s Syndrome


and


four


with


known


etiology.


suspected


to be


One

Cogan


these


unknown


s Syndrome.


etiologies


Unaided


pure


was

tone


averages


(500,


1000.


2000


the


better


ear


ranged


from


75 to 115+


HL with


a mean


of 102


dB HL.


The


duration


hearing


loss


the


subjects


ranged


from


to 41


years


with


a mean


of 20


years.


One


subject


wore


a hearing


aid


unimplanted


ear


no longer


wore


hearing


aids,


and


one


never


worn


one.


The


subjects


reported


having


worn


the


cochlear


implant


a minimum


of five


and


one-half


months


to a maximum


five


and


one-half


years.


Estimates


of daily


wearing


time


ranged


from


18 hours


a day,


with


a mean


wearing


time


of 13 hours


a day.


The


subjects


reported


having


worn


the


cochlear


implant


a minimum


five


and


one-half


years.


Estimates


of daily


wearing


time


ranged


from


18 hours


a day


with


a mean


wearing


time


13 hours


day.


Information


the


individual


subjects


summarized


Table


Test


Procedures


Preliminary


Procedure


After


the


subj ect


read


and


signed


informed


consent


consent












51







N
*u N
r C -



Cd) N

44 N N
N ~~~I 6 i
Ml N~ 0' *



I' Lw SD IA -




N
.c N '
A0

*I 0 0 CIo



3U N C
4' 63 4C N\ -


*, Le .'
Lu a- N N S -
I m01- '0 C


C
a-a



4' ND r

cu OO0. 0. N
z Ml 'S N N S K -





I C 4

U- a

~I -
** N

1O '0crO

V] u cnN










obtain


her


evaluation


of each


subjects


' family


support


system.


The


majority


subjects


reportedly


enjoyed


"good"


family


support


system.


To elicit


an objective


picture


the


subject


performance


using


the


implant,


each


was


given


subtests,


the


4-Choice


Spondee


Test


and


10 sentences


from


the


CID


Everyday


Sentence


Test,


from


The


Minimal


Auditory


Capabilities


Battery


(Owens,


Kessler,


Telleen,


Schubert,


1981).


These


materials


and


the


test


configuration


were


selected


based


on the


Cochlear


Corporation


test


protocol


individuals


receiving


their


device.


Scoring


both


tests


was


based


on a percentage


correct


score.


The


percentage


correct


score


sentences


was


based


on key


word


scoring


Both


tests


were


administered


using


only


auditory


cues


only


with


the


speech


processor


worn


at its


usual


setting.


The


tests


were


administered


a sound-treated


audiological


test


suite,


sound


field,


with


the


subject


facing


the


output


speaker


at a distance


one


meter.


The


material


was


presented


an intensity


70 dB SPL.


Results


these


tests


can


found


Table


To obtain


a subjective


estimate


of performance


using


the


implant,


the


subj ects


were


asked


to complete


Performance


Inventory


Profound


and


Severe Loss


(PIPSL)










Table


Percentage


Choice
Tests.


correct


Spondee


results


and


CID


obtained
Everyday


on the


Sentences


4-Choice Spondee CID Everyday Sentences*
Ss % Correct % Correct


1 90 22

2 100 53

3 100 30

4 100 75

5 60 31

6 100 67

7 70 3

8 100 1


* Ten


sentences


were


presented










profound


hearing


loss.


It consists


of written


responses


items


from


siX


different


categories:


Understanding


SDeech


with


Visual


Cues,


which


evaluates


ability


understand


connected


speech


when


Intensity


speaker


, which


face


describes


vis


person


s ability


detect


everyday


sounds


and


their


relative


loudness


Respons e


to Auditory


Failure,


which


addresses


how


the


person


manages


communication


breakdowns;


Environmental


Sounds


, which


addresses


issues


recogni


zing


familiar


sounds


auditorily


only


Understanding


Speech


with


Visual


Cues,


which


evaluates


speech


understanding


without


visual


cues


Personal,


designed


to eli


cit


the


patient


feelings


about


their


hearing


loss.


Two


other


categories


Occupation


and


General


are


evaluated


separately


The


former


is evaluated


in an employment


setting


three


areas:


Understanding c


Speech


with


Visual


Cme,


Response


to Auditory


Failure,


and


Personal


The


latter


is made


up of "general


communication"


items


that


did


not


into


any


the


other


categories,











question


based


on the


type


amplification


worn,


, for


that


particular


tuation.


For


each


question


there


are


possible


responses


that


are


assigned


a value


in the


following


manner:


- Never


- Almos

- Occas


Never


ionally


About


Half


The


Time


- Frequ


ently


- Practically


Always


the


person


responds


with


"Does


not


apply"


, the


item


scored.


mean


score


obtained


each


category


with


exception


General


and


Occupation


categories


which


are


interpreted


separately


In fact,


authors


state


that


"all


items


can


treated


singly


in a rehabilitative


program.


" (Owens


Raggio,


1988,


Owens


and


Raggio


(1988)


also


make


following


recommendations.


Regarding
profile,
warranted


each


results


the
the


use


because


scale
with


of percentage
of the small


Preferably,


a client,


the


mean


scores
number


discussing
numerical


not
items


the


scores


the


scales


corresponding
0-0.4 (never),


cas


esc


treated
riptive


0.5-1.4
ionally)


according
terms, as


(practically


.5-3


to the
follows
never),


half


the


tinsi


(fram1inH lf\


(nranti rall


Always


interpretation


individual


.5-2


(about


e;,d


d


.


-


1


.










56
mc1
UtYI ?~~7~i] 04
H





01


0 30
Cl) N.~Z

CT) ~C
V~

C C

(4-4

Ca 0

LU -

SC '4-4
t C3 -

z C'
C~OO




0
oP 0
(4-
*, $4
a0
r 4



rdU
a,~


-J

C-,










and


there


was


wide


variability


between


subjects.


Group


means


and


ranges


are


illustrated


Fig.


and


four


of the


areas


describe


group


s perception


their


communication


abilities


using


thecochlear


implant.


They


reported


that


they


frequently


understood


speech


with


visual


cues,


were


able


to detect


everyday


sounds


and


their


relative


loudness,


and


recognized


familiar


sounds


using


audition


alone.


addition,


they


were


able


to understand


speech


without


visual


cues


approximately


half


the


time.


Experimental


Procedures


The


experimental


test


procedures


consisted


reading


Grandfather


70 under


Passage


simultaneous


counting


delayed


backwards


auditory


from


feedback


(SAF/DAF).


These


tasks


were


performed


three


different


conditions


speech


proce


ssor


off,


speech


processor


and


conventional


amplification


the


unimplanted


ear.


To insure


uniformity


subjects


were


given


written


instructions


experimental


tasks


and


an opportunity


to clarify


any


questions


regarding


the


tasks.


Eauioment


Equipment


research


to conduct


consisted


test


protocol


following:


the


this


subject











58



(U






a.o 0
04


So 5.
0,
C



C C?




1C
C a H


CO~n 0C


LU eL
\w 1w
arEc




Li. c0




0 ~ 'a
aO C

r m,
F-' >0 On4I



no

*a 0

V.) ~ C. 0>

*( 1 40t
21 ;zc O
OC.- '
n5a










input


boot,


Unitron


Direct


Audio


Input


Kit,


AIWA


cassette


tape


recorder


with


one-half


inch


microphone.


The


functioning


the


cochlear


implant


was


checked


using


the


Nucleus


22-Channel


Cochlear


Implant


System.


This


system


is composed


an IBM


compatible


microcomputer,


dual


This


processor


system


interface


used,


(DPI),


initially,


and


customized


to create


a "map"


software.


, or


prescription,


the


individual


subject.


The


"map"


created


attaching


speech


processor


to the


programming


each


electrode


separately.


The


parameters


are


the


patients'


threshold


(the


softest


level


sound


they


reported


hearing)


and


maximum


comfort


level,


the


loudest,


comfortable


level


of sound.


Each


subj ect


s implant


was


checked


prior


their


participating


the


study.


This


was


accomplished


attaching


the


speech


processor


and


sweeping


the


electrodes,


allowing


comparison


and


confirmation


the


speech


processor


"map"


and


the


last


prescribed


"map.


According


the


manufacturer' s


product


information


sheet,


the


Phonic


Mirror


miniDAF


has


a delay


control


setting


of approximately


25 milliseconds


(ms)


at position


number


to a maximum


effect


delay


varying


of 220


auditory


ms at position


feedback


number


duration


The


was


"map"


.e.










possibility


ms delay


produced


the


miniDAF


unit


dial


selector


control)


influencing


subjects'


responses,


subjects


were


tested


the


SAF


condition


without

miniDAF.


their

This


speech


processor


necessitated


being


using


connected

separate


to the

microphones.


The


recording


microphone


was


connected


AIWA


cassette


tape


recorder


throughout


test


conditions.


additional


crophone


(Lapel


Microphone


AT 164),


provided


with


miniDAF,


was


added


during


DAF


test


procedures.


was


recognized


that


the


necess


using


microphones--one


recording


and


one


routing


speaker'

it could


addition,


output


there


through

overcome


was


the

with


DAF--was


an impedance


a limitation;


available


mismatch


however,


equipment.


between


cochlear


implant


speech


processor


and


miniDAF


unit.


overcome


Ear.


this,


It should


a custom

also be


patch


noted


cord

that


was

their


provided

e was no


Phonic


frequency


shaping


of the


signal


taking


place


as that


signal


was


routed


through


the


DAF;


therefore,


the


only


change


occurring


was


the


preset


delay


Mendoza,


personal


communication,


1990).


To calibrate


miniDAF


unit,


a signal


was


generated


using


Coulbourn


instrument


modules


This


signal


had


duration


of 20 milliseconds


(ms)


with


an instantaneous










electrical


output


was


into


a Tektronix


Type


oscilloscope.


The


delay


time


miniDAF


was


adjusted


visually


observing


output


displayed


oscilloscope.


Once


a 200


ms signal


rate


was


obtained


delay


control


knob


on the


miniDAF


was


secured


fixed


position.


end


data


collection,


delay


time


signal


was


confirmed


using


equipment


and


method


described


above.


Tes


ting


with


conventional


amplification


was


accomplished


using


a Unitron


UE PPL


behind


-the-ear


hearing


and


direct


audio


input.


The


use


the


boot


enabled


experimenter


to connect


hearing


aid


miniDAF.


The


boot


was


to exclude


the


hearing


aid


microphone


thereby


preventing


amplification


any


environmental


sounds.


result


this


arrangement


was


that


only


the


subject'


speech


was


transmitted


through


hearing


aid


and


miniDAF


unit.


The


subjects


wore


a stock,


regular


style,


earmold


during


testing


exception


this


was


one


subject


who


still


wears


a hearing


unimplanted


ear.


He used


own


regular


style


earmold.


An electroacoustical


analysis


the


hearing


aid


(ANSI,


1976)


the


was


study


conducted


to confirm


beginning


performance


and


conclusion


based


the


manufacturer'


specification


data.


Initially,


coupling










Seamans,


personal


communication,


1990).


This


was


achieved


using


the


attentuator


provided


with


the


direct


audio


input


kit.


The


attenuator


output


was


measured


electoacoustically,


and


maximum


attentuation


recorded


was


58 dB SPL.


hearing


aid


volume


control


was


to midlevel


(volume


setting


was


coupled


to the


attenuator.


The


attentuator


output


was


again


measured


output


control


knob


was


rotated,


the


knob


was


secured


fixed


position


when


attenuation


provided


was


18 dB


attenuation


. Seamans,


personal


communication,


1990).


Conditions


Three


experimental


conditions


were


evaluated:


cochlear


implant


off,


cochlear


implant


and


conventional


amplification


unimplanted


ear.


Two


experimental


treatments


were


administered


each


experimental


condition:


simultaneous


auditory


feedback


(SAF) ,


and


delayed


auditory


feedback


(DAF).


These


two


treatments


were


alternated


the


following


manner:


SAF


with


processor


off/on,


DAF


with


processor


off/on,


SAF


with


conventional


amplification.


amplification


SAF


always


and


preceded


DAF


with


DAF


conventional


each


condition


and


was


used


as baseline


data.










"off"),


the


"on"


condition


was


designated


an odd


number


"off"


condition


was


designated


an even


number.


Using


random


numbers


subject.


table,


that


a number


number


was


ended


selected


an odd


first


number,


the


condition


was


selected


as the


first


condition


and


"off"


condition


was


number


two.


number


selected


was


even


number


"off"


condition


was


selected


as the


first


condition,


was


second


subsequent


condition


subjects,


speech


conditions


processor


were


counterbalanced


based


on the


first


subject


(Hegde,


1987).


Written


instructions


reading


and


counting


tasks


were


given


to each


subject.


They


were


asked


to read


the


Grandfather


Passaae


silently


order


to familiarize


themselves


with


the


material.


Any


questions


were


answered


and


they


then


read


passage


aloud


SAF


condition.


Following


reading


they


were


given


the


written


instructions


counting


task.


They


were


initially


asked


to count


themselves


silently


and


then


count


aloud.


The


purpose


the


silent


counting


was


to allow


them


practice


and


so startled


request


that


any


difficulties


might


be due


surprise


the


request


and


to the


test


condition.


The


use


of backwards


counting


obtain


uniform


speech


sample


material


was


selected


based


"on"


I, on"










difficulty


with


this


type


task


than


younger


people.


Interpretation


counting


results


this


study


need


to be


made


with


these


limitations


mind.


Pfeiffer


(1975)


included


backward


counting


his,


Short


Potable


Mental


Status


Questionnaire


LSPMSO)


test


cognitive


functioning


older


adults,


since


this


skill


might


related


to cognitive


decline


the


elderly.


During


testing,


the


selected


volume


on the


miniDAF


was


based


on the


subject


s reported


comfort


level.


The


implant

miniDAF


speech


processor


unit


to-microphone


was


previous


distance


connected

v describe


inches


was


directly

d manner.


maintained


to the

A mouth-


using


head

164)


held


microphone


provided


with


set-up


miniDAF


the

unit.


lapel

This


microphone


allowed


subj ect


s speech


productions


to be passed


through


miniDAF


resulting


their


hearing


with


a 200


ms delay.


Following


completion


test


protocol


with


the


cochlear


implant


the


conditions


were


repeated


with


the


subject


wearing


the


Unitron


UE 12-PPL


behind-the-ear


hearing


aid


their


unimplanted


ears.


The


hearing


aid


internal


controls


were


set


to provide


widest


response,


and


the


volume


setting


was


selected


based


on the


subject


s reported


comfort


level.


To determine


test-retest


reliability


the










All


the


cassette


subjects


tape


recorder


esponses

using a


were


recorded


one-half


inch


on an AIWA

clip


microphone.


A mouth-to-microphone


distance,


the


recording


microphone,


.5 inches


was


maintained


during


testing.


Analysis


Data


Transcription


Consensus


transcription


(Shriberg


, Kwiatkowski,


Hoffmann,


1988),


using


the


International


Phonetic


Alphabet


(IPA)


was


performed


a speech-language


pathologist


and


experimenter


productions.


fourth


on all


A sound-by-sound


sentences


subject


analysis


Grandfather


s experimental


the


Passaae


speech


second


dresses


himself


an old


back


frock


coat


usually


several


buttons


missing


A long


beard


clings


to his


chin


giving


those


who


observe


him


a pronounced


feeling


the


utmost


respect.)


and


the


following


additional


words


from


the


passage:


cracked,


ilp)fl,


the,


short,


more,


mod


ern


was


performed


on the


subject


s speech


productions


in the


SAF


and


DAF


conditions.


This


sample


was


selected


based


on the


frequency


occurrence


the


phonemes


included


conversational


English

analysis


(Mines,

was pe


Hanson,


formedd


Shoup,


the


1978) .


Programs


The


phonological


to Examine


Phonetic


e


- -










input


phonetically


transcribed


speech


samples


and


analyze


them


type


and


frequency


of speech


sound


error.


subj ect


s utterances


during


counting


were


also


transcribed


analyzed


same


manner.


Following


transcription


the


samples,


data


were


entered


into


the


PEPPER


program


analysis.


The


program


scanned


consonant


and


vowel


productions


errors


based


on a standard


adult


model.


model


had


been


entered


from


a transcript


an adult,


female,


General


American


speaker


who


read


the


Grandfather


Passage


and


counted


backwards


from


Possible


errors


detected


included


omissions.


substitutions,


distortions


that


differed


significantly


from


standards


typical


production


suggested


Shriberg


(1986).


Intertranscriber


reliability


Swas


determined


the


samples.


These


were


selected


ass


signing


each


taped


test


condition


a number


between


and


and


then


drawing


them


out


of a box.


These


samples


were


independently


transcribed


another


speech-language


pathologist.


Two


commonly


used


techniques


calculating


interobserver


reliability


have


been


discussed


McReynolds


and


Kearns


(1983).


"total


method"


based


percentage


a study.


agreement


total


"Point-to-point"


number


reliability


observations


tallies


observer











may


agree


on the


total


number


responses


obtained,


agreement


lacking.


regarding


The


the


weakness


occurrence


the


target


"point-to-point"


behavior


method


"relates


the


fact


that


the


agreement


level


obtained


direct


behavior"


reflection


(McReynolds


rate


& Kearns,


of production


1983,


of the


150-151)


target


The


result


that


instances


extremely


high


or low


rates


of observed


behaviors


, there


a "high


probability


that


agreement


(McReynolds

pitfalls me


independent


will


be obtained


Kearn


ntioned


samples


s, p.

above


was


based


151) .

a 2X2


performed


on chance


an effort


chi-square


alone"


to avoid


analysis


to determine


reliability


between


the


transcribers


(Huck,


Cormier,


Bounds,


1974).


The


obtained


1738


.41,


df = 1,


was


significant


beyond


the


.001


level


confidence.


Statistical


Analysis


The


small


sample


= 8)


this


study


prohibited


use


parametric


statistics


therefore,


a Friedman


two-way


analysis


variance


(ANOVA)


was


performed


In addition,


effect


size


(ES)


was


calculated


duration,


consonant


errors,


and


vowel


errors


reading


and


counting


as well


the


cochlear


implant


and


hearing


aid


conditions.


Null


1~ ~rnn~hocn aim race 4-a nn't. i nf a narirn ~rrha n nn


rhnt-


r~CC


h vnnt h a C!a a


fih~


nn~ i nn


nk a n nm a n an


EIYn










convenient


use


the


phrase


'effect


size'


mean


'the


degree


which


phenomenon


is present


the


population.


'the


degree


to which


the


null


hypothesis


false'"'


. 9-10)


Regardless


the


type


of phenomenon


being

effect


studied,


size


"the

zero"


null

and


hypothesis


...the


always


larger


means


this


that


value,


the

the


greater


degree


to which


the


phenomenon


under


study


manifested


(Cohen,


1977


, p.


10) .


This


relationship


also


carries


over


to sample


size


where


...the


larger


posited,


other


things


significance


criterion,


desired


power)

detect


being

it" (


equal,


Cohen,


the

1977


smaller


, p.


the


sample


Effect


size


size


necessary


can


described


relative


terms


of "small"


, "medium,


and


large"


To describe


these


terms


further


, numerical


values


have


been


defined:


a "small"


effect


size


, (b)


"medium"


effect


size


"large


size


0.8.


summary,


the


res


earch


experiment


was


a factorial


design


with


independent


variables:


SAF,


DAF,


conventional


amplification,


speech


processor


, and


speech


processor


on.


The


dependent


variables


were:


reading


rate,


counting


rate,


and


occurrence


speech


the


errors.


subject


The


s speech


experimental


processor


design


two


can


conditions


described


(off,










and


the


repeated


measures


taken


on each


subject,


data


were


subjected


a Friedman


two-way


analysis


variance


(Marascuilo


& McSweeney


, 1977


Huck


et al


1974


and


to calculations


test-retest


of effect


reliability


the


(Cohen,


experimental


1977)


Finally


conditions


was


evaluated


using


Pearson


Product-Moment


statistic














CHAPTER ]
RESULTS


The


purpose


this


study


was


investigate


effect


of delayed


auditory


feedback


(DAF)


on the


speech


production


post-lingual


cochlear


implant


users.


In addition,


subjects


were


also


tested


wearing


conventional


amplification


their


unimplanted


ear.


The


specific


research


questions


formulated


this


study


were:


there


a difference


between


Simultaneous


Auditory


Feedback


(SAF)


and


Delayed


Auditory


Feedback


(DAF)


conditions


the


duration


reading


and


counting


cochlear


implant


users?


there


a difference


between


the


SAF


and


DAF


conditions


the


total


number


consonant


and


vowel


errors


made


cochlear


implant


users


during


reading


and


counting


tasks?


there


a difference


between


the


SAF


and


conditions


duration


reading


and


counting


cochlear


implant


users


when


wearing


a hearing


aid


alone?


there


a difference


between


the


SAF


and










vowel


errors


cochlear


implant


users


when


wearing


a hearing


aid


alone


and


performing


reading


and


counting


tasks?


Effect


on Reading


and


Counting


Duration


Means


and


Rances


Reading


and


counting


duration


means


and


ranges


seconds)


as a


function


amplification


type


and


feedback


condition


are


reported


Tabl


and


illustrated


Figure


and


Figure


Examination


reading


duration


means


and


ranges


revealed


no differences


between


SAF


and


DAF


the


absence


amplification


(SAF


= 50.2


, DAF


= 49.0


However,


there


were


duration


differences


between


the


SAF


DAF


means


when


subjects


were


tested


with


the


cochlear


implant


(SAF


= 50.5


, DAF


= 64.0


and


the


hearing


aid


(SAF


= 49.1


, DAF


= 63


. Wide


ranges


reading


duration


were


also


noted


when


subjects


wore


either


cochlear


implant


or the


hearing


aid


Counting


duration


means


remained


similar


under


SAF/DAF


the


absence


of amplification


SAF


= 32


, DAF


with


the


cochlear


implant


(SAF


33.0


, DAF


36.8)


and


with


hearing


aid


(SAF


= 32


, DAF


35.1)


However,


counting


duration


varied


among


subjects











Table


Duration


means


conditions
hearing aid
auditory fe
Grandfather


and


ranges


amplification,


)


under


edback


ssaqe


second


cochlear


simultaneous


SAF/DAF)


and


when


counting


and


) in three
implant,
delayed


reading
backwards


SAF


Amplification


Time


Ranges
Seconds


Time/


Ranges
Seconds


Reading


None


49.0


40 57


- 57


- 58


64.0


102

107


63.0


Counting


None


- 49


33.0


**HA


- 48


- 49


20 50


23 59


35.1


18 49


*CI
**HA


= Cochlear
= Hearing


Implant
Aid


**HA









73




.QI


Ir ~
""41-4
0.0

o Li
u-c


O


CC
o> 2
Z 00
Th 00


"et
.4 P

O me N.
o- W~QiC
00CC r
0) kHU

I IB'
Cd V
CO4


O~)B


kOC
00Q


0 0"

oo 0 0
C'J
1~ ( 0Q)ZC










74



S01

r4Wr
00a




LL o U



C:



___~ C 0
o O


MC k
4i~r%


-JC
HO C
-~ 'coo


WCO

I r-
401
1C 43
QIW 0 M r4
4rW






SC--

o 0 0 0 0 0
o 0 (0 N
r: 46










ANOVA


A Friedman


performed


two-way


answer


analysis


each


variance


research


(ANOVA)


questions.


was


This


statistical


a group


method


data,


assigns r

yielding


anks


to each


a mean


rank


the


score


mean

for e


scores


ach


condition.


rank


The


scores,


greater


greater


differences


difference


between


between


mean


test


measures.


Duration


There


were


no significant


differences


either


duration


duration


of reading--X2(5,

of counting--X2(5,


= 8)


= 8)


.23,


= 9.96,


.14--or


<.07--in


the


the


SAF


and


conditions,


regardless


absence


presence


of amplification.


The


mean


rank


scores


reading


counting


duration


are


summarized


Table


and


Figure


and


Table


and


Figure


, respectively.


Examination


reading


mean


rank


scores


Table


revealed


that


the


absence


amplification


subj ects


scores


were


larger


.19)


SAF


condition


than


DAF


condition


.63)


indicating


that


their


reading


duration


was


longer


under


SAF.


Opposite


results


were


obtained


with


the


cochlear


implant


hearing


aid.


Mean


rank


scores


when


the


cochlear


implant


was


worn


were


(SAF)










Table


Mean


rank


reading
implant,
delayed


scores


and


conditions


hearing
auditory


effect


size


values


amplification,


aid)


under


feedback


simultaneo


(SAF/


three


cochlear
us and


DAF)


Reading


Duration


Amplification


Devi


None


HA**


SAF


2.88


DAF


5.00


0.96


0.99


(B) Consonant Errors


Amplification Device

SAF 5.87 1.69 3.94

DAF 4.44 4.13 3.00

ES 0.22 0.76 0.44


(C) Vowel Errors


Amplification Device

SAF 3.75 2.44 3.25

DAF 5.50 3.81 2.94

ES 0.24 0.65 0.41


ES***










































SAF


DAF










Table


Mean


rank


counting
implant,
delayed


scores


and


conditions


hearing
feedback


aid)
(SAF


effect


size


values


amplification,


under


simultaneou


three


cochlear
s and


/DAF)


Counting


Duration


Amplification


Devi


None


HA**


SAF


0.28


(B) Consonant Errors


Amplification Device

SAF 3.31 3.44 3.50

DAF 3.19 4.19 3.38

ES 0.19 0.52 0.17


(C) Vowel Errors


Amplification Device

SAF 4.06 2.88 3.38

DAF 3.06 4.31 3.31

ES 0.48 1.02 0.08


ES***

















5.13


SAF


DAF










further


evaluate


these


data,


calculations


effect


size


(ES)


were


performed


(Cohen,


1977)


was


stated


Chapter


, effect


size


not


intended


to imply


a cause


and


effect


relationship


but


rather


degree


to which


a relationship


between


two


measures


may


exist


(Cohen,


1977)


A large


effect


size


would


suggest


that


differences


did


exist


between


conditions,


e.g


SAF


and


Effect


size


values


reading


duration


are


summarized


Table


A small


value


was


seen


without


amplification,


very


large


values


were


seen


both


the


cochlear


implant


.96)


and


hearing


aid


.99)


conditions


Counting


The


counting


duration


mean


rank


scores


are


summarized


Table


5 and


illustrated


Figure


These


results


indicate


that


subjects


took


longer


to perform


counting


task


under


DAF


.81)


than


under


SAF


.88)


The


same


pattern


was


observed


when


cochlear


implant


was


worn,


(DAF)


and


(SAF)


Thi


pattern


continued


when


the


subjects


wore


the


hearing


aid


under


DAF


.06)


under


SAF


duration


.69)


tasks


can


results


seen


yielded


the


Tabi


same


counting


pattern


values


absence


of amplification,


i.e.


a small


value


.30),


that


was


noted


reading


data.


Unlike


* r


rrl -I~


W


--










DAT


Effect


on Number


of_ CosnntRll oe -nr
- - -. wrr U.. U W- p -..*


Means


and


RanQes


Means


and


ranges


consonant


and


vowel


errors


during


reading


and


counting


are


presented


Table


6 and


Table


and


illustrated


omissions,


Figures


substitutions,


through


and


deletions


All


nonstandard


of consonants


and


vowels


were


counted


as errors.


Suggestions


from


the


PEPPER


manual


(1986)


were


used


to determine


variations


productions


were


standard,


casual


speech


or errors.


Readinca.


Examination


of Table


6 revealed


small


differences


in the


number


of reading


consonant


errors


the


absence


amplification


(SAF


= 6.8),


and


with


the


hearing


aid


(SAF


- 6.3,


.8) .


There


were


large


differences


range


scores


, however,


indicating


wide


variability


among


the


subj ects


under


SAF


DAF


Vowel


reading


errors


followed


the


same


pattern,


i.e.,


small


differences


between


the


SAF/DAF


condition.


The


means


were


(SAF)


and


(DAF)


absence


amplifica-


tion,


(SAF)


and


(DAF)


with


the


cochlear


implant,


(SAF)


and


(DAF)


with


hearing


aid.


The


ranges


were


smaller


vowel


errors,


with


the


exception


amplification


condition


(SAF


- 17,


DAF


- 25).


of Consnn~n+


;rnrl


VnwP 1


kl~mr~










Table


Mean


and


ranges


three conditions
implant, hearing
delayed auditory


Grandfather


for c
(no
aid)


:onsonant


and


amplification,


under


feedback


(S


vowel


errors


cochlear


simultaneous
AF/DAF) when


Passage.


and
reading


Consonant


Errors


SAF


Amplification


Ranges
Errors


Ranges
Errors


None


- 18

- 15

- 9


**HA


Vowel


Errors


SAF


DAF


Amplification


Ranges
Errors


Ranges
Errors


None


- 17


- 25


**HA


*CI
**HA


= Cochlear
= Hearing


Implant
Aid










Table


Means


and


ranges


three conditions
implant, hearing
delayed auditory
backwards.


for
(no
aid)


consonant


and


amplification,


under


feedback


(S


vowel


errors


cochlear


simultaneous
AF/DAF) when


and
counting


Consonant


Errors


Amplification


Ranges
Errors


Ranges
Errors


None


14.1


- 34


- 22


- 22


**HA


- 21


16.2


- 24

- 24


14.7


Vowel


Errors


SAF


DAF


Amplification


Ranges
Errors


Ranges
Errors


None


- 17


**HA


- 14


*CI
**HA


= Cochlear
= Hearing


Implant
Aid









84







Cc


LCQ

CON
'C

C U (/




And





o 'a

Sal


OC

Os~r

(C O





COWr
0'0
SOCr


LL
-
00Q


CO


Va






















VC


ON~
*~c

.aCC


Cc4


ra,
2rQ
d C 0I
rdC1 J)
Sc~l
0.0
0-*4'-
O I V
1.4Z
0000 c


-CkU),
0~)000

OWVf

00


tacoI
000
rs'-CO




Cr40
0.te
CClr


0CCE
Sc










86



ft.

C1

I Via

Cr4
LL Q00

S0)
0'



C
Cud
C:

o CH ~
Z0
Z~ ov 4)
SO'4
Ct MC Wki



040)
0200


d04)
CD>.
-- 000,



0CC
00r
""'2



(I)lll C4J





CO V
OWOC


4) O









87







I~C *rII





C r



Co v

Et)
8i l

IFV
CWOO
m 1.40
r (0,-Ea
O 1.10>1
Z r WoO0

>CC


cno



WBUL

z m 0 V

CSk 0H











12.8,


DAF


= 16.2


with


hearing


aid


(SAF


= 13


DAF


= 14.7)


However


the


ranges


consonant


counting


errors


were


very


large


in the


SAF


and


condition.


Means


vowel


counting


errors


also


yielded


small


differences


between


SAF


without


amplification


(SAF


, DAF


with


cochlear


implant


(SAF


= 1.1


= 6.1),


and


with


hearing


aid


(SAF


, DAF


The


pattern


of a wide


range


scores


among


subjects


seen


previous


experimental


conditions


continued


vowel


counting


errors.


ANOVA


Reading


. The


results


the


Friedman


two-way


ANOVA


were


= 8)


11.64,


and


= 8)


= 7.75,


, for


reading


consonant


vowel


errors,


respectively


These


findings


revealed


that


the


introduction


of DAF


number


did


result


consonant


a significant


or vowel


errors


difference


reading.


Mean


rank


scores


reading


consonant


errors


are


summarized


Table


4 and


Figure


They


indicate


greater


number


errors


under


SAF


.87)


versus


.44)


the


absence


amplification.


pattern


was


repeated


when


the


subjects


wore


hearing


aid


under


SAF


.94)


under


DAF


I -


.001 .


However. th


pattern


was


reversed


when














5,87


SAF


DAF










more


consonant


errors


were


produced


under


.13)


than


under


SAF


69).


Examination


of effect


size


reading


consonant


errors


revealed


small


ES values


.22)


the


absence


amplifica-


tion


and


with


the


hearing


aid


A moderate


ES value


was


noted


with


cochlear


implant.


Mean


rank


Table


scores

Figure


of vowel


reading


DAF


errors


condition


are summarized

yielded larger


mean


rank


scores


reading


vowel


errors


absence


amplification


than


SAF


condition


.75)


This


was


the


case


when


subj ects


wore


the


cochlear


implant


reading


under


vowel


DAF


errors


.81)


were


versus


smaller


under


with


SAF


the


.44)


However,


hearing


under


.94)


than


SAF


.25)


Small


effect


size


values


reading


.24)


vowel


and


errors


with


were


hearing


obtained


aid


without


.41).


amplification


A moderate


value,


0.65,


was


obtained


with


the


cochlear


implant.


Countina.


Statistical


analysis


counting


errors


revealed


, E.


<.92


consonant


errors


and


= 8)


, p.


vowel


errors.


This


lack


significant


differences


between


the


SAF


and


conditions


the


counting


task


followed


the


same


pattern


shown


the


reading


task.


X2~5,



















5.5


SAF


DAF


---


























4.19


3 31 3 .4 4 3 .5
3sq 4

H re
II
b+=+ u

ii+i$4


3,19


3.38


SAF


DAF