Citation
Effects of experimental palatal Fistulas on speech and resonance

Material Information

Title:
Effects of experimental palatal Fistulas on speech and resonance
Creator:
Richtner, Ulla E. M., 1941-
Copyright Date:
1986
Language:
English

Subjects

Subjects / Keywords:
Air pressure ( jstor )
Cleft palate ( jstor )
Fistulas ( jstor )
Hard palate ( jstor )
Palatal consonants ( jstor )
Palate ( jstor )
Pressure ( jstor )
Soft palate ( jstor )
Spoken communication ( jstor )
Vowels ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
AEK2515 ( LTUF )
29789590 ( ALEPH )
15300479 ( OCLC )

Downloads

This item has the following downloads:


Full Text














EFFECTS OF EXPERIMENTAL PALATAL FISTULAS
ON SPEECH AND RESONANCE By


ULLA E.M. RICHTNER


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










UNIVERSITY OF FLORIDA


1986





Copyright 1986

by

Ulla E.M. Richtner





To Maj and Nils





ACKNOWLEDGMENTS


My appreciation and gratitude are expressed to my

committee chairman, Dr. William N. Williams, for suggesting the topic and for kindly making available to me needed laboratory facilities and the resources of a Biomedical Research Support grant.

Gratitude is also extended to the other committee

members, all of whom showed trust in me and the project. Dr. Thomas B. Abbott also has been my adviser and his encouragement and assistance have helped me overcome many hurdles, not only in my doctoral research but throughout my graduate career. Dr. Kenneth R. Bzoch has shared generously his vast knowledge in the area of cleft palate. Dr. G. Paul Moore has provided me with invaluable clinical insights in the area of voice disorders. Dr. Nikzad S. Javid offered his professional skill and cooperation in construction of the research prosthesis. Special thanks go to Dr. Paul W. Wharton who volunteered his time and effort and without whom this work could not have been completed.

I owe special gratitude to Dr. Ronald G. Thomas for

his generous and invaluable help in guiding me through the





labyrinth of statistics and computer use. Special thanks also are due the six judges who donated their time and expertise for the perceptual listening task. I extend especially warm thanks to Marcia Buchanan not only for her editorial assistance but for her personal interest and concern. I would also like to take the opportunity to thank everyone from the Departments of Speech and Oral Biology who shared their professional skills, positive attitudes, and research equipment.

Rotary International and khl6n-stiftelsen of Sweden contributed to the completion of my graduate studies by their generous stipend contribution.
















TABLE OF CONTENTS


Page

ACKNOWLEDGMENTS. iv

ABSTRACT. viii

CHAPTER

I INTRODUCTION AND PURPOSE. 1

Introduction. 1 Surgical Treatment of Oronasal Openings. 5 Prosthetic Treatment of Oronasal Openings . 9
Cleft Palate and Speech. 10 Velopharyngeal Insufficiency and Speech. 10 Palatal Fistulas and Speech. 13 Statement of Purpose. 16

II LITERATURE REVIEW AND RESEARCH QUESTIONS. 20

Introduction. 20 Palatal Embryology and Anatomy. 21
The Effects of Velopharyngeal Insufficiency
on Speech. 31
Perceptual Judgment. 34 Instrumentation.39
The Effects of Controlled Velopharyngeal
Insufficiency on Speech. 51
The Effects of Palatal Fistulas on Speech .58 Research Questions. 64

III METHODS. 68

Introduction. 68 Materials and Procedures. 69

IV RESULTS. 88

Perceptual Measurements. 89 Instrumental Measurements. 116 Reliability. 133










Page

138 139 151


V DISCUSSION AND CONCLUSIONS .

Discussion .
Conclusions . APPENDIX

A AUDIOMETRIC EVALUATION .

B ORDER OF THE EXPERIMENTAL TEST CONDITIONS .

C CRANIOFACIAL CENTER SPEECH TESTS .

D ARTICULATION RATING TEST .

E HYPERNASALITY RATING TEST .

F SEQUENCE OF THE THREE TEST TAPES PRESENTED FOR
PERCEPTUAL JUDGMENT OF THE 95 TESTS (FIVE
SPEECH TESTS AT 19 FISTULA CONDITIONS) .

G INSTRUCTIONS FOR THE PERCEPTUAL JUDGMENT
PERFORMANCE . BIBLIOGRAPHY . BIOGRAPHICAL SKETCH .


155 156 157 159 160



161 162

164 174


vii











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

EFFECT OF EXPERIMENTAL PALATAL FISTULAS ON SPEECH AND RESONANCE

by

Ulla E.M. Richtner

August, 1986

Chairman: William N. Williams Major Department: Speech

Palatal fistulas are frequently the result of congenital anomalies, trauma, or tissue breakdown following cleft palate repair. It has been observed clinically that speech articulation and voice resonance are impaired as the result of an opening between the oral-nasal cavities. Extensive documentation of velopharyngeal insufficiency and its influence on speech suggest that a 10 mm 2to 20 mm opening through the velar port results in significant distortion of speech. However, little quantitative information is available concerning similar effects of palatal fistulas.

The purpose of this study was experimentally to manipulate size and location of openings through a subject's prosthetic palatal prosthesis and to measure the effect of these openings on the individual's speech articulation, resonance quality, nasal airflow, intraoral air pressure, and differential pressure. Four experimental openings (fistulas) of the respective sizes 5 im2, 10 mm20 20 mm2


viii





and 30 mm 2 were drilled in a removable palatal appliance covering a subject's complete cleft of the hard and the soft palates. The four locations of these openings were in the anterior, middle, and posterior part of the hard palate and one velar opening, all along the midline. The effect that each hole size and location had on the individual's speech and resonance was evaluated perceptually by six judges. Nasal airflow, intraoral air pressure, and differential pressure were measured instrumentally.

Analyses of the data revealed that both hole size and

location of a fistula determine the degree of speech impairment. The findings indicate that the greatest distortion of articulation and resonance quality was perceived by the judges when the fistula was located in the velar port and at the most anterior position of the hard palate (approximately at the incisive foramen). The size of a fistula was also found to be a significant variable with greater distortions occurring with increased openings. Even the smallest fistula of 5 mm 2 caused mild distortion to the speech signal which was statistically significant in the three instrumental measurements. The five perceptual tests indicated that speech, in terms of articulation proficiency and nasality, was significantly distorted when the fistula size was between 10 mm 2 and 20 MM2 .





CHAPTER I
INTRODUCTION AND PURPOSE



Introduction


The production of oral speech sounds is dependent upon a complete separation between the nasal and oral cavities, thus directing the vocal stream from the larynx through the oral cavity alone. This separation between the oral and nasal cavities is primarily anatomical. The structure separating the two is the palate. It consists of bone, muscle, and mucosal tissue which forms the roof of the mouth and the floor of the nasal cavity. Complete closure depends on the action of the soft palate or velum. The soft palate is elevated through muscular action to make complete contact with the posterior pharyngeal wall. When the soft palate is in relaxed position, as during nasal breathing or production of nasal sounds, air is directed into the nasal cavity. Thus, an opening in this separation, either in the stationary hard palate or in the movable soft palate, will be detrimental to normal oral speech sound production, since it will allow undesirable airflow into the nasal cavity. This airflow results in a reduction of intraoral air pressure which is an important physiological component for the production of different speech sounds









(Brown & McGlone, 1969). Leakage of air into the nasal cavity also causes audible, inappropriate nasal air emission during articulation of consonant sounds. In addition, a shift in resonance quality, due to acoustic damping related to the coupling of the nasal cavity with the oropharyngeal cavities, may occur which is perceived as hypernasality. Clinical observation has revealed that the magnitude of the effect of these two factors (hypernasality and nasal air emission) on speech intelligibility is negatively influenced by the increasing size of a palatal opening. This means, then, that the larger the palatal opening, the greater the degree of speech impairment. Articulation problems likely to occur concomitant with palatal openings are misarticulations, ranging from indistinct production due to slight air emission to complete omission of the sound. Distorted sound production, simple sound substitution, gross sound substitution, and complete omission of the sound are influenced by nasal air emission and hypernasality (Bzoch, 1979).

There are three conditions that could result in a physical breakdown of the palatal structure. The first condition is a cleft occurring in the hard and/or the soft palate. The type of cleft is dependent upon the location and size of the cleft itself. The most common cause of cleft palate is when atypical embryological development prevents normal fusion along the midline of the palate. Congenital overt clefts of the lip and/or palate occur at









an incidence as frequently as one in every 700 live births (Bzoch & Williams, 1979). Two other causes known to cause cleft palate would be trauma or ablative surgery. Head trauma, or accidents that cause orofacial injuries, such as fractures of skull bones, may include fractures of the palatal bones, or a cleft along the palatal midline. Ablative surgery, or removal of tissue, is usually due to disease typically following removal of tumors.

The second condition of palatal malfunction is velopharyngeal insufficiency (VPI) which is defined as the inability of the soft palate, or velum, to make adequate closure with the posterior pharyngeal wall for normal speech production. Velopharyngeal insufficiency can be a result of congenital or acquired anomaly and the speech disorders that typically occur can be compared to those impairments associated with cleft palate and palatal fistulas. Velopharyngeal insufficiency can also result from neurological disorders. It is important to remember the physiological differences between the hard and the soft palates. The function of velopharynx involves intricate movements and interactions of a number of muscles active in velopharyngeal closure. The hard palate, on the other hand, does not display any myofunctional movement since it consists mainly of bone and tissue lining and serves as a static border between the two cavities. It is difficult to determine the actual incidence of VPI, as it varies with the concomitant disorder. The incidence of VPI, as it occurs









after primary cleft palate repair, is reported to be between 20 percent and 40 percent (Bradley, 1979).

The third condition responsible for a palatal breakdown is a fistula in the hard and/or the soft palate. An oronasal fistula is defined, in this study, as a hole in the palate connecting the oronasal cavities. It may develop subsequent to surgical repair of a cleft palate. It may also occur from traumatic perforation of ablative surgery. Fistulas also may result from pathologic abnormalities such as noma (gangrenious sore), syphillitic gummas, leprosy, and leishmaniasis (infection due to the parasite Leishmania) (Gordon & Brown, 1980).

It is clear that these three different palatal openings or insufficiencies are almost exclusively caused by a morphological abnormality. Only VPI can be diagnosed as resulting from a functional disorder. Patients who suffer from any of these abnormalities are referred for either surgical closure of the palatal openings, reconstruction of muscular tissue to restore velopharyngeal function, or for prosthetic treatment. When the differential diagnosis signals a function disorder, speech therapy should be the treatment of choice. Sometimes a combination of treatments is needed, depending on the extent of the physical abnormality. Therefore, a need exists for evaluation of the impact of palatal disorders on speech intelligibility, as well as for the quantification of the effect of these palatal insufficiencies.





-5-


Surgical Treatment of Oronasal Openings Cleft Palate


A variety of surgical procedures is available for

palatal closure and correction of VPI. An excellent indepth description of several techniques in cleft palate surgery is given by Millard (1980), and in several chapters of Cleft Lip and Palate (Grabb, Rosenstein, & Bzoch, 1971). Selection of the appropriate surgical technique is dependent upon the pathological extent of the condition, individual need, and optimal goal. Important information regarding morphology, tissue reaction, and optimal timing for surgery are taken into consideration. There are several reasons why obtaining an adequate closure between the oronasal cavities is important. The primary goal in cleft palate surgery, and perhaps the task that is most difficult to achieve, is to establish normal speech production. As a direct result of atypical airflow into the nasal cavity, the necessary intraoral air pressure cannot be maintained and speech is affected in the form of hypernasality and nasal air emission. Normal hearing is important for the individual who is learning how to speak. Hearing loss that is caused primarily by otitis media (middle ear inflammation) occurs in almost all cleft palate individuals (Millard, 1980). The frequency of otitis media in these patients is caused by impaired palatal






-6-


muscle activity involved in the opening and closing of the auditory tube (Eustachian tube). This is due to upper respiratory disease which is easily transmitted to the ear via the Eustachian tube. The tube, being the single dynamic connection between nasopharynx and the middle ear, serves as an equalizer for air pressure across the tympanic membrane (ear drum) and loses its function when infected (Davis, 1970). Speech habilitation is, therefore, complicated by inadequate auditory feedback. Another important reason for surgical closure of a cleft palate is to allow the individual normal feeding. Correction of dentition and establishing of normal swallowing habits are, therefore, additional important goals in palatal surgery.



Velopharyngeal Insufficiency


Velopharyngeal insufficiency occurs when the soft

palate is unable to adequately separate nasopharynx from oropharynx during speech. Regardless of the cause, resulting from either morphologic or functional disorder, VPI is an opening between the oral and the nasal cavities which allows airflow into the nasal cavity during the production of oral speech sounds.

Several different surgical techniques are available for the establishment of normal velopharyngeal function. Some examples of these procedures are palatal pushback (a palal lengthening procedure), pharyngeal flap (a skin






-7-


graft), teflon pharyngoplasty (augmentation by injection). A combination of these techniques may be used in order to obtain optimal results (Millard, 1980). In addition to surgical repair, treatment may involve prosthetics and/or speech therapy.



Palatal Fistula


Palatal fistulas, defined holes in the palate connecting the oronasal cavities, may appear secondary to cleft palate surgery or can occur as the result of trauma or ablative surgery. The frequency of occurrence of fistulas following palatal repair has been reported to range from 9 percent to 34 percent (Henningsson, 1983; Ross & Johnston, 1972). Abyholm, Borchgrevink, and Eskeland (1979) report that frequency of occurrence seems to be dependent upon surgical technique. They report an overall incidence of 18 percent. Other frequency occurrences of fistulas resulting from trauma or ablative surgery are not readily available in the literature.

Surgical repair of fistulas is undertaken only after comprehensive evaluation of the anatomical structures. if speech is distorted and a fistula is present, it is necessary to differentially evaluate the cause of the distortion. By temporary obturation of the fistula, the significance of its relationship to speech and the presence





-8



of any velopharyngeal insufficiency can be diagnosed (Bless, Ewanowski, & Dibbell, 1980; Bloch, 1979; O'Neal, 1971; Skolnick, Glaser, & McWilliams, 1980). If the speech sounds normal while the fistula is obturated, it can be concluded that the fistula is the cause of the speech distortion. However, if speech continues to be impaired, this would be an indication of VPI. If it were determined that fistulas closure was necessary,, a concomitant decision also would be made with regard to surgical or prosthetic treatment. O'Neal (1971) has argued that consideration should be given to delaying or even eliminating the need for surgical repair of an existing fistula in growing children with ongoing development of the orofacial and orthodontic structures.

A number of surgical techniques are available for fistula closure. The preferred techniques appear to be dependent upon fistula size and location. James (19.80) reported on buccal flap surgery to repair small fistulas. Additionally, he also reported on the use of large buccal flaps in conjunction with bone grafts for reinforcement in large fistulas. He reported large fistulas to be of

0.5 centimeter and 2.5 centimeters in size. Harris (1980) reported that pharyngoplasty was preferred when closing a posteriorly located fistula. For anteriorly located fistulas, Harris reports the use of bone grafts. The bone graft is subsequently placed between flaps that are






-9-


taken from the nasal and the palatal soft tissues. Further, a bone graft always would be indicated when the fistula occurs in the alveolar ridge.



Prosthetic Treatment of Oronasal Openings


Individuals with incomplete palatal closure who are not candidates for or who do not elect surgical treatment generally are treated with prosthetics (Adisman, 1971). Different kinds of prosthetic appliances are available, depending upon the physical defect being alleviated by obturation. A removable prosthetic appliance is used to cover the complete cleft of the hard and the soft palate, with the posterior part serving as the speech bulb. Closure between the oral and the nasal cavities is established as the posterior pharyngeal wall makes contact with the speech bulb. A simpler appliance, such as a speech bulb, is used for patients with a cleft of the soft palate only. Velopharyngeal insufficiency can be treated prosthetically in a similar manner, resulting in appropriate velopharyngeal closure. Reconstructive surgical repair of fistulas is always the treatment of choice (Drane, 1973), but prosthetic rehabilitation may sometimes be indicated. A prosthetic obturation may be recommended especially with recurring fistulas, soft palate inadequacy, or when serving as an interim obturator while awaiting surgery or orthodontic treatment. The traditional





-10-


prosthodontic treatment includes either a fixed or a removable obturator (Reisberg, Gold, & Dorf, 1985). Small fistulas that do not cause a significant speech disorder may be temporarily obturated to prevent food and liquids from entering the nasal cavity during eating and drinking. These materials may include dental wax, acrylic resin, or chewing gum (Reisberg et al., 1985).



Cleft Palate and Speech


A cleft palate does not provide the separation between the oronasal cavities necessary for normal speech production. Therefore, it will cause hypernasal voice quality and excessive nasal air emission. These two factors, hypernasality and excessive nasal air emission, negatively affect speech intelligibility. The severity of the speech disorder resulting from this pathology depends upon the size of the cleft and the extent of involvement of the cleft with tissues of the hard and/or soft palates. The speech intelligibility of cleft palate patients is impaired by a wide range of errors, including indistinct production, simple and gross substitution, and complete omission of sounds.



Velopharyngeal Insufficiency and Speech


Velopharyngeal insufficiency has been proven to cause disorders of both nasality and intelligibility of speech,









depending on its size. Thorough investigations of the size of such openings between the oral and nasal cavities and their influence on speech has been published. The objective has been to quantify the magnitude of openings in the velar port and the influence of these openings on hypernasality and nasal air emission.

The earliest attempts at this quantification were

performed using direct observational techniques, such as lateral still roentgenography or cineradiography (Benson, 1972; Bj~rk, 1961; Bzoch, 1968; Graber, Bzoch, & Aoba, 1959; Moll, 1960, 1962, 1964; Shelton, Brooks, & Youngstrom, 1964; Subtelny, Koepp-Baker, & Subtelny, 1961). Much information, therefore, is available regarding the pattern of velopharyngeal kinesiology and configuration during speech. This has contributed greatly to the knowledge base about velopharyngeal function and closure (Aram & Subtelny, 1959; Bj6rk, 1961; McKerns, 1968/1969; Moil, 1965; Nylen, 1961; Warren & Hoffmann, 1961). The radiologic measures obtained were compared to indirect measures, such as perceptual judgments of speech intelligibility and voice quality, to acoustical measures (sound spectrography) (Andrews & Rutherford, 1972; Bj6rk, 1961), or to measurements found in the pneumatic pressure-flow (aerodynamic) technique (Warren, 1964a,b; Warren & DuBois, 1964).

Further quantification attempts regarding velopharyngeal openings led to research where the velar port






-12-


openings were experimentally controlled in normal subjects. Different sized polyvinyl tubing was inserted through the nose into the velopharynx to create artificial velopharyngeal openings (Isshiki, Honjow, & Morimoto, 1968; Bernthal & Beukelman, 1977). Similarly, variable aperture speech appliances with inserts were employed by Andrews and Rutherford (1972), Liebman (1964), and Watterson and Emanuel (1981a,b). By manipulating the velopharyngeal port size in this manner, differing degrees of hypernasality could be created and computations of the critical size of openings in the velopharyngeal area performed. Generally, these reported studies on VPI agree that distortion of speech quality is introduced when there is an opening
2 2.
between 10 mm and 20 mm in the velar port. This is perceived as hypernasality, with audible nasal air emission causing distortion of oral consonants.

Based on the literature, it can be concluded that the larger the velopharyngeal opening, the more distortions occur in speech characteristics of intelligibility and resonance quality. However, a complete linear relationship is not always present (Subtelny et al., 1961; Isshiki et al., 1968; Andrews & Rutherford, 1972; Dalston, 1982). Liebman (1964) found that posteriorly positioned velar openings caused fewer speech problems than those anteriorly positioned. The two factors of size and position have proven important with velar insufficiencies and are





-13-


thought to play an important role with other oral-nasal openings such as palatal fistulas.



Palatal Fistulas and Speech


Fistulas, like cleft palate and VPI, cause inadequate closure between the oronasal cavities. Therefore, patients having fistulas exhibit the same speech impairments as do cleft palate and VPI patients.



Fistula Location


It is unclear as to the influence of fistula location on speech. Ross and Johnston (1972) state that fistulas occurring posteriorly in the hard palate or in the soft palate cause a cleft palate-like speech, i.e., hypernasal voice quality and audible, inappropriate nasal air emission. No clear description is offered regarding the influence that anteriorly positioned fistulas may have on speech. It is stated that speech may be affected because of air escaping into the nasal cavity, influencing both articulation and voice quality (Ross & Johnston, 1972). Cosman and Falk (1980) reported that defects of speech production were significant in the presence of anterior fistulas. These palatal defects located anterior to the place of articulation of pressure consonants caused sound substitution. For example, a /t/ sound, an unvoiced





-14-


anterior plosive, was substituted for by /k/, an unvoiced posterior plosive, and the /d/ sound, a voiced anterior plosive, was substituted for by /g/, a voiced positive plosive.



Fistula Size


The importance of fistula size and its influence on speech is evident in the literature. It has been proven clinically that speech impairment increases with fistula size. Morley reported in 1962 that even a small fistula may affect the development of normal speech. This occurs because of a prevention of the build-up of intraoral air pressure necessary for the production of oral consonants. In the literature published twenty years later, it is still stated that speech is increasingly and adversely affected by palatal fistulas of increasing size (Henningsson, 1983; Shelton & Blank, 1984). This information, however, only describes the fistulas as being either small, medium, or large without any quantitative definitions of size.

In a review of the literature reporting on palatal fistulas, no study was found that dealt with quantification of palatal fistulas and their influence on speech. There were only two different studies which defined large versus small fistulas. Proctor (1969) defined large oronasal fistulas as ranging from 0.5 centimeter to





-15-


2.5 centimeters, while Henderson (1982) defined small fistulas as ranging from 0.5 centimeter to 1.5 centimeters in width and 0.5 centimeter to 2.0 centimeters in length. One pertinent study (Shelton & Blank, 1984) deals with palatal fistulas and their influence on speech. Shelton and Blank reported on six patients with oronasal fistulas. The sizes of the fistulas were classified as small, moderate, or large, without further specification. Patients with a small or moderate fistula maintained sufficient intraoral air pressure for adequate oral sound production. Patients with large oronasal fistulas demonstrated a reduction in intraoral air pressure. Nasal air flow, however, was detected with all fistula sizes and was found to generate some noise.

With this meager information and inadequate guidelines on palatal fistulas, there seems to be a demonstrated need for more quantitative information regarding the effect of fistula size and location on speech production. Knowledge about how differently sized and positioned fistulas would influence speech would enhance patient advisement regarding surgery. In addition, such knowledge would aid in the selection of prosthetic appliances, including those for temporary obturation of the opening while eating and drinking. In this present study, the first steps have been taken to quantitatively define the impact of palatal fistula size and location on speech, as it was measured






-16-


perceptually (evaluation of articulation and nasality), and instrumentally (nasal airflow and intraoral and differential air pressure). It appears that this new quantitative data, together with earlier well-established

qualitative information, will contribute to better and necessary guidelines in the clinical management of patients with palatal fistulas and speech disorders.



