• TABLE OF CONTENTS
HIDE
 Title Page
 Acknowledgement
 Table of Contents
 Abstract
 Main














Title: comparison of articulatory performance over three observations in apraxia of speech utilizing CVC words weighted for phonemic complexity /
CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00102814/00001
 Material Information
Title: comparison of articulatory performance over three observations in apraxia of speech utilizing CVC words weighted for phonemic complexity /
Physical Description: Book
Language: English
Creator: Andersen, Helen Graham, 1942-
Copyright Date: 1981
 Record Information
Bibliographic ID: UF00102814
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 07872477
ltuf - ABS1662

Table of Contents
    Title Page
        Page i
    Acknowledgement
        Page ii
    Table of Contents
        Page iii
        Page iv
    Abstract
        Page v
        Page vi
    Main
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
Full Text












A COMPARISON OF ARTICULATORY PERFORMANCE
OVER THREE OBSERVATIONS IN APRAXIA OF SPEECH
UTILIZING CVC WORDS WEIGHTED FOR PHONEMIC COMPLEXITY








By

HELEN GRAHAM ANDERSEN


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY




UNIVERSITY OF FLORIDA


1981















ACKNOWLEDGMENTS


I wish to express my deep appreciation to my Supervisory Com-

mittee Chairman Dr. Thomas B. Abbott for his guidance throughout my

graduate career. I wish to acknowledge the guidance and support given

to me by my co-chairman Dr. Leonard LaPointe. This manuscript is the

result of his encouragement and extraordinary commitment of time

devoted to supervising the completion of this dissertation.

I also wish to thank Dr. Paul Moore, Dr. William Wolking,

and Mrs. Margaret Nattress for their assistance and suggestions in the

review of this manuscript. Special thanks go to Dr. William Cutler

whose expertise and assistance in the production of the stimulus tape

were invaluable. In appreciation for the assistance with the statis-

tical analysis, I wish to acknowledge Dr. Ron Marks, Sandee, and the

Department of Biostatistics.

To my parents, Bill and Helen Graham, and my grandmother,

Florine 0. Smith, who have encouraged me and sustained faith in my

completion of this project and to my family Torsten, Karen and Kathleen

who have persevered, encouraged and supported me throughout this

research project, I dedicate this dissertation.
















TABLE OF CONTENTS




ACKNOWLEDGMENTS . . . . . . . . . . . . .

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

CHAPTER

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

Terminology . . . . . . . .
Differential Features of Neurogenic
Disorders . . . .
Characteristics of the Disorder . . . . . .
Variability in Phoneme Production . . . . .
Treatment . . . . . . . . . . .
Statement of the Problem . . . . . . . .
Statement of the Purpose . . . . . . . .

II MATERIALS AND PROCEDURES . . . . . . . . .

Selection of Subjects . . . . . . . .
Audiological, Speech and Language
Assessment . . . . . . . . . .
Development of Stimulus Words . . . . . .
Stimulus Word Recording . . . . . . .
Testing Procedures . . . . . . . . .
Transcription of Tape-Recorded Responses . . . .
Intrajudge and Interjudge Reliability . . . .

III RESULTS . . . . . . . . . . . . .

Word and Initiation Errors: Research
Question 1 . . . . . . . . . . .
Initial and Final Position Errors:
Research Question 2 . . . . . . . .
Individual Phoneme Errors: Research
Question 3 . . . . . . . . . . .
Error Categories: Research Question 4 . . . .
Word Error Frequency: Research
Question 5 . . . . . . . . . . .
Consistent and Inconsistent Word Errors:
Research Question 6 . . . . . . . .
Weighted Word Errors: Research Question 7 . . .


Page

ii

v



1

2

5
9
14
19
20
26

28

28

28
33
35
37
39
45

47











CHAPTER

IV


DISCUSSION . . . . . . . . .

Consistency of Quantitative Performance
Consistency of Qualitative Performacne
Performance on Weighted Words . . .
Suggestions for Further Research . .
Summary . . . . . . . .


APPENDIX

A BIOGRAPHICAL AND MEDICAL INFORMATION SHEET

B SUBJECT SELECTION CRITERIA SUMMARY SHEET .

C PORCH INDEX OF COMMUNICATIVE ABILITIES
INDIVIDUAL SCORES PERCENTILESS) . . .

D CONSENT FORM . . . . . . . .

E SPEECH SAMPLE TASKS . . . . . .

F DIRECTIONS FOR THE EXPERIMENTAL TASK . .

G WORD ERROR RECORD SHEET . . . . .

H SUMMARY DATA SHEET . . . . . .

I TRANSCRIPTION RECORD SHEET ........

J PHONEME ERROR PATTERN RECORD SHEET ....


K WORDS ELIMINATED FROM THE WORD ERROR RATE COUNT .

L STIMULUS WORDS USED IN THE EXPERIMENTAL TASK .

M STIMULUS WORDS WITH WEIGHTED VALUES . . . .

N FREQUENCY OF OCCURRENCE OF THE 120 STIMULUS WORDS

0 FREQUENCY OF OCCURRENCE OF THE 18 PHONEMES
USED IN THE 120 STIMULUS WORDS . . . . .

P RANK ORDER OF DIFFICULTY OF THE 120 STIMULUS
WORDS BASED ON TOTAL ERROR RATES FOR SIX SUBJECTS

REFERENCE LIST . . . . . . . . . . .

BIOGRAPHICAL SKETCH . . . . . . . . . .


Page


. . 85

. . 86


. . 88

. . 89

. . 90

. . 92

. . 94

. . 96

. . 98

. . 100

. . 101

. 102

. . 103

. . 109


. . 110


. 111

. . 113

. 117















Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



A COMPARISON OF ARTICULATORY PERFORMANCE
OVER THREE OBSERVATIONS IN APRAXIA OF SPEECH
UTILIZING CVC WORDS WEIGHTED FOR PHONEMIC COMPLEXITY

By

Helen Graham Andersen

June 1981

Chairman: Thomas B. Abbott
Co-Chairman: Leonard L. LaPointe
Major Department: Speech

A comparison of the articulatory performance over three obser-

vations separated by 48 hours for a group of six neurologically stable

apraxic adults revealed a consistent pattern of articulatory performance.

There existed no significant differences over the three observations

for the mean number of word and initiation errors, initial and final

position phoneme errors, errors for each of the 18 phonemes and the six

error categories. The mean number of words produced in error once over

three observations was significantly higher than the mean number of

words produced in error two or three times over three observations.

There were no significant differences between the mean number of con-

sistent and inconsistent word errors.

This study evaluated the relative difficulty of six weighted-

word lists designed from a set of phonetic criteria for the purpose of

hierarchically arranging CVC stimulus words for use in therapeutic

v











management. Word list A was weighted as the most difficult with the

difficulty of lists B, C, D, E, and F decreasing in descending order.

Analysis revealed that the mean word errors for word list A was sig-

nificantly higher than word lists D, E, and F. Word list C had a

significantly higher mean than word lists D, E, and F. Word list B

had a significantly higher mean than word lists D and E, but not sig-

nificantly higher than F.

The results indicate that quantitatively, the articulatory

performance of this group of apraxic subjects is predictable from

observation to observation suggesting a certain lawfulness to the

apraxic's articulatory behavior. One can more confidently trust that

a single measurement of articulatory performance in the stable adult

apraxic is representative of the subject's articulatory performance

from day to day. Qualitatively, the pattern of consistency and predic-

tability of articulatory performance is less evident.

The set of phonetic criteria selected for weighting CVC stimulus

words did not successfully predict the relative difficulty of the

six word lists weighted for phonemic complexity beyond indicating that

words containing fricatives and/or affricates were difficult to produce.

The results suggest that certain combinations of the phonetic variables

(phoneme, place, manner, voicing shift) appear to interact to influence

articulatory accuracy but specifically the question of which phonetic

variables and in what combination remains speculative.















CHAPTER I
REVIEW OF THE LITERATURE


A century of controversy and confusion has surrounded the

nature and terminology of the neurogenic phonological selection and

sequencing impairment referred to by some as apraxiaa of speech." A

myriad of terms describing this speech disorder can be found in the

literature and includes such labels as aphemia, motor aphasia, anarth-

ria, phonetic disintegration, cortical dysarthria, Broca's aphasia,

apraxic dysarthria and apraxia of speech. Despite the terminological

confusion, the descriptions of the resultant speech problem are strik-

ingly consistent. For example, Lebrun, Buyssens, and Henneaux (1973)

described anarthria as a condition in which ". . pronunciation is

severely impaired, or even impossible, though there is no paralysis of

the speech musculature. Comprehension, reading and writing, on the

other hand, are unabated" (p. 126). In a description of phonetic dis-

integration, Shankweiler and Harris (1966) noted that ". . in spite

of the difficulties in emitting speech sounds, expressive abilities

(e.g., as shown in writing) may be well preserved, indicating continued

access to vocabulary and continued adherence to the rules of syntax"

(p. 277). Johns (1968) used the term apraxia of speech to describe

the patient who ". . knows exactly what he wishes to say but he

cannot say it, in the absence of any problem ofsymbolic nature or paraly-

sis or incoordination of the speech musculature adequate to account for

the speech disorder" (p. 18).










During the past decade some degree of order has emerged from

this conflict, but complete agreement on the nature and appropriate

terminology of the disorder have not been reached.


Terminology

The controversy regarding the appropriate terminology for

describing a neurogenic phonological impairment appears to focus on

whether motor programming of volitional speech production can be selec-

tively impaired or whether there is no discrete separation of speech

motor programming from other language processes. Some researchers

agree that a motor speech disorder can occur in isolation resulting

in selective impairment of the programming of phoneme production without

impairment of language functioning (Shankweiler and Harris, 1966; Aten,

Darley, Deal and Johns, 1975; Lecours and Lhermitte, 1976; Itoh,

Sasanuma, and Ushijima, 1979). Shankweiler and Harris (1966) wrote,

". speech may be selectively impaired at the phonetic level" (p. 277-

278).

In an explanation of anarthria, Lecours and Lhermitte (1976)

wrote, ". . The neuronal nets governing the third articulation in

spoken linguistic expression are endowed with at least relative anatoma-

physiological autonomy and that their localization in the brain,

although not precisely known, is discrete enough that focal damage to

brain tissue can result in their isolated or nearly isolated dysfunc-

tion" (p. 106).

Itoh, Sasanuma and Ushijima (1979) observed that apraxia of

speech "occurs in combination with various degrees of symbolic dis-

turbances due to aphasia. However, it is also true that a relatively










'isolated' form of such a disorder can occur although it is much less

frequent" (p. 237).

Deal and Darley (1972) define apraxia of speech as

S. an articulatory disorder resulting from impairment,
as a result of brain damage, of the capacity to program
the positioning of speech musculature and the sequencing
of muscle movement for the volitional production of
phonemes. The speech musculature does not show signifi-
cant weakness, slowness, or incoordination when used for
reflex and automatic acts. Prosodic alterations may be
associated with the articulatory problem, perhaps in
compensation for it. (p, 639)

Recent research (Freeman, Sands and Harris, 1978; Itoh, Sasanuma,

and Ushijima, 1979) investigating timing errors in apraxia of speech

suggests that poor temporal coordination of airflow, phonation, velar

movement and articulation is a significant component of apraxia of

speech and perhaps underlies a significant portion of the phonemic

errors that constitute apraxia of speech.

LaPointe (.in press) in the following definition and descrip-

tion of apraxia of speech alludes to the impaired timing aspect of

coordinated movement of the speech musculature during purposive speech.

Apraxia of speech is a neurogenic phonologic disorder
resulting from sensorimotor impairment of the capacity
to select, program, and/or execute in coordinated and
normally timed sequence, the positioning of the speech
musculature for the volitional production of speech
sounds. The loss or impairment of the phonologic rules
of the native language is not adequate to explain the
observed pattern of deviant speech, nor is the dis-
turbance attributable to weakened or slowed actions of
specific muscle groups.

A rationale for the appropriateness of the term apraxia of

speech was proposed by LaPointe and Johns (1975):










The term apraxia of speech, focuses on the underlying
dynamics of the disorder. Apraxia of speech explicitly
(1) directs one's attention to the motor aspects of
speech, (2) emphasizes the volitional execution of
articulation, (3) excludes significant weakness,
paralysis, and incoordination of the speech musculature,
and (4) indicates a discrepancy between execution of
the speech act and relative linguistic intactness.
(p. 259)

Martin (1974) challenged the validity of the term apraxia of

speech. His major objection to the term was ". . the implication that

the observed phonological impairment is a motor impairment separate and

distinct from other language systems" (p. 53). Research by Martin and

Rigrodsky (1974) indicated that the ". presence of a semantic com-

ponent significantly aided in phonemic production" (p. 323). These

investigators concluded that ". . the evident effect of the higher

level semantic component upon phonological performance argues against

the concept of impaired phonemic performance operating separate from

other language processes" (p. 323).

In response, Aten, Darley, Deal and Johns (1975) defended their

choice of terminology arguing that apraxia of speech is a useful and

descriptive term.

It carries useful connotations as well, suggesting that the
disorder is not primarily due to sensory or intellectual
or higher language function impairment. It does not assert
that speech and language are unrelated or that phonologic
impairment is wholly uninfluenced by language functioning,
but it does suggest that programming of phoneme production
can be selectively impaired without impairment of language
function. It identifies a behavior that requires a thera-
peutic approach different from that required for
aphasia. (p. 419)

Although the interrelationship of phonological impairment with

other language.processes and the definitive description and terminology










to be applied remains an issue inviting further research, for the

purpose of this study the term apraxia of speech will be used to

denote the neurogenic phonological selection and sequencing impair-

ment described in the definitions by Deal and Darley (1972) and La

Pointe (in press).


Differential Features of Neurogenic Disorders

Apraxia of Speech and Dysarthria

Although apraxia implies impaired motor functioning, the disorder

is quite distinct from that produced by the motor impairment due to

faulty innervation of the speech musculature. In contrast, dysarthria

is characterized by reduced range, speed or direction of movement

of the speech musculature and the motor impairment is evident both

in articulatory errors and involuntary and reflexive acts.

Consistency of performance is a major distinguishing feature

between apraxia of speech and dysarthria. Johns and .LaPointe (1976)

give the following illustration of the predictability of performance of

the dysarthric patient:

S If the dysarthric patient is unable to produce a
correct phoneme in isolation, it is predictable that he
will be unable to produce it in a word, regardless of
whether the act is volitional or a reactive, automatic
speech production. If the dysarthric patient cannot
lick his lips upon request, it is to be expected that
he will be unable to perform this motor act in an
attempt to remove food or lick his lips after drinking
water. On the other hand, a patient with an apraxia
of speech might demonstrate an inability to perform
a motor sequence on request but do so reflexively. The
apraxic patient unable to produce the /sh/ phoneme in
the word "ship" may produce a perfectly acceptable /sh/
in an emotional response such as cursing. (p. 178)










Johns and LaPointe (1976) state that, ". a major distinc-

tion that can be made between the apraxic patient and the dysarthric

patient is that of volitional, purposive movement versus reflexive or

emotional responses in motor speech acts. Another is that of con-

sistency versus inconsistency, the former being associated with certain

types of dysarthria, the latter with apraxia" (p. 179).


Apraxia of Speech and Literal Paraphasia

In an effort to differentiate between Broca's aphasia, Wernicke's

aphasia and conduction aphasia, analysis of phonological errors of

these clinical groups has been performed (Lecours and Lhermitte, 1969;

Blumstein, 1973; Halpern, Keith and Darley, 1976; Burns and Canter, 1977).

The fluency-nonfluency concept also has been applied in an effort to

further delineate the speech characteristics of these clinical groups

(Kerschensteiner, Poeck and Brunner, 1972; Wagenaar, Snow and Prins,

1975).

Blumstein's (1973) phonological analysis of three aphasic groups

(Broca's, Wernicke's and conduction) revealed uniformity of error types

and error directions. An analysis of phonological errors (phoneme

substitution, simplification, addition and environment) revealed the

same distributional pattern for each group. A distinctive feature analy-

sis of phoneme substitution errors revealed that, in all groups,

errors of one distinctive feature occurred significantly more often

than errors of more than one distinctive feature, and there was a

common tendency in all groups to substitute unmarked values for marked

values.










The phonological analysis revealed no consistent differences

among the three aphasic groups. "And yet, clinically, these three

aphasic groups can clearly be distinguished" (p. 134). Blumstein sug-

gests that in classifying the various aphasic syndromes, it is important

to describe the profile of language abilities and disabilities and it

is the "totality" of patient's performance on various linguistic tasks

and the relative impairment of one language modality over another

that determines the patient's profile. Blumstein noted that the improper

control of articulatory movements and the flattened intonational pattern

often associated with Broca's aphasia affects the total configuration

and quality of speech production. She observed that what qualitatively

distinguishes Broca's aphasia is the presence of this dysarthric quality

of speech.

Perhaps what Blumstein is referring to when she uses the

descriptive phrase "dysarthric quality of speech" is the nonphonemic

hesitant, searching behavior typical of apraxia of speech characterized

by audible and silent groping in an effort to achieve correct position-

ing of the speech musculature. Compensatory adjustments such as slow

and even rate of speech in an effort to overcome expected articulatory

difficulty affect prosody and perhaps also contribute to the impres-

sion of a "dysarthric quality of speech."

Kerschensteiner et al. (1972) characterized fluent (sensory or

amnesic aphasia) and nonfluent (motor aphasia) groups. The nonfluent

group speaks at a very slow rate (50 wpm or less), word choice is

predominantly nominal or fragmentary, articulation is impaired, and










phrases are generally less than four utterances. There are marked

effort, many pauses, and poor prosody. Perseverations are frequent

as are phonemic paraphasias as opposed to verbal paraphasias. The

fluent group speaks faster often within normal limits. They use many

relational words and cliches. Articulation is relatively well pre-

served, effort is minimal with no marked augmentation of pauses.

Phrases are usually of four or more utterances with speech being

melodious and of normal rhythm. Perservations are few, phonemic

paraphasias relatively rare, and language is distorted by frequent

verbal paraphasias.

Canter (1973) refers to literal (or phonemic) paraphasia as

. . an articulatory pattern characterized by frequent phonemic and

syllabic errors in a context of more-or-less fluent speech--a pattern

which stands in marked contrast with the laborious, hesitant articula-

tory behavior of the individual with apraxia of speech" (p. 2). Canter

(1973) further outlined the salient differentiating characteristics.

Apraxics have difficulties in initiation and phonemic and syllabic

transitions whereas paraphasics do not. Paraphasics have marked

difficulties in phonemic and syllabic sequencing resulting in frequent

errors of metathesis whereas apraxics make these errors much less

frequently. Since Canter does not present data to support the differ-

entiating characteristic described above, it must represent speculation

based on impressions gained from clinical observation.

In apraxia, phonemic errors increase as articulatory complexity

increases. More errors are made on consonant clusters followed by

singleton consonants and vowels. In contrast, paraphasics demonstrate










roughly equal difficulty in producing consonant clusters, singleton

consonants and vowels. Phoneme substitutions by apraxic subjects

are usually reasonable approximations to the target phoneme whereas

paraphasic phonemic substitutions are often phonetically illogical.

Burns and Canter (1977) observed the complicated alteration

of target phonemes in aphasics with posterior cerebral lesions

and noted that when considered within a total response, these compli-

cated alterations appeared the result of semantic rather than phonemic

confusions. Take for example a subject's response to the stimulus

picture of a bag:

"Bushel, no . bas, no, not a basket . . It's
a funny word, isn't it? . Bas?" The final word
chosen for analysis was "bas," described phonemically
as an /s/ for /g/ substitution, a highly unusual error.
However, from inspection of the total response it
appears that the patient was confusing two semantically
and phonemically similar words, "bushel" and "basket,"
with the stimulusitem of "bag." This tendency for real
words which are semantically and phonemically related to
the target to intrude upon a final response was observed
repeatedly throughout the subjects' responses. (p. 504)

Goodglass (1975) noted that when tabulating the distance across

which consonant assimilation errors occur, Broca's aphasics substitute

an earlier or later appearing consonant for a target phoneme within the

same word twice as often as across word boundaries. Wernicke's

aphasics, on the other hand, make twice as many assimilations across

word boundaries as within and conduction aphasics are almost equal in

intra-morphmic and cross-morphemic assimilation errors.

Characteristics of the Disorder

Apraxia of speech may coexist with aphasia, dysarthria, an

oral-facial apraxia or any combination of these disorders. Its










occurrence in pure form is a relatively rare phenomenon (Shankweiler

and Harris, 1966; Deal and Darley, 1972; Trost and Canter, 1974). Trost

and Canter (1974) noted that "more typically, apraxia of speech in the

neurologically impaired adult is seen as an element of an aphasic

symptom complex" (p. 63).

Incidence studies by Wertz, Rosenbek, and Deal (1970) estimated

that, of the adult cases they reviewed, approximately 65 percent demon-

strated apraxia of speech complicated by aphasia. When this articu-

latory disorder does appear in isolation as a pure entity, the patient

S. knows exactly what he wishes to say but he cannot say it"

(Johns and LaPointe, 1976, p. 172).

Subjects displaying apraxia of speech show a discrepancy between

volitional, purposive and spontaneous, reflexive speech performance.

That is, automatic, reactive speech is better than volitional, pur-

posive speech (Johns and LaPointe, 1976).

LaPointe (1969) observed that one of the salient character-

istics of the disorder is the demonstration of islands of error-free

production during automatic or emotional utterances.

The apraxic speaker is usually aware of his errors but often

is unable to anticipate or correct them. Deal and Darley (1972)

observed that subjects with apraxia of speech can predict errors beyond

pure chance but the ability to predict errors is an individual rather

than a group characteristic.

Initiation difficulties are evident in the apraxic speaker's

groping behavior resulting in numerous retrials and repetitions. Johns










and Darley (1970) describe the groping, struggle behavior and its

resultant effect upon prosody:

They appeared to be--and they reported that they were--
attempting to slow down and approach articulatory
adjustments cautiously. Their speech was often effort-
ful. They appeared to be groping for the right produc-
tion, Facial grimacing was common, with silent and
phonated movements of the articulators as if the speakers
were trying to set up the sequence to follow. Their
rate tended to be slow, even, and hesitant as they used
these devices to compensate for their expected articula-
tory difficulties. But they still made phonemic errors;
and whatever partial success may be attributed to these
compensatory adjustments to gain articulatory control,
the prosody and inflectional patterns suffered. (p. 581)

Of the numerous speech and non-speech behaviors that have been

identified as characteristic of apraxia of speech, LaPointe (in press)

selects initiation difficulty, speech sound substitution, and variabil-

ity in production pattern on immediate repeated trials of the same

target as the cardinal features of apraxia of speech.

A reawakening of interest in the 1970s prompted researchers

(LaPointe, 1.969; Deal and Darley, 1972; Johns and Darley, 1972;

Martin and Rigrodsky, 1974; Trost and Canter, 1974; Dunlop and Marquardt,

1977) to further delineate the features of apraxia of speech. As a

result, a composite of two decades of research has yielded a fundamen-

tal understanding of the phonologic and linguistic parameters that

influence phonemic behavior in this neurogenic phonological selection

and sequencing impairment.

Errors vary with the complexity of articulatory adjustment.

Vowels are produced with greater accuracy than consonants and singleton

consonants are produced with greater accuracy than consonant clusters









(Shankweiler and Harris, 1966; LaPointe, 1969; Larimore, 1970; Trost

and Canter, 1974; La Pointe and Johns, 1975).

Feature analysis utilizing categories of place, manner and

voicing revealed a relatively high occurrence of errors involving in-

correct placement of the articulators (LaPointe, 1969; Trost and Canter,

1974) with lingua alveolar and bilabial phonemes being less susceptible

to error than other place categories ,(LaPointe and Johns, 1975). Pho-

nemes belonging to the fricative and affricate manner categories are more

susceptible to error (Shankweiler and Harris, 1966; Johns and Darley,

1970; Trost and Canter, 1974; LaPointe and Johns, 1975; Dunlop and

Marquardt, 1977). Analysis of voicing errors indicated no significant

differences between error percentages of voiced and unvoiced sounds

(LaPointe-and Johns, 1975) though subjects may be deficient in the

ability to voice or unvoice appropriately. Trost and Canter (1974)

found a predominance of voiceless-for-voiced substitution.

Phoneme substitution errors predominate, significantly out-

ranking the error categories of distortion, omission, addition, repe-

tition and prolongation (Shankweiler and Harris, 1966; LaPointe, 1969;

Johns and Darley, 1970; Deal and Darley, 1972; Trost and Canter, 1974).

Substitutions of consonant clusters for single consonants were common

and a distinguishing feature of apraxia of speech (Shankweiler and

Harris, 1966; Johns and Darley, 1970). Although sequential errors

(anticipatory, reiterative and metathesis) represent a relatively

small percentage of substitution errors, anticipatory errors signifi-

cantly outnumber reiterative errors by a ratio of 6 to 1 (LaPointe

and Johns, 1975).










Distinctive feature analysis of phoneme substitutions indicates

that substitution errors are generally close approximations to the

target sound (Blumstein, 1973; Trost and Canter, 1974). Contradictory

results (Shankweiler and Harris, 1966; LaPointe and Johns, 1975)

prompted LaPointe and Johns (1975) to suggest that conclusions regard-

ing the articulatory distance between error and target sounds must be

interpreted cautiously since they are dependent upon the number of

features used in the analysis.

Markedness analysis has been applied to phonemic errors of the

apraxic speaker (Blumstein, 1973; Marquardt, Reinhart, and Peterson,

1979). Marquardt et al. (1979) define markedness as a theory that

. . assigns a marked or unmarked value to each distinctive feature

ofa phoneme based upon articulatory and perceptual factors. The

resulting complexity of a phoneme is determined by the sum of its

marked features" (p. 482). Apraxic speakers produce more errors on

phonemes high in markedness (Marquardt et al., 1979) and there is a

tendency to substitute unmarked values for marked values (Blumstein,

1973; Marquardt et al., 1979).

No single position in a word (initial, medial, final) appears

to be more difficult (Johns and Darley, 1970; Larimore, 1970; LaPointe

and Johns, 1975; DunlopandMarquardt, 1977). Contradictory results in

the literature on this point (Shankweiler and Harris, 1966; Trost

and Canter, 1974) may be definitional depending on whether retrials

and repetitions are tallied as a separate error type or are included

in those errors made on the initial phonemes of words (LaPointe and Johns,

1975).










Word length (Johns and Darley, 1970; Larimore, 1970; Deal and

Darley, 1972) and word abstractness (Dunlop and Marquardt, 1977) are

influential factors in phonemic accuracy. Words of five phonemes or

less are produced more accurately than words of seven or more phonemes

(Deal and Darley, 1972). Grammatical class alone does not appear to

influence phonemic accuracy (Deal and Darley, 1972; Dunlop and Marquardt,

1977) but word length combined with grammatical class does negatively

influence phonemic accuracy (Deal and Darley, 1972). Articulatory

accuracy is better for meaningful than nonmeaningful utterances (Johns

and Darley, 1970; Martin and Rigrodsky, 1972).


Variability in Phoneme Production

Phonemic variability or inconsistency is one of the frequently

reported features of the disorder (Shankweiler and Harris, 1966; La-

Pointe, 1969; Johns and Darley, 1970; Lebrun et al., 1973; Johns and

LaPointe, 1976; LaPointe and Horner, 1976). Shankweiler and Harris

(1966) cited "variability" as one of the striking features of phonetic

disintegration.

Lebrun et al. (1973) analyzed the speech of two French subjects

who had sustained brain injury resulting in anarthria. These investi-

gators noted that mispronunciations in the two cases were random and

unpredictable in that words may be articulated correctly at one time

and mispronounced a few minutes later. A phoneme was neither syste-

matically dropped nor added.

In a clinical description of dysarthria and apraxia, Johns and

Darley (1970) wrote:












Whereas the key phrases for describing the speech
patterns of the dysarthric group were "consistency
of error" and "predictability of error patterns,"
the opposites of these terms "inconsistency" and "unpre-
dictability" best describe the efforts of the apraxic
group. Their speech productions--from sound to sound,
from word to word, from stimulus presentation to stimu-
lus presentation, in different modes of response, and
in contextual-conversational speech--were characterized
by a high degree of variability. (p. 580)

Until the early work of LaPointe (1969), Johns and Darley

(1970), and later Deal (1974) and LaPointe and Horner (1976), the

descriptions of the unpredictability and inconsistency of the error

pattern in the apraxic subjectweremore a clinical impression rather

than a description based on investigation. These researchers directly

addressed the issue of variability of articulatory performance in the

apraxic speaker.

LaPointe (1969) examined the articulatory performance of 28

brain-injured adults to determine whether sounds were consistently

misarticulated during each occurrence and to determine whether error

sounds were misarticulated in the same way each time they were produced

incorrectly. laPointe found that no subject demonstrated entirely

consistent articulatory deviation in regard to the consistency of phoneme

error and consistency of phoneme error pattern. In total, 13 of 28 sub-

jects demonstrated partial consistency in either consistently mis-

articulating some sounds during each occurrence in the test or con-

sistently producing some errors in the same way during each occurrence

or both. LaPointe advocated further exploration of articulatory error










consistency and urged reliability studies of behavior on articulatory

measures.

Johns and Darley (1970) first investigated the influence of a

repeated trials condition on phonemic accuracy. They analyzed the

consistency of phoneme production under three response conditions.

Although the results were not statistically significant, the apraxic

subjects did perform more accurately when they were given the opportun-

ity to repeat the stimulus word three times successively after being

spoken by the examiner. Johns and Darley attributed this success to the

subjects' ability to "zero in" on target often progressing from incorrect

to correct production. Variability of phoneme production was more

evident in the task Johns and Darley used to assess the integrity of

the apraxic subject's production of difficult consonantal sounds. Sub-

jects were asked to repeat words of varying length in which the test

phonemes were variably located. Inconsistency of articulatory perfor-

mance both between subjects and by a given patient appeared to be the

more typical speech behavior of the apraxics.

Deal (1974) explored adaptation and consistency during the

reading of a short passage. His purpose was to determine whether sub-

jects demonstrating apraxia of speech made fewer word errors in five

successive readings of the same 100-word paragraph and to determine

whether these same subjects made consistent word errors during succes-

sive readings of the paragraph. Analysis revealed that the apraxic

subjects as a group did make consistent word errors and did adapt after

five successive readings.











In an effort to gain further insight into the issue of variabil-

ity and consistency of error patterns over repeated trials in patients

presenting with a neurogenic phonological selection and sequencing

impairment, LaPointe and Horner (1976) extended the study of earlier

investigators. They described the error pattern and phonemic integrity

of seven subjects on a repeated trials task and assessed the effect on

error pattern and phonemic integrity of two response conditions and

four modes of stimulus presentation. Of special interest to this in-

vestigator were the results of the analysis of the variability of pro-

duction over 10 repetitions of each word. This study examined a substan-

tial number of repetitions as to the presence or absence of variability

relative to the immediate preceding response and whether or not the

response was closer, further, or the same distance from the phonologic

target as the previous response. The range of variability for the group

was 23 percent to 51 percent. These authors interpreted this high degree

of variability of production over repeated trials of a word as clinically

significant and viewed phonemic variability as perhaps the ". . most

potent discriminator of the patient with phonological selection and

sequencing impairment or apraxia of speech" (p. 9).

Studies investigating the consistency of articulatory performance

of verbal apraxic subjects have produced conflicting results. On the

one hand, Johns and Darley (1970) report a trend toward accurate phoneme

production when a subject is given the opportunity to produce a stimulus

word three times in succession. The subjects more consistently produced

an accurate response in the repeated trials condition.










Deal (1974) also reported consistency in terms of word errors

on five successive readings of a paragraph. The mean consistency scores

for the five subjects were all higher than 60 percent.

LaPointe. (1969) utilizing a single-word picture-naming task

reported that 35 percent of his subjects demonstrated partial consistency

in either misarticulating some sounds during each occurrence in the test

or consistently producing some errors in the same way during each

occurrence or both.

In contrast, LaPointe and Horner (1976) in a later study

reported a high degree of articulatory variability by apraxic subjects

on a repeated trials task where the judgment of the presence or absence

of variability was relative to the immediately preceding response.

Perhaps the difference in procedural methods may account for

the discrepancy in research results regarding variability of articulatory

performance in the apraxic speaker. The studies differed in the number

of stimulus words, the phonemic complexity of the stimulus words, the

modes of stimulus presentation and the types of response conditions.

The conflicting results also, in part, may be definitional. In order

to compare studies describing the variability of articulatory perfor-

mance, it is important to define the term "variability" and to determine

if all of the studies are operating under the same definition.

The studies were similar in one aspect. All data were collected

during a single session and assumptions were generated regarding con-

sistency of phonemic performance in the apraxic speaker.










Treatment

According to some, therapy for the apraxic speaker should con-

centrate on the disordered articulation by directing therapeutic

attention to the regaining of adequate points of articulation and the

sequencing of articulatory gestures. This approach is quite different

from the language stimulation and auditory and visual processing thera-

pies used with the aphasic (Rosenbek, Lemme, Ahern, Harris and Wertz,

1973).

The basic goals of therapy are to (1) relearn the adequate points

of articulation beginning with mastery of individual phonemes and

progressing to C+V utterances; (2) strengthen the ability to leave a

correct articulatory posture and return to it efficiently and accurately

using rapid, reduplicating syllabic utterances; (3) establish the

ability to accurately sequence and blend phonemes into progressively

longer utterances using CV CV sequences and a nucleus of CVC monosyllabic

words; (4) provide conditions such that the patient can progress from

automatic-reactive speech to volitional-purposive communication (Rosenbek

et al., 1973; Dabul and Bollier, 1976). Utilization of teaching

strategies such as phonetic placement techniques, progressive approxi-

mation from a key articulation gesture, verbal instruction, graphic cues

and auditory-visual stimulation has been claimed to be effective in

accomplishing basic goals of therapy (Chappel, 1973; Rosenbek et al.,

1973; Dabul and Bollier, 1976).

Rosenbek et al. (1973) generated five guiding principles of

therapy. The foundation of these is the principle that ". all therapy










activities should be organized according to task continue so that

the patient does not struggle and so that he works at a high level of

success during each session" (p. 462). Basic phonetic conditions that

influence the apraxic patient's articulatory accuracy such as manner

of articulation, speech-sound position, distance between successive

speech sounds, word length and word frequency have been identified and

basic therapeutic principles have been derived from these phonetic con-

ditions.

Therapy programs have been developed (Chappel, 1973; Rosenbek

et al., 1973; Skelly, Schinsky, Smith and Fust, 1974; Dabul and Bollier,

1976; Sparks and Holland, 1976; Holtzapple and Marshall, 1977; Deal and

Florance, 1978). Many of these programs have devoted attention to

hierarchically arranging levels of therapy activities/tasks, and have

established guidelines for an orderly progression from one activity/task

level to the next. However, less consideration has been devoted to

developing a hierarchical arrangement of stimulus material, especially

beyond the phoneme level, and establishing guidelines for an orderly

progression from easy to more complex stimulus material.


Statement of the Problem

Variability of Articulatory Performance

When describing variability of articulatory performance of the

adultapraxicspeaker, two clarifications seem important: (1) definition

of the term "variability," (2) observation of articulatory behavior on

more than one occasion.

To understand the articulatory performance in the apraxic speaker,

it is necessary to define the term "variability," since an operational










definition functions as a point of reference which allows more accurate

between study comparisons. What is meant by the description that ar-

ticulatory performance in the apraxic speaker is highly variable?

Does the description of variability refer to the variability of articu-

latory performance on a repeated trials task that is characteristic of

apraxia of speech? On such a task, an utterance may be produced

accurately the first trial but misarticulated in two entirely different

ways on the two succeeding trials. Does variability of articulatory

performance at this level of observation mean that there is no law-

fulness or predictability to the articulatory behavior of the apraxic

in general?

Perhaps a quantitative and qualitative analysis of error patterns

on a stimulus-response imitative task repeated over several observations

might yield a different pattern of articulatory performance with regard

to the issue of variability of articulatory performance in the apraxic

speaker. The comparison of articulatory behavior over several obser-

vations might reveal a certain lawfulness to the apraxic's overall

articulatory behavior that is not apparent in the analysis of variabil-

ity of articulatory performance on a repeated trials task.

The reliability of gathering data during a single test session

for the purpose of accurately describing the nature of consistency of

articulatory performance in verbal apraxia has not been determined.

Studies investigating phonemic variability in the apraxic speaker

(LaPointe, 1969; Johns and Darley, 1970; Deal, 1974; LaPointe and

Horner, 1976) were similar in one respect. All data were collected during









a single test session and assumptions regarding consistency of articu-

latory performance in the apraxic speaker were generated from that

single observation of each subject.

It is important to know whether data gathered during a single

test session is in fact representative of a subject's articulatory

performance from day to day. The measuring of consistency of articula-

tory performance during a single data-gathering session most likely

characterizes the subject's range of variability at that moment in time

under the particular subject, examiner, and environmental conditions

operating. It is unclear whether or not within subject generalizations

can be made based on data collected during a single test session. It is

also speculative whether data gathered during a single test session

accurately describe or characterize the subject as he will be the next

day, the next week, or the next month. Can we be confident that the

pattern of errors and range of variability demonstrated during a particu-

lar data-collecting session accurately describe the quality and variabil-

ity of performance that will exist the next day or at any other time in

the future?

If data gathered during a single test session arerepresentative

of a subject's articulatory performance from day to day, then we can

have confidence in the results of past studies that have based descrip-

tions of variability of articulatory performance on a single observation

of the articulatory behavior. If articulatory performance on a particu-

lar verbal task is highly variable from day to day, then the validity

of past studies is undermined resulting in a misunderstanding of the

entire disorder. Therapeutic management and the organizing of therapy










activities according to a task continue presupposes that the patient

will demonstrate a predictable pattern relative to the nature and

severity of the disorder, and that changes will be the result of progress

in therapy rather than changes due to variability of performance from

day to day.

In order to make assumptions regarding consistency of articu-

latory performance in apraxic speakers who are beyond the stage of

spontaneous remission, it appears important to observe their behavior

on more than one occasion to determine if the initial measurement is

representative of the quality of performance and the range of variabil-

ity on subsequent observations.


Hierarchical Arrangement of Stimulus Words

The importance of a hierarchically organized therapy program to

insure a high level of success applies not only to the appropriate

selection of levels of therapy activities but includes the appropriate

selection of stimulus material utilized at each level of therapy. When

developing therapeutic strategies, consideration of basic phonetic con-

ditions that influence articulatory accuracy also is essential in order

to organize therapy according to task continue.

The study of phoneme errors in apraxic speakers has provided

knowledge regarding the difficulty of individual phonemes and has

influenced the order of presentation for acquisition. For example,

plosives and vowels are less difficult to produce than fricatives and

affricates and thus are the earliest phonemes to be focused upon. As

therapy progresses from individual phonemes and CV sequences to the










production of monosyllabic CVC words, criteria for selection of stimulus

items in an effort to hierarchically organize stimulus material becomes

less clear.

A therapeutic strategy that would provide a system for evaluat-

ing and quantifying the motoric or articulatory complexity of CVC words

would permit a hierarchical arrangement of CVC stimulus words thereby

maximizing the potential for the appropriate sequencing of material

during treatment.

A word-weighting technique was used to evaluate the effect of

certain linguistic variables on phoneme accuracy by Deal and Darley

(1972). To date, similar word-weighting techniques for the purpose of

evaluating the motoric complexity of CVC monosyllabic words have not

been used. By assigning a numerical value to CVC stimulus words based

on a set of criteria derived from phonologic principles, words could be

hierarchically arranged and then tested on a sample of apraxic subjects.

The therapeutic advantage of an orderly presentation of stimulus words

progressing from less motorically complex words to words requiring more

complex articulatory gestures is that it allows for the control of

fatigue and frustration by selecting levels of performance with high

success rates and then gradually increasing difficulty as the patient

progresses.

Motoric complexity may be described in terms of the number of

different articulatory maneuvers required to accomplish a CVC sequence.

For example, the number of different phonemes, different manners of

production, and of laryngeal adjustments (the number of times the vocal










folds are required to initiate adduction and abduction) in an utterance

contributes to the overall motoric complexity of that utterance. Smith

and Ruder (1975) have observed that sequences requiring minimal adjust-

ments of the larynx are less complex since a minimum of physical effort

is required. This, however, may not hold for longer or more complex

articulatory sequences. For example, normal speakers seem to have less

articulatory agility on utterances with minimal change in articulation.

This is the basis for the design of most "tongue-twisters," which normals

find difficult to produce. Dabul and Bollier (1976) noted that some

patients perform more poorly on voiced-voiceless combinations such as

"pa ba" than on voiced-voiced combinations.

The utilization of a set of criteria based on the interaction

of the subphonemic features of manner, place and voicing and selected

phonetic conditions that influence phonemic accuracy permits the

weighting of words to reflect motoric complexity. For example, the

CVC word "kiss" requires a different place and manner of articulation

for the initial and final consonant, two laryngeal adjustments to

accomplish the voiceless-voiced-voiceless sequence, and contains a

fricative. This word would be motorically more complex than the word

"none" that requires the same place and manner of articulation for the

initial and final consonant, no laryngeal adjustments to accomplish the

voiced-voiced-voiced sequence, and contains no fricatives or affricates.

More articulatory maneuvers are required for the production of "kiss"

than for "none" and thus the former word may be considered more motori-

cally complex than the latter.










The development and evaluation of a weighted-word technique

utilizing a set of criteria based on the interaction of subphonemic

features of place, manner, and voicing to quantify motoric complexity

of CVC monosyllabic stimulus words for hierarchical arrangement appears

to have potential usefulness for understanding more about relative diffi-

culty of stimulus items.


Statement of the Purpose

Although two decades of research have greatly expanded the

knowledge of the phonetic and linguistic characteristics of apraxia of

speech, certain areas continue to warrant investigation for the purpose

of resolving inconsistent research findings reported by investigators and

for the purpose of furthering our understanding of the nature of this

impairment. The purpose of this study is twofold: (1) to assess the

consistency of articulatory performance over three observations separated

by 48 hours in the neurologically stable apraxic adult, and (2) to evalu-

ate the relative difficulty of a weighted-word list designed from a set

of phonetic criteria.

The research questions were as follows:

1. Are there significant differences among the group mean

total errors (word and initiation errors combined), mean

word errors, and mean initiation errors for observations

one, two, and three?

2. Are there significant differences among the group mean

total phoneme errors (initial and final position errors

combined), mean initial position phoneme errors, and


I










mean final position phoneme errors for observations

one, two, and three?

3. Are there significant differences among the group mean

errors for each of the 18 phonemes for observations one,

two, and three?

4. Are there significant differences among the group mean

errors for each of the six error-categories for observa-

tions one, two, and three?

5. Are there significant differences among the group mean

number of words produced in error one, two, and three

times over three observations?

6. Are there significant differences between the group mean

number of consistent errors and the group mean number

of inconsistent errors for the three observations?

7. Are there significant differences among the group mean

total errors (word and initiation errors combined), mean

word errors, and mean initiation errors for word lists

A, B, C, D, E, F which increase in phonemic complexity?














CHAPTER II
MATERIALS AND PROCEDURES


Selection of Subjects

Six neurologically impaired subjects were selected from the

clinical records of the Audiology and Speech Pathology Services of the

Veterans Administration Medical Center in Gainesville, Florida, for

participation in the study.

All subjects were male, native speakers of English, between

the ages of 45 and 66 years with a mean age of 58 years and with eight

or nine years of formal education. All subjects had sustained medically

documented left-hemisphere cerebral vascular accidents with no history

of dementia. At the time of the study, at least 15 months had elapsed

since the onset of the neurological disorder with a range of 15 to 150

months and mean of 44.6 months post-onset. Upon initial evaluation by

the Speech Pathology Service, all subjects were diagnosed as having

apraxia of speech with concomitant aphasia and were subsequently en-

rolled in speech and language therapy for varying lengths of time

ranging from one month to 24 months with a mean of 13 months. A summary

of the above descriptive data is presented in Table 1.


Audiological, Speech and Language Assessment

Prior to inclusion in the study, subjects were evaluated for

integrity of the speech and hearing mechanism in addition to an assess-
ment of speech and language function.
















TABLE 1
SUBJECT DESCRIPTIVE DATA


Years of Months Months
Formal Post of
Subject Age Education Onset Therapy

1 57 8 39 24

2 60 8 18 1

3 66 9 150 20

4 45 9 23 2

5 63 8 18 17

6 59 8 15 14

Range 45-66 8-9 15-150 1-24

Mean 58 8.3 44.6 13










Audiological Evaluation

Subjects were screened at 25 dB HL (hearing level) for frequen-

cies 250 Hz, 500 Hz, 1 kHz, 2 kHz, 4 kHz using a portable Maico Audiometer

Model MA 27, calibrated to ANSI 1969 standard. The criterion for elimi-

nation from the study was failure to respond to two frequencies in both

ears at the selected intensity levels. Two subjects were tested in a

sound-treated room and four subjects were tested in relatively quiet

surroundings at home.


Evaluation of the Speech Mechanism

Assessment of the integrity of the speech musculature involved

observing isolated voluntary movements of the speech mechanism. No

subject demonstrated apparent weakness, slowness, or reduced range of

motion of the speech musculature.


Speech and Language Assessment

The Porch Index of Communicative Abilities picaA), a battery of

18 subtests of speech and language function, was administered. This

test was standardized on 280 randomly selected aphasic patients and is

scored on a 16-point scale of adequacy with 16 being an accurate, complex,

complete, prompt response and a score of 1 indicating no attention to the

stimulus. Criterion for inclusion in the study was an overall percentile

ranking on the 18 subtests of between 45 percent and 90 percent. The

range for the group was 49 percent to 82.5 percent with a mean of 66.25

percent. Overall percentiles for the 18 subtests for each subject are

presented in Table 2.












TABLE 2
SUMMARY OF SPEECH AND LANGUAGE
(PERCENTAGES)


EVALUATION


Substitution Omission
Overall Variability Initiation Addition Distortion
Subject PICA Score Errors Errors Errors

1 79.0 75.0 0.0a 66.0 33.0

2 82.5 65.0 50.0 50.0 0.0

3 49.0 50.0 55.0 27.0 16.0

4 63.0 65.0 33.0 48.0 19.0

5 71.0 70.0 25.0 51.0 22.0

6 53.0 95.0 27.0 64.0 9.0

Ranae 49.0-82.5 50.0-95.0 0.0-55.0 27.0-66.0 0.0-33.0

Mean 66.3 70.0 32.0 51.0 16.5

aSubject number 1 was included in the study even though he did not
demonstrate initiation difficulties during the spontaneous speech
sample. During two years of therapy, this subject's apraxic impair-
ment was well documented and therapeutic intervention focused on
reducing the initation difficulty by utilizing silent pause times
prior to verbal production rather than audible groping and retrials.










To judge the presence or absence of apraxia of speech, a

speech sample was tape recorded and analyzed for the predominant charac-

teristics of the disorder.

First, a 50-word spontaneous speech sample was elicited and

analyzed for the frequency of occurrence of phonemic errors of substi-

tution, addition, distortion, omission, and evidence of initiation

errors characterized by repetition of the first sound or syllable and/or

retrials with varying sound combinations each trial.

Next, five polysyllabic words (administration, refrigerator,

impossibility, tornado, condominium) were presented as a repeated

trials task to assess variability of phonemic production. The subject

repeated each word five times consecutively. Variability was determined

by judging each repetition as to the presence or absence of variabil-

ity relative to the immediately preceding response. A variability

percentage score was assigned to each of the five polysyllabic words

based on the number of times each repetition of the word varied from

the preceding response. The first production served as the reference

and there were four opportunities for variability. For example, if

the word was produced differently four times, the word was judged to

be 100 percent variable. If the word was produced differently twice,

it was assigned a variability score of 50 percent. The variability

scores for the five polysyllabic words were averaged and a mean variabil-

ity score was derived for each subject.

A subject was included in the study if analysis of the 50-word

speech sample and the polysyllabic repeated trials task indicated the

following:










1. A predominance of substitution and addition errors

over distortion and omission errors

2. Variability of phonemic error pattern of at least

10 percent

3. Evidence of initiation errors characterized by repe-

tition of the first sound or syllable and/or retrials

with varying sound combinations each trial

A summary of the results of each subject's speech sample analysis is

presented in Table 2.


Development of Stimulus Words

To assess the validity of utilizing a weighted word strategy

to hierarchically arrange stimulus words according to anticipated

difficulty of production, a list of 120 monosyllabic, CVC (Consonant,

Vowel, Consonant) predominantly high frequency, real words were

selected and assigned a numerical value based on the following precon-

structed criteria:

1 point was assigned if the CVC stimulus word contained

a different phoneme in the initial and final

position

1 point was assigned for each occurrence of an affricate

or fricative in the CVC stimulus word

1 point was assigned if the manner of production was

different for the initial and final phonemes in

the CVC stimulus word

1 point was assigned if the initial and final phoneme in

the CVC stimulus word utilized a different place

of articulation












1 point was assigned for each occurrence of a voicing

shift. A voicing shift was defined as a change

from voicing to unvoicing or the reverse in suc-

cessive phonemes. For example, "dog" (voiced-

voiced-voiced) has no voicing shifts whereas

"cat" (voiceless-voiced-voiceless) has two voicing

shifts

The points credited to each stimulus word were totaled and the

sum total represented the weight of the stimulus word (Appendices L and

M). The most difficult words based on the preconstructed criteria

described above received a weighted word score of 6 and the least

difficult words received a weighted score of 2. Based on the weighted

value, each stimulus word was assigned to one of six word lists comprised

of 20 words each. Stimulus words weighted 6, 5, 4a, 4b, 3, 2 were

assigned to word lists A, B, C, D, E, F, respectively. The distinction

between word list C and word list D both with an assigned value of 4

is that words in list C have at least one occurrence of a fricative or

affricate and words in list D have no occurrence of a fricative or

affricate. This distinction is also true for word lists A, B, C and

D, E, F. Stimulus words in lists A, B, and C have at least one occur-

rence of a fricative or affricate whereas words in lists D, E, and F

have no occurrence of a fricative or affricate.

No attempt was made to obtain equal frequency of occurrence in

the initial or final position for the 18 phonemes represented, and










no restriction was placed on the grammatical class of the 120 stimulus

words.

The vowel nucleus was not considered in the weighting of the

stimulus words. Trost and Canter (1974) observed that the majority

of the errors made on vowels by the apraxic speakers they studied

suggested that vowels were misarticulated in relation to the articula-

tory difficulty on contiguous consonants, that is, difficulty in

articulating the CV or VC transition and/or selecting the correct

consonant phoneme(s) seemed to cause the vowel nucleus to become dis-

torted.

Based on a word count by Thorndike and Lorge (1972), 88 of

the 120 stimulus words had a frequency of occurrence of 50 or 100 per

one million printed words and received a rating of A or AA, respec-

tively. Thirty-two words had a frequency of occurrence of less than

50 per million words. To insure that the lower frequency words were

not a variable influencing the error rates for the 120 stimulus words,

an equal number of high frequency words matched for weighted value

were randomly selected and their error rates compared to the error

rates for the 32 lower frequency words. A two-sample t-test was

administered with the results being non-significant (p > .939).


Stimulus Word Recording

A random numbers table was used to randomize the 120 stimulus

words prior to recording the stimulus words on low-noise high-fidelity

recording tape.










A Sony all-silicon transistor tape recorder model TC 377 and

a Sony Condensor ECM 260F microphone were used in recording the stimulus

words on a master tape.

The speaker was comfortably seated in a sound-treated room

with the microphone placed 12 inches from the speaker's mouth. The

words were recorded at a speed of 7.5 ips. A BrUel and Kjaer type

1616/2209 sound level meter with a SLM setting of 50 dB fast was

placed in view of the speaker for monitoring voice level. A second

individual situated outside the sound treated room monitored the vu

meter of the tape recorder. The tape recorder vu meter range during

recording was -10 to -20 dB. The stimulus words were recorded with a

5-second interval between each word.

To insure the quality of the recording, five normal listeners

judged the intelligibility of the tape recorded words. A criterion

level of 100 percent accuracy was established. Any word not identified

accurately by all five normal listeners was retaped and submitted to

five different normal listeners for judgment of intelligibility of

the recorded word. This procedure was followed until all 120 words had

been identified with 100 percent accuracy by five normal listeners. The

retaped stimulus words were spliced onto the master tape.

To control for any bias that might result from following the

same order of stimulus word presentation each of the three observa-

tions, the order of stimulus word presentation was counterbalanced.

The master tape of 120 stimulus words was divided into three equal

sections of 40 words each. The first 40 words of the master tape were










assigned to section A, the second 40 words to section B, and the third

40 words to section C. The three sections were rearranged and retaped

to produce three separate tapes of 120 stimulus words each.


Tape Order of Sections

Tape 1 ABC

Tape 2 B C A

Tape 3 C A B


In addition to counterbalancing the order of stimulus word

presentation, the order of presentation of the three tapes to the sub-

jects was also counterbalanced in the following manner:


Order of Tape
Subject Presentation

1 1, 2, 3

2 1, 2, 3

3 2, 3, 1

4 2, 3, 1

5 3, 1, 2

6 3, 1, 2


Testing Procedures

Two Sony all-silicon transistor tape recorders Model TC 270,

a pair of Telephonic earphones TDH 49 fitted with 41/AR cushions and

a Sony F-27 Dynamic microphone were used in the presentation of the

stimulus words to the subjects. One tape recorder was used for the

presentation of the taped stimulus words and the other for recording










the subject's responses. Two subjects were tested in a sound-treated

room at the VA Medical Center and four subjects were tested at home in

quiet surroundings.

The subject was comfortably seated with the microphone placed

approximately 6 inches from the subject's mouth. The stimulus words

were presented through earphones.

Each subject was presented the list of 120 stimulus words three

times with a time interval of 48 hours between each observation. The

time interval between observations one and two and observations one

and three was 48 hours and 96 hours, respectively.

Every effort was made to duplicate the experimental conditions

for each of the three observations. Each subject was tested at approxi-

mately the same time, in the same location and under the same environ-

mental conditions, and all subjects appeared to be in the same

emotional and physical state each of the three tests sessions.

Prior to each of the three presentations of the stimulus words,

an explanation of the nature of the experimental task was read to each

subject (Appendix F). A brief training session followed in which tape-

recorded words similar to the test stimulus words were presented through

earphones and the subject was asked to repeat the words. The purpose of

the training session was to acquaint the subject with the nature of

the experimental task and to calibrate the recording equipment.

Immediately following the training session, the 120 stimulus

words were presented to the subject through earphones and the subject's

responses were preserved on magnetic tape for future analysis.










Transcription of Tape-Recorded Responses

The International Phonetic Alphabet (IPA) plus selected

allophonic symbols and modifiers from the IPA as presented by Fisher

and Logemann (1971) was used in the transcription of each subject's

tape-recorded responses.

At least three weeks intervened between the testing of a sub-

ject and the transcription of the subject's tape-recorded responses.

Each tape was listened to in its entirety at least three times and

individual words were listened to as many as 15 to 20 times each.

The first evaluation of a subject's tape-recorded responses

involved a plus/minus judgment of word production accuracy. A

response was scored plus if the entire stimulus word was produced

accurately and promptly. A response was scored minus if one or more

of the three phonemes comprising the CVC stimulus word was produced

inaccurately. In addition to the plus/minus scoring judgment, initia-

tion errors were assigned a numerical score based on a scoring system

modified after the PICA scoring system.

The following scoring system was used to describe three types

of initiation errors:

13 = A 13 was assigned if there was a significant silent

delay prior to the attempted production of the

stimulus word suggesting need for additional proces-

sing time to formulate the motor pattern. A score of

13 was also assigned to a response that was preceded

by an audible utterance denoting a pause such as

"uh" or "um."










(1J = A ()was assigned if audible groping was evident in
which one or more productions of the initial phoneme

or syllable was produced followed by an attempted

production of a stimulus word such as "/f/ /f/ face."

Prolongation of the initial phoneme or syllable was

also scored (3.

10 = A 10 was assigned if audible groping was evident in

which the initial phoneme or syllable produced was

different from the initial phoneme or syllable of

the stimulus word followed by an attempted production

of the stimulus word such as "/t/ /p/ face." Also

included were unintelligible vocalizations followed

by an attempted production of the stimulus word.

The plus/minus judgments and the numerical scores assigned to

initiation errors were recorded jointly on the Vord Error Record Sheet

(Appendix G). This recording procedure enables one to observe whether

or not the initiation error was followed by an accurate production of

the stimulus word.

The recorded data from each subject's Word Error Record Sheet

were utilized to complete the Summary Data Sheet (Appendix H) which

allowed compilation and comparison of the following data over three

observations:

1. An error rate for each of the six weighted word lists

(A,B,C,D,E,F) for each of the three observations

2. A total word error rate (errors for six weighted word

lists combined) for each of the three observations










3. An overall total word error rate for the three obser-

vations combined

4. An error rate for each of the three types of initia-

tion errors by word weight over three observations

5. A total initiation error rate for each of the three

observations

6. An overall total initiation error rate for the three

observations combined

The second evaluation of a subject's tape-recorded responses

involved the complete transcription of those stimulus words produced

inaccurately plus transcription of the initiation errors. The tran-

scribedresponses were recorded on the Transcription Record Sheet

(Appendix I) which was designed to readily allow the comparison of a

subject's three productions of a stimulus word.

After transcribing all stimulus words produced in error, each

misarticulated phoneme was examined and the error classified according

to one of the six error categories (substitution, distortion, omission,

addition, distorted substitution, other).

The definitions of the six error categories are as follows:

Substitution--The replacement of a target phoneme by

another phoneme. The phoneme substitution in error

was recorded on the Trancription Record Sheet.

Distortion--Indistinct production of a phoneme. Even

though the production of the phoneme is imprecise,

the desired phoneme is recognizable. Distortions were

noted on the Transcription Record Sheet by assigning










the number 14 to the distorted phoneme and when

possible noting the direction of the distortion by

using the IPA allophonic symbols and modifiers.

Omission--Failure to produce the desired phoneme and no

other phoneme is substituted in its place.

Addition--The addition of phoneme(s) that precede or follow

the target phoneme. To differentiate an initiation

error from an additive initial consonant cluster,

evidence of coarticulation was the key judgment. A

brief pause between two phonemes would indicate an

initiation error whereas coarticulation of the two

phonemes would indicate a consonant cluster.

Distorted Substitution--The combination of the above

substitution and distortion error categories.

Other--Any phonemic error that does not fit into the other

five error categories.

A final notation made on the Transcription Record Sheet was

that of "no response." If a subject demonstrated initiation difficulty

in which the subject attempted the production of the target word several

times resulting in several initiation attempts but no final production

of the target word appeared to be selected, the initiation attempts were

transcribed and scored and a minus score was recorded on the record

sheet with the notation of "no response."

After each phoneme error had been classified into one of the

six error categories, the error type was then recorded on the Phoneme










Error Pattern Record Sheet (Appendix J). This data sheet displayed a

confusion matrix on which the phoneme substituted in error for the

target phoneme was indicated in the body of the matrix. In addition,

the data sheet summarized the frequency of occurrence of the six error

categories in the initial and final positions for each of the 18

phonemes represented for each of the three observations and for the

three observations combined.

Each subject's tape recording of responses was listened to in

its entirety for a third time, and a second plus/minus scoring judgment

was recorded and compared to the first plus/minus scoring judgment.

A discrepancy in scoring between the first and second judgments for a

word resulted in that word being assigned to a separate list for fur-

ther evaluation.

A second trained listener with 15 years experience in speech

pathology listened simultaneously with the researcher to the list of

words designated as requiring further evaluation. The two trained

listeners reviewed each word and came to agreement as to the accuracy

or inaccuracy of the production and the most descriptive and acceptable

transcription of the word in error.

A research question posed in the study required a final review

of each subject's tape-recorded responses. The research question

attempted to determine whether a word produced in error two or more

times over three observations was produced in error the same way each

time by the subject.

Since there are often subtle changes in the way a word is

produced in error from one production to another, it was sometimes










difficult to identify subtle production differences with the trans-

cription symbols used in this study. If a word was inaccurately pro-

duced by a subject two or more times over three observations and the

Transcription Record Sheet indicated that subtle differences in

production over the three trials might be present, then the three pro-

ductions of the word were taped consecutively onto a second tape for

ease in determining with greater accuracy the presence or absence of

subtle differences in production.

A ratio of the number of times a word was produced in error

to the number of times the word was produced in error in the same

way was utilized in quantifying the judgment of consistency of error

production. For example, a ratio of 3:3 indicated that of the three

times the word was produced in error over three observations the word

was produced in error in the same way each of the three times. A ratio

of 2:0 indicated that of the two times the word was produced in error

over three observations the word was produced in error differently each

of the two times. The ratio scores were recorded on the Transcription

Record Sheet.

The Transcription Record Sheet was examined for error produc-

tions that were attributed to auditory perceptual confusion rather

than the inability to articulate the stimulus word accurately. Any

attempted production of a stimulus word that resulted in a well-

articulated acoustically similar real word which was different from

the stimulus word was suspected of being heard inacurately by the sub-

ject and thus represented an error due to auditory perceptual confusion.










Of the 2,160 total responses produced by the six subjects

(120 stimulus words x three observations x six subjects), 13 responses

were suspected of being in error due to auditory perceptual confusion

or paraphasia. Though we can never really be sure, it seems a reason-

able inference that the responses were actually perceptual confusions

because of the acoustic similarity to the target word. The 13 respon-

ses eliminated from the word error rate count represented 13 different

words and are listed in Appendix K.


Intrajudge and Interjudge Reliability

Three subjects representative of the range of articulatory

abilities were selected and a representative sample of their tape-

recorded responses were retranscribed for intrajudge and interjudge

reliability. The first 40 stimulus words presented to each of the

three subjects during the initial test session were retranscribed.

The researcher transcribed the 40 stimulus words for each of

the three subjects totaling 120 words and compared the results to the

original transcription. A percentage of agreement was computed for

phoneme and word evaluation using the formula:

Agreements
Reliability Agreements
Agreements + differences


Intrajudge reliability was 91 percent for phoneme evaluation and

96 percent for word evaluation.

A trained listener with 15 years experience in speech pathology

transcribed the same 120 words and compared the results to the research-

er's transcription. A percentage of agreement was computed. Interjudge










reliability was 76 percent for phoneme evaluation and 85 percent for word

evaluation.

Of the 29 disagreements between the two judges, 15 were related

to disagreements of the presence or absence of distortion in a phoneme.

This fine distinction between when a production of a phoneme is within

normal limits and when that phoneme is perceived as being distorted

accounted for 52 percent of the disagreements in transcription between

the two judges. The more salient indicator of reliability for this

study is the intrajudge reliability measure. Since this study investi-

gated the consistency of a subject's responses over three observations,

it is important that the individual transcribing all of the data

demonstrate high intrajudge reliability. The levels of 91 percent for

phoneme and 96 percent for word judgments appear to indicate that for

the purpose of this study a high level of reliability was achieved.















CHAPTER III
RESULTS


The purpose of this study was to assess the consistency of

articulatory performance over three observations/sessions separated by

48 hours in the neurologically stable apraxic adult, and to evaluate the

relative difficulty of a weighted-word list designed from a set of

phonetic criteria. Six subjects were tested and the results were

examined and compared.


Word and Initiation Errors: Research Question 1

Are there significant differences among the group mean total

errors (word and initiation errors combined), mean word errors, and mean

initiation errors for observations one, two, and three?

The Barlett's Test and the Hartley Test (Neter and Wassermann,

1974) were conducted to determine if there existed a significant differ-

ence in variance for the individuals from the random sample. The

assumption of the analysis of variance is that the variance for the in-

dividuals is equal. The tests were negative for a significant differ-

ence in variance.

To determine if there were significant differences among the

group mean total errors (word and initiation errors combined) for obser-

vations one, two, and three, a two-way analysis of variance (ANOVA) was

performed. Total errors for individual subjects for the three observations










and group observation means are reported in Table 3. The group means

for observations one, two, and three are 41.00, 38.66, and 36.33, respec-

tively. The analysis of the data revealed that the subjects differed

significantly in their performance at the .05 level of confidence, but

the mean total errors for the group for observations one, two, and three

are not significantly different at the .05 level (F = 0.63; df = 5,2;

p > .05).

A two-way ANOVA was performed to determine if there were sig-

nificant differences among the group mean word errors and the group mean

initiation errors for observations one, two, and three. Word error rates

for individual subjects for the three observations and observation means

are reported in Table 4. Individual initiation error rates and group

means over the three observations are reported in Table 5.

The group means for word errors for observations one, two, and

three were 32.16, 30.66, and 28.50, respectively. The group means for

initiation errors for observations one, two, and three were 8.83, 8.00,

and 7.83, respectively. For both analyses, the subjects were signifi-

cantly different at the .05 level of confidence. Group mean word errors

for observations one, two, and three were not significantly different

at the .05 level of significance (F = 0.58; df = 5,2; P > .05); nor were

group mean initiation errors (F = 0.41; df = 5,2; p > .05).


Initial and Final Position Errors: Research Question 2

Are there significant differences among the group mean total

phoneme errors (initial and final position errors combined), mean ini-

tial position phoneme errors, and mean final position phoneme errors

for observations one, two, and three?













TABLE 3
GROUP AND INDIVIDUAL TOTAL ERRORS
(WORD AND INITIATION ERRORS COMBINED)
FOR THREE OBSERVATIONS FOR SIX SUBJECTS


Observations Individual

Subjects 1 2 3 Total Mean SD

1 16 16 12 44 14.66 2.30

2 30 28 25 83 27.66 2.51

3 50 31 37 118 39.33 9.71

4 39 35 44 118 39.33 4.50

5 60 52 43 155 51.66 8.50

6 51 70 57 178 59.33 9.71

Group Total 246 232 218 696 --

Group Mean 41.00 38.66 36.33 38.66 --


Note: SD = 7.21.
















TABLE 4
WORD ERRORS
FOR THREE OBSERVATIONS FOR SIX SUBJECTS


Observations

Subjects 1 2 3

1 5 7 5

2 20 17 14

3 46 29 32

4 32 31 36

5 50 44 39

6 40 56 45

Group Total 193 184 171

Group Mean 32,16 30.66 28.50


Note: SD = 5.94

















TABLE 5
INITIATION ERRORS
FOR THREE OBSERVATIONS FOR SIX SUBJECTS


Observations

Subjects 1 2 3

1 11 9 7

2 10 11 11

3 4 2 5

4 7 4 8

5 10 8 4

6 11 14 12

Group Total 53 48 47

Group Mean 8.83 8.00 7.83


Note: SD = 2.04










Three two-way ANOVAs were performed to determine if significant

differences existed among the group mean total phoneme errors (initial

and final positions combined), the group mean initial position phoneme

errors, and the group mean final position phoneme errors. Total phoneme

error rates, initial position phoneme error rates, and final position

error rates for individual subjects and group means for the three obser-

vations are reported in Tables 6, 7, and 8.

The group means for total phoneme errors for observations one,

two, and three were 34.00, 32.66, and 30.83, respectively. Group means

for initial position phoneme errors for observations one, two, and

three were 20.00, 20.00, and 19.50. Group means for total final position

phoneme errors for observations one, two, and three were 14.00, 12.66,

and 11.33. For all three analyses, a significant difference existed

among the subjects at the .05 level of confidence. Group means for total

phoneme errors for observations one, two, and three were not signifi-

cantly different at the .05 level of confidence (F = 0.30; df = 5,2;

p > .05); nor were group means for initial position phoneme errors

(F = 0.01; df = 5,2; p > .05). Group means for final position phoneme

errors for observations one, two, and three were not significantly dif-

ferent at the .05 level of confidence (F = 3.64; df = 5,2; p > .05).


Individual Phoneme Errors: Research Question 3

Are there significant differences among the group mean errors

for each of the 18 phonemes for observations one, two, and three?

Eighteen one-way ANOVAs were performed to determine if there

were significant differences among the group mean errors for each of
























TABLE 6
TOTAL PHONEME ERRORS
(INITIAL AND FINAL POSITION ERRORS COMBINED)
FOR THREE OBSERVATIONS FOR SIX SUBJECTS


Observations

Subjects 1 2 3

1 5 4 4

2 22 16 12

3 38 37 41

4 51 32 36

5 52 51 42

6 36 56 50

Group Total 204 196 185

Group Mean 34.00 32.66 30.83


Note: SD = 1.86


















TABLE 7
INITIAL POSITION PHONEME ERRORS
FOR THREE OBSERVATIONS FOR SIX SUBJECTS


Observations

Subjects 1 2 3

1 0 1 2

2 10 7 6

3 21 22 23

4 43 27 30

6 25 24 19

5 21 39 37

Group Total 120 120 117

Group Mean 20.00 20.00 19.50


Note: SD = 6.09



















TABLE 8
FINAL POSITION PHONEME ERRORS
FOR THREE OBSERVATIONS FOR SIX SUBJECTS


Observations

Subject 1 2 3

1 5 3 2

2 12 9 6

3 17 15 18

4 8 5 6

5 27 27 23

6 15 17 13

Group Total 84 76 68

Group Mean 14.00 12.66 11.33


Note: SD = 1,71









the 18 phonemes for observations one, two, and three. Total errors and

means for each phoneme for the three observations are reported in Table

9. An analysis of the data revealed thatthe means for each of the 18

phonemes for observations one, two, and three were not significantly

different at the .05 level of confidence.


Error Categories: Research Question 4

Are there significant differences among the group mean errors for

each of the six error categories for observations one, two, and three?

To determine if there were significant differences among the

group mean errors for each of the six error-categories for observations

one, two and three, six one-way ANOVAs were performed. Total errors and

means for each error-category for observations one, two and three are

reported in Table 10.

An analysis of the data revealed that

1. The mean substitution errors for observations one, two

and three were not significantly different at the .05

level of confidence (F = 0.02; df = 2, p > .05)

2. The mean distorted substitution errors for observations one,

two, and three were not significantly different at the .05

level of confidence (F = 0.91; df = 2; p > .05)

3. The mean distortion errors for observations one, two,

and three were not significantly different at the .05

level of confidence (F = 0.16; df = 2; p > .05)

4. The mean addition errors for observations one, two, and

three were not significantly different at the .05 level

of confidence (F = 0.61; df = 2; p > .05)











TABLE 9
ERRORS AND MEANS FOR EIGHTEEN PHONEMES FOR THREE OBSERVATIONS
FOR THE GROUP OF SIX SUBJECTS


Observation

Phonemes 1 2 3


Errors
Mean


17
2.83


Errors
Mean


9
1.50


Errors
Mean


Errors
Mean


Errors
Mean


Errors
Mean


Errors
Mean


14
2.33


11
1.83


18
3.00


6
1.00


20
3.33


Errors
Mean


12
2.00


11
1.83


7
1.16


17
2.83


11
1.83


17
2.83


8
1.33


16
2.66


12
2.00


4
0.66


12
2.00


9
1.50


14
2.33


11
1.83


20
3.33


9
1.50


11
1.83










TABLE 9--Continued


Observation

Phonemes 1 2 3


Errors
Mean



Errors
Mean


Errors
Mean



Errors
Mean


10
1.66



13
2.16


23
3.83



7
1.16


Errors
Means


Errors
Mean



Errors
Means


Errors
Mean



Errors
mean


4
0.66


6
1.00



12
2.00


15
2.50



7
1.16


11
1.83


17
2.83


25
4.16


7
1.16


9
1.50


12
2.00


18
3.00


25
4.16


5
0.83


4
0.66


4
0.66


11
1.83


3
0.50


12
2.00


6
1.00


9
1.50


7
1.16


6
1.00




59





TABLE 9--Continued


Observation

Phoneme 1 2 3



Errors 0 1 0
Mean 0.00 0.16 0.00













ERRORS AND MEANS


TABLE 10
FOR SIX ERROR-CATEGORIES FOR THREE OBSERVATIONS
FOR THE GROUP OF SIX SUBJECTS


Observations

Error Categories 1 2 3

Substitution:


Errors
Mean
SD = 13.22

Distorted Substitution:
Errors
Mean
SD = 1.86

Distortion:

Errors
Mean
SD = 3.50

Addition:

Errors
Means
SD = 2.90

Omission:

Errors
Means
SD = 0.45

Other:

Errors
Means
SD = 0.61


118
19.66



12
2.00


39
6.50


32
5.33


1
0.16


2
0.33


112
18.66


109
18.16


19
3.16


20
3.33


33
5.50


33
5.50


28
4.66


21
3.50


2
0.33


2
0.33


1
0.16


1
0.16










5. The mean omission errors for observations one, two,

and three were not significantly different at the .05

level of confidence (F = 0.28; df = 2; p > .05)

6. The mean other errors for observations one, two, and

three were not significantly different at the .05 level

of confidence (F = 0.15; df = 2; p > .05)


Word Error Fre uency: Research Question 5

Are there significant differences among the group mean number

of words produced in error one, two, and three times over three observa-

tions?

To determine if significant differences existed for this question,

a two-way ANOVA was performed. Individual subject data and the mean

number of words in error one, two, and three times over three observations

are reported in Table 11. Letters A, B, and C were used to designate

words in error one, two, and three times over three observations,

respectively. The mean for words in error once over three observations

was 22.50. The mean for words in error twice over three observations

was 15.66, and the mean for words in error three times over three obser-

vations was 12.50.

A significant difference existed among the subjects at the .05

level of confidence. There also was a significant difference among the

means for words produced in error one, two, and three times over three

observations at the .05 level of significance (F = 7.59; df = 5,2;

p < .05). A Duncan's Multiple Range Test revealed that the mean number


















Sn
4- S- (0
o >




CC

3*
*^3
^^ S
.?
4J U I
'0 E U
1-^=
| 5r
5^


LU
>-<
H-



- -
LU






OaL

















LUJ
0.
o
al a-



o")





I.
z'
T-




QL


o s-


S0 -0


-- S .- 4-'
a.



a Ln









3 >






S 0 S-
o :-g--a
L 0
o ca






O Eo-

















=












U)
.0
E Q1








(n
-V)'
-o


CM,. 0 00 O i-
r ) a0) 1 O0--
-- mn L







































i-- iO C -
aj CM CM


















ci *- n M
i- CM i o CO
CU~Tj


Uo
0.0


















mC)

cr"
- JCM
CM


CL



S.-
(D









of words produced in error once over three observations was significantly

higher than the mean number of words produced in error two or three times

over three observations at the .05 level of confidence.


Consistent and Inconsistent Word Errors:
Research Question 6

Are there significant differences between the group mean number

of consistent errors and the group mean number of inconsistent errors?

To evaluate the consistency or lack of consistency of error

patterns in words produced in error over three observations, the error

responses of the group of six subjects were examined. Each word produced

in error more than once was assigned a ratio. The first number of the

ratio represents the number of times a word was produced in error over

three observations. The second number of the ratio represents the number

of words produced in error, but produced in the same fashion. Words

assigned a ratio of 2:2 or 3:3 were assigned to the consistent error

group and represented words that were produced in error in the same way

either two or three times over three observations. Words assigned a

ratio of 2:0, 3:0, 3:2 were assigned to the inconsistent group and

represented words that were not consistently misarticulated in the same

way over the three observations.

Individual subject performance, total number of words assigned

to the consistent and inconsistent error groups, and group means for

the consistent and inconsistent error groups are reported in Table 12.

Group means for the consistent and inconsistent error groups were 3.66

and 6.77, respectively. To determine if there were significant differ-

ences between the means of the two groups, a two-sample t-test was

















TABLE 12
NUMBER OF CONSISTENT AND INCONSISTENT ERRORS PRODUCED
OVER THREE OBSERVATIONS FOR SIX SUBJECTS


Consistent Inconsistent
Error Groupa Error Groupsa

Subjects 2:2 3:3 2:0 3:0 3:2

1 1 0 0 1 1

2 5 1 7 1 2

3 6 2 15 6 6

4 2 9 8 4 7

5 5 4 21 9 1

6 8 1 17 5 11

Column Total 27 17 68 26 28

Group Total 44 122

Group Mean 3.66 6.77

aThe first number of the ratio represents the number of
times a word is produced incorrectly over three observa-
tions. The second number of the ratio represents the
number of times a word is produced in error in the same
way each error occurrence.










performed. The means for the consistent error group and the inconsistent

error group were not significantly different at the .05 level of confi-

dence.


Weighted Word~Errors:. Research Question 7

Are there significant differences among the group mean total

errors (word and initiation errors combined), mean word errors, and

mean initiation errors for word lists A, B, C, D, E, F, which increase

in phonemic complexity?

Three two-way ANOVAs were performed to determine if significant

differences existed among the group mean total errors (word and initia-

tion errors combined), the group mean word errors, and the group mean

initiation errors for word lists A, B, C, D, E, F.

Total error rates, word error rates, and initiation error rates

for individual subjects and group means for the six word lists are

reported in Tables 13, 14, and 15. For all three analyses, a significant

difference existed among the subjects at the .05 level of confidence.

The group means for total errors for the six word lists were

A = 26.33, B = 24.00, C = 24.16, D = 14.67; E = 12.00, F = 14.83. These

group means were significantly different at the .05 level of confidence

(F = 5.96; df = 5,5; p < .05).

A Duncan's Multiple Range Test revealed that the means for

word lists A, B, C were not significantly different, and the means for

word lists D, E, F were not significantly different. However, word lists

A, B, C had significantly higher mean errors than word lists D, E, F

at the .05 level of confidence.




















TABLE 13
TOTAL ERRORS (WORD AND INITIATION ERRORS COMBINED)
OVER THREE OBSERVATIONS FOR WORD LISTS
A, B, C, D, E, F FOR SIX SUBJECTS


Word Lists

Subjects A B C D E F

1 10 8 7 7 3 9

2 12 21 19 8 8 15

3 27 23 21 18 13 16

4 32 26 38 7 8 7

5 30 32 34 19 15 25

6 47 34 26 29 25 17

Group Total 158 144 145 88 72 89

Group Mean 26.33 24.00 24.16 14.66 12.00 14.83


Note: SD = 6.18















TABLE 14
WORD ERRORS FOR THREE OBSERVATIONS
FOR WORD LISTS A, B, C, D, E, F
FOR SIX SUBJECTS


Word Lists

Subjects A B C D E F

1 2 3 1 5 1 5

2 6 13 14 4 4 10

3 25 22 18 18 11 13

4 27 20 33 5 8 6

5 26 24 29 16 14 24

6 39 29 23 19 17 14

Group Total 125 111 118 67 55 77

Group Mean 20.83 18.50 19.66 11.16 9.16 12.00


Note: SD = 5.68

















TABLE 15
INITIATION ERRORS FOR THREE OBSERVATIONS
FOR WORD LISTS A, D, C, D, E, F
FOR SIX SUBJECTS


Word Lists

Subjects A B C D E F

1 8 5 5 3 2 4

2 6 8 5 4 4 5

3 2 1 3 0 2 3

4 5 6 5 2 0 1

5 4 8 5 3 1 1

6 8 5 3 10 8 3

Group Total 33 33 26 22 17 17

Group Mean 5.50 5.50 4.33 3.66 2.93 2.83


Note: SD = 2.04










The group means for word errors for the six word lists were

A = 20.83, B = 18.50, C = 19.66, D = 11.16, E = 9.16, F = 12.00. These

group means were significantly different at the .05 level of confidence

(F = 4.60; df = 5,5; p< .05). A Duncan's Multiple Range Test revealed

that word list A has a significantly higher mean than word lists D, E,

F. Word list C has a significantly higher mean than word lists D, E,

F. Word list B has a significantly higher mean than word list D and E,

but not significantly higher than F.

The group means for initiation errors for the six word lists

were A = 5.50, B = 5.50, C = 4.33, D = 3.66, E = 2.83, F = 2.83. These

group means were not significantly different at the .05 level of confi-

dence (F = 2.12; df = 5,5; p > .05).















CHAPTER IV
DISCUSSION


A comparison of articulatory performance over three observations

separated by 48 hours for a group of six neurologically stable apraxic

adults revealed a remarkably consistent pattern of articulatory per-

formance. No significant differences existed over the three observa-

tions for

1. Mean number of total errors (word and initiation errors

combined)

2. Mean number of word errors

3. Mean number of initiation errors

4. Mean number of total phoneme errors (initial and final

position errors combined)

5. Mean number of initial position phoneme errors

6. Mean number of final position phoneme errors

7. Mean number of errors for each of the six error-

categories

8. Mean number of errors for each of the 18 phonemes


Consistency of Quantitative Performance

When variability is viewed from a quantitative perspective

(number of errors produced), conclusions regarding consistency of

articulatory performance in this group of apraxic subjects become ap-

parent. Quantitatively, the articulatory performance of this group of










subjects appears predictable from observation to observation, suggesting

a certain lawfulness to their articulatory behavior and predictability

from observation to observation.

The apparent predictability of the apraxic subject's articula-

tory behavior suggests that one can be more confident that the articu-

latory error rate observed during a particular data-gathering session

in fact accurately represents a subject's articulatory performance, at

least quantitatively, as it will exist the next day or some time in the

near future. It appears that data gathered during a single test session

arerepresentative of a subject's overall articulatory performance.

The indication that quantitative articulatory performance is

predictable and relatively consistent significantly aids the interpre-

tation of data collected during therapeutic management. One can be more

confident that quantitative changes observed during the course of

therapy are probably attributable to progress in therapy rather than to

changes reflecting variability or inconsistency of articulatory perfor-

mance from day to day.


Consistency of Qualitative Performance

The fact that one can predict an apraxic subject's quantitative

articulatory performance does not suggest that it is possible to predict

with the same confidence the precise nature or quality of an apraxic

subject's qualitative articulatory performance. It does not appear

that one can predict with the same confidence (1) the specific phonemes

or words that will be produced in error on successive observations

(consistency of error frequency), or (2) whether or not phonemes or










words will be produced in error the same way each error occurrence

on successive observations (consistency of error pattern).


Consistency of Error Frequency

To determine if words produced in error during the first obser-

vation were produced in error on subsequent observations, mean error

rates for words produced inaccurately one, two, and three times over

three observations were compared. Analysis of the data revealed that

the mean rates for words produced in error one time over three observa-

tions was significantly higher than the mean rates for words produced

in error two or three times over three observations. Of the 304 words

produced in error one or more times over three observations by the

group, 44 percent were misarticulated only once over three observations,

31 percent were misarticulated twice, and 25 percent were misarticulated

all three times. Since only 25 percent of the words were produced in

error each occurrence, it suggests that this group of apraxic subjects

did not consistently misarticulate the same words each occurrence over

three observations.

Deal (1974) investigated consistency of word errors on five

successive readings of a 100-word paragraph in subjects demonstrating

apraxia of speech. Deal concluded that as a group his subjects made

consistent word errors. The mean consistency scores of his subjects

all were higher than 60 percent.

The disparity between the present study and that of Deal on

consistency of word errors appears partially definitional and partially

attributable to differences in subjects and experimental design. Deal's










subject selection criteria were more stringent and in fact only two of

the six subjects in the present study would have been eligible for

inclusion in Deal's study. Perhaps subjects functioning at a higher

level of articulatory proficiency, as those in Deal's study, demonstrate

a higher degree of consistency.

The experimental designs differed with respect to stimulus

material, stimulus presentation mode, response condition, and the time

interval between observations. Deal's study required the subjects to

read a 100-word paragraph five times in succession with an interval of

five seconds between readings. This study presented a 120-item word

list three times with a 48-hour interval between presentations. Perhaps

consistency of word errors is more evident on successive presentations

of stimulus material than when a substantial period of time elapses

between stimulus presentations.

The lack of agreement between the two studies relative to the

consistency of word errors also might be definitional. The present

study based the judgement of consistency of word error on a particular

word being produced in error each occurrence. Twenty-five percent of

the words produced in error in this study were produced in error three

times over the three observations. Deal, in comparing consistency of

word errors on successive readings of a paragraph, based his definition

of consistency on the total number of errors made from reading to read-

ing. For example, a mean word error consistency of 60 percent meant that

60 percent of the same words were produced in error on two successive

readings of a paragraph. This means that a particular word was not neces-

sarily produced in error each time it occurred in the paragraph.










LaPointe (1969) investigated the consistency of misarticulations

on successive occurrences of phonemes on a modified version of a standard

articulation test. Although differences existed in experimental design

and data analysis, the operational definition of consistency used by

LaPointe was similar to the one used in the present study. The judgment

of consistency was based on a phoneme being in error each time it

occurred.

No subject demonstrated an entirely consistent error pattern

in which all 25 sounds were misarticulated each occurrence. Thirteen

subjects demonstrated partial consistency in that some of the 25 sounds

were misarticulated each occurrence. Twenty-eight percent of the

sounds produced by the 13 subjects who demonstrated partial consistency

were consistently misarticulated each occurrence, a figure that is com-

parable to the 25 percent consistency of word error found in this study.

Hatfield and Walton (1975) analyzed phonemic substitutions on

a syllable repetition task of a subject with severeBroca's aphasia to

determine if substitution errors were systematic enough to guide the

design of remedial programs. These authors were able to discern a

pattern to the subject's phonological behavior which they described in

terms of "rules." Although the patient's phonological behavior appeared

to be rule governed, these researchers noted that the subject's per-

formance was not so strict as to allow prediction of every occurrence

of a misarticulated phoneme.

Conclusions regarding the predictability of word or sound

errors may be dependent upon the operational definition of the term










"consistency." If consistency is viewed as words or sounds being pro-

duced in error each occurrence, then the subjects in this study did

not make consistent word errors.


Consistency of Error Pattern

To evaluate the consistency of error patterns in words produced

in error two or more times over three observations, the error responses

of the subjects were examined. A ratio was assigned to each error word

and based on the ratio, error words were assigned to either the consis-

tent error word group or the inconsistent error word group (Table 10).

The consistent error word group (ratios 2:2, 3:3) was comprised of

words produced in error in the same fashion on each occurrence. The

inconsistent error word group (ratios 2:0, 3:0, 3:2) was comprised of

words produced in error in a different fashion each occurrence or pro-

duced in a different fashion two out of the three occurrences.

Analysis of the data revealed that no significant differences

existed between the means of the consistent and inconsistent error word

groups. In LaPointe's study (1969), no subject consistently misarticu-

lated all error sounds in the same way each error occurrence. Sixteen

of the 28 subjects demonstrated partial consistency in that a subject

misarticulated some words the same way each time they occurred but mis-

articulated other sounds in a variety of ways.

It appears that apraxic subjects do not misarticulate sounds

or words in the same way each time they are produced incorrectly. Some

sounds or words are misarticulated in the same way each error occurrence

whereas other sounds or words are misarticulated in a variety of ways.










Is the articulatory behavior of the neurologically stable

apraxic adult best described as consistent or inconsistent? Both

descriptions are accurate depending upon the level of observation and

the operational definition of "consistency."

If the apraxic subject's articulatory performance is viewed

from a quantitative perspective (the number of misarticulations), then

it would be described as consistent. The mean number of word errors,

initiation errors, phoneme errors, and the mean number of errors for

each of the six error-categories werenot significantly different from

observation to observation suggesting predictability and consistency of

the articulatory error rate pattern for this group of apraxic subjects.

If one is referring to whether or not these subjects misarticu-

lated the same words in error each occurrence or misarticulated the

error words in the same way each error occurrence, then the apraxic

subject's articulatory performance would be described as inconsistent.

The fact that mean error rates at the word, phoneme and error

category level are consistent over successive observations is not in-

compatible with the finding that the same words are not consistently

produced in error each occurrence nor are misarticulated words produced

in error in the same fashion each error occurrence by this group of

apraxic subjects. It appears that the group's articulatory performance

when analyzed from a quantitative perspective is predictable and con-

sistent but when analyzed from a qualitative perspective, the pattern

of consistency and predictability of articulatory performance is less

evident.










Performance on Weiahted Words

The development of a weighted-word list based on a set of

phonetic criteria and the evaluation of the relative difficulty of this

word listwere attempted in an effort to develop a meaningful arrangement

of stimulus words that might be used in therapy. Significant differ-

ences existed among the group mean errors (word and initiation errors

combined) for the word lists. The means for word lists A, B, and C were

not significantly different and the means for the word lists D, E, and

F were not significantly different. However, word lists A, B, and C

had significantly higher mean errors than word lists D, E, and F, indi-

cating that as a group word lists A, B, and C were more difficult for.

this group of apraxic subjects than word lists D, E, and F. The feature

that distinguished word lists A, B, and C from word lists D, E, and F is

the presence of one or more fricatives and/or affricates in word lists

A, B, and C. Previous studies have documented that fricatives and

affricates are the most difficult sounds for the apraxic to produce

accurately.

The group means for initiation errors for the six word lists

were not significantly different but the group means for word errors

for the six word lists were significantly different. Word lists A and

C had significantly higher means than word lists D, E, and F, and word

list B had a significantly higher mean than D and E, but not F.

The initial impression when viewing the mean total errors (word

and initiation errors combined) is that the only potent discriminator

between the six weighted word lists is the presence of one or more










fricatives and/or affricates. Analysis of the results of word errors

alone reveals that this conclusion may not hold. If the presence of

a fricative or affricate is the only important variable distinguishing

the difficulty of the word lists, then it becomes difficult to explain

why the word error means for lists B and F are not significantly differ-

ent.

It appears that some or all of the remaining four variables

(phoneme, manner, place, voicing shift) used to weight the stimulus

words for relative difficulty may have some influence on articulatory

accuracy but perhaps not to the degree anticipated. For example,

it appears that over twice as many errors (26 percent) were made on CVC

stimulus words with a weighted value of 3 in which a different phoneme,

manner of articulation and place of articulation were utilized for the

initial and final consonants. In contrast, words weighted 3 and com-

prised of a different phoneme, place of articulation and the presence

of a voicing shift were in error half as many times (12 percent).

Similarly, words weighted 3 and comprised of a different phoneme,

manner and the presence of a voicing shift also were in error half as

many times (11 percent).

One might speculate that the interaction of the three variables

of different phoneme, manner and place for the initial and final

consonants is more influential in determining articulatory accuracy

than the interaction of phoneme, place and voicing shift or the inter-

action of phoneme, manner, and voicing shift. Error patterns within a

weighted word group, as the one described above, suggest that certain










combinations or interactions of the variables selected for weighting

CVC stimulus words are more influential in determining articulatory

accuracy than others. It would be beneficial to develop a list of

stimulus words that would evaluate independently the influence of each

of the variables of phoneme, manner, place and voicing shift on articu-

latory accuracy. Such a strategy would serve to identify the relative

influence on an articulatory accuracy of each of the phonetic variables

used in the present study.

The influence of word weight might interact differently with a

different sample of subjects, as well. This study examined the effect

of word weight on articulatory performance using a broad range of

subjects. Perhaps a large homogeneous sample of mildly involved or a

sample of moderately to severely involved subjects would perform charac-

teristically and differently on selected word lists. Also, the influence

of therapy might affect performance on the weighted words lists. Thera-

peutically naive subjects who have not been influenced by compensatory

suggestions may perform differently.

The results of this study suggest the following conclusions:

1. Quantitatively (the number of errors) the articulatory

performance of this group of apraxic subjects is pre-

dictable from observation to observation at the phoneme,

word, initiation, and error-category levels suggesting

a certain lawfulness to the apraxic's articulatory

behavior.

2. One can more confidently trust that a single measurement

of articulatory performance for this group of stable










apraxic adults is representative of the subjects'

articulatory performance from day to day.

3. Qualitatively, the pattern of consistency and predic-

tability of articulatory performance is less evident.

4. It does not appear possible to predict with accuracy for

this group of subjects which specific words will be in

error on successive observations.

5. Words were not consistently produced in error in the

same fashion each occurrence by this group of subjects.

Some words were misarticulated in the same way each

error occurrence whereas other words were misarticulated

in a variety of ways.

6. The set of phonetic criteria selected for weighting

CVC stimulus words did not successfully predict the

relative difficulty of the six word lists weighted for

phonemic complexity beyond indicating that words con-

taining fricatives and or affricates were more diffi-

cult to produce.

Although these apraxic subjects demonstrated consistency of ar-

ticulatory performance from a quantitative perspective, variability of

immediate productions of target sounds or words remains a hallmark

characteristic of apraxia of speech and distinguishes it from other

phonological disorders.


Suggestions for Further Research

One purpose of this study was to investigate consistency of

articulatory performance in the neurologically stable subject. The










subjects in this study ranged from 15 to 150 months post-onset with a

mean of 44.6. A longitudinal study initiated early in the recovery

process for the purpose of measuring the range of variability of articu-

latory performance during this unstable period of spontaneous remission

might prove valuable. By systematically sampling the range of varia-

bility during the recovery process a pattern of performance of predictive

value might emerge. Perhaps the spontaneous recovery process for the

apraxic subject could be described in terms of changes in the range of

variability of articulatory performance. The ability to predict the

termination of the spontaneous recovery period based on predictable

changes in the range of variability of articulatory performance would be

of significant therapeutic importance.

The impairment of apraxia of speech can be viewed as a continuum

of severity with the mute subject on one end of the continuum and the

imperceptibly impaired subject on the opposite end of the continuum.

Although the subjects in this study were different in their range of

performance, the full range of the severity continuum was not sampled.

A fuller appreciation of consistency of articulatory performance rela-

tive to severity levels of the impairment would be valuable. Perhaps

the consistency of performance demonstrated by the subjects in this

study is not representative of the consistency of articulatory perfor-

mance or lack of consistency of articulatory performance at the more

extreme ends of the performance continuum. A study similar in purpose

to the present study designed to sample the full range of the severity

continuum would serve to document the consistency of articulatory per-

formance relative to severity levels.










The second purpose of this study was to evaluate the relative

difficulty of a weighted-word list designed from a set of phonetic

criteria. The results of this study demonstrates the need for continued

evaluation and refinement of the weighted-word technique for the purpose

of hierarchically organizing stimulus words. A fruitful approach might

be to develop lists of stimulus words that would evaluate independently

or separately the influence of the selected variables (phoneme, manner,

place, voicing shift) on articulatory accuracy and to determine if

combinations of these variables interact to influence articulatory

accuracy.

Efforts to delineate the speech characteristics of apraxia of

speech and literal paraphasia have not been conclusive in differentiating

these two phonological impairments. Perhaps a fruitful area for future

research would be the comparison of the consistency of articulatory

performance both quantitatively and qualitatively between these two

pathological groups. Such a comparison might aid in determining whether

or not these two phonoloigcal impairments are in fact distinct and

distinguishable.


Summary

This study compared the articulatory performance over three

observations separated by 48 hours of a group of six neurologically

stable apraxic adults. In addition, the study evaluated the relative

difficulty of six weighted-word lists developed from a set of phonetic

criteria for the purpose of hierarchically arranging CVC stimulus words

for use in therapeutic management. The results of the study suggest

the following conclusions:










1. Quantitatively (the number of errors) the articulatory

performance of this group of apraxic subjects is pre-

dictable from observation to observation at the phoneme,

word, initiation, and error-category levels suggesting

a certain lawfulness to the apraxic's articulatory

behavior.

2. One can more confidently trust that a single measure-

ment of articulatory performance for this group of

apraxic adults is representative of the subjects'

articulatory performance from day to day.

3. Qualitatively, the pattern of consistency and predicta-

bility of articulatory performance is less evident.

4. It does not appear possible to predict with accuracy

for this group of subjects which specific words will be

in error on successive observations.

5. Words were not consistently produced in error in the

same fashion each occurrence by this group of subjects.

Some words were misarticulated in the same way each

error occurrence whereas other words were misarticulated

in a variety of ways.

6. The set of phonetic criteria selected for weighting CVC

stimulus words did not successfully predict the relative

difficulty of the six word lists weighted for phonemic

complexity beyond indicating that words containing

fricatives and/or affricates were more difficult to

produce.


































APPENDICES















APPENDIX A
BIOGRAPHICAL AND MEDICAL INFORMATION SHEET


Name

Birthdate

Social Security Number

Address

Relatives

Native Language

Occupation


Race



Referral Source

Phone

Therapist

Education

Physical Status


Medical History

Etiology and Site of Lesion

Onset of Disorder Months Post-Onset

Previous CNS Involvement Previous Sp/Lang.Involvement

Additional Medical Information


Language Evaluation

Initial Language Evaluation Date

Test Scores:


Description of Speech/Language Impairment:



Enrolled in Therapy Dates

Therapy Goals:


~















APPENDIX B
SUBJECT SELECTION CRITERIA SUMMARY SHEET


PASSED CRITERIA


Name

Months Post-Onset (6 months)

Etiology (Cerebral Vascular Accident)

History of Dementia

Native Speaker of English

Normal Hearing Acuity (passed screening test)

Overall PICA Score of Between 45%-90%_

Evidence of Phonemic Variability (10%)
(5 polysyllabic words)

Evidence of Initiation Errors (10%)
(50-word spontaneous speech sample)

Predominance of Substitution and Addition Errors
over Omission and Distortion Errors
(50-word spontaneous speech sample)


Yes No

Yes No

Yes No

Yes No

Yes No

Yes No


Yes No


Yes No



Yes No


Percentage of Substitution Errors___
Combined
Percentage of Addition Errors

Percentage of Omission Errors
Combined
Percentage of Distortion Errors

Error Rate on 30 Monosyllabic Words










Check When Completed:

Pica

Oral Peripheral

Polysyllabic Word List

Monosyllabic Word List


Spontaneous Speech Sample

Hearing Screening

Biographical and Medical
Information

Signed Consent Form


Subject Passed Selection Criteria


Yes No














APPENDIX C
PORCH INDEX OF COMMUNICATIVE ABILITIES
INDIVIDUAL SCORES
PERCENTILESS)


Subject Overall Score Gestural Score Verbal Score Graphic Score

RH 79.0 78.5 79.0 78.5

CF 82.5 74.0 79.5 86.0

BT 49.0 39.0 49.5 60.0

RS 63.0 62.5 71.0 60.5

FB 71.0 92.0 52.5 72.0

FE 54.0 74.0 49.5 48.0


Range 82.5-49.0 92.0-39.0 79.5-49.5 86.0-48.0

Mean 66.3 70.0 63.5 67.5













APPENDIX D
CONSENT FORM


Apraxia is an articulation problem resulting from brain injury. A
person with apraxia of speech has difficulty saying words correctly.
Not all is known about the most effective methods of therapy for apraxia
of speech. A list of 120 words that might be used in therapy has been
selected. This study will determine which of the 120 words are most
difficult for a person with apraxia of speech to say correctly. The
types of errors a person makes when trying to say the list of 120 words
will be examined. Also the study will determine if the person makes the
same types of errors each time the person says the words.

The list of 120 words has been recorded on the tape recorder. You
will hear the words through earphones. You will repeat each word after
you hear it. Your responses will be recorded on a tape recorder. You
will be asked to say the 120 words three times. A period of 48 hours will
separate each of your three attempts at saying the words.

Your continued participation in this study is voluntary. This
study is done in cooperation with the V.A. Hospital.



I understand the nature of this study and consent to participating
in this study. I consent to the use of the results of this study in
future publications.






Date















APPENDIX E
SPEECH SAMPLE TASKS


SPONTANEOUS SPEECH TASK


Questi


)ns used to elicit a spontaneous speech sample of at least 50 words.

1. Tell me about your family. How many members are there in
your family?

2. Tell me about your speech difficulties.

3. Tell me what happened when you had your stroke. How did
you feel?

4. Describe your occupation, the type of work you did before you
became ill.

5. Tell me about your hobbies. What do you enjoy doing most?

6. Where do you live? Describe the type of town/city it is.

7. Describe an ideal or favorite vacation.


REPEATED TRIALS TASK
Five polysyllabic words each repeated 5 times consecutively for the pur-
pose of measuring phonemic variability.

Administration

Refrigerator

Impossibility

Tornado

Condominium









PREDICTIVE ERROR RATE TASK
Repetition of 30 monosyllabic words (representative of the stimulus words
utilized in the experimental task) for the purpose of predicting a sub-
ject's ability to complete the experimental task. Five words were
selected from each of the six weighted word categories.

Word Weight Word Word Weight Word

6 Tough 4b Comb

6 Cough 4b Nap

6 Soap 4b Meek

6 Choke 4b Peck

6 Sash 4b Cope

5 Zip 3 Mop

5 Sag 3 Bang
5 Bus 3 Nag

5 Vase 3 Code

5 Soothe 3 Gap

4a Buzz 2 Pep

4a Beige 2 Gong

4a Zoom 2 Keg

4a Jug 2 Pub

4a Sin 2 Mob















APPENDIX F
DIRECTIONS FOR THE EXPERIMENTAL TASK


Oral instructions read to each subject prior to presentation of the

120 stimulus words:

"Mr. I want you to listen to words

I have recorded on the tape recorder. You will hear the

words through these earphones. (Show earphones.) I want

you to repeat each word that you hear. There will be a

short time period between each word so listen carefully

so that you will hear the word. If you do not understand

a word, raise your hand and I will replay the word. Do

not raise your hand if you are having trouble saying the

word only if you did not hear the word you are to say.

Let's practice a few words so that you will be familiar

with how the words are presented and so that we can find

a comfortable listening level for you."

"We are finished practicing. Do you have any ques-

tions? lam now starting the tape. Ready!"
































APPENDIX G
WORD ERROR RECORD SHEET


















S++++4- 4-

* C -4+-+ 4+4. 1 4

+
4-







=E









o + 4. + 4 I +
e-' ",- o~.- -+ 4-+ --








.- 4I + o + +.










a -a I*- ri i as 4.
E 0 P Vy 45 C
I/"














C - v 4-+ s + 4- +
oo










4-- ++t tI I 0 + 3- +











4- + --+ + 4 I +

S-- 1 I I W -








E: In ..0 P 0 i
o u


S -4 N l + ++ 1 0














L n I r C I 0
0- - - =t 1 -











-; -- C 4'o N N N N C -
LO




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.8 - mvs