Statement of Purpose


The primary objective of this study was to quantitatively and qualitatively analyze speech in terms of nasality and articulation, as well as some of the associated aerodynamic correlates in the presence of artificially introduced openings through the hard palate leading directly into the nasal cavity. Of particular interest is the dependence of the speech characteristics on the size and position of such openings. Definition of the requisite critical size of VPI is an ongoing endeavor and is of much interest in the literature. However, there is little information available on defects in the hard or soft palate anterior to the velopharyngeal port and the resulting impacts on speech. Although clinical observations suggest that oronasal fistulas contribute to hypernasality and nasal emission, there is little quantitative information describing the influence that fistula size and location have on speech.






-17-


This present study is an attempt to close the gap regarding this lack of knowledge and the purpose is to begin defining the influence of size and location of an opening through the hard palate on speech in terms of intelligibility and resonance quality. In order to perform this quantification, the size and location of openings anterior to the velopharyngeal port were experimentally manipulated. Openings were introduced in the hard palate part of a removable palatal prosthetic appliance (research appliance). A 28-year-old Caucasian male with a cleft of the hard and soft palate and who wears a removable prosthetic palatal appliance was the only subject in this study.

A combination of three locations (anterior, middle, and posterior) and four sizes (5 mm2 10 MM2, 20 mmand 30 mm-) of these oronasal openings were tested for a series of clinical and physiological speech tests. The clinical tests incorporated tests of nasal air emission, articulation, hypernasality, an articulation rating, and a hypernasality rating. The nasal air emission test was evaluated by the examiner taking a repeated measurement of two readings. The other clinical tests were presented to six speech-language pathologists specially trained in evaluating articulation and hypernasality in patients with palatal dysfunction. These tests were evaluated by the six judges who listened to and subjectively scored





_18-


the remaining four clinical tests. The speech physiology testing was conducted by employing an aerodynamic instrument that computed intraoral air pressure and differential pressure.

Since previous research has been concerned basically with velar incompetencies and with controlled velar port openings, a replication study was conducted to assess the influence of velopharyngeal inadequacy of speech. Openings of the same sizes as in the experimental study were drilled in the velar part of the speech appliance which would imply a resemblance to VPI. The same test battery of clinical and physiological speech tests was administered. Data obtained from this replication study were compared to the experimental results, as well as to previous research reports. Such comparisons were conducted in order to evaluate the results from the present experimental study on oronasal openings in the velar region and previously published results.

It is known from previous research efforts that speech is negatively influenced by an increased size of VPI. Furthermore, there are clinical implications stating that increasing fistula size is detrimental to speech intelligibility. Fistula location and its influence on speech

have been discussed only vaguely but it is believed that posteriorly positioned fistulas cause more speech problems than anteriorly positioned fistulas. It was anticipated,






-19



therefore, that the results from this present study would be negatively influenced also by the magnitude of the hole size, resulting in imprecise articulation, excessive nasality and nasal emission, and a decrease in intraoral air pressure and differential pressure as the oronasal openings were increased in size. It was anticipated also that these speech variables, likewise, would be influenced as the oronasal opening was moved posteriorly.















CHAPTER II
LITERATURE REVIEW AND RESEARCH QUESTIONS



Introduction


Maintaining appropriate intraoral air pressure during the production of oral speech sounds is an important physiological component of normal speech. Normal speech production will be adversely affected when an opening occurs between the oronasal cavities.

Oronasal openings, such as a cleft palate, VPI, or

palatal fistulas permit air to escape from the oral cavity into the nasal cavity. This air escape results in transmission of the air and sound energy into the nasal cavity, changing the physiological event of normal speech (Daniloff, Schuckers, & Feth, 1980; Spriestersbach & Powers, 1959). This process results in a reduction in intraoral air pressure and increased nasal air emission. These factors, reduced intraoral air pressure and increased nasal air emission, cause abnormal speech production by (a) giving speech a hypernasal voice quality and (b) reducing articulation proficiency. In addition, there are other palatal abnormalities that often are undiagnosed at birth, but which are discovered later, which impair normal palatal


-20-






-.21-


anatomy and physiology. Such abnormalities interfere with speech proficiency. These later diagnosed abnormalities occur in the form of submucous clefts, cleft uvula, unusually high and/or narrow palate, and paresis of the soft palate muscles. Occurrence of VPI in the non-cleft population may be due to several conditions, including a neurological disorder, a function disorder, or tissue ablation, or tissue reduction following a surgical procedure, such as a tonsillectomy or adenoidectomy (Bradley, 1979). In order to more fully understand oronasal openings and their effect on speech, a brief description of relevant embryology and anatomy will be addressed. A discussion of the speech characteristics that occur with various oronasal openings and how they are measured will follow. A review of experimental studies on VPI will be considered, followed by a discussion of the research questions which are the subject of the present research.



Palatal Embryology and Anatomy


For an in-depth study of embryology and anatomy of

the palatal structures and the oronasal cavities the reader is urged to consider the works by Bateman (1977), Brescia (1971) , Bzoch and Williams (1979) , DuBruhi (1980) , and Zemlin (1968). The following material is excerpted from the above references.





-22-


Palatal Embryology


The initial stage of formation of the embryonic face including the primary and secondary palates can be seen as early as in the second month of gestation. The face and mouth including the lips and the hard palate have completed their basic development process at the end of the third month.

The development of the orofacial structures is dependent on the merging and fusion of five prominences, the single frontal prominence, the paired maxillary prominence, and the paired mandibular prominence. These are present in the third week of pregnancy. The mandibular processes merge at the end of the fourth week and give rise to the lower lip and the chin.

The frontal prominence grows downward and forward and becomes the frontonasal process. On either side of this process the two olfactory pits are formed and as the frontonasal process grows further downward these pits will divide it into the median nasal and the two lateral nasal processes (Bzoch & Williams, 1979). This occurs by the end of the sixth week (Figure 1). It is the nasomedial process that forms the midportion of the upper lip, the medial portion of the alveolar ridge, and the primary palate.

The primary palate is developed as the nasomedial

process grows down and back to form the upper anterior roof





-23-


Frontonasal prominence (process) Nasomedial process Nasolateral process Olfactory pits /1

Maxillary process Mandibular arch



6 wk (12 mm)



Figure 1. Embryo at 6 weeks (12 mm).


(a) (b)


Figure 2. Palatal structure. (a) Before fusion (7 to 8
weeks), (b) Fused palate (8 to 10 weeks).





-24-


of the mouth and it extends vertically to the floor of the nasal cavity. This plate projects into the oral cavity as far as to where the incisive foramen is later located. The nasomedial process then extends laterally to the alveolar ridge and encompasses the four incisor teeth (Bateman, 1977). Formation of the primary and secondary palates in the human embryo has been observed as early as the second month of gestation (Crelin, 1976).

The secondary palate is the main part of the palate and includes both the hard and the soft palates. It is formed from paired shelf-like outgrowths that arise from the maxillary processes. Further contributors are the lateral palatine processes or the lateral palatal shelves that grow medially and inferiorly during the eighth week. When these palatal shelves first start to develop, the mandible lies high in the mouth and the tongue must rise between them. During this time the lateral processes grow downward on each side of the tongue. In the ninth week of development a sudden growth spurt of the mandible increases the vertical distance between it and the maxillary arch and the tongue therefore drops down. This allows the palatine processes to rise up and start to fuse. Fusion of the palates is complete between the end of the 10th and the beginning of the 12th week (Figure 2) (Crelin, 1976).





-25-


Palatal Anatomy and Physiology


The palatal structures, or the roof of the mouth,

consist of the hard and the soft palates. The hard palate forms a shelf between the oral and the nasal cavities. These cavities can be completely separated by the normal function of the soft muscular palate. Adequate anatomy of both palates is necessary for normal function.

The hard palate. The hard skeletal palate consists

of two sets of paired bones, the palatine processes and the palatal bones (Figure 3). The palatine processes are formed by horizontal, medially directed projections from the body of the maxilla (upper jaw bones) at the boundary between the body and the alveolar process. The alveolar process forms the thick spongy part of the maxilla where the teeth are positioned. The palatine processes (in anteroposterior direction) are shorter than the maxilla and form only about three-fourth of the hard palate. They terminate posteriorly with a rough edge, to which the horizontal plates of the palatal bones articulate in the transverse palatine suture. The posterior borders of these bones are free and form the posterior nasal spine at midline. The midline suture terminates anteriorly at the incisive foramen. In young children, an irregular suture line can be seen extending bilaterally from this foramen to the alveolar process. The small triangular-shaped structure anterior to this suture line is called the






-26-


Alveolar process
Incisive foramen < Premaxilla


Median palatine
-suture
Palatine process
of maxilla
--Transverse / palatine suture
-Palatine bone .Posterior nasal spine


Figure 3. The bones of the hard palate.


papilla


zone


Soft palate


/ N Uvula


Figure 4. The hard and the soft palate.






-27-


premaxilla. The premaxilla represents the two parts of the maxillary bone that unite early in embryonic life. The suture line indicating the existence of the premaxilla is almost erased on the adult skull (DuBruhl, 1980).

The hard palate is covered with mucosa of a grayish pink color and various areas of the hard palate can be identified due to the varying structure of the submucous layer. Anteriorly, there is a series of transverse ridges called rugae, posterior to which we find the midline raphe. The gingiva and the glandular zone can also be clearly identified (Figure 4). The hard palate is thick anteriorly and laterally but becomes progressively thinner toward its midline. This results in an arching of the palate both transversely and anteroposteriorly. This arching differs from one individual to the next and contributes to the acoustic properties of the oral cavity and probably also to individual voice characteristics (Zemlin, 1968).

The soft palate. The soft palate, or velum (Figure 5), consists of muscular and fibrous tissues, nerves, and blood vessels. It is positioned as a continuum from the posterior border of the palatine bones of the hard palate. It is attached bilaterally to the pharyngeal wall and extends posteriorly into the oropharynx where it ends in a free margin. The anterior part of the velum, slightly posterior to the articulation with the hard palate, is the palatal aponeurosis, a point where the palatal muscles are





-28-


--V2



V vi















HP = Hard palate Vl = Velum at rest V2 = Velum at closure PPW : Posterior pharyngeal wall


Figure 5. Lateral view of the hard and the soft palates.


PPW





-29-


attached. The posterior middle of the free margin of the velum consists mainly of the uvula. Bilaterally to the uvula,, two curved folds of mucous membrane arch downward and outward. These are the anterior and the posterior faucal pillars (DuBruhl, 1980; Zemlin, 1968; Morley, 1962).

Normal function of the hard and the soft palates. The function of the hard palate is to serve as the boney wall separating the nasal cavity from the oral cavity. It thus builds the floor of the nasal cavity and the roof of the mouth. A review of the literature finds a prevalence of the conclusion that the primary muscle active in velopharyngeal closure is the levator veli palatini muscle. The contraction of this muscle elevates and pulls the soft palate posteriorly. It has also been suggested that this

muscle exerts posteromec(ial movement of the lateral pharyngeal walls, aiding velopharyngeal closure proficiency (Dickson, 1972).

The middle third of the velum demonstrates the greatest range of movement. It achieves closure with the posterior pharyngeal wall during speech, thus serving as a functional valve during velopharyngeal closure. This elevation also is thought to be substantially assisted by the uvular muscle (Dickson, 1972).





-30-


Abnormal Palatal Function and Speech


The palatal anomalies affecting speech intelligibility, which are the bases in this work, are limited to palatal fistulas and VPI. The severity of a speech impairment is dependent upon the size/location of a fistula, or the size of a velopharyngeal opening. In other words, the severity is dependent upon the magnitude of the coupling of the oronasal cavities.

Extensive research has determined that a velopharyngeal
2 2.
opening between 10 mm and 20 mm in size will result in hypernasality with audible nasal air emission, thus distorting the production of oral consonants (Isshiki et al., 1968; Shprintzen, Croft, Levin, & Rakoff, 1977; Warren, 1964a,b). With the critical size of VPI being defined there is, however, little information and no quantifiable data available on the effect of palatal fistulas. It appears that the severity of a speech disorder--in the presence of a palatal fistula--will depend on the size and location of the fistula. However, it has been concluded that the same components of speech that are present with VPI (hypernasality with audible nasal air emission) also occur with palatal fistulas (Lindsay, 1971; Morley, 1962; Palmer, Hamlen, Ross, & Lindsay, 1969; Reid, 1962; Ross & Johnston, 1972; Skolnick, Glaser, & McWilliams, 1980).






-31-


The Effects of Velopharyngeal Insufficiency on Speech


The importance of adequate velopharyngeal valving and

its influence on speech have long been recognized. Although it is difficult to determine the exact frequency of VPI in the non-cleft population, it has been suggested that between 20 percent and 40 percent of patients who undergo cleft palate repair suffer a concomitant VPI (Bradley, 1979). The major anatomical defects which occur are related to velopharyngeal disproportion, muscular deficiency, or surgical removal of tissue (Bradley, 1979).

Several authors have established the existence of a close relationship between the degree of velopharyngeal opening and the severity of speech impairments in normal and in cleft palate patients (Bzoch, 1968; Moll, 1962; Shprintzen et a!., 1975; Subtelny et al., 1961). The major speech distortions associated with VPI can be measured by employing standardized tests and instruments or by the subjective judgment rating of resonance quality and articulation. The characteristics of these speech errors help us to differentially diagnose VPI. Secondary components of speech behavior, like nasal and facial grimacing, sometimes accompany VPI (Bzoch, 1979).

Bradley (1979) lists specific speech sound errors as being weak production of plosive sounds (together with nasal emission of air), accompanied by a nasal non-speech






-32-


sound, called a nasal snort. She also reports that the high vowels /u/ and /i/ can be expected to sound hypernasal with accompanied nasal air emission. Bj6rk (1961) has observed that the velar port is not completely closed when nasal consonants and their neighboring vowels are produced in normal speech. Findings by Moll (1962) showed the high vowels to exhibit more velopharyngeal closure than their low counterparts, /ae/ and /a/. Furthermore, when the vowels were uttered in connection with a nasal sound such as In!, he observed incomplete velopharyngeal closure, especially if the vowel preceded the nasal sound. This was confirmed by Benson (1972), who found open velopharyngeal ports in 63 percent of the subjects phonating /a/, in 23 percent when phonating /i/, and in 9 percent phonating /u/. Fritzell (1973) also concluded that slight nasal airflow could be measured in the normal population during the production of /a/ but not with lii, or with the fricatives /f/ or Is/. He pointed out that individuals with VPI could not attain the intraoral air pressure necessary for the distinct production of plosive and fricative sounds. The speech of these individuals, therefore, contained hypernasal voice quality and compensatory speech errors, such as glottal stops and pharyngeal fricatives.

Shelton et al. (1964) added to this information that velopharyngeal closure was found to be complete when testing isolated vowels in one subject group. However, this





-33-


same group of subjects later presented abnormal closure patterns when tested in connected speech. This would seem to indicate that syllable or sentence testing should be employed when evaluating velopharyngeal closure. These findings were investigated by Bzoch (1968) , who found in the testing of repeated syllables containing the plosive /p/, combined with either of the high vowels, /i/ and /u/, that complete velopharyngeal closure was obtained. This was also the case when the low vowel /a/ was tested in the same manner. However, a discrepancy of one millimeter in the configuration of lower palatal elevation for the lower vowel /a/ was observed.

Articulatory distortions in individuals with palatal disorders were investigated by Bzoch (1965). He reported that two-consonant blends, fricatives, and plosives--in that order--were the most misarticulated sounds. The error types included substitutions, distortions, and omission of sounds. Substitutions can be recognized as simple or gross substitutions. Gross substitutions are characteristic for cleft palate patients and for the VPI population. Gross substitutions included such errors as pharyngeal fricatives, glottal stops, and nasal snorts. These occurred as substitutions for plosive and fricative sounds and occurred in the pronunciation of affricate and glide speech sounds. The nasal snort was found to substitute for the nasal sounds /m/ and /n/.





-34-


Perceptual Judgment


In order to evaluate different speech characteristics, especially nasality, perceptual judgment has been and is still used extensively both clinically and in research. It is a subjective method of determining the degree of speech impairment using different rating scales. There have been, for example, 3-, 4-, 5-, and 10-point scales in use (Counihan & Cullinan, 1970; Isshiki et al., 1968; Subtelny et al., 1961; Warren, 1964b). Scaling procedures with values from 0 to 100 has been utilized also (Fletcher & Bishop, 1970). Different studies employed panels of judges or single judges. Some studies claim higher reliability using one judge, while other studies claim higher reliability using a panel of judges (Bradford, Brooks, & Shelton, 1964). The number of judges also has varied considerably, ranging from 3 judges to 9 judges to 48 judges (Counihan & Cullinan, 1970; Liebman, 1964; Sherman, 1970; Subtelny et al., 1961).

It appears especially difficult to obtain reliable perceptual judgment of nasality. This difficulty is influenced by many factors, including speaker variations due to phonation, articulatory precision, regional dialect, and language (Counihan & Cullinan, 1970; Fletcher, Sooudi, & Frost, 1974; Warren, Hall, & Davis, 1981). Great variance in sample selection, using normal and cleft palate subjects,was also demonstrated. The difficulty to






-35-


obtain a homogenous sample was another problem, due to differences in type and severity of the clefts (Bradford et al., 1964; Coleman, 1963/1964; Andrews & Rutherford, 1972; Watterson & Emanuel, 1981a,b).

The selection of different speech samples and the clinician's training and experience also contributed to the difficulty of obtaining valid and reliable nasality ratings (Bradford et al., 1964; Bzoch, 1979; Counihan & Cullinan, 1970). Speech samples included isolated sounds, syllables, and/or connected speech in forward or backward play (Counihan & Cullinan, 1970).

It is difficult to compare the results and to compute accurate reliability because of the disparity in perceptual testing of nasality. However, it has been documented (Philips & Bzoch, 1969) that perceptual judgment of articulation offers high intrarater reliability averaging 85 percent, although interrater reliability is lower with an average of 74 percent. Furthermore, Sherman (1970) found articulation and hypernasality to be functionally related in children with cleft palate.

Different instruments have been employed to measure

nasality, sometimes in combination with perceptual judgment testing. Up to this point, however, no single objective test instrument seems to do so adequately and exclusively (McWilliams, Glaser, Philips, Lawrence, Lavorato, Beery, & Skolnick, 1981; Van Demark, Bzoch, Daly, Fletcher,






-36-


McWilliams, Pannbacker, & Weinberg, 1985). The traditional measures used to augment perceptual judgment testing have been radiographics, aerodynamics, and acoustics (Reich & Redenbaugh, 1985).

In conclusion, it appears that it has been, and continues to be, a difficult task to obtain a reliable measure of nasality. Therefore, a combination of methods needs to be considered to secure the reliability of nasality judgment.

The influence of examiner expectancy on subjective evaluations, such as with perceptual judgments, has been documented (Ramig, 1982). Physical disabilities, orofacial anomalies, and other visual as well as auditory cues have been noted to influence the examiner's judgments in faceto-face examinations. Most studies reviewed here, however, have employed tape recorded speech samples.

Subtelny et al. (1961) employed perceptual judgment rating of nasality and intelligibility in combination with lateral roentgenography measures. They reported on the size of the velopharyngeal port opening as it was changing during speech. The nasality ratings were derived by using a four-point rating scale. The speech of 70 adult subjects with cleft palate revealed that a distance between the posterior pharyngeal wall and the soft palate of 3.5 mm. to 7.0 mm (3.5 mm equals 9.62 mm 2) resulted in the auditory






-37-


perception of hypernasal speech. Smaller distances (0.5 mm to 3.0 mm) resulted in the perception of only moderately nasalized speech.

Word-syllable intelligibility evaluations were assessed by the same judges and compared to aperture size (Subtelny et al., 1961). Subjects with complete closure scored on the average a 10 percent loss of intelligibility. Subjects with the 0.5 mm to 3.0 mm distance showed a 14.3 percent loss, and the 3.5 mm to 7.0 mm group had a 27.31 percent loss. Even greater openings than these were not accompanied by significant errors of nasality or intelligibility. Only when the openings became as large as 11 mm to 18 mm was intelligibility reduced significantly. This nonlinear relationship, as described here, has been evidenced when comparing judgments of hypernasality with velopharyngeal port size that were measured with lateral still X-ray (Subtelny et al., 1961), with cineradiography (Andrews & Rutherford, 1972), or with aerodynamic measurements (Isshiki et al., 1968). It appears, from the results of the study by Subtelny and co-workers (1961), that small gaps have a greater effect on nasality than on intelligibility. Subjects with a 3 mm diameter velopharyngeal port opening still produced intelligible speech but were judged as moderately nasal. This information agrees with Bjork's study (1961), in which subjects who displayed a

4 mm large gap were assessed as hypernasal.





-38-


Hypernasality Testing


Evaluation of hypernasality can be executed easily by employing a test described by Bzoch (1979). This test is the Hypernasality Test which is standardized and convenient for use both clinically and in research. The results from the test are derived from a pass-fail procedure of 10 trials where the correct productions are added together and presented as a single score, where a maximum correct score would be 10.

As the subject produces 10 one-syllable words (beet, bit, bait, bet, bat, bought, boat, boot, but, Bert) once with the nares open and again with them gently pinched closed, a shift of voice quality will determine if hypernasality is present. If the velopharyngeal port is inappropriately open during production of any of these words, closing of the nares will result in changing this openended resonator to a closed-end resonator and, consequently, a change in resonance quality.



Articulation Testing


The Bzoch Error Pattern Screening Articulation Test (BEPSAT) was designed to evaluate individuals with incomplete velopharyngeal closure (Bzoch, 1979). This test was compared to the Iowa Pressure Articulation Test by Erickson (1984). A correlation study indicated significant






-39-


relationship between the two tests. The two tests also showed high validity as a tool for appropriately identifying articulatory errors in the speech of cleft palate individuals. The BETSAT, however, appears to provide more valuable information on the total articulatory behavior, because it includes all of the distinctive feature groups. There are 31 test words in the BEPSAT, with 23 words containing the individual sound element in medial position and eight consonant blends where it occurs in initial position.



Instrumentation


Clinicians have tried, over the years, to correlate the perceptual judgment of velopharyngeal competence or incompetence with objective testing employing various instruments designed to measure velopharyngeal closure. There has been a concerted effort to find such a tool that would measure the magnitude of VPI and to relate it to the changes of voice resonance and articulation (Bjark, 1961; Bzoch, 1979; Dalston, 1982; Edgerton, Sadove, Compton, Bull, Blomain, McDonald, & Bralley, 1981; McWilliams et al., 1981; Warren & DuBois, 1964). One of the reasons for designing a measurement instrument was the search for objective and reliable means to test nasality, since it had been found that both interjudge and intrajudge reliability of perceptual judgment testing were low (Bradford et al., 1964; Counihan & Cullinan, 1970).





-40-


Observational techniques, such as lateral still

roentgenography and cineradiography, have been employed in a variety of research studies (Benson, 1972; Bj6rk, 1961; Bzoch, 1968; Graber et al., 1959; Moll, 1960, 1962, 1964; Shelton et al., 1964; Subtelny et al., 1961). These techniques have also been used extensively in clinical diagnostics.

Other direct observational techniques include tomography (Bj6rk, 1961; Kuehn & Dolan, 1975) and the employment of fiberoptics in photodetector assessment (Dalston, 1982) and in nasopharyngoscopy with videopharyngoscopy (Shprintzen et al., 1977). Some measures of palatal function and nasality would include intraoral air pressure using oral manometers (Morris, 1966; McWilliams et al., 1981), acoustic analysis with spectrography (Andrews & Rutherford, 1972; Bj~rk, 1L61; Coleman, 1963/1964; House & Stevens, 1956; Watterson & Emanuel, 1981a), and nasal airflow testing (Bzoch, 1979). Another instrument for acoustic analysis is called Tonar (Fletcher, 1970; Fletcher & Bishop, 1970; Fletcher et al., 1974). Two additional techniques are the accelerometric technique also called nasal vibration analysis (Edgerton et al., 1981; Horii, 1980; Reich & Redenbaugh, 1985; Stevens, Kalikow, & Willemain, 1975), and respirometrics (Shaw & Gilbert, 1982).

Derivation of the hydraulic principle was used in an attempt to measure pressure-flow relationships (Gorlin &






-41-


Gorlin, 1951). Warren and DuBois (1964) extended the attempt by employing the resulting hydrokinetic equation to obtain indirect measures of the velopharyngeal port area. The changes in size when different speech sounds were produced could thus be determined (Warren, 1964a,b). Quantitative measurements from the pneumatic pressure-flow technique (Gorlin & Gorlin, 1951; Warren, 1964a,b; Warren & DuBois, 1964) have been employed to obtain indirect measures of the velopharyngeal port area and to determine its changes in size when different speech sounds are produced.

Spectrographic analysis of the acoustic signal is

another method employed to assess the degree of nasalized speech (Andrews & Rutherford, 1972; Bjbrk, 1961; House & Stevens, 1956; Watterson & Emanuel, 1981a). At other times a standardized test of hypernasality was administered (Bzoch, 1979, Bzoch et al., 1984). However, a combination of methods has been employed to ensure reliability of the tested material. Although several instruments are available for the measurement of nasality, it has, nevertheless, continued to be recorded by using perceptual judgment procedures (Isshiki et al., 1968; Subtelny et al., 1961; Watterson & Emanuel, 1981a,b).


Lateral Still Roentgenography and Cineradiography


Both lateral still roentgenography and cineradiography have been used to examine visually the function of the






-42-


velopharyngeal port and to determine the relationship between the velar length and the depth of the nasopharynx (Figure 5). The patterns of the paldtopharyngeal movements and their changes, during speech, give much information about velopharyngeal function and closure. When inspecting the still X-ray or the cinefilm during speech production, the elevated soft palate can be seen attempting to make closure with the posterior pharyngeal wall. The least linear distance between these two anatomical structures, the soft palate and the posterior pharyngeal wall, is measured in millimeters. Bj~5rk (1961) demonstrated a strong linear relationship between the width of this gap as measured from the lateral cinefilms and the area measures of the gap in the velopharyngeal port as determined in the cross-sections obtained from tomography. With these two measures, he developed a diagram that can be employed in order to determine the size of the coupling area by use of cinefilms only (Bj6rk, 1961).

Occasionally, nasalization of speech is noted in spite of obvious closure visible on the lateral film. Therefore, it must be assumed that although closure might take place at one point as the velum stretches toward the posterior pharyngeal wall, it might do so only at that point which can be noted on the lateral film. One or more parts of the velum, laterally or medially to the touching point, may not attain complete closure. This lack of complete closure pattern, or uneven closure, would result in an





-43-


opening of the velopharyngeal port. This can be observed with basal view cineradiography (Skolnick, 1970; Williams, 1986) or through tomography (Bj~rk, 1961). The limitation of the lateral measurement and the radiation danger has led to the development of other instrumental methods of velopharyngeal quantifiable assessment such as nasoendoscopy (Dalston, 1982).



Aerodynamic Technique


Intraoral air pressure is a physiological component

related to the production of different speech sounds. The normal production of plosive and fricative sounds is dependent upon adequate oral cavity pressure. Depending on the size of a velopharyngeal deficiency and airflow into the nasal cavity, interference with the pressure sounds can be expected (Shelton et al., 1973). A number of instruments have been developed to measure airflow and air pressure with the basic components being flowmeters and pressure transducers. The flowmeter records the volume rates of airflow, and pressure transducers record air pressure within the vocal tract (Lubker, 1970; Warren, 1973).

The pneumotachograph sensing element is the most

common and most reliable type of flowmeter. The principle of this instrument is that there exists a proportionality between the airflow rate in a tube and the pressure difference between its endpoints. This instrument typically





-44-


measures the volume of air in cubic centimeter per second (Counihan, 1979; Hardy, 1965; Lubker, 1970).

It has been established that there is no simple relationship between the nasal airflow and the size of the velopharyngeal port. There are factors such as respiratory effort and nasal airway resistance that influence airflow in subjects with VPI (Hutters, 1982; Warren, 1973). Hutters (1982) claims that her findings indicate that nasal airflow alone will provide reliable information about velopharyngeal closure and that testing of nasal airflow would be valuable as a diagnostic tool.

In order to obtain data on intraoral air pressure, it is necessary to use a sensitive pressure transducer which is connected to an amplifier and a recording device. Its connecting tube, or catheter, is placed in the mouth perpendicular to the airflow with the requirement that it is neither too long, nor that it has a too large or too small a diameter (Lubker, 1970).

One of the aerodynamic techniques, the pressure-flow technique, has been described by Warren (1964a,b) and by Warren and DuBois (1964) as a method of calculating the area size of VPI. It involves recording the amount of airflow through the velopharyngeal orifice simultaneously with the differential pressure across it during speech. Airflow is measured with a pneumotachograph and differential pressure with a pressure transducer. By using these






-45-


measures, the velopharyngeal area can then be calculated by using a hydrokinetic equation (Warren & DuBois, 1964, p. 52; Warren & Devereux, 1966, p. 105).



VPA Vn
.65 2


where

VPA = velopharyngeal area in mm 2 Vn = rate of airflow in cm 3/sec

AP = differential pressure in cm H 20, converted to
dynes/cm2

D = density of air (0.001), gm/cm3

.65 = correlation coefficient



Warren (1964a) employed this method and equation to establish the relationship between oropharyngeal pressure, nasal airflow, and velopharyngeal port size. He examined the critical size openings in the velopharyngeal port in 10 normal-speaking subjects. The velopharyngeal opening was seldom found to be larger than 10 mm2 and was usually smaller than 3 mm 2 under these conditions. Only once was it open more than 20 mm In a succeeding experiment with 10 speakers with cleft palate, Warren (1964b) found that those speakers who were considered to be rehabilitated, that is, exhibited acceptable voice quality, rarely evidenced
2
an orifice larger than 10 mm .With the nasal speakers,






-46-


however, he frequently recorded the aperture to be larger than 20 mm 2 ,and the extremely nasal speakers displayed
2
openings as large as 100 mm .He concluded that the critical range of closure, and where speech is perceived as hypernasal, starts at about 20 mm.

Warren (1979) developed an instrument, the Palatal

Efficiency Rating Computed Instantaneously (PERCI). It was designed to evaluate palatal efficiency (pressure drop across an oral/nasal orifice) both for clinical use and for research. This instrument permits instantaneous computing of measures such as intraoral air pressure, nasal airflow, and the differential pressure between the oral and the nasal cavities. Other measurements that can be obtained from its second version, PERCI II, are airway conductance and the area of the orifice related to the palatal mechanism (Microtronics Corporation, n.d.). Intraoral pressure is detected by a pressure transducer and the nasal airflow by a pneumotachograph. Input to these instruments is transmitted via catheters inserted into the mouth and nose, respectively. When complete separation exists between the oral and the nasal cavities, the nasal pressure will be zero while the oral pressure will range between 3 to 7 cm H 20. With incomplete separation, such as when a palatal fistula is present, the difference in pressure would typically vary with the size of that opening (Figure 6).





-47-


Pressure Transducer


(a) Intraoral Air Pressure


Pressure Transducer


(b) Differential Pressure



Figure 6. Palatal efficiency rating computed instantaneously
(PERCI). (a) Intraoral air pressure, (b) Differential pressure.





-48-


When recording airflow with a pneumotachograph, such as the one used in the PERCI instrumentation, the use of a face mask is required to channel oral and nasal airflow. Such a mask would typically be an anesthesia--or a diver's-mask with built-in openings for the necessary catheters. The use of face masks has been criticized. Lubker and Moll (1965) cautioned that it may restrict articulatory movements. Furthermore, the amount of pressure that is required for the mask to obtain an airtight fit also may influence the normal speech behavior. The face mask has been criticized also in that the air storage in it would provide a resistance to the expiratory airstream and, therefore, induce elevated intraoral air pressure and airflow (Counihan, 1979). However, these considerations have not been substantially confirmed. With the use of a short, straight catheter with a wide opening, resistance is not clearly affecting speech measurement data.



Sound Spectrography


Sound spectrography has been widely employed to measure nasalized speech. The spectrograph analyzes complex signals, such as the speech signal, as a function of frequency, time, and intensity. The different speech sounds create certain patterns on the visual output, the spectrogram. The most important characteristics in evaluating presence of nasality are the formants, which can be seen as dark





-49-


areas on the spectrogram. They consist of regions of accentuated intensity and represent the resonances of the vocal tract (Hadding & Petersson, 1970). The three first formants (Fl, F2, and F3) usually can be clearly distinguished, and sometimes there is even a clear display of F4 and F5.

Schwartz (1979) described the nasalized feature

characteristics as a reduction in the intensity of the first formant, antiresonances, extra resonances, and a change in the center frequencies of the formants. Other spectral features that can be associated with nasality also may occur. These have, however, been observed to be too unpredictable or too sporadic to deserve mentioning. Even the four characteristics just described may not always occur. However, whenever a nasalized vowel is produced, it is likely at least one of them will be seen on the spectrogram.

In an attempt to evaluate velopharyngeal area size during normal speech and during simulated nasal speech, simultaneous cineradiography and sound spectrography were employed by Bj5rk (1961). Thirty adults and 26 children, all normal speakers, served as subjects. The findings showed that when relatively small changes of nasality were registered by the spectrograph, the cineradiographic readings would indicate a velopharyngeal opening of
2
approximately 10 mm In a second study, these normal





-50-


speaking subjects were asked to mimic hypernasal speech. The coupling areas were then found to have increased to between 100 mm2 and 280 MM 2



Nasal Airflow


Velopharyngeal and/or palatal openings will permit

airflow into the nasal cavity and nasal air escape can be recorded. A convenient and simple test, the Nasal Emission Test, can be employed for evaluation of nasal air emission (Bzoch, 1979). Bzoch, Kemker, and Dixon-Wood (1984) stated that clinical experience has proven it to be a valuable speech evaluation tool of velopharyngeal function in cleft palate patients. Bzoch et al. (1984) continued to report that, in a comparison study between cineradiographic measures of the velopharyngeal port and an analysis of the spoken test words, there was a 96 percent agreement of palatal function when the speech of 40 patients was tested. The standardized Nasal Emission Test tests 10 two-syllable words (people, paper, puppy, pepper, piper, baby, Bobby, bubble, B.B., bye-bye). As each word is repeated by the subject, the examiner observes any airflow through the nose using a simple indicator, such as a small paper paddle about the size and shape of a toothbrush held under the nose (Bzoch, 1979).

The results are recorded from a pass-fail procedure of 10 trials where the correct productions are added






-51-


together and presented as a single score, with a maximum correct score of 10. This test should always yield a perfect score of 10 to indicate complete closure of the velopharyngeal mechanism.



The Effects of Controlled Velopharyngeal
Insufficiency on Speech


Some experimental studies were performed where the investigators created different size velar port openings in normal speaking subjects. This involved (a) inserting of different sized polyvinyl tubing in the velopharynx (Bernthal & Beukelman, 1977; Isshiki et al., 1968) or

(b) using a variable aperture speech appliance with inserts (Figure 7) (Andrews & Rutherford, 1972; Liebman, 1964; Watterson & Emanuel, 1981a,b). By manipulating the velopharyngeal port size, different degrees of hypernasality could be created and computations of the critical size of openings in the velopharyngeal area were performed.

In order to determine whether there is a critical size of the velopharyngeal insufficiency at which point speech becomes unacceptable, Isshiki and co-workers (1968) conducted an experimental test with 11 young adult normal speakers. Polyvinyl tubes were inserted into the velopharynx where artificial openings were created at diameters of 5 mm, 7 mm, 9 mm, and 12 mm. Articulation and hypernasality were perceptually judged by five speech specialists






-52-


ZZD


(b)



Figure 7. Two examples of tools for introducing velopharyngeal insufficiency in normal speakers.
(a) Polyvinyl tubing (Isshiki et al., 1968), (b) Variable palatal prosthesis with inserts
(Andrews & Rutherford, 1972).


0


(0






-53-


on a three-point and a five-point rating scale, respectively. The findings indicated that at no specific point along the range of dimensions did speech suddenly sound abnormal. The authors claim, instead, that there is a gradual change in both intelligibility and nasality. Speech quality was judged as acceptable--although slightly nasal-up to a critical aperture size of 5 mm in diameter which
2
equals 19.6 mm . The articulatory errors did not exceed 25 percent at this size opening. When the larger size aperture of 7 mm (38.5 mm 2) was tested, nasality was unquestionable. Articulation errors increased to more than 60 percent, and overall speech was unacceptable. Both nasal airflow and intraoral air pressure were measured for all of the above tube size conditions. The results showed increasing nasal air excape and decreasing intraoral pressure with increasing hole size.

Andrews and Rutherford (1972) designed a study aimed at assessing the contribution of the nasally emitted sound to the perception of nasality. Through the use of special instrumentation they were able to separate the orally emitted sound from the combined orally and nasally emitted sound as they are produced in normal speech. By applying a variable aperture palatal prosthesis with inserts (Figure 7) they created the following velopharyngeal
2 2 2
conditions: no opening, 60 mm , 120 mm , 180 mm , and 240 mm openings. In order to evaluate hypernasality,





-54-


data from perceptual judgment and spectrographic recordings were collected. Perceptual findings show a clear trend of more hypernasality ratings for the combined nasal and oral signal than for the orally emitted alone, especially for the vowel /u/. Spectrographic data supported these findings by displaying clear nasal features on the spectrogram, such as a combination of increased intensity and reduced frequency of the third format for /u/, and extra resonances between F 1 and F 2 for /i/.

The effects of velar perforations on speech were

studied by Liebman (1964). She employed an experimental prosthetic speech appliance that was constructed to simulate these perforations in a variety of sizes and locations, singly and in pairs. Listeners' judgments of nasality and intelligibility in one speaker served as indicators of the influence of these artificial velopharyngeal openings on speech. The general findings indicated that speech became worse when the apertures were larger and placed in a more forward position. Furthermore, if the hole was divided into two openings of equal combined size and located in the same general area, speech was noted to be improved. This finding has not been explained satisfactorily but was suggested to have clinical implications.

The studies that have been discussed up to this point have focused primarily on critical area size of VPI and its influence on nasality and intelligibility. The





-55-


following studies to be discussed will deal with how changes of velopharyngeal port size affect special speech characteristics, such as vowel intensity (Bernthal & Beukelman, 1977), vowel identification and nasality, and whispered vowel spectra (Watterson & Emanuel, 1981a,b).

It is observed repeatedly in the clinic that subjects with VPI speak with a reduction of an overall intensity level, because of increased coupling between the oral and nasal cavities. Bernthal and Beukelman (1977) were able to demonstrate that vowel intensity is a factor of velopharyngeal portal opening. One normal speaking subject was fitted with a palatal prosthesis in which controlled openings at the position of the velopharyngeal port could be manipulated. The result showed that when velopharyngeal openings of 7 mm 2 and 50 mm 2 were present vowel intensity was reduced, especially at the 50 mm aperture.

It was stated earlier that loudness reduction is frequently observed in subjects with VPI. It is, however, uncertain whether this is a result of the intensity reduction or if it is caused by a behavior to try and hide a speech disorder (Morris, 1968). With Bernthal and Beukelrafl's results at hand, it appears that there may exist a close relationship between intensity and loudness in this population. To compensate for this loss of sound energy, speakers with VPI often increase their vocal






-56-


effort, which might abuse their vocal folds (Bernthal & Beukelman, 1977).

It is well known that presentation of loud noise alters auditory feedback and causes changes in speech. The most noticeable change is the increase in vocal intensity. Articulation, however, seems to be unaffected by the introduction of noise, an assumption that was confirmed by Garber, Speidel, Siegel, Miller, and Glass (1980).

A number of authors studied the acoustics of nasalized speech by employing analog methods (Watterson & Emanuel, 1981a). It has been suggested also that an increase in oral-nasal coupling would show reliable and predictable acoustic effects on the spectrogram. When these theories were tested empirically, however, the acoustic effects read from the spectrograms were sometimes much less reliable and predictable (Curtis, 1970; Schwartz, 1979; Watterson & Emanuel, 1981a). It was, therefore, reasoned that one way of avoiding these difficulties would be to test whispered vowels (Watterson & Emanuel, 1981a). By employing whispered vowels, continuous spectral energy would be produced rather than harmonic, which is the result when using voiced speech. Watterson and Emanuel expected to observe systematic formant changes in the spectrograms as the velopharyngeal aperture was manipulated. A variable speech appliance with inserts allcwed for 12.57 mm 2
2 2 2 2
28.27 mm , 50.26 mm , 73.53 mm , and 153.94 mm openings






-57-


in the velopharynx. The results indicated that whispered speech samples did not establish clearly a relationship between aperture size and degree of hypernasality when it was measured with spectrography, nor was there a clear difference between whispered or voiced vowel samples.

Watterson and Emanuel (1981b), therefore, decided to

include a perceptual listening study to measure the effects of oral-nasal coupling on vowel identification and vowel nasality. Both voiced and whispered speech were tested while using the same variable prosthesis as in their previous study. The results indicate that ratings of nasality were more reliable with the voiced than with the whispered test vowels.

The research reviewed above has dealt with determining velopharyngeal port size, introducing artificial VPT through a variable prosthetic appliance, and measuring their effect on speech. They have evaluated speech in terms of perceptual testing of articulation and hypernasality. Measurements on some correlates of speech, such as differential air pressure and nasal airflow, vocal intensity of vowels, vowel identification, vowel nasality, and vowel spectra also were included. All studies concluded that the larger the velopharyngeal port size the more distortion is detected in all of the above speech characteristics, although there is not always a complete linear relationship present.






-58-


Subtelny and co-workers (1961) observed that nasalized speech increased in severity as velopharyngeal openings increased. When, however, sizes were larger than 7.0 rm in size, they saw a leveling off in the nasality ratings. Similar observations were recorded with the intelligibility ratings. Again, speakers with openings from 7.5 mm to 11 nm did not score significantly different from the group having openings from 3.5 mm to 7.0 mm. However, speakers with a gap exceeding 11.5 mm (11.5 mm to 18 mm) showed an increase in loss of intelligibility. The major common finding of all these studies, thus, infersthat an opening between 10 mm 2 and 20 mm 2 through the velar port results in significant distortion of speech intelligibility and resonance quality.



The Effects of Palatal Fistulas on Speech


Orofacial surgical skills have greatly improved over the past decade. Occasionally, however, an opening along the suture line of a repaired palatal cleft will occur (Figure 8). This opening or fistula between the oral and nasal cavities results in inappropriate loss of air through the nose and excessive nasal resonance.

The incidence of fistulas after surgery in cleft

palate patients has varied considerably. Ross and Johnston (1972) concluded that the incidence range was from





-59-


Figure 8. Palatal fistula.






-60-


10 percent to 20 percent, depending on different surgical techniques and cleft type, as well as different sample selections in the available studies. Postsurgical occurrence of fistulas in cleft palate individuals was reported to be as high as 18 percent by Xbyholm and co-workers (1979), while a range between 9 percent and 34 percent was indicated by Henningsson (1983). She mentioned no quantitative indications of fistula size or location, except that these fistulas usually occur along the midline and that size and position determine the degree of the symptoms. In addition to the reported incidence of fistulas following surgery in cleft palate patients, there are also fistulas resulting from ablative surgery and from trauma. No information appears to be available on incidence resulting from congenital anomalies, nor has any information been found on its factual occurrence.

In this context, it should be mentioned, that indi-viduals with severe clefts, such as bilateral complete clefts, would develop fistulas secondary to surgical repair more frequently than patients with a simpler type of cleft. More residual palatal fistulas were reported to have occurred when the von Langenbeck palatoplasty had been performed than after pushback repair (Krause, Tharp, & Morris, 1976; Lindsay, 1971; Palmer et al., 1979). These residual fistulas that occurred after the von Langenbeck procedure were found to be slightly larger than those






-61-


resulting from the pushback technique. The two reports by Lindsay (1971) and Palmer et al. (1969) were in agreement that the residual fistulas seemed to be occurring mostly in the anterior and alveolar parts of the hard palate.

In contrast to this information, Reid (1962) described fistulas as being small, those less than two centimeters in diameter, while a large hole would be greater than two centimeters, with its location typically along the midline. He further commented that such a large residual fistula would follow pushback surgery. Witzell, Clarke, Lindsay, and Thomson (1979) also compared results following the von Langenbeck palatoplasty and the pushback procedure. The two surgical procedures compared positively when no residual fistulas occurred.

With the availability of ample information on VPI and its effect on speech, it might be concluded that similar data would be available for palatal fistulas and their effect on speech. However, in a review of the literature, the only studies found were descriptive and qualitative reports. These studies mainly dealt with incidence and descriptions of surgical techniques. Case histories were also reported, including general descriptions of fistulas. It was suggested that fistulas alone, or in association with VPI, contribute to speech-crippling effects similar to that of an unrepaired cleft palate (Millard, 1980).





-62-


As mentioned in Chapter I, it is unclear whether

fistula location is important. It has been suggested that anteriorly located fistulas exert significant negative influence on speech production (Cosman & Falk, 1980). Other authors merely indicated a possible existence of speech disorders resulting from anteriorly located fistulas (Ross & Johnston, 1972). Posteriorly located fistulas were, however, declared to cause cleft palate-like speech (Ross & Johnston, 1972).

Palmer et al. (1969) and Lindsay (1971) stated in their respective studies that all postsurgicaily occurring fistulas were situated in the anterior portion of the hard palate. These anteriorly located fistulas were said to be a source of audible nasal air emission, impairing articulation proficiency. These fistulas also were said to cause slight nasality (Lindsay, 1971; Palmer et al., 1969). A report by Musgrave and Bremner (1960) indicated that many of the complications following the second stage of palatal repair occurred as midpalatal fistulas. No indication was given as to how or if any of these fistulas influenced normal speech production.

Fistula size has been described in a qualitative and descriptive manner. Clinical experience indicates a definite increase of speech unintelligibility and nasal resonance with increasing fistula size. This has been substantiated in the literature (Gordon & Brown, 1980;





-63-


Henningsson, 1983; Morley, 1962; Reid, 1962; Shelton & Blank, 1984). However, few authors have defined the size of fistulas in relation to speech disorders. Those authors who quantitatively defined fistula size were all describing fistula size in relation to surgical techniques (Henderson, 1982; Proctor, 1969; Reid, 1962).

Palatal fistulas were categorized grossly as small, moderate-sized, or large. Reid (1962) described small holes as being less than two centimeters in diameter and a large hole as being more than two centimeters in diameter. Speech defects, apparently, were registered only with the larger fistulas. These large fistulas were reported to cause nasal air escape and an increased nasal voice quality. Proctor (1969) defined large oronasal fistulas to be 0.5 centimeter to 2.5 centimeters in size. No other sizes were, however, discussed. Henderson (1982), on the other hand, reported an approximal definition of small fistulas (0.5 centimeter to 1.5 centimeters in width and 0.5 centimeter to 2.0 centimeters in length), but failed to offer any information on large fistulas. Gordon and Brown (1980), who discussed very small and moderate-sized fistulas, explained the very small hole to be less than a few millimeters, but declined to define the moderate-size defects.

An article by Shelton and Blank (1984) describes the

single relevant study where palatal fistula size is related






-64-


to speech impairments. However, the hole sizes are not quantified but simply described as small, moderate, and large. These authors determined that small palatal fistulas did not exhibit any nasal airflow during speech. The intraoral air pressure necessary for normal speech production could be maintained in the presence of small fistulas. It was found that the subjects having moderate fistulas could retain the necessary intraoral air pressure at the expense of nasal air emission. All subjects with large palatal fistulas developed nasal air emission and a reduction of intraoral air pressure. The presence of nasal air emission and the reduction of intraoral air pressure were found to be detrimental to the production of adequate articulation (Shelton & Blank, 1984).



Research Questions


Both clinical observation and much research indicate

that palatal defects resulting in abnormal openings between the oral-nasal cavities do, in fact, greatly influence articulation and resonance quality of speech. This, of course, depends largely on the size, shape, and location of the opening. Little research is reported which defines the influence of incomplete separation of the oronasal cavities anterior to the velopharyngeal port. It is, however, clear that with a fistula present, leaving all other structures intact or fully repaired, we can expect





-65-


a breakdown of intraoral air pressure, as well as air excape into the nasal cavity and through the nose. This affects both resonance quality and articulation of speech.

With the inference from above reviewed research one must ask the following questions.

QUESTION ONE

Can a fistula in the hard palate cause similar
speech impairments as those resulting from VPI?

It has been shown that an opening in the velopharyngeal port negatively affects normal speech production. One would consequently assume that any other abnormal opening, such as a palatal fistula, connecting the oronasal cavities, would likewise impair normal speech.

QUESTION TWO

To what extent would a speech impairment vary
with differing sizes of a palatal fistula?

Previous research studies of velopharyngeal closure have shown that speech becomes progressively more distorted as the velopharyngeal port size is increased. It would seem reasonable to predict, and it is clinically known, that this is also the case with palatal fistulas. However, since there are no definable quantitative data available, it would seem pertinent to begin defining fistula size, measured in square millimeter, and to measure the effect that different size fistulas exert on speech production.

QUESTION THREE

Is there a critical fistula size at which point
speech can be measured or judged as being distorted?





-66-


Intraoral air pressure has been found to decrease rapidly when velopharyngeal openings exceeded 10 mm2 The critical size of velopharyngeal closure was suggested to occur at the point which distinguishes velopharyngeal adequacy from inadequacy. This was found to begin at approximately

22
a 10 to 20 mrm range (Warren, 1964a,b). It seems important to find out if a similar range exists with palatal fistulas or if the critical size here would be smaller or larger.

QUESTION FOUR

To what extent would a speech impairment vary with
different positions of a palatal fistula?
Anterior positions of velar openings have been demonstrated

to cause flower speech distortions than posteriorly positioned ones. Regarding fistulas in the hard palate, there seems to be no relevant information indicating the influence of fistula location. one report indicated that a posterior location of a fistula caused more speech problems than an anterior location. However, in this context there has been no description on the exact fistula location and how it would influence speech. First of all, it is of interest to examine the significance of the localization between VPI and hard palate fistulas. Secondly, a definition of significant differences among fistulas located in the hard palate would seem pertinent.

Millard (1980) reported that surgical closure of a fistula often can be extremely difficult to obtain.






-67-


Patients with residual palatal fistulas are fitted instead with an obturating device to achieve normal speech and to eliminate food and liquids from passing to the nasal cavity, as well as preventing nasal fluids from entering the mouth. It is the intention that, with the results from this study, quantified definitions of palatal fistulas will facilitate the choice of treatment for these individuals. The primary goal of this work, however, is to perform basic research to obtain quantitative definitions of fistulas in the hard palate and their influence on speech. In other words, the purpose of this study is to begin defining the influence that size and location of an opening through the hard palate has on speech in terms of articulation and resonance quality.















CHAPTER III
METHODS



Introduction


Measurement of oral speech in the presence of a palatal defect was discussed in the previous chapter. The basic speech characteristics related to palatal disorders are nasality and articulation. Validity and reliability, as reviewed in Chapter II, were both difficult to obtain in the severity ratings of hypernasality. Both intra- and interrater reliability were evidenced to be poor. Regarding the perceptual judgment of articulation, intrarater reliability was good but interjudge reliability were low-.

Due to the ambiguity of perceptual judgment-testing of nasality, additional tools were developed and tested in the hope to obtain one single objective measure of resonance quality. The consensus of the literature reviewed above was that no single instrument exists that exclusively documents nasality (Van Demark et al., 1985; Fletcher et al., 1974). The recommendation has been to employ a combination of methods to obtain adequate nasality measurements (Reich & Redenbaugh, 1985). These combinations usually contained subjective testing, employing rating scales or standardized tests of articulation




_68-






-69-


and hypernasality, together with an objective instrumental technique.



Materials and Procedures


To answer the questions posed in Chapter II, palatal openings mimicing fistulas had to be experimentally created. To provide a range of sizes and locations of palatal openings, a research prosthesis for one subject was constructed. The subject has an unrepaired cleft palate and wears a removable palatal appliance. To define and quantify the effects of different palatal openings on speech, a battery of tests was included. These tests can be divided into two categories: (a) subjective perceptual testing and (b) instrumental testing.



Subject


One young adult male with complete bilateral cleft of the hard and the soft palates was employed for this study. This individual is a 28-year-old Caucasian professional. He speaks English with general American dialect, holds a master's degree in speech pathology and a Ph.D. degree in education. He wears a removable palatal appliance and his speech is perceived as normal. The subject was born with a bilateral cleft of the hard and the soft palates, the alveolar ridge, premaxilla, and the





-70-


upper lip (Wharton, 1978). Surgical closure of the lip was performed shortly after birth with several secondary reconstructive repairs at later ages. The hard and the soft palates have never been closed. It was explained that the hard palate was never surgically closed because of the width of the palatal cleft. Instead, the subject was fitted with a removable palatal prosthesis at age three in order to obtain separation between the nasal and the oral cavities. His speech was judged to be extremely good, only four months after fitting of his palatal appliance, by the cleft palate team at the University of Florida teaching hospital. Today the subject still wears a removable palatal appliance and his speech is perceived as normal. An audiometric evaluation of the subject's hearing prior to conducting any testing indicated normal hearing bilaterally (Appendix A).



Research Prosthesis


A duplicate of the subject's own presently existing

removable palatal prosthesis was constructed by a prosthodontist in the Department of Removable Prosthodontics at the University of Florida J. Hillis Miller Health Center. The appliance consists of two parts. One is a metal frame with a special design for retention of the appliance in the subject's mouth. The other part is an acrylic resin





-71-


portion which fits accurately to the cleft area. After the construction, which required three fitting sessions, the subject was instructed to wear and adjust to the appliance, in order to overcome any functional differences that may have yet been present. After 10 days, it was decided that the subject had adapted well to this new appliance. This was independently judged by two speechlanguage pathologists. The subject's speech was free from articulation errors and his resonance quality was normal. The subject himself expressed satisfaction with his speech intelligibility, resonance quality, and the ease with which normal speech could be produced.



Test Conditions


To permit manipulation of the size and location of openings anterior to the velopharyngeal port, holes were drilled through the appliance, resulting in an opening between the oronasal cavities. Speech was then tested, using one hole at a time. The locations of the openings were positioned along the midline, in the front onethird (A), in the middle one-third (B), and in the back one-third (C) of the hard palate part of the prosthesis. A fourth location of holes was drilled in the middle of the speech bulb (D), which is also the velar part of the prosthesis (Figure 9). The area sizes of the openings were 5 mm 2 (1), 10 mm. 2 (2), 20 mm. 2 (3), and 20 mm 2 (4). The area sizes were selected on the






-72-


o 5 mm o 10 mm2


0


2 0 mmn 30 rm,2


A =Anterior of the hard palate B = Middle of the hard palate C = Posterior of the hard palate D = Velar


Figure 9. Sizes and positions of experimental fistulas.





-73-


grounds of previous research on critical points of velopharyngeal insufficiency and speech. Warren, Dalston, Trier, and Holder (1985) concluded that these sizes ranged from 1 mm 2 to 9 mm 2 for adequate closure, from 10 mm 2 to 19 mm 2 for borderline closure, and inadequate closure was
2
registered when the VPI was larger than 19 mm

Each opening was drilled one at a time and resealed after the entire test battery had been completed for that particular opening. Self-curing resin was used to reseal the holes to secure an airtight seal. Altogether there were 16 experimental test conditions, combining the four area sizes (1, 2, 3, and 4) with the four different positions (A, B, C. and D). All holes were drilled in a random order, taking into consideration the physical limitations of the prosthesis. There was concern due to the possibility of fracture of the partition between the largest sizes of positions B and C. This may have influenced the randomization. The testing order for all conditions is presented in Appendix B. All tests were also conducted under three baseline conditions. The baselines included testing with the subject's normal, original prosthesis (BN) in place. Baseline testing of the research prosthesis (BR) was performed before any of the experimental openings were drilled. A post-test baseline measurement (PBR) was also included. At the PBR testing, all openings had been securely plugged with resin.





-74-


Human Experimentation Approval


The study was granted Human Experimental Approval by the Institutional Review Board at the Shands Medical Center. The subject was informed that his participation in this study would not include any known discomfort or anticipated risks. An informed consent to participate in research form was read and signed by the subject.



Test Environment and Audio Recording


Testing was performed with the subject comfortably

seated in an acoustically sound-proof research unit at the University of Florida J. Hillis Miller Health Center. A tape recorder (Sony, TC-158 SD, Dolby system), with its 300 ohm dynamic microphone (Thoro Test) and cassette tapes (Scotch AVX-50),was used to record all speech samples. The Bioengineering Service tested the recording unit through the microphone and found it to present a "MIC" input constant of 0.005V RMS with a drive level of -3 dB. As the unit was tested using a sound level meter, it was found that all values up to a frequency of 6500 Hertz were normally represented by the tape recording with a tapering off to 8000 Hertz, whereafter there was a sharp drop in the quality of tape and recording reproduction.

The microphone was securely fastened to a microphone stand and placed at a distance of 30 centimeters from the





-75-


mouth and at a 90-degree angle of incidence. This position

of the microphone minimized distortions of the acoustic voice signal by respiratory airflow noise. The VU-meter or volume indicator on the tape recorder was adjusted and monitored to control the intensity level (loudness) by the investigator. No attempt was made to control the loudness or pitch of the subject's voice. The subject was able to hold a general level of vocal production (syllables, words, and sentences) with only a negligible change of decibels throughout the testing, which was sufficient for this purpose. It was necessary that all the equipment be located within the testing booth, together with the subject and the investigator.



Perceptual Measurements


There was a total of five perceptual tests in the speech evaluation. These consisted of (1) one standardized test of articulation, (2) one articulation rating test, (3) one standardized test of hypernasality without auditory masking noise, (4) one standardized test of hypernasality with auditory masking noise, and (5) one hypernasality rating test. The standardized tests listed above, (1) and (3), are tests employed at the University of Florida's Cranlofacial Center and the Department of Communicative Disorders. The two perceptual rating tests were constructed in an






-76-


attempt to evaluate perceptual scaling judgments and to compare them to the other testing tools.

Articulation. The Bzoch Error Pattern Screening Articulation Test (BEPSAT) (Bzoch, 1979), previously described in Chapter II, was employed for the articulation testing (Appendix C). The BEPSAT was selected because it is a standardized test and provides valuable information of the articulatory proficiency of all distinctive feature groups.

The perceptual rating of articulation proficiency in connected speech was performed by judging four sentences. These sentences are listed in Appendix D. The sentences represent a sample of sentences frequently used in the evaluation of speech disorders related to palatal disorders at the University of Florida Craniofacial Center. Judgments of these sentences were conducted by employing a five-point ordinal rating scale (Appendix D).

Hypernasality. For the evaluation of nasality, the

standardized Cul-de-sac Test of Hypernasality was administered (Appendix C). This test was described in detail in Chapter II. This test was repeated a second time while the subject was exposed to auditory masking noise. The reason for this was to rule out any effect from compensation for nasality due to auditory feedback (Coleman, 1963/1964; Andrews & Rutherford, 1972). To obtain this new test environment, a second tape recorder, a Sharp Educator





-77-


#376895 TLR with headphones (Thoro-Test SHP-2, 8 ohm), was employed in the experiment. The masking noise consisted of a prerecorded multitalker tape. Twenty speakers were recorded simultaneously on this tape. It was stated that "the recording is unintelligible and the level is stable over time, having a relatively flat spectrum (-2.5 dB) from 200 to 500 Hz" (Holmes, Frank, & Stoker, 1983, p. 88). The tape was obtained from a University of Florida, Department of Speech, audiologist. The multitalker noise tape was presented to the subject binaurally at a sound pressure level of 90 decibels as it was adjusted with a sound-level meter (Davis, 1970).

The second test of hypernasality consisted of the

perceptual rating of eight paired consonant-vowel-syllables (CVC-syllables). The CVC-syllable "pin" was paired with another CVC-syllable that contained the plosive consonant sound /p/, the vowel sound /i/ followed by one each of the six plosive sounds /p/, /b/, /t/, /d/, /k/, and /g/. There were, thus, six different syllable combinations. The vowel /i/ was selected because, as a high vowel, it has been demonstrated to be the sound where the least velopharyngeal opening was observed, thus indicating a high level of intraoral air pressure (Moll, 1962; Watterson & Emanuel, 1981a). Two of the pairs "pin-pib" and "pin-pip" were repeated, but not in consecutive order. This repetition was included to perform a reliability test among and





-78-


between the judges. The six judges rated nasality according to a five-point ordinal rating scale (Appendix E).



Procedures of Data Collection


The speech samples obtained through the administration of the articulation and the hypernasality tests were taperecorded. There were five different tests altogether--two articulation tests and two nasality tests, of which one nasality test was repeated in a masking noise environment. With the above five tests and the 19 different conditions under which they were tested, a total of 95 test objectives was to be perceptually judged. A test tape was then arranged with the 95 tests occurring in random order. Due to the repetitious nature and the bulk of this test battery, it is not included in its entirety in the dissertation.



Judges


The recorded speech samples were subject to perceptual judgment. Six speech-language pathologists served as judges. They were selected because of their experience in working with resonance and articulation disorders associated with VPI. All but one of the judges were able to evaluate all 95 tests in one sitting. Each judge spent an average time of one hour and 30 minutes for this





-79-


evaluation. The range was from one hour and 20 minutes to one hour and 45 minutes. The 95 tests were recorded on three tapes. Tape one contained tests one through 38, tape two tests 39 through 78, and tape three the tests 79 through 95. Each judge was presented with a different sequence of the three tapes, as shown in Appendix F. This was arranged to avoid a tiring influence on the judges' response.

Prior to testing, each judge was familiarized with

the four tests and read a set of instructions (Appendix G). The judges were told that they could stop the test tape, using a stop button on the tape recorder, should they need to stop for replay or for taking a rest. The judges were not informed of the masking feature in some of the material.

Each judge was required to make a total of 589

(31 x 19) articulation test judgments and to rate articulatory proficiency from 76 (4 x 19) connected speech samples (sentences) with a five-point scale. Additionally each judge was asked to make 380 (2 x 10 x 19) judgments on the hypernasality test (190 responses from the unmasked version and 190 responses from the masked version). The rating of hypernasality included judgments of 171 (8 x 19) paired CVC-syllables based on a five-point scale. Variation in the listening environment was minimized and at all






-80-


times judging was performed in a closed room with only the judge and the experimenter present.



Instrumentation


Three instrumental tests were employed to further assess the effect of palatal openings on speech. These tests represented measurements of nasal airflow, aerodynamic measurements of intraoral air pressure, and differential pressure. The first test, the Nasal Emission Test, supplies the airflow data. The second and third tests were performed by employing the PERCI II instrument which provides the measurements of intraoral air pressure and differential pressure.

Nasal airflow. The data on nasal emission of air were collected by employing the Nasal Emission Test as seen in Appendix C. This is a standardized test employing simple and accurate means to measure the presence or absence of nasal air emission. The test is employed frequently at the University of Florida's Cranofacial Center and the Department of Communicative Disorders. The test was described in detail in Chapter II.

Aerodynamic measurements. objective instrumental

measurements were next applied to measure the effects of palatal openings. The aerodynamic entities that were selected for testing, which are closely related to normal






-31-


speech production, were intraoral air pressure and differential pressure. Measurement of nasal emission of air was also attempted but the results were inconclusive. They were, therefore, excluded from the study (see section of PERCI below, page 82).

The data from the intraoral air pressure and differential pressure measurements were gathered to increase the reliability of the perceptual tests. Furthermore, they were collected to determine whether any of these measures could be singly employed for nasality detection.

To compute intraoral air pressure and differential

pressure an instrument, PERCI II, was employed. The function of this instrument was described in Chapter II. To make PERCI II operative, two separate pressure transducers and a pneumotachograph had to be connected to the instrument via catheters. These catheters had to be of different diameters, depending on what part it would serve.

For testing of intraoral air pressure during speech,

an oral catheter, two millimeters in diameter, was attached to a Stratham PM5ETC pressure transducer, which in turn was connected to the pressure channel on PERCI II. Before testing was begun, the pressure channel of the unit was calibrated with a water manometer, constructed by the investigator.

For the differential pressure test, an additional

catheter was added to the pressure transducer. The catheter






-82-


size was four millimeters in diameter. It was threaded onto a nasal olive that was to be placed in one of the subject's nostrils during testing (Figure 6).

It was planned to include an additional test of nasal emission with this instrument. It was thought that such results obtained by instrumental testing would enhance the results from the perceptual testing of nasal emission. To measure nasal emission, the flow channel of the instrument had to be made operative. Nasal flow is measured with a nasal catheter with an airtight fit into one nostril. The other end of the catheter is connected to a Fleisch pneumotachograph (pneumotach) Type One. The pneumotach is, in turn, connected to a heater power supply (filament transformer of 6.3 Volt at 1.2 ampere), used to heat the mesh screen inside the pneumotach cylinder. This is necessary to prevent condensation from forming on the screen, which would lead to an erroneous increase in the airflow signal (Microtronics Corporation, n.d.). The small pressure differential developed across the screen in the pneumotach is then detected by a Stratham PM15ETC pressure transducer and the signal is sent to PERCI II for read-out.

Calibration of the flow channel was performed by connecting the pneumotach to a captured air source. A flow gauge (Fischer and Porter Company Precision Bore Flowrator Tube No. FP-1/2-17-G-10/27 scc/sec-air met. @ 14.7 P.S.I.A. & 700) with a scale range of 500 cubic





-83-


centimeters per second was connected to the output side of the pneumotach. The calibration of the flow channel was not linear. Deviations were occurring in the lower ranges. The instrument could be calibrated only at the ranges from 500 cubic centimeters through 300 cubic centimeters per second. At the level of 250 cubic centimeters the calibration errors were measured at 10 units on PERCI II. At 200 cubic centimeters, 100 cubic centimeters, and at 50 cubic centimeters the errors were 20, 40, and 50 units offf respectively. It was expected that the results could be adjusted to this error in calibration. However, during testing, it was noted that the errors were too grave and erratic. Therefore, due to this internal calibration instability, this measurement of nasal air emission was excluded from the study.



Procedures of Data Collection


Data from the Nasal Emission Test were collected by

the examiner at the time of the original testing. A small paper paddle was held to the subject's nose while he produced each of the 10 words from the test. The test was repeated twice for reliability. Any movement of the paper paddle, indicating nasal air flow, was registered. This test should yield a score of zero movement of the paddle to record no nasal air leakage. The data from this test






-84-


of nasal emission were collected by the examiner and repeated twice for the three baselines and the 16 experimental test combinations.

The procedures for testing intraoral air pressure and differential pressure called for employing the instrument PERCI II. The two tests were measured during speech for each of the 19 palatal conditions. The test with PERCI II followed the perceptual testing for each baseline and palatal opening.

The speech sample employed was one of those suggested in the PERCI manual (Microtronics Corporation, n.d.). The speech sample for both intraoral air pressure and differential pressure was the word "pop." Five readings were recorded by the investigator. It was explained in the PERCI manual that words and phrases containing the plosive sound /p/ were preferred. When these sounds are produced, the mouth is closed and a stagnant column of air is created. This is desirable because a normal measurement during the baseline tests would rule out any velopharyngeal incompetence, which was especially necessary for the accuracy in the experimental tests with the research prosthesis. These sounds, furthermore, eliminated any problems that could occur from tongue placement. For example, if the place of articulation would occur behind a palatal opening, the true speech effect of that opening would not be accurately documented. By employing plosive sounds the effect of respiratory effort





-85-


is also controlled in the presence of palatal openings. In the testing of differential pressure, a ratio of oral and nasal pressure is measured and the respiratory effort is cancelled out, because it affects both pressures equally. It was expected that, with a negligible VPI, the measurements of intraoral air pressure and differential pressure could be used accurately to measure the influence that defects in the hard palate, such as fistulas, would have on these parameters important for normal speech production.



Reliability Testing


The conclusion of the reviewed literature on perceptual judgment rating of nasality showed low intra- and interreliability (see Chapter II). Evidence from perceptual

rating of articulation had been found to indicate good intrareliability but less reliable interagreement. Other perceptual tests, such as the employment of standardized tests for articulation and hypernasality, had not demonstrated these reliability problems.

The knowledge of these findings was the prerequisite for the reliability testing, on the hypernasality rating test that was performed in this study. As described above, there were six different paired CVC-syllables in the test. Two of these CVC-syllable pairs were selected for repetition for the purpose of reliability testing, thus increasing





-86-


the total number of tested syllable pairs to eight. The repeated syllables were "pin-pip" and "pin-pib." These CVC-syllables were nestled into the list of the original test syllables. The judges were not instructed about this procedure. No one of the judges commented on the fact that they had identified the repeated syllable-pairs in the test material. The reliability data were extracted from the total number of responses on the hypernasality rating test and were treated separately. Both intrareliability and interreliability were analyzed with the Spearman rank order correlation procedure.



Replication Study


The velar opening (D) has been treated as one of the experimental palatal openings. The same hole sizes were drilled and the same series of tests were employed for each of the D-openings. These D-openings were, furthermore, intermixed in the sampling regarding test order and no differences in data collection were considered.

Nevertheless, D was originally included for comparison studies with previous results from research on VPI. It was intended that if the D-results compared favorably to previous results on VPI and speech, a common ground could be established for comparing the effects of VPI on speech with those from palatal fistulas.





-87-


Data Analysis


The data collected from the above-described testing

procedures were subjected to statistical analysis. The data were organized to permit comparisons between the different test implementations. The choice of appropriate statistical tests was discussed with a statistics consultant from the Department of Biostatistics at the J. Hillis Miller Health Center. To accurately answer the research questions in Chapter II, the Statistical Analysis System (SAS) was employed (Statistical Analysis System Institute, Inc., 1982ab). Because of the complicated nature of the data, each of the perceptual tests was computed independently. The aerodynamic test scores of intraoral air pressure and differential pressure were computed together. The program of statistical analyses included means, standard deviations, and analysis of variance (Marks, 1982ab). Duncan's multiple range test was employed for a posteriori pairwise comparison among the means (Kirk, 1968). Its purpose is to further identify which means are different from each other after the analysis of variance has been performed. if significant differences can be identified with this a posteriori test, there is a high degree of confidence in the conclusions. To determine the degree of relationship between the variables Spearman's rank order correlation coefficients were computed.















CHAPTER IV
RESULTS



This study represents an attempt to start defining and quantifying the influence of palatal fistulas on speech. Specific sizes and locations were obtained by systematically manipulating artificial openings in a removable palatal prosthesis. Four locations along the palatal midline were utilized and four artificial fistula sizes were used. These ranged from 5 mm 2 to 30 mm 2. To address the questions posed in Chapter II, data were gathered from several tests of speech and from measurements of the speech correlates of nasal air emission, intraoral air pressure, and differential pressure. There were seven types of data obtained from these tests and measures. Five of these data sets were collected from two perceptual tests of articulation, two perceptual tests of hypernasality, and one set of data originating from an instrumental pass-fail test of nasal air emission. The last two data sets were obtained from instrumental test measurements of intraoral air pressure and differential pressure.

The purpose of this study was to quantify and define palatal openings by their influence on speech through descriptive statistics, rather than to analyze speech as


-88-






-89-


it was affected by apalatal openings. The objectives were to define the critical size/location of the different openings, whether speech was negatively influenced in any way, and to quantify each opening size/position by the influence it exerted on speech behavior. Therefore, each separate test was analyzed relative to error production on the perceptual listening tests. The instrumental tests were analyzed relative to the necessary intraoral air pressure and differential pressure for maintaining normal speech production and to quantitatively correlate each size/location to its deviation. The presence of any nasal emission of air was recorded, and the data from two readings were analyzed.



Perceptual Measurements


There were four perceptual tests in the test battery

judged by six judges. The tests employed were (1) the Bzoch Error Pattern Screening Articulation Test, (2) the Articulation Rating Test, (3) the Cul-de-sac Hypernasality Test, and (4) the Hypernasality Rating Test. The Cul-de-sac Hypernasality Test was used twice, the second time with auditory masking, thus increasing to five the actual number of evaluated perceptual tests.

Each of these five tests was assessed with the three

baselines (EN, BR, and PER) and the 16 experimental openings





_90-


2 2
(At B, C, and D locations and the sizes 5 mm , 10 mm
2 2
20 mm " and 30 mm ) in random order.



Articulation Test


The response variables from the BEPSAT were determined by assessing all test sounds contained in each of the 31 words. The number of correct articulations was extracted and used in the data analysis as the cutoff score that was to identify complete articulatory proficiency versus nonproficiency. No differentiation of error type was determined.

The means and standard deviations of correct articulation among the six judges for all 19 test conditions are shown in Table 1 and Table 2. Figure 10 and Figure 11 illustrate these results graphically by size and position.

The analysis of variance indicated significant differences between the means (F = 2.12, df = 15, p < 0.01). Further analysis of the independent variables, position and size, did not show significant differences between the positions (F = 0.15, df = 63, p = 0.92). The differences between the sizes were, however, significant (F = 8.69, df = 3, p < 0.01). On a size-by-size analysis, the a posteriori Duncan Multiple Range Test was employed to identify further differences among the sizes. It indicated that the 5 mm 2 openings were not significantly different from the baseline measurements. However, all larger











Table 1. Means of normal articulation scores among the
six judges (Articulation Test; n 31).


Size (mm 2

Position 5 10 20 30

A 28 23 15 23
B 29 26 18 20
C 27 24 21 20
D 26 25 20 21


Baselines.

BN BR PBR

27 29 30


Table 2. Standard deviations of the normal articulation
scores among the six judges (Articulation Test;
n = 31).


Size (mm 2

Position 5 10 20 30

A 3.57 6.83 8.95 6.24
B 2. 73 4.11 9. 83 7.90
C 4.49 6.25 7.03 6.33
D 3.31 5. 89 7.17 9.21


Baselines

BN BR PBR

3.88 3.09 0. 81


-91-




Full Text

PAGE 1

EFFECTS OF EXPERIMENTAL PALATAL FISTULAS ON SPEECH AND RESONANCE By ULLA E.M. RICHTNER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UN I V ERSIT Y OF FLOR IDA I N PARTI A L FULF I L LMEN T OF THE REQU I R EMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1986

PAGE 2

Copyright 1986 by Ulla E.M. Richtner

PAGE 3

To Maj and Nils

PAGE 4

ACKNOWLEDGMENTS My appreciation and gratitude are expressed to my committee chairman, Dr. William N. Williams, for suggest ing the topic and for kindly making available to me needed laboratory facilities and the resources of a Biomedical Research Support grant. Gratitude is also extended to the other committee members, all of whom showed trust in me and the project. Dr. Thomas B. Abbott also has been my adviser and his encouragement and assistance have helped me overcome many hurdles, not only in my doctoral research but throughout my graduate career. Dr. Kenneth R . Bzoch has shared generously his vast knowledge in the area of cleft palate. Dr. G. Paul Moore has provided me with invaluable clinical insights in the area of voice disorders. Dr. Nikzad s. Javid offered his professional s kill and cooperation in construction of the research prost h esis. Special thanks go to Dr. Paul W. Wharton who volunteered his time and effort and without whom this work could not have been completed~ I owe special gratitude to Dr. Ronald G. Thomas for his generous and invaluable help in guiding me through the iv

PAGE 5

labyrinth of statistics and computer use. Special thanks also are due the six judges who donated their time and expertise for the perceptual listening task. I extend especially warm thanks to Marcia Buchanan not only for her editorial assistance but for her personal interest and con cern . I would a lso like to take the opportunity to thank everyone from the Dep artments of Speech and Oral Biology who shared their professional skills, positive attitudes, and research equipment. Rotary International and ihlen -stiftelsen of Sweden contributed to the completion of my graduate studies by their generous stipend contribution. V

PAGE 6

TABLE OF CONTENTS Page ACKNOWLEDGMENTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv ABSTRACT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii CHAPTER I II III IV INTRODUCTION AND PURPOSE ...................... . 1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Surgical Treatment of Oronasal Openings........ 5 Prosthetic Treatment of Oronasal Openings...... 9 Cleft Palate and Speech........................ 10 Velopharyngeal Insufficiency and Speech........ 10 Palatal Fistulas and Speech. . . . . . . . . . . . . . . . . . . . 13 Statement of Purpose.......................... 16 LITERATURE REVIEW AND RESEARCH QUESTIONS ...... . 20 Introduction................................... 20 Palatal Embryology and Anatomy................. 21 The Effects of Velopharyngeal Insufficiency on Speech. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Perceptual Judgment............................ 34 Instrumentation................................ 39 The Effects of Controlled Velopharyngeal Insufficiency on Speech...................... 51 The Effects of Palatal Fistulas on S p eech...... 58 Research Questions............................. 64 METHODS ....................................... . 68 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 8 Materials and Procedur e s....................... 69 RESULTS ....................................... . 88 Perceptual Measurements........................ 89 Instrumental Measurements...................... 116 Reliability .................................... 133 vi

PAGE 7

V DISCUSSION AND CONCLUSIONS ..................... 138 Discussion..................................... 139 Conclusions.................................... 151 APPENDIX A AUDIOMETRIC EVALUATION. . . . . . . . . . . . . . . . . . . . . . . . . 155 B ORDER OF THE EXPERIMENTAL TEST CONDITIONS ...... 156 C CRANIOFACIAL CENTER SPEECH TESTS ............... 157 D ARTICULATION RATING TEST. . . . . . . . . . . . . . . . . . . . . . . 159 E HYPERNASALITY RATING TEST...................... 160 F SEQUENCE OF THE THREE TEST TAPES PRESENTED FOR PERCEPTUAL JUDGMENT OF THE 95 TESTS (FIVE SPEECH TESTS AT 19 FISTULA CONDITIONS) ......... 161 G INSTRUCTIONS FOR THE PERCEPTUAL JUDGMENT PERFORMANCE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2 BIBLIOGRAPHY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 4 BIOGRAPHICAL SKETCH. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 vii

PAGE 8

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EFFECT OF EXPERIMENTAL PALATAL FISTULAS ON SPEECH AND RESONANCE by Ulla E.M. Richtner August, 1986 Chairman: William N. Williams Major Department: Speech Palatal fistulas are frequently the result of congenit al anomalies, tr auma , or tissue breakdown following cleft palate repair. It has been observed clinically that speech articulation and voice resonance are impaired as the result of an opening between the oral-nasal cavities. Extensive documentation of ve lophar yngea l insufficiency and its influence on speech suggest that a 10 mm 2 to 20 mm 2 opening through the ve lar port results in significant distortion of speech. However, little quantitative information is available c once rnin g similar effects of palatal f istulas. The purpose of this study was exp erimentally to manip ulate size and location of open i ngs through a subject's prosthetic palatal prosthesis and to measure the effect of these openings on the individual's speech articulation, resonance quality, nasal airflow, intraoral air pressure, and differential pressure. Four experimental openings (fistulas) of the respective sizes 5 mm 2 , 10 mm 2 , 20 mm 2 , viii

PAGE 9

and 30 mm 2 were drilled in a removable palatal appliance covering a subject's complete cleft of the hard and the soft palates. The four locations of these openings were in the anterior, middle, and posterior part of the hard palate and one velar opening, all along the midline. The effect that each hole size and location had on the individual's speech and resonance was evaluated perceptually by six judges. Nasal airflow, intraoral air pressure, and differential pressure were measured instrum en tally. Analyses of the data revealed that both hole size and location of a fistula determine the degree of speech impair ment. The findings indicate that the g reat es t distortion of articulation a nd resonance quality was perceived by the jud ges when th e f istula was loc a t ed in the velar port and at the most anterior position of the hard palate (approximately at the incisi ve foramen). The size of a fistula was also found to be a significant variable with greater distortions occurring with increased openings. Even the smallest fistula of 5 mm 2 caused mild distortion to the speech signal which was statistically significant in the three instrum ental measurements. The five perceptual tests indicated that speech, in terms of articulation proficiency and nasality, was significantly distorted when the fistula size was between 10 mm 2 and 20 mm 2 . ix

PAGE 10

CHAPTER I INTRODUCTION AND PURPOSE Introduction The production of oral speech sounds is dependent upon a complete separation between the nasal and oral cavities, thus directing the vocal stream from the larynx through the oral cavity a lone. This separ a tion between the oral and nasal cavities is primarily a natomic a l. The structure separating the two is the palate. It consists of bone, muscle, and mucosal tissue which forms the roof of the mouth and th e floor of the nasa l cavity. Complete closure depends on the a ction of the soft pa late or velum. The soft palate is elevated through muscular action to make complete contact with the posterior pharyngeal wall. When the soft palate is in relaxed position, as during nasal breathing or production of nasal sounds, air is directed into the nasal cavity. Thus, an opening in this separation, either in th e stationary hard palate or in the movable soft palate, will be detrimental to normal oral speech sound production, since it will allow undesirable airflow into the nasal cavity. This airflow results in a reduction of intraoral air pressure which is an important physiological component for the production of different speech sounds -1

PAGE 11

-2(Brown & McGlone, 1969). Leakage of air into the nasal cavity also causes audible, inappropriate nasal air emission during articulation of consonant sounds. In addition, a shift in resonance quality, due to acoustic damping related to the coupling of the nasal cavity with the oropharyngeal cavities, may occur which is perceived as hypernasality. Clinical observation has revealed that the magnitude of the effect of these two factors (hypernasality and nasal air emission) on speech intelligibility is negatively influ enced by the increasing size of a palatal opening. This means, then, that the larger the palatal opening, the greater the degree of speech impairment. Articulation problems likely to occur concomitant with palatal openings are misarticulations, ranging from indistinct production due to slight air emission to complete omission of the sound. Distorted sound production, simple sound substitu tion, gross sound substitution, and complete omission of the sound are influenced by nasal air emission and hyper nasality (Bzoch, 1979). There are three conditions that could result in a physical breakdown of the palatal structure. The first condition is a cleft occurring in the hard and/or the soft palate. The type of cleft is dependent upon the location and size of the cleft itself. The most common cause of cleft palate is when atypical embryological development prevents normal fusion along the midline of the palate. Congenital overt clefts of the lip and/or palate occur at

PAGE 12

-3an incidence as frequently as one in every 700 live births (Bzoch & Williams, 1979). Two other causes known to cause cleft palate would be trauma or ablative surgery. Head trauma, or accidents that cause orofacial injuries, such as fractures of skull bones, may include fractures of the palatal bones, or a cleft along the palatal midline. Ablative surgery, or removal of tissue, is usually due to disease typically following removal of tumors. The second condition of palatal malfunction is velo pharyngeal insufficiency (VPI) which is defined as the inability of the soft palate, or velum, to make adequate closure with the posterior pharyngeal wall for normal speech production. Velopharyngeal insufficiency can be a result of congenital or acquired anomaly and the speech disorders that typicall y occur can be compared to those impairments associated with cleft palate and palatal fistulas. Velo pharyngeal insufficiency can also result from neurological disorders. It is important to remember the physiological differences between the hard and the soft palates. The function of ve lopharynx involves intricate movements and interactions of a number of muscles active in velopharyngeal closure. The hard palate, on th e other hand, does not display any myofunctional mov e ment since it consists mainly of bone and tissue lining and serves as a static border between the two cavities. It is difficult to determine the actual incidence of VPI , as it va ries wi th the con comitant disorder. The incidence of VPI, as it occurs

PAGE 13

-4after primary cleft palate repair, is reported to be between 20 percent and 40 percent (Bradley, 1979). The third condition responsible for a palatal break down is a fistula in the hard and/or the soft palate. An oronasal fistula is defined, in this study, as a hole in the palate connecting the oronasal cavities. It may develop subsequent to surgical repair of a cleft palate. It may also occur from traumatic perforation of ablative surgery. Fistulas also may result from pathologic abnormalities such as noma (gangrenious sore), syphillitic gummas, leprosy, and leishmaniasis (infection due to the parasite Leishmania) (Gordon & Brown, 1980). It is clear that these three different palatal openings or insufficiencies are almost exclusively caused by a morphological abnormality. Only VPI can b e diagnosed as resulting from a functional disorder. Patients who suffer from any of these abnormalities are referr ed for either surgical closure of the palatal openings, reconstruction of muscular tissue to restore velopharyngeal function, or for prosthetic treatment. When the differential diagnosis signals a function disorder, speech therapy should be the treatment of choice. Sometimes a combination of treatments is needed, depending on the ex t e nt of the physical abnormality. Ther efore , a n ee d exists for evaluation of the impact of palatal disorders on speech intelligibility, as well as for the quantification of the effect of thes e palatal insufficiencies.

PAGE 14

-5Surgical Treatment of Oronasal Openings Cleft Palate A variety of surgical procedures is available for palatal closure and correction of VPI. An excellent in depth description of several techniques in cleft palate surgery is given by Millard (1980), and in several chapters of Cleft Lip and Palate (Grabb, Rosenstein, & Bzoch, 1971). Selection of the appropriate surgical technique is depen dent upon the pathological extent of the condition, individual need, and optimal goal. Important information regarding morphology, tissue reaction, and optimal timing for surgery are taken into consideration. There are several reasons why obtaining an adequate closure between the oronasal cavities is important. The primary goal in cleft palate surgery, and perhaps the task that is most diffi cult to achieve, is to establish normal speech production. As a direct result of atypical airflow into the nasal cavity, the necessary intraoral a ir pressure cannot be maintained and speech is affected in the form of hyper nasality and nasal air emission. Normal hearing is important for the individual who is learning how to speak. Hearing loss that is caused primarily by otitis media (middle ear inflammation) occurs in almost a ll cleft palate individuals (Millard, 1980). The frequency of otitis media in these patients is caused by impair e d palatal

PAGE 15

-6muscle activity involved in the opening and closing of the auditory tub e (Eustachian tube). This is due to upper respiratory disease which is easily transmitted to the ear via the Eustachian tube. The tube, being the single dynamic connection between nasopharynx and the middle ear, serves as an equalizer for air pressure across the tympanic mem brane (ear drum) and loses its function when inf e ct ed (Davis, 1970). Speech habilitation is, therefore, complicated by inadequate auditory feedback. Another important reason for surgical closure of a cleft palate is to allow the individual normal feeding. Correction of dentition and establishing of normal swallowing habits are, there fore, additional important goals in palatal surgery. Velopharvngeal Insufficiency Velopharyngeal insufficiency occurs when the soft palate is unable to adequately separate nasopharynx from oropharyn x during speech. Regardless of the cause, resulting from either morphologic or functional disorder, VPI is an opening b e tween the oral and the nasal cavities which allows airflow into the nas a l cavity during the production of oral spee ch sounds . Sever a l different surgical techniques are avai labl e for the establishment of normal velopharyngeal function . Some examples of th e se procedures are palatal pushback (a palal lengthening procedure), pharyngeal flap ( a skin

PAGE 16

-7graft), teflon pharyngoplasty (augmentation by injection). A combination of these techniques may be used in order to obtain optimal results (Millard, 1980). In addition to surgical repair, treatment may involve prosthetics and/or speech therapy. Palatal Fistula Palatal fistulas, defined holes in the palate con necting the oronasal cavities , may appear secondary to cleft palate surgery or can occur as the result of trauma or ablative surgery. The frequency of occurrence of fis tulas following palatal r epair has been reported to range from 9 percent to 34 percent (Henningsson, 1983; Ross & Johnst on, 1972). Abyholm, Borchgrevink, and Eskeland (197 9) report that frequency of occurrence seems to be dependent upon surgical technique. Th ey report an overall incidence of 18 percent. Other frequency occurrences of fistulas resulting from trauma or ablat i ve surgery a re not r ea dily available in the literature . Surgical rep a ir of fistulas is undertaken only after comprehensi ve evaluation of the anatomical structures . If speech is distorted an d a fistula i s present, it is necessary to differentially eva luate the cause of the distortion. B y temporary obturation of th e fistula, the significance of its relat i onship to speech and the presence

PAGE 17

-8of any velopharyngeal insufficiency can be diagnosed (Bless, Ewanowski, & Dibbell, 1980; Bloch, 1979; O'Neal, 1971; Skolnick, Glaser, & McWilliams, 1980). If the speech sounds normal while the fistula is obturated, it can be concluded that the fistula is the cause of the speech dis tortion. However, if speech continues to be impaired, this would be an indication of VPI. If it were determined that fistula closure was necessary, a concomitant decision also would be made with regard to surgical or prosthetic treat ment. O'Neal (1971) has argued that consideration should be given to delaying or even eliminating the need for surgical repair of an existing fistula in growing children with ongoing development of the orofacial and orthodontic structures. A number of surgical techniques are available for fistula closure. The preferred techniques appear to be dependent upon fistula size and location. James (1980) reported on buccal flap surgery to repair small fistulas. Additionally, he also reported on the use of large buccal flaps in conjunction with bone grafts for reinforcement in large fistulas. He reported large fistulas to be of 0.5 centimeter and 2.5 centimeters in size. Harris (1980) reported that pharyngoplasty was preferred when closing a posteriorly located fistula. For anteriorly located fistulas, Harris reports the use of bone grafts. Th e bone graft is subs eq uently placed between flaps that are

PAGE 18

-9taken from the nasal and the palatal soft tissues. Further, a bone graft always would be indicated when the fistula occurs in the alveolar ridge. Prosthetic Treatment of Oronasal Openings Individuals with incomplete palatal closure who are not candidates for or who do not elect surgical treatment generally are treated with prosthetics (Adisman, 1971). Different kinds of prosthetic appliances are available, depending upon the physical defect being alleviated by obturation. A removable prosthetic appliance is used to cover the complete cleft of the hard and the soft palate, with the posterior part serving as the speech bulb. Closure between the oral and the nasal cavities is estab lished as th e posterior pharyngeal wall makes contact with the speech bulb. A simpler appliance , such as a speech bulb, is used for patients with a cleft of the soft palate only. Velopharyngeal insufficiency can be treated prosthetically in a similar manner, resulting in appro priate velopharyngeal closure. Reconstruct ive surgical repair of fistulas is always the treatment of choice (Drane, 1973), but prosthetic rehabilitation may sometimes be indicated. A prosthetic obturation may be recommended especially with recurring fistulas, soft palate inadequacy, or when serving as an interim obturator while awaiting surgery or orthodontic treatment. The traditional

PAGE 19

-10prosthodontic treatment includes either a fixed or a removable obturator (Reisberg, Gold, & Dorf, 1985). Small fistulas that do not cause a significant speech disorder may be temporarily obturated to prevent food and liquids from entering the nasal cavity during eating and drinking. These materials may include dental wax, acrylic resin, or chewing gum (Reisberg et al., 1985). Cleft Palate and Speech A cleft palate does not provide the separation between the oronasal cavities necessary for normal speech produc tion. Therefore, it will cause hypernasal voice quality and excessive nasal air emission. These two factors, hypernasality and excessive nasal air e mission, negatively affect speech intelligibility. The severity of the speech disorder resulting from this pathology depends upon the size of the cleft and the extent of involvement of the cleft with tissues of the hard and / or soft palates. The speech intelligibility of cl e ft palate patients is imp a ired by a wide range of errors, including indistinct produc tion, simple and gross substitution, and complete omission of sounds. Velopharyngeal Insu f ficiency and Speech Velopharyngeal insufficiency has been proven to c ause disorders of both nasality and intelligibility of speech,

PAGE 20

-11depending on its size. Thorough investigations of the size of such openings between the oral and nasal cavities and their influence on speech has been published. The objec tive has been to quantify th e magnitude of open ings in the velar port and the influence of th ese openings on hyper nasality and nasal a ir emission. The earliest attempts at this quantification were performed using direct observational techniques, such as lateral still roentg enog raphy o r cineradiography (Benson, 1972; Bj6rk, 1961; Bzoch, 1968; Gr aber , Bzoch, & Aoba, 1959; Moll, 1960, 1962, 1964; Shelton, Brooks, & Youngstrom, 1964; Subteln y , Koepp-Baker, & Subtelny, 1961). Much information, therefore, is available regarding the pattern of velopharyngeal kinesiology and configuration during speech. This has contributed greatly to the knowledge base about velopharyngeal function and closure (Aram & Subtelny, 1959; Bj6rk, 1961; McKerns, 196 8/ 1 969 ; Mo ] l, 1965; Nylen, 1961; Warren & Hoffmann, 1961). The radiologic measures obtained were compar ed to indirect measures, such as perceptual jud gmen ts of speech intelligibility and voice quality, to acoustical measures (sound spectrography) (Andrews & Rutherford, 1972; Bjork, 1961), or to measure ments found in the pneumatic pressure-flow (a erodyna mic) technique (Warren, 1964a,b; Warren & DuBois, 1964). Further quantification attempts regarding velo pharyngeal openings led to research where the velar port

PAGE 21

-12openings were experimentally controlled in normal subjects. Different sized polyvinyl tubing was inserted through the nose into the velopharynx to create artificial velopharyn geal openings (Isshiki, Honjow, & Morimoto, 1968; Bernthal & Beu kelman , 1977). Similarly, variable aperture speech appliances with inserts were e mployed by Andrews and Rutherford (1972), Liebman (1964), and Watterson and Emanuel (1981 a ,b). By manipulating the velopharyngeal port size in this manner, differing degrees of hypernasality could be created and computations of the critical size of openings in the velopharyngeal area performed. Generally, these reported studies on VPI agree that distortion of speech quality is introduced when there is an opening between 10 rnrn 2 a nd 20 rnrn 2 in the velar port. This is perceived as hypernasality, with audible nasal air emission causing distortion of oral consonants. Based on the literature, it can be concluded that the larger the velopharyngeal opening, the more distortions occur in speech characteristics of intelligibility and resonance quality. However, a complete linear relationship is not a l ways present (Subt e ln y et al., 1961; Isshiki et al., 1968; Andrews & Rutherford, 1972; Dalston, 1 982 ). Liebman (1964) found that posteriorly positioned velar openings caused fewer speech problems than those anteriorly position ed . The t wo factors of size and position have proven import an t with velar insufficiencies and are

PAGE 22

-13thought to play an important role with other oral-nasal openings such as palatal fistulas. Palatal Fistulas and Speech Fistulas, like cleft palate and VPI, cause inadequate closure between the oronasal cavities. Therefore, patients having fistulas exhibit the same speech impairments as do cleft palate and VPI patients. Fistula Location It is unclear as to the influence of fistula location on speech. Ross and Johnston (1972) state that fistulas occurring posteriorly in the hard palate or in the soft palate cause a cleft palate-like speech, i.e., hypernasal voice quality and audible, inappropriate nasal air emis sion. No clear description is offered regarding the influence that anteriorly positioned fistulas may have on speech. It is stated that speech may be affected because of air escaping into the nasal cavity, influencing both articulation and voice quality (Ross & Johnston, 1972). Cosman and Falk (1980) reported that defects of speech production were significant in the presence of anterior fistulas. These palatal defects locat e d anterior to the place of articulation of pressure consonants caused sound substitution. For example, a /t/ sound, an unvoiced

PAGE 23

-14anterior plosive, was substituted for by /k/, an unvoiced posterior plosive, and the /d/ sound, a voiced anterior plosive, was substituted for by /g/, a voiced positive plosive. Fistula Size The importance of fistula size and its influence on speech is evident in the literature. It has been proven clinically that speech impairment increases with fistula size. Morley reported in 1962 that even a small fistula may affect the development of normal speech. This occurs because of a prevention of the build-up of intraoral air pressure necessary for the production of oral consonants. In the literature published twenty years later, it is still stated that speech is increasingly and adversely affected by palatal fistulas of increasing size (Henningsson, 1983; Shelton & Blank, 1984). This information, however, only describes the fistulas as being either small, medium, or large without any quantitative definitions of size. In a review of the literature reporting on palatal fistulas, no study was found that dealt with quantifica tion of palatal fistulas and their influence on speech. There were only two different studies which defined large versus small fistulas. Proctor (1969) defined large oronasal fistulas as ranging from 0.5 centimeter to

PAGE 24

-152.5 centimeters, while Henderson (1982) defined small fistulas as ranging from 0.5 centimeter to 1.5 centimeters in width and 0.5 centimeter to 2.0 centimeters in length. One pertinent study (Shelton & Blank, 1984) deals with palatal fistulas and their influence on speech. Shelton and Blank reported on six patients with oronasal fistulas. The sizes of the fistulas were classified as small, moderate, or large, without further specification. Patients with a small or moderate fistula maintained sufficient intraoral air pressure for adequate oral sound production. Patients with large oronasal fistulas demonstrated a reduction in intraoral air pressure. Nasal air flow, however, was detected with all fistula sizes and was found to generate some noise. With this meager information and inadequate guide lines on palatal fistulas, there seems to be a demonstrated need for more quantitative information regarding the effect of fistula size and location on speech production. Know ledge about how differently sized a nd positioned fistulas would influence speech would enhance patient advisement reg a rding surgery. In addition, such knowledge would aid in the selection of prosthetic appliances, including those for temporary obturation of the opening while eating and drinking. In this present study, the first steps have been taken to quantitatively define the impact of palatal fistula size and location on speech, as it was measured

PAGE 25

-16perceptually (evaluation of articulation and nasality), and instrumentally (nasal airflow and intraoral and differential air pressure). It appears that this new quantitative data, together with earlier well-established qualitative information, will contribute to better and necessary guidelines in the clinical management of patients with palatal fistulas and speech disorders. Statement o f Purpose The primary objective of this study was to quanti tatively and qualitatively anal y ze speech in terms of nasality and articulation, as well as some of the associated aerodynamic correlates in th e presence of artificially introduced openings through the hard palate leading directly into the nasal cavity. Of particular interest is the dependence of the speech characteristics on the size and position of such openings. Definition of the requisite critical size of VPI is an ongoing endeavor and is of much interest in the literature. However, there is little information available on defects in the hard or soft palate anterior to the velopharyngeal port and the result ing impacts on speech. Although clinical observations suggest that oronasal fistulas contribute to hypernasality and nasal emission, there is little quantitative informa tion describing the influence that fistula size and location have on speech.

PAGE 26

-17This present study is an attempt to close the gap regarding this lack of knowledge a nd the purpose is to begin defining the influence of size and location of an opening through the hard palate on speech in terms of intelligibility and resonance quality. In order to perform this quantification, the size and location of openings anterior to the velopharyngeal port were experimentally manipulated. Openings were introduced in the hard palate part of a removable palatal prosthetic appliance (research appliance). A 28-year-old Caucasian m a le w ith a cleft of the hard and soft palate and who wears a removable pros thetic palatal appliance was the only subject in this study. A combination of three locations (ant e ri or , middle, and posterior) and four sizes (5 mm 2 , 10 mm 2 , 20 mm 2 , and 30 mm 2 ) of these oro nasal openings we r e t es t ed for a series of clinical and physiological speech t es ts. Th e clinical tests incorporated tests of nasal air em ission, articula tion, hypernasality, an articulation rating, and a h ype rnasalit y rating. The nasal air emission t e st was evaluated by the examiner t ak in g a repeated measurement of two readin gs . The othe r clinical tests were presented to six speech-l ang uag e pathologists specially trained in evaluating articulation and hypernasality in patients with palatal dysfunction. These tests were evaluated by th e six jud ges who listened to and sub j ectively scored

PAGE 27

-18the remaining four clinical tests. The speech physiology testing was conducted by employing an aerodynamic instrument that computed intra oral air pressure and differential pressure. Since previous research has been concerned basically with velar incompetencies and with controlled velar port openings, a replication study was conducted to assess the influence of velopharyngeal inadequacy of speech. Openings of the same sizes as in the experimental study were drilled in the velar part of the speech appliance which would imply a resemblance to VPI. The same test battery of clinical and physiological speech tests was administered. Data obtained from this replication study were compared t o the experimental results, as well as to previous research reports. Such comparisons were conducted in order to evaluate the results from the present experimental study on oronasal openings in the ve l ar region and previously published results. It is known from previous research efforts that speech is negati vely influenced by an increased size of VPI . Furthermore, there a re clinical implications stating that increasing fistula size is detrimental to speech intel ligibility. Fistula location and its influence on speech have been discussed only vaguely but it is believed that posteriorly positioned fistulas cause more speech problems than an terior ly positioned fis tul as . It was anticipated,

PAGE 28

-19therefore, that the results from this present study would be negatively influenced also by the magnitude of the hole size, resulting in imprecise articulation, excessive nasality and nasal emission, and a decrease in intraoral air pressure and differential pressure as the oronasal openings were increased in size. It was anticipated also that these speech variables, likewise, would be influenced as the oronasal opening was moved posteriorly.

PAGE 29

CHAPTER II LITERATURE REVIEW AND RESEARCH QUESTIONS Introduction Maintaining appropriate intraoral air pressure during th e production of oral speech sounds is an important physiological component of normal speech. Normal speech production will be adve rsely affected when an opening occurs between the oronasal cavities. Oronasal openings , such as a cleft palate, VPI, or palatal fistulas permit air to escape from the oral cavity into the nasal cavity. This air escape results in transmission of the air and sound energy into the nasal cavity, changing the physiological event of normal speech (Daniloff, Schuckers, & Feth, 1980; Spriestersbach & Powers, 1959). This process results in a reduction in intr aora l a ir pressure and increased nasal air em ission. These factors, reduced intr aora l air press ur e and i ncreased nasal a ir emission, cause abnorma l speech production by (a) giving speech a hypernasal voice quality and (b) reducing articulation profici ency . In addition, th e re are other palatal abnormalities that often are undiagno sed at birth, but which are discovered later, which im pa ir normal palatal -20

PAGE 30

-21anatomy and physiology. Such abnormalities interfere with speech proficiency. These later diagnosed abnormalities occur in the form of submucous clefts, cleft uvula, unusuall y high and/or narrow palate, and paresis of the soft palate muscles. Occurrence of VPI in the non-cleft population may be due to several conditions, including a neurological disorder, a function disorder, or tissue ablation, or tissue reduction following a surgical procedure, such as a tonsillectomy or adenoidectomy (Bradley, 1979). In order to more fully understand oronasal openings and their effect on speech, a brief description of relevant embryology and anatomy will be addressed. A discussion of the speech characteristics that occur with various oronasal openings and how they are measured will follow. A review of experimental studies on VPI will be considered, followed by a discussion of the research questions which are the subject of the present research. Palatal Embryolog y a nd A natomy For an in-d e pth study of embryology and anatomy of the palatal structures and the oronasal cavities the reader is urged to consider the works by Bat e man (1977), Brescia (1971), Bzoch and Williams (1979), DuBruhl (1980), a nd Zemlin (1968). The following mat e ri a l is ex cerpted from the above ref ere nces.

PAGE 31

-22Palatal Embryology The initial stage of formation of the embryonic face including the primary and secondary palates can be seen as early as in the second month of gestation. The face and mouth including the lips and the hard palate have com pleted their basic development process at the end of the third month. The development of the orofacial structures is depen dent on the merging and fusion of five prominences, the single frontal prominence, the paired maxillary prominence, and the paired mandibular prominence. These are present in the third week of pregnancy. The mandibular processes merge at the end of the fourth week and give rise to the lower lip and the chin. The frontal prominence grows downward and forward and becomes the frontonasal process. On either side of this process the two olfactory pits are formed and as the frontonasal process grows further downward these pits will divide it into the median nasal and the two lateral nasal processes (Bzoch & Williams, 1979). end of the sixth week (Figure 1). This occurs by the It is the nasomedial process that forms the midportion of the upper lip, the medial portion of the alveolar rid ge , and the primary palate. The primary palate is developed as the n asomedial process grows down and back to form the upper anterior roof

PAGE 32

-23Frontonasal prominence (process) Nasomedial process Nasolateral process Olfactory pits Mandibular arch 6 wk (12 mm) Figure 1. Embryo at 6 weeks (12 mm). (a) Nose I Medial palatine __ __,'---r-----r""' _.:,.__~--proc es s (prema x illa) Left lateral pro cess (b) Figur e 2. Palatal structure. (a) Before fusion (7 to 8 weeks), (b) Fused palate (8 to 10 weeks).

PAGE 33

-24of the mouth and it extends vertically to the floor of the nasal cavity. This plate projects into the oral cavity as far as to where the incisive foramen is later located. The nasomedial process then extends laterally to the alveolar ridge and encompasses the four incisor teeth (Bateman, 1977). Formation of the primary and secondary palates in the human embryo has been observed as early as the second month of gestation (Crelin, 1976). The secondary palate is the main part of the palate and includes both the hard and the soft palates. It is formed from paired shelf-like outgrowths that arise from the maxillary processes. Further contributors are the lateral palatine processes or the lateral palatal shelves that grow medially and inferiorly during the eighth week. When these palatal shelves first start to develop, the mandible lies high in the mouth and the tongue must rise between them. During this time the lateral processes grow downward on each side of the tongue. In the ninth week of development a sudden growth spurt of the mandible increases the vertical distance between it and the maxillary arch and the tongue therefore drops down. This allows the palatine processes to rise up and start to fuse. Fusion of the palates is complete between the end of the 10th and the beginning of the 12th week (Figure 2) (Crelin, 1976).

PAGE 34

-25Palatal Anatomy and Physiology The palatal structures, or the roof of the mouth, consist of the hard and the soft palates. The hard palate forms a shelf between the oral and the nasal cavities. These cavities can be compl e tely separated by the normal function of the soft muscular palate. Adequate anatomy of both palates is necessary for normal function. The hard palate. The hard skeletal palate consists of two sets of paired bones, the palatine processes and the palatal bones (Figure 3). The palatine processes are formed by horizontal, medially directed projections from the body of the maxilla (upper jaw bones) at the boundary between the body and the alveolar process. The alveolar process f o rms the thick spongy part of the maxilla where the teeth are positioned. The palatine processes (in anteroposterior direction) are short e r th a n the maxilla and form only about three-fourth of the hard palate. They terminate posteriorly with a rough edge, to which the horizontal pl a tes of the palatal bones articulate in the transverse pal a tine suture. The posterior bord e rs of these bones are free and form the posterior nasal spine at midline. The midline suture terminates ant e riorly at the incisi v e foramen. In young chi l dren, an irr e gul a r suture line can be seen extending bilaterally from this foramen to the alveolar process. The small triangular-shaped structure anterior to this suture line is called the

PAGE 35

-26Alveolar process Transverse ~--_.__ __ palatine suture Palatine bone Posterior nasal spine Figure 3. The bones of the hard palate. Incisive papilla zone I ~-__'\.-----Uvula Figure 4. The hard and the soft palate.

PAGE 36

-27premaxilla. The premaxilla represents the two parts of the maxillary bone that unite early in embryonic life. The suture line indicating the existence of the premaxilla is almost erased on the adult skull (DuBruhl, 1980). The hard palate is covered with mucosa of a grayish pink color and various areas of the hard palate can be identified due to the varying structure of the submucous layer. Anteriorly, there is a series of transverse ridges called rugae, posterior to which we find the midline raphe. The gingiva and the glandular zone can also be clearly identified (Figure 4). The hard palate is thick anteriorly and laterally but becomes progressively thinner toward its midline. This results in an arching of the palate both transversely and anteroposteriorly. This arching differs from one individual to the next and contributes to the acoustic properties of the oral cavity and probably also to individual voice characteristics (Zemlin, 1968). The soft palate. The soft palate, or velum (Figure 5), consists of muscular and fibrous tissues, nerves, and blood vessels. It is positioned as a continuum from the posterior border of the palatine bones of the hard palate. It is attached bilaterally to the pharyngeal wall and extends posteriorly into the oropharynx where it ends in a free margin. The anterior part of the velum, slightly posterior to the articulation with the hard palate, is the palatal aponeurosis, a point where th e pal a tal muscles are

PAGE 37

-28V2 ' ' I HP= Hard palate Vl = Velum at rest V2 = Velum at closure PPW = Posterior pharyngeal wall I I Figure 5. Lateral view of the hard and the soft palates.

PAGE 38

-29attached. The posterior middle of the free margin of the velum consists mainly of the uvula. Bilaterally to the uvula, two curved folds of mucous membrane arch downward and outward. These are the anterior and the posterior faucal pillars (DuBruhl, 1980; Zemlin, 1968; Morley, 1962). Normal function of the hard and the soft palates. The function of the hard palate is to serve as the boney wall separating the nasal cavity from the oral cavity. It thus builds the floor of the nasal cavity and the roof of the mouth. A review of the literature finds a prevalence of the conclusion that the primary muscle active in velo pharyngeal closure is the levator veli palatini muscle. The contraction of this muscle elevates and pulls the soft palate posteriorly. It has also been suggested that this muscle exerts post e rom2dial movement of the lateral pharyngeal walls, aiding velopharyngeal closure pro ficiency (Dickson, 1972). The middle third of the velum demonstrates the greatest range of movement. It achieves closure with the posterior pharyngeal wall during speech, thus serving as a functional valve during velopharyngeal closure. This elevation also is thought to be substantially assisted by the uvular muscle (Dic kson, 1972).

PAGE 39

-30Abnormal Palatal Function and Speech The palatal anomalies affecting speech intelligibility, which are the bases in this work, are limited to palatal fistulas and VPI. The severity of a speech impairment is dependent upon the size/location of a fistula, or the size of a velopharyngeal opening. In other words, the severity is dependent upon the magnitude of the coupling of the oronasal cavities. Extensive research has determined that a velopharyngeal opening between 10 mrn 2 and 20 mm 2 in size will result in hypernasality with audible nasal air emission, thus dis torting the production of oral consonants (Isshiki et al., 1968; Shprintzen, Croft, Levin, & Rakoff, 1977; Warren, 1964a,b). With the critical size of VPI being defined there is, however, little information and no quantifiable data availabl e on the effect of palatal fistulas. It appears that the severity of a speech disorder--in the presence of a palatal fistula--will depend on the size and location of the fistula. However, it has been concluded that the same components of speech that are present with VPI (hypernasality with audible nasal air emission) also occur with palatal fistulas (Lindsay, 1971; Morley, 1962; Palmer, Hamlen, Ross, & Lindsay, 1969; Reid, 1962; Ross & Johnston, 1972; Skolnick, Glaser, & Mcwilliams, 1980).

PAGE 40

-31The Effects of Velopharyngeal Insufficiency on Speech The importance of adequate velopharyngeal valving and its influence on speech have long been recognized. Although it is difficult to determine the exact frequency of VPI in the non-cleft population, it has been suggested that between 20 percent and 40 percent of patients who undergo cleft palate repair suffer a concomitant VPI (Bradley, 1979). The major anatomical defects which occur are related to velopharyngeal disproportion, muscular defi ciency, or surgical removal of tissue (Bradley, 1979). Several authors have established the existence of a close relationship between the degree of velopharyngeal opening and the severity of speech impairments in normal and in cleft palate patients (Bzoch, 1968; Moll, 1962; Shprintzen e t a l., 1975; Subtelny et al., 1961). The major speech distortions associated with VPI can be measured by employ ing standardized tests and instruments or by the subjective judgment rating of resonance quality and articulation. The characteristics of these speech errors help us to differentially diagnose VPI. Secondary components of speech behavior, like nasal and facial grimacing, sometimes accompany VPI (Bzoch, 1979). Bradley (1979) lists specific speech sound errors as being weak production of plosive sounds (together with nasal emission of air), accompanied by a nasal non-speech

PAGE 41

-32sound, called a nasal snort. She also reports that the high vowels /u/ and /i/ can be expected to sound hypernasal with accompanied nasal air emission. Bjork (1961) has observed that the velar port is not completely closed when nasal consonants and their neighboring vowels are produced in normal speech. Findings by Moll (1962) showed the high vowels to exhibit more velopharyngeal closure than their low counterparts, /ae/ and /a/. Furthermore, when the vowels were uttered in connection with a nasal sound such as /n/, he observed incomplete velopharyngeal closure, especially if the vowel preceded the nasal sound. This was confirmed by Benson (1972), who found open velopharyngeal ports in 63 percent of the subjects phonating /a/, in 23 percent when phonating /i/, and in 9 percent phonating /u/. Fritzell (1973) also concluded that slight nasal airflow could be measured in the normal population during the production of /a/ but not with /i/, or with the fricatives /f/ or /s/. He pointed out that individuals with VPI could not attain the intraoral air pressure necessary for the distinct production of plosive and fricative sounds. The speech of these individuals, therefore, contained hyper nasal voice quality and compensatory speech errors, such as glottal stops and pharyngeal fricatives. Shelton et al. (1964) added to this information that velopharyngeal closure was found to be complete when testing isolat ed vowels in one subject group. However, this

PAGE 42

-33same group of subjects later presented abnormal closure patterns when tested in connected speech. This would seem to indicate that syllable or sentence testing should be employed when evaluating velopharyngeal closure. These findings were investigated by Bzoch (1968), who found in the testing of repeated syllables containing the plosive /p/, combined with either of the high vow els, /i/ and /u/, that complete velopharyngeal closure was obtained. This was also the case when the low vowel /a/ was tested in the same manner. However, a discrepancy of one millimeter in the configuration of lower palatal elevation for the lower vowel /a/ was o bserved. Articulatory distortions in individuals with palatal disorders were investigated by Bzoch ( 1965 ). He report e d that two-consonant blends, fricatives, and plosives -in that orderwe re the most misarticulated sounds. The error types included substitutions, d istortions, and omission of sounds. Substitutions can be recognized a s simple or gross substitutions. Gross substitutions are characteristic for cleft pala t e patients and for the VPI population. Gross substitutions included such e rrors a s pharyngeal fricatives, glottal stops , and nasal snorts. These occurred as substitutions for plosive and fricative sounds and occurred in the pronunciation of affricate and glide speech sounds . The nasal snort was found to substi tute for th e nasal sounds /m/ and /n/ .

PAGE 43

-34Perceptual Judgment In order to evaluate different speech characteristics, especially nasality, perceptual judgment has been and is still used extensively both clinically and in research. It is a subjective method of determining the degree of speech impairment using different rating scales . There have been, for example, 3-, 4-, 5-, and 10-point scales in use (C ounihan & Cullinan, 1970; Isshiki et al., 1968; Subtelny et al. , 1961; Warren, 1964b). Scaling procedures with values from Oto 100 has been utilized also (Fl etcher & Bishop, 1970). Different studies employed pane ls of judges or single judges. Some studies claim higher reliability using one judge, while other studies claim higher relia bility using a panel of judges (Bradford, Brooks, & Shelton, 1964). The number of judges also has varied considerably, ranging from 3 jud ges to 9 judges to 48 judges (Counihan & Cullinan, 1970; Liebman, 1964; Sherman, 1970; Subtelny et al., 1961). It appears especially difficult to obtain reliable perceptual judgment of nasality. This difficulty is influenced by many factors, including speaker variations due to phonation, articu l atory precision, regional dialect, and language (Counihan & Cullinan, 1970; Fletcher, Sooudi, & Frost, 1974; Warren, Hall, & Davis, 1981). Great variance in sample selection, using normal and cleft palate subjects,was also demonstrated. The difficulty to

PAGE 44

-35obtain a homogenous sample was another problem, due to differences in type and severity of the clefts (Bradford et al., 1964; Coleman, 1963/1964; Andrews & Rutherford, 1972; Watterson & Emanuel, 1981a,b). The selection of different speech samples and the clinician's training and experience also contributed to the difficulty of obtaining valid and reliable nasality ratings (Bradford et al., 1964; Bzoch, 1979; Counihan & Cullinan, 1970). Speech samples included isolated sounds, syllables, and/or connected speech in forward or backward play (Counihan & Cullinan, 1970). It is difficult to compare th e results and to compute accurate reliability because of the disparity in perceptual testing of nas a lity. However, it has been documented (Philips & Bzoch, 1969) that percep tu al judgment of articulation offers high int rarate r reliability a veraging 85 percent, a lthough interrat er reliability is lower with an average of 74 percent. Furthermore, Sherman (1970) found articul a tion and hypernasality to be functionally related in children with cleft pala te. Different instruments have been e mployed to measure nasality, sometimes in combination with perceptual judgment testing. Up to this point, however, no single objective test instrument seems to do so adequately and ex clusively (McWilliams, Glaser, Philips, Lawrence, Lavorato, Beery, & Skolnick, 1981; Van Demark, Bzoch, Daly, Fletcher,

PAGE 45

-36Mcwilliams, Pannbacker, & Weinberg, 1985). The traditional measures used to augment perceptual judgment testing have been radiographies, aerodynamics, and acoustics (Reich & Redenbaugh, 1985). In conclusion, it appears that it has been, and con tinues to be, a difficult task to obtain a reliabl e measure of nasality. Therefore, a combination of methods needs to be considered to secure the reliability of nasality judgment. The influence of examiner e x pectancy on subjective evaluations, such as with perceptual judgments, has been documented (Ramig, 1982). Physical disabilities, orofacial anomalies, and other visual as well as auditory cues have been noted to influence the examiner's judgments in face to-face e x aminations. Most studies reviewed here, however, have employed tape recorded speech samples. Subtelny et a 1. ( 1961) employed perceptual judgment rating of nasality and intelligibility in combination with lateral roentgenography measures. They reported on the size of the velopharyngeal port opening as it was changing during speech. The nasality ratings were derived by using a four-point rating scale. The speech of 70 adult subjects with cleft palate revealed that a distance between the po~terior pharyngeal wall and the soft palate of 3.5 mm to 7.0 mm (3.5 mm equals 9.62 mm 2 ) resulted in the auditory

PAGE 46

-37perception of hypernasal speech. Smaller distances (0.5 mm to 3.0 mm) resulted in the perception of only moderately nasalized speech. Word-syllable intelligibility evaluations were assessed by the same judges and compared to aperture size (Subtelny et al., 1961). Subjects with complete closure scored on the average a 10 percent loss of intelligibility. Sub jects with the 0.5 mm to 3.0 mm distance showed a 14.3 percent loss, and the 3.5 mm to 7.0 mm group had a 27.31 percent loss. Even greater openings than these were not accompanied by significant errors of nasality or intel ligibility. Only when the openings became as large as 11 mm to 18 mm was intelligibility reduced significantly. This nonlinear relationship, as described here, has been evidenced when comparing judgments of hypernasality with velopharyngeal port size that were measured with lateral still X-ray (Subtelny et al., 1961), with cineradiography (Andrews & Rutherford, 1972), or with aerodynamic measurements (Isshiki et al., 1968). It appears, from the results of the study by Subtelny and co-workers (1961), that small gaps have a greater effect on nasality than on intelli gibility. Subjects with a 3 mm diameter velopharyngeal port opening still produced intelligible speech but were judged as moderately nasal. This information agrees with Bjork's stud y (1961), in which subjects who displayed a 4 mm large gap were assessed as hypernasal.

PAGE 47

-38Hypernasality Testino Evaluation of hypernasality can be executed easily by employing a test described by Bzoch (1979). This test is the Hypernasality Test which is standardized and convenient for use both clinically and in research. The results from the test are derived from a pass-fail procedure of 10 trials where the correct productions are added together and presented as a single score, where a maximum correct score would be 10. As the subject produces 10 one-syllable words (beet, bit, bait, bet, bat, bought, boat, boot, but, Bert) once with the nares open and again with them gently pinched closed, a shift of voice quality will determine if hyper nasality is present. If the velopharyngeal port is inap propriately open during production of any of these words, closing of the nares will result in changing this open ended resonator to a closed-end resonator and, consequently, a change in resonance quality. Articulation Testing The Bzoch Error Pattern Screening Articulation Test (BEPSAT) was designed to evaluate individuals with incom plete velopharyngeal closure (Bzoch, 1979). This test was compared to the Iowa Pressure Articulation Test by Erickson (1984). A correlation study indicated significant

PAGE 48

-39relationship between the two tests. The two tests also showed high validity as a tool for appropriately identifying articulatory errors in the speech of cleft palate indiviuals. The BETSAT, however, appears to provide more valuable infor mation on the total articulatory behavior, because it includes all of the distinctive feature groups. There are 31 test words in the BEPSAT, with 23 words containing the individual sound element in medial position and eight con sonant blends where it occurs in initial position. Instrumentation Clinicians have tried, over the years, to correlate the perceptual judgment of velopharyngeal competence or incompetence with objective testing employing v arious instruments designed to measure velopharyngeal closure. There has be e n a concerted effort to find such a tool that would measure the magnitude of VPI and to relate it to the changes of voice resonance and articulation (Bjork, 1961; Bzoch, 1979; Dalston, 1982; Edgerton, Sadove, Compton, Bull, Blomain, McDonald, & Bralley, 1981; McWilliams et al., 1981; Warren & DuBois, 1964). One of the reasons for designing a measurement instrument was th e search for obj e cti v e and reliable means to test nasality, since it had b e en found that both interjud ge and intrajudge reliability of percep tual judgment testing were low (Bradford et al., 1964; Counihan & Cullinan, 1970).

PAGE 49

-40Observational techniques, such as lateral still roentgenography and cineradiography, have been employed in a variety of research studies (Benson, 1972; Bjork, 1961; Bzoch, 1968; Graber et al., 1959; Moll, 1960, 1962, 1964; Shelton et al., 1964; Subtelny et al., 1961). These tech niques have also been used extensively in clinical diagnostics. Other direct observational techniques include tomography (Bjork, 1961; Kuehn & Dolan, 1975) and the employment of fiberoptics in photodetector assessment (Dalston, 1982) and in nasopharyngoscopy with videopharyngoscopy (Shprintzen et al., 1977). Some measures of palatal function and nasality would include intraoral air pressure using oral manometers (Morris, 1966; Mcwilliams et al., 1981), acoustic analysis with spectrography (Andrews & Rutherford, 1972; Bjork, 1S61; Co l eman, 19 6 3/1964; House & Stevens, 1956; Watterson & Emanuel, 1981a), and nasal airflow testing (Bzoch, 1979). Another instrument for acoustic analysis is call3d Tonar (Fletcher, 1970; Fletcher & Bishop, 1970; Fletcher et al., 1974). Two additional techniques are the accelerometric technique also called nasal vibration analysis (Edgerton et al., 1981; Horii, 1980; Reich & Redenbaugh, 1985; Stevens, Kalikow, & Willemain, 1975), and respirometrics (Shaw & Gilbert, 1982). Derivation of the hydraulic principle was used in an attempt to measure pressure-flow relationships (Gorlin &

PAGE 50

-41Gorlin, 1951). Warren and DuBois (1964) extended the attempt by employing the resulting hydrokinetic equation to obtain indirect measures of the velopharyngeal port area. The changes in size when different speech sounds were pro duced could thus be determined (Warren, 1964a,b). Quanti tative measurements from the pneumatic pressure-flow technique (Gorlin & Gorlin, 1951; Warren, 1964a,b; Warren & DuBois, 1964) have been employed to obtain indirect measures of the velopharyngeal port area and to determine its changes in size when different speech sounds are produced. Spectrographic analysis of the acoustic signal is another method employed to assess the degree of na salized speech (Andrews & Rutherford, 1972; Bjork, 1961; House & Stevens, 1956; Watterson & Emanuel, 1981a). At other times a stand a rdized test of hypernasality was administered (Bzoch, 1979, Bzoch et al., 1984). However, a combination of methods has been employed to ensure reliability of the tested material. Although several instruments are available for the measurement of nasality, it has, never theless, continued to be recorded by using perceptual judgment procedures (Isshiki et al., 1968; Subtelny et al., 1961; Watterson & Emanuel, 1981a,b). Lateral Still Roentgenography and Cin e radiography Both lateral still roentgenography and cineradiography have been used to examine visually the function of the

PAGE 51

-42velopharyngeal port and to determine the relationship between the velar length and the depth of the nasopharynx (Figure 5). The patterns of the palatopharyngeal move ments and their changes, during speech, give much information about velopharyngeal function and closure. When inspecting the still X-ray or the cinefilm during speech production, the elevated soft palate can be seen attempting to make closure with the posterior pharyngeal wall. The least linear distance between these two anatomical struc tures, the soft palate and the posterior pharyngeal wall, is measured in millimeters. Bjork (1961) demonstrated a strong linear relationship between the width of this gap as measured from the lateral cinefilms and the area measures of the gap in the velopharyngeal port as determined in the cross-sections obtained from tomograph y . With these two measures, he d e veloped a diagram that can be employed in order to determine the size of the coupling area by use of cinefilms only (Bjork, 1961). Occasionally, nasalization of speech is noted in spite of obvious closure v isible on the lateral film. Therefore, it must be assumed that although closure might take place at one point as the velum stretches toward the posterior pharyngeal wall, it might do so only at that point which can be noted on the lateral film. One or more parts of the velum, laterally or medially to the touching point, may not attain complete closure. This lack of complet e closure pattern, or uneven closure, would result in an

PAGE 52

-43opening of the velopharyngeal port. This can be observed with basal view cineradiography (Skolnick, 1970; Williams, 1986) or through tomography (Bjork, 1961). The limitation of the lateral measurement and the radiation danger has led to the development of other instrumental methods of velopharyngeal quantifiable assessment such as naso endoscopy (Dalston, 1982). Aerodynamic Technique Intraoral air pressure is a physiological component related to the production of different speech sounds. The normal production of plosive and fricative sounds is depen dent upon adequate oral cavity pressure. Depending on the size of a velopharyngeal deficiency and airflow into the nasal cavity, interference with the pressure sounds can be expected (Shelton et al., 1973). A number of instruments have been developed to measure airflow and air pressure with the basic components being flowmeters and pressure transducers. The flowmeter records the volume rates of airflow, and pressure transducers record air pressure within the vocal tract (Lubker, 1970; Warren, 1973). The pneurnotachograph sensing element is the most common and most reliable type of flowrneter. The principle of this instrument is that there exists a proportionality between the airflow rate in a tube and the pressure dif ference between its endpoints. This instrument typically

PAGE 53

-44measures the volume of air in cubic centimeter per second (Counihan, 1979; Hardy, 1965; Lubker, 1970). It has been established that there is no simple rela tionship between the nasal airflow and the size of the velopharyngeal port. There are factors such as respiratory effort and nasal airway resistance that influence airflow in subjects with VPI (Butters, 1982; warren, 1973). Butters (1982) claims that her findings indicate that nasal airflow alone will provide reliable information about velopharyngeal closure and that testing of nasal airflow would be valuable as a diagnostic tool. In order to obtain data on intraoral air pressure, it is necessary to use a sensitive pressure transducer which is connected to an amplifier and a recording device. Its connecting tube, or catheter, is placed in the mouth per pendicular to the airflow with the requirement that it is neither too long, nor that it has a too large or too small a diameter (Lubker, 1970). One of the aerodynamic techniques, the pressure-flow technique, has been described by Warren (1964a,b) and by Warren and DuBois (1964) as a method of calculating the area size of VPI. It involv e s recording the amount of airflow through the velopharyngeal orifice simultaneously with the d iff e rential p ressure across it during speech. Airflow is measur e d with a pneumotachograph and differen tial pressure with a pressure tr a nsducer. By usin g these

PAGE 54

-45measures, the velopharyngeal area can then be calculated by using a hydrokinetic equation (Warren & DuBois, 1964, p. 52; Warren & Devereux, 1966, p. 105). where VPA = Vn .65j2~p VPA = velopharyngeal area in rnrn 2 Vn = rate of airflow in cm 3 /sec 6 P = differential pressure in cm H 2 o, converted to dynes/cm 2 D = density of air 3 ( 0 . 0 0 1 ) , gm/ cm .65 = correlation coefficient Warren (1964a) employed this method and equation to establish the relationship between oropharyngeal pressure, nasal airflow, and velopharyngeal port size. He examined the critical size openings in the velopharyngeal port in 10 normal-speaking subjects. The velopharyngeal opening was seldom found to be larger than 10 rnrn 2 and was usually smaller than 3 mrn 2 under these conditions. Only once was it open more than 20 mrn 2 . In a succeeding experiment with 10 speakers with cleft palate, Warren (1964b) found that those speakers who were considered to be rehabilitated, that is, exhibited acceptable voice quality, rarely evidenced an orifice larger than 10 mrn 2 . With the nasal speakers,

PAGE 55

-46however, he frequently recorded the aperture to be larger 2 than 20 mm, and the extremely nasal speakers displayed openings as large as 100 mm 2 . He concluded that the critical range of closure, and where speech is perceived 2 as hypernasal, starts at about 20 mm. Warren (1979) developed an instrument, the Palatal Efficiency Rating Computed Instantaneously (PERCI). It was designed to evaluate palatal efficiency (pressure drop across an oral/nasal orifice) both for clinical use and for research. This instrument permits instantaneous computing of measures such as intraoral air pressure, nasal airflow, and the differential pressure between the oral and the nasal cavities. Other measurements that can be obtained from its second vers ion, PERCI II, are airway conductance and the area of the orifice related to the palatal mechanism (Microtronics Corporation, n.d.). Intraoral pressure is detected by a pressure transducer and the nasal airflow by a pneumotachograph. Input to these instruments is transmitted via catheters inserted into the mouth and nose, respectively. When complete separation exists between the oral and the nasal cavities, the nasal pressure will be zero w hile the oral pressure will range between 3 to 7 cm H 2 o. With incomplete separation, such as when a palatal fistula is present, th e difference in pressure would typic a lly vary w ith the size of that opening (Fig ure 6).

PAGE 56

-47(a) Intraoral Air Pressure (b) Differential Pressure Pressure Transducer Pressure Transducer PERCI PERCI Figure 6. Palatal efficiency rating computed instantaneously (PERCI). (a) Intraoral air pressure, (b) Dif ferential pressure.

PAGE 57

-48When recording airflow with a pneumotachograph, such as the one used in the PERCI instrumentation, the use of a face mask is required to channel oral and nasal airflow. Such a mask would typically be an anesthesia--or a diver'smask with built-in openings for the necessary catheters. The use of face masks has been crit icized. Lubker and Moll (1965) cautioned that it may restrict articulatory movements. Furthermore, the amount of pressure that is required for the mask to obtain an airtight fit also may influence the normal speech behavior. Th e face mask has been criticized also in that the air storage in it would provide a resistance to the expiratory a irstream and, therefore, induce elevated intraoral air pressure and airflow (Counihan, 1979). However, these considerations have not been substantially confirmed. With the use of a short, straight catheter with a wide opening, resistance is not clearly affecting speech measurement data . Sound Spectrography Sound spectrography has been widely employed to measure nasalized speech. The spectrograph analyzes complex sig nals, such as the speech signal , a s a f unction of frequency, time, and intensity. The different speech so unds create certain patterns on the visual output, the spectrogram. The most important characteristics in evaluating presence of nasality are the formants, which can be seen as dark

PAGE 58

-49areas on the spectrogram. They consist of regions of accentuated intensity and represent the resonances of the vocal tract (Hadding & Petersson, 1970). The three first formants (F 1 , F 2 , and F 3 ) usually can be clearly dis tinguished, and sometimes there is even a clear display of F 4 and F 5 . Schwartz (1979) described the nasalized feature characteristics as a reduction in the intensity of the first formant, antiresonances, extra resonances, and a change in the center frequencies of the formants. Other spectral features that can be associated with nasality also may occur. These have, however, been observed to be too unpredictable or too sporadic to deserve mentioning. Even the four characteristics just described may not always occur. However, whenever a nasalized vowel is produced, it is likely at l east one of them will be seen on the spectrogram. In an attempt to evaluate velopharyngeal area size during normal speech a nd during simulated nasal speech, simultaneous cineradiography and sound spectrography were employed by Bjork (1961). Thirty adults and 26 children, all normal speakers, served as subjects. The findings showed that when r ela tively small changes of nasality were registered by the spectrograph , the cineradiographic readings would indicate a ve lopharyngeal opening of approximately 10 mm 2 . In a second study, these normal

PAGE 59

-50speaking subjects were asked to mimic hypernasal speech. The coupling areas were then found to have increased to between 100 rnrn 2 and 280 rnrn 2 . Nasal Airflow Velopharyngeal and/or palatal openings will permit airflow into the nasal cavity and nasal air escape can be recorded. A convenient and simple test, the Nasal Emission Test, can be employed for evaluation of nasal air emission (Bzoch, 1979). Bzoch, Kemker, and Dixon-wood (1984) stated that clinical experience has proven it to be a valu able speech evaluation tool of velopharyngeal function in cleft palate patients. Bzoch et al. (1984) continued to report that, in a comparison study between cineradiographic measures of the velopharyngeal port and an analysis of the spoken test words, there was a 96 percent agreement of palatal function when the speech of 40 patients was tested. The standardized Nasal Emission Test tests 10 two-syllable words (people, paper, puppy, pepper, piper, baby, Bobby, bubble, B.B., bye-bye). As each word is repeated by the subject, the examiner observes any airflow through the nose using a simple indicator, such as a small paper paddle about the size and shape of a toothbrush held under the nose (Bzoch, 1979). The results are recorded from a pass-fail procedure of 10 trials where the correct productions are added

PAGE 60

-51together and presented as a single score, with a maximum correct score of 10. This test should always yield a perfect score of 10 to indicate complete closure of the velopharyngeal mechanism. The Effects of Controlled Velopharyngeal Insufficiency on Speech Some experimental studies were performed where the investigators created different size velar port openings in normal speakin g subjects. This involved (a) inserting of different sized polyvinyl tubing in the velopharynx (Bernthal & Beukelman, 1977; Isshiki et al., 1968) or (b) usin g a variable aperture speech appliance with inserts (Figure 7) (Andrews & Rutherford, 1972; Liebman, 1964; Watterson & Emanuel, 1981a,b). By manipulating the velopharyngeal port size, different degrees of hyper nasality could be created and computations of the critical size of openings in the velopharyngeal area were performed. In order to determine whether there is a critical size of the velopharyngeal insufficiency at which point speech becomes unacceptable, Isshiki and co-workers (1968) conducted an experimental test with 11 young adult normal speakers. Polyvinyl tubes were inserted into the velo pharynx where artificial openings were created at diameters of 5 mm, 7 mm, 9 mm, and 12 mm. Articulation and hyper nasality were perceptually judged by five speech specialists

PAGE 61

-52C (a) (b) Figure 7. Two examples of tools for introducing velo pharyngeal insufficiency in normal speakers. (a) Polyvinyl tubing (Isshiki et al., 1968), (b) Variable palatal prosthesis with inserts (Andrews & Rutherford, 1972).

PAGE 62

-53on a three-point and a five-point rating scale, respec tively. The findings indicated that at no specific point along the range of dimensions did speech suddenly sound abnormal. The authors claim, instead, that there is a gradual change in both intelligibility and nasality. Speech quality was judged as acceptable--although slightly nasalup to a critical aperture size of 5 mm in diameter which equals 19.6 mm 2 . The articulatory errors did not exceed 25 percent at this size opening. When the larger size aperture of 7 mm (38.5 mm 2 ) was tested, nasality was unquestionable. Articulation errors increased to more than 60 percent, and overall speech was unacceptable. Both nasal airflow and intraoral air pressure were measured for all of the above tube size conditions. The results showed increasing nasal air excape and decreasing intraoral pres sure with increasing hole size. Andrews and Rutherford (1972) designed a study aimed at assessing the contribution of the nasally emitted sound to the perception of nasality. Through the use of special instrumentation they were able to separate the orally emitted sound from the combined orally and nasally emitted sound as they are produced in normal speech. By applying a variable aperture palatal prosthesis with inserts (Figure 7) they created the following velopharyngeal 2 2 2 conditions: no opening, 60 mm, 120 mm, 180 mm, and 240 mm 2 openings. In order to evaluate hypernasality,

PAGE 63

-54data from perceptual judgment and spectrographic recordings were coll e cted. Perceptual findings show a clear trend of more hypernasality ratings for the combined nasal and oral signal than for the orally emitted alone, especially for the v owel /u/. Spectrographic data supported these findings by displa y ing clear nasal features on the spectrogram, such as a combination of increased intensity and reduced fre quency of the third formant for /u/, and extra resonances between F 1 a nd F 2 for /i/. The effects of velar perfor a tions on speech were studied by Liebman (1964). She employed an e x perimental prosthetic speech appliance that w as constructed to simulate these per f orations in a vari e ty of sizes and locations, singly and in p a irs. List e ners' judgments of nasality and intelligibility in one speaker served a s indicators of the influence of these artificial v eloph a ryn ge al o p e nings on speech. The general findings indic a ted that speech became worse when the apertures were larger and placed in a more forward position. Furthermore, if the hol e w a s divided into two openings of e qual combin e d size and locat e d in the same ge n e ral a rea, s peech w as n oted to be improved. This finding has not b ee n ex pl a in e d satisfac torily but was suggest e d to h a ve clinical implications. The studies that have b ee n d iscuss e d up to this point have focus e d primarily on critic a l area siz e of VPI and its influence on nasality and int e lligibility. The

PAGE 64

-55following studies to be discussed will deal with how changes of velopharyngeal port size affect special speech characteristics, such as vowel intensity (Bernthal & Beukelman, 1977), vowel identification and nasality, and whispered vowel spectra (Watterson & Emanuel, 198la,b). It is observed repeatedly in the clinic that subjects with V PI speak with a reduction of an overall intensity level, because of increased coupling between the oral and nasal cavities. Bernthal and Beukelman (1977) were able to demonstrate that vowel intensity is a factor of velo pharyngeal portal opening. One normal speaking subject was fitted with a palatal prosthesis in which controlled openings at the position of the velopharyngeal port could be manipulated. The result showed that when v elopharyngea l . 2 2 . . openings of 7 mm and 50 mm were present vowel intensity was r e duc e d, especially at the 50 mm 2 aperture. It w as stated earlier that loudness r e duction is frequently observed in subjects with VPI. It is, however, uncertain whether this is a result of the int e nsity reduc tion or if it is caused by a behavior to try a nd hide a speech disorder (Morris, 1968). With Bernthal and Beukelman's results at hand, it appears that there may exist a close r e l a tionship between int e nsity a nd loudn e ss in this population. To comp e nsate for this l o ss of sound energy, speakers with VPI o f t e n increase th e ir vocal

PAGE 65

-56effort, which might abuse their vocal folds (Bernthal & Beukelman, 1977). It is well known that presentation of loud noise alters auditory feedback and causes changes in speech. The most noticeable change is the increase in vocal intensity. Articulation, however, seems to be unaffected by the introduction of noise, an assumption that was confirmed by Garber, Speidel, Siegel, Miller, and Glass (1980). A number of authors studied the a coustics of nasalized speech by employing analog methods (Watterson & Emanuel, 1981a). It has been suggested also that an increase in oral-nasal coupling would show reliable and predictable acoustic effects on the spectrogram. When these theories were tested empirically, however, the acoustic effec ts read from the spectrograms were sometimes much less reliable and predictable (Curtis, 1 9 70; Schwartz, 1979; Watterson & Emanuel, 1981a). It was, therefor e , reasoned that one way of avoiding these difficulties would be to test whispered vowels (Watterson & Emanuel, 1981a). By employing whispe r ed vowe ls, continuous spectral energy would be produced rather than harmonic, which is the result when using voiced speech. Watterson and Emanuel expected to observe systematic fo rmant changes in the spectrograms as the vel op haryng eal aperture was manipulat ed . sp eech appliance w ith insert s al l cwed for 12.57 A variable 2 mm ' 2 2 2 2 28.27 mm, 50.2 6 mm, 73.53 mm, a nd 153.94 mm op en ings

PAGE 66

-57in the velopharynx. The results indicated that whispered speech samples did not establish clearly a relationship between aperture size and degree of hypernasality when it was measured with spectrography, nor was there a clear dif ference between whispered or voiced vowel samples. Watterson and Emanuel (1981b), therefore, decided to include a perceptual listening study to measure the effects of oral-nasal coupling on vowel identification and vowel nasality. Both voiced and whispered speech were tested while using the same variable prosthesis as in their pre vious study. The results indicate that ratings of nasality were more reliable with the voiced than with the whispered test vowels. The research reviewed above has dealt with determining velopharyngeal port size, introducing artificial VPI through a variable prosthetic appliance, and measuring their effect on speech. They have evaluated speech in terms of per ceptual testing of articulation and hypernasality. Measure ments on some correlates of speech, such as differential air pressure and nasal airflow, vocal intensity of vowels, vowel identification, vowel nasality, and vowel spectra also were included. All studies concluded that the larger the velopharyngeal port size the more distortion is detected in all of the above speech characteristics, although there is not always a complete linear relation ship present.

PAGE 67

-58Subtelny and co-workers (1961) observed that nasalized speech increased in severity as velopharyngeal openings increased. When, however, sizes were larger than 7.0 mm in size,they saw a leveling off in the nasality ratings. Similar observations were recorded with the intelligibilit y ratings. Again, speakers with openings from 7.5 mm to 11 mm did not score significantly different from the group having openings from 3.5 mm to 7.0 mm. However, speakers with a gap e x ceeding 11.5 mm (11.5 mm to 18 mm) showed an increase in loss of intelligibility. The major common finding of all these studies, thus, infersthat an opening between 2 2 10 mm and 20 mm through the velar port results in significant distortion of speech intelligibility and resonance quality. The Effects of Palatal Fistulas on Speech Orofacial surgical skills have greatly improved over the past decade. Occasionally, however, an opening along the suture line of a repaired palatal cleft will occur (Figure 8). This opening or fistula between the oral and nasal cavities results in inappropriate loss of air through the nose and e x cessive nasal resonance. The incidence of fistulas after surgery in cleft palate patients has varied considerably. Ross and Johnston (1972) concluded that the incidence range was from

PAGE 68

-59Figure 8. Palatal fistula.

PAGE 69

-6010 percent to 20 percent, depending on different surgical techniques and cleft type, as well as different sample sel e ctions in the available studies. Postsurgical occurrence of fistulas in cleft palate individuals was reported 0 to be as high a s 18 percent by Abyholm and co-workers (1979), while a range between 9 percent and 34 percent was indicated by Henningsson (1983). She mention e d no quantitati v e indications of fistula size or location, except that these fistulas usually occur along the midline and that size and position determine the degree of the symptoms. In addition to the report e d incidence of fistulas following sur g ery in cleft palate patients, there are also fistulas resulting from ablative s urg e ry and from trauma. No information a ppears to b e av ail a ble on incidence resulting from con genital anomalies, nor has an y information been found on its factual occurrence. In this conte x t, it should be mentioned, that indi viduals with s e vere clefts, such as bilateral complete clefts, would develop fistulas secondary to surgical repair more frequentl y than patients with a simpler type of cl e ft. More residual p a lat a l fistulas were r e ported to ha v e occurred when th e von Langenb e ck p a l a toplasty had been performed than aft e r pushback rep a ir (Krause, Tharp, & Morris, 1976; Lindsay, 1971; Pa lm e r e t al., 1979). These residual fistul a s that occurr e d af t e r the van Lan ge nbeck procedure were found to be sli g htly larger than those

PAGE 70

-61resulting from the pushback technique. The two reports by Lindsay (1971) and Palmer et a l. (1969) were in agreement that the residual fistulas seemed to be occurring mostly in the anterior and alveolar parts of t h e hard palate. In contrast to this information, Reid (1962) described fistulas as being small, those less than two centimeters in diameter, while a large hole would be greater than two centimeters, with its location typically along the midline. He further commented that such a large residual fistula would follow pushback surgery. Witzell, Clarke, Lindsay, and Thomson (1979) a lso compared results following the von Langenbeck palatoplasty and the pushback procedure . The two surgical procedures compared positively when no residual fistulas occurred. With the availability of a mple information on VPI and its effect on speech, it might be concluded that similar data would be available for palatal fistulas and their effect on speech. However, in a review of the literature, the only studies found were descriptive and qualitative reports. These studies mainly dealt with incidence and descriptions of surgical techniques. Case histories were also reported, including general descriptions of fistulas. It was suggested that fistulas alone, or in association with VPI, contribute to speech-crippling effects similar to that of an unrepaired cleft palate (Millard, 1980).

PAGE 71

-62As mentioned in Chapt e r I, it is uncle a r whether fistula location is important. It has b e en suggested that anteriorly located fistulas e x ert significant negative influence on speech production (Cosman & Falk, 1980). Other authors merely indicated a possibl e ex istence of speech disorders resulting f rom ant e riorl y loc a ted fistulas (Ross & Johnston, 1972). Posteriorl y l o c a t e d fistulas were, however, declared to cause cleft palate-like speech (Ross & Johnston, 1972). Palmer et a l. (1969) a nd Li nds a y (1971) stated in their respective studies that all post s ur g ic a lly occurring fistulas were situat e d in th e a nt e ri o r p ortion of the hard palate. These a nt e ri o rly loc a t ed f istul a s w e re s a id to be a source of audible n a s a l a ir e mi s si o n, i mpairin g articu lation profici e nc y . Th e se f i s tul as also wer e said to cause slight nasality (Linds ay , 1 9 71; Pa lmer e t a l., 1969). A report by Musgrave and Bremner (1 9 60) indicated that many of the complications following the second stage of palatal repair occurred as midpalatal fistulas. No indic a tion was given as to how or if a ny of these fistulas influenc e d normal speech production. Fistula size has b e en described in a qualitati v e and descriptive manner. Clinical exp e rience indicates a definite increase of speech unintelli g ibility and nasal resonance with increasing fistula size. This has been substanti a ted in the literature (Gordon & Brown, 1980;

PAGE 72

-63Henningsson, 1983; Morley, 1962; Reid, 1962; Shelton & Blank, 1984). However, few authors have defined the size of fistulas in relation to speech disorders. Those authors who quantitatively defined fistula size were all describing fistula size in relation to surgical techniques (Henderson, 1982; Proctor, 1969; Reid, 1962). Palatal fistulas were categorized grossly as small, moderate-sized, or large. Reid (1962) described small holes as being less than two centimeters in diameter and a large hole as being more than two centimeters in diameter. Speech defects, apparently, were registered only with the larger fistulas. These large fistulas were reported to cause nasal air escape and an increased nasal voice quality. Proctor (1969) defined large oronasal fistulas to be 0.5 centimeter to 2.5 centimeters in size. No other sizes were, however, discussed. Henderson (1982), on the other hand, reported an approximal definition of small fistulas (0.5 centimeter to 1.5 centimeters in width and 0.5 centimeter to 2.0 centimeters in length), but failed to offer any inform a tion on large fistulas. Gordon and Brown (1980), who discussed very small and moderate-sized fistulas, explained the very small hole to be less than a few millimeters, but declined to define the moderate-size defects. An article by Shelton and Blank (1984) describes the single relevant study where palatal fistula size is related

PAGE 73

-64to speech impairments. However, the hole sizes a re not quantified but simply described as small, moderate, and large. These authors determined that small palatal fistulas did not exhibit any nasal airflow during speech. The intraoral air pressure necessary for normal speech production could be maintained in the presence of small fistulas. It was found that the subjects having moderate fistulas could retain the necessary intraoral air pressure at the e x pense of nasal air emission. All subjects with large palatal fistulas developed nasal air emission and a reduction of intr a oral air pressure. The presence of nasal air emission and the reduction of intraoral air pressure were found to be detrimental to the production of adequate articulation (Shelton & Blank, 1984). R e search Questions Both clinical observation and much research indicate that palatal defects resulting in abnormal openings between the oral-nasal ca v ities do, in f act, greatly influ e nce articulation and r e sonance qu a lit y o f speech. This, of course, depends larg e ly on th e size, shape, and location of the opening. Little research is r e ported which d e fines the influence of incomplete s e p a r a tion of the oron a sal cavities anterior to the velopharyng e al port. It is, however, clear th a t with a fistula present, leaving all other structures intact or fully r e p a ir e d, we can ex pect

PAGE 74

-65a breakdown of intraoral air pressure, as well as air excape into the nasal cavity and through the nose. This affects both resonance quality and articulation of speech. With the inference from above reviewed research one must ask the following questions. QUESTION ONE Can a fistula in the hard palate cause similar speech impairments as those resulting from VPI? It has been shown that an opening in the velopharyngeal port negatively affects normal speech production. One would consequently assume that any other abnormal opening, such as a palatal fistula, connecting the oronasal cavities, would likewise impair normal speech. QUESTION TWO To what extent would a speech impairment vary with differin g sizes of a palatal fistula? Previous research studies of velopharyngeal closure have shown that speech becomes progressively more distorted as the velopharyngeal port size is increased. It would seem reasonable to predict, and it is clinically known, that this is also the case with palatal fistulas. However, since there are no definable quantitative data available, it would seem pertinent to begin d e fining fistula size, measured in square millimeter, and to measure the effect that different size fistulas exert on speech production. QUESTION THREE Is there a critical fistula size at which point speech can be measured or judged as being dis torted?

PAGE 75

-66Intraoral air pressure has been found to decrease rapidly when velopharyngeal openings exceeded 10 2 rrun The critical size of velopharyngeal closure was suggested to occur at the point which distinguishes velopharyngeal adequacy from inadequacy. This was found to begin at approximately 2 20 mm . From this it was concluded that VPI may begin at a 10 to 20 mm 2 range (Warren, 1964a,b). It seems important to find out if a similar range exists with palatal fistulas or if the critical size here would be smaller or larger. QUESTION FOUR To what extent would a speech impairment vary with different positions of a palatal fistula? Anterior positions of velar openings have been demonstrated to cause f e w e r speech distortions than posteriorly positioned ones. Regarding fistulas in the hard palate, there seems to be no relevant information indicating the influence of fistula location. One report indicated that a posterior location of a fistula caused more speech problems than an anterior location. However, in this context there has been no description on the exact fistula location and how it would influence speech. First of all, it is of interest to examine the significance of the localization between VPI and hard palate fistulas. Secondly, a definition of significant differences among fistulas located in the hard palate would seem pertinent. Millard (1980) reported that surgical closure of a fistula often can be extremely difficult to obtain.

PAGE 76

-67Patients with residual palatal fistulas a re fitted instead with an obturating device to achieve normal speech and to eliminate food and liquids from passing to the nasal cavity, as well as pr ev enting nasal fluids f rom entering the mouth. It is the int e ntion that, with the results from this study, quantified definitions of palatal fistul a s will facilitate the choice of treatment for these individuals. The primary goal of this work, however, is to perform basic research to obtain quantitative definiti o ns of f istulas in the hard palate and their in f luence on s p ee ch. In other words, the purpose of this stu dy is to b eg in d e fining the influence that size and location of an opening through the hard palate has on sp e ech in terms of articulation and resonance quality.

PAGE 77

CHAPTER III METHODS Introduction Measurement of oral speech in the presence of a palatal defect was discussed in the previous chapter. The basic speech characteristics related to palatal disorders are nasality and articulation. Validity and reliability, as reviewed in Chapter II, were both difficult to obtain in the severity ratings of hypernasality. Both intraand interrater reliability were evidenced to be poor. Regarding the perceptual judgment of articulation, int ra rater reliability was good but interjudge reli ab ility were low . Due to th e ambiguity of perceptual jud gme nt-testing of nasality, additional tools were deve loped a nd tested in the hope to obtain one single objective measure of resonance quality. The consensus of the literature reviewed above was that no single instrument exists that exclusively documents nasality (Van D e mark et al., 1985; Fletcher et al., 1974). The recommendation has been to employ a combination of methods to obtain adequate nasality measurements (Reich & Redenbaugh, 1985). Th e se combinations usually contained subjective testing, employ ing rating scales or standardized tests of articulation -68

PAGE 78

-69and hypernasality, together with an objective instrumental technique. Materials and Procedures To answer the questions posed in Chapter II, palatal openings mimicing fistulas had to be experimentally created. To provide a range of sizes and locations of palatal openings, a research prosthesis for one subject was constructed. The subject has an unrepaired cleft palate and wears a removable palatal appliance. To define and quantify the effects of different palatal openings on speech, a battery of tests was included. These tests can be divided into two categories: (a) subjective perceptual testing and (b) instrumental testing. Subject One young adult male with complete bilateral cleft of the hard and the soft palates was employed for this study. This individual is a 28-year-old Caucasian pro fessional. He speaks English with general American dialect, holds a master's degree in speech pathology and a Ph.D. degree in education. He wears a removable palatal appliance and his speech is perceived as normal. The subject was born with a bilateral cleft of the hard and the soft palates, the alveolar ridge, premaxilla, and the

PAGE 79

-70upper lip (Wharton, 1978). Surgical closure of the lip was performed shortly after birth with several secondary reconstructive repairs at later ages. The hard and the soft palates have never been closed. It was explained that the hard palate was never surgically closed because of the width of the palatal cleft. Instead, the subject was fitted with a removable palatal prosthesis at age three in order to obtain separation between the nasal and the oral cavities. His speech was judged to be extremely good, only four months after fitting of his palatal appliance, by th e cl e ft pal a te team at the University of Florida teaching hospital. Tod ay the subj e ct still wears a removable palatal appliance and his speech is perceived as normal. An audiometric evaluation of the subject 's hearing prior to conducting any testing indicated normal hearing bilaterally ( Appendix A) . Research Prosthesis A duplicate of the subject ' s own presently e xis ting removable palatal prosthesis was constructed by a prostho dontist in the D epa rtment of Removabl e Prosthodontics at the University of Florida J. Hillis Miller Health Center . The appliance consists of two parts . One is a metal frame with a special design for retention of the appliance in the subject's mouth. The other part is an acrylic resin

PAGE 80

-71portion which fits accurately to the cleft area. After the construction, which required three fitting sessions, the subject was instructed to wear and adjust to the appliance, in order to overcome any functional differences that may have yet been present. After 10 days, it was decided that the subject had adapted well to this new appliance. This was independently judged by two speech language pathologists. The subject's speech was free from articulation errors and his resonance quality was normal. The subject himself expressed satisfaction with his speech intelligibility, resonance quality, and the ease with which normal speech could be produced. Test Conditions To permit manipulation of the size and location of openings a nterior to the velopharyngeal port, holes were drilled through the appliance, resulting in an opening between the oronasal cavities. Speech was then tested, using one hole at a time. The locations of the openings were positioned along the midline, in the front one third (A), in the middle one-third (B), and in the back one-third (C) of the hard palate part of the prosthesis. A fourth location of holes was drilled in the middle of the speech bulb (D), which is also the velar part of the prosthesis (Figure 9). The area sizes of the openings were 5 mm 2 (1), 10 mm 2 (2), 20 mm 2 (3), and 20 mm 2 (4). The area sizes were selected on the

PAGE 81

-720 0 5 0 10 @ 0 20 0 30 A= Anterior of the hard palate B = Middle of the hard palate C = Posterior of the hard palate D = Velar 2 mm mm2 2 mm 2 mm Figure 9. Sizes and positions of e x perimental fistulas.

PAGE 82

-73grounds of previous research on critical points of velo pharyngeal insufficiency and speech. Warren, Dalston, Trier, and Holder (1985) concluded that these sizes ranged from 1 mm 2 to 9 mm 2 for adequate closure, from 10 rnm 2 to 19 mm 2 for borderline closure, and inadequate closure was registered when the VPI was larger than 19 mm 2 . Each opening was drilled one at a time and resealed after the entire test battery had been completed for that particular opening. Self-curing resin was used to reseal the holes to secure an airtight seal. Altogether there were 16 experimental test conditions, combining the four area sizes (1, 2, 3, and 4) with the four different positions (A, B, C, and D). All holes were drilled in a random order, taking into consideration the physical limitations of the prosthesis. There was concern due to the possibility of fracture of the partition between the largest sizes of positions Band C. This may have influenced the randomization. The testing order for all conditions is presented in Appendix B. All tests were also conducted under three baseline conditions. The baselines included testing with the subject's normal, original prosthesis (BN) in plac e . Baseline testing of the research prosthesis (BR) was performed before any of the experimental openings were drilled. A post-test base line measurement (PBR) was also included. At the PBR testing, all openings had been securely plugged with resin.

PAGE 83

-74Human Experimentation Approval The study was granted Human Experimental Approval by the Institutional Review Board at the Shands Medical Center. The subject was informed that his participation in this study would not include any known discomfort or anticipated risks. An informed consent to participate in research form was read and signed by the subject. Test Environment and Audio Recording Testing was performed with the subject comfortably seated in an acoustically sound-proof research unit at the University of Florida J. Hillis Miller Health Center. A tape recorder (Sony, TC-158 SD, Dolby system), with its 300 ohm dynamic microphone (Thoro Test) and cassette tapes (Scotch AVX-50), was used to record all speech samples. The Bioengineering Service tested the recording unit through the microphon e and found it to present a "MIC" input constant of 0.00SV RMS with a drive level of -3 dB. As the unit was tested usin g a sound level meter, it was found that all values up to a frequency of 6500 Hertz were normally represented by the tape recording with a tapering off to 8000 Hertz, whereafter there was a sharp drop in the quality of tape and recording reproduction. The microphone was securel y fasten ed to a microphon e stand and placed at a distance of 30 centimeters from the

PAGE 84

-75mouth and at a 90-degree angle of incid e nce. This position of the microphone minimiz e d distortions of the acoustic voice signal by r e spiratory airflow noise. The VU-meter or volume indicator on the tape recorder was adjusted and monitored to control the intensity level (loudness) by the investigator. No attempt was made to control the loudness or pitch of the subject's voice. The subject was able to hold a general level of vocal production (syllables, words, and sentences) with only a ne g ligible change of decibels throughout the testing, which was sufficient for this purpose. It was necessary that all the equipment be located within the testing booth, together with the sub ject and the investi g ator. Perceptual M e asurements Ther e wa s a tot a l o f fi ve p e rce p tu a l t e st s in th e speec h evaluation. Thes e consisted of (1) one s t a ndardized test of articulation, (2) on e a rticul a ti o n ratin g t e st, (3) o n e standardized test of h y p e rn a s a lit y w ith o ut a uditor y mask ing noise, (4) one stand a rdiz e d t e st o f hype rn a sality with auditory masking noi s e, a nd (5) on e h y p e rna s ality rating test. Th e st a ndardized t e st s li s t e d a b ove , (1) and (3), are tests empl oye d a t the Uni ve rsity o f Florid a 's Cr a nio facial Center and the D e partm e nt of Communic a ti v e Disord e rs. The two perceptual rating tests were construct e d in an

PAGE 85

-76attempt to evaluate perceptual scaling judgments and to compare them to the other testing tools. Articulation. The Bzoch Error Pattern Screening Articulation Test (BEPSAT) (Bzoch, 1979), previously described in Chapter II, was employed for the articulation testing (Appendix C). The BEPSAT was selected because it is a standardized test and provides valuable information of the articulatory proficiency of all distinctive feature groups. The perceptual rating of articulation proficiency in connected speech was performed by judging four sentences. These sentences are listed in Appendix D. The sentences represent a sample of sentences frequently used in the evaluation of speech disorders related to palatal dis orders at the University of Florida Craniofacial Center. Judgments of these sentences were conducted by employing a five-point ordinal rating scale (Appendix D). Hypernasality. For the evaluation of nasality, the standardized Cul-de-sac Test of Hypernasality was adminis tered (Appendix C). This test was described in detail in Chapter II. This test was repeated a second time while the subject was exposed to auditory masking noise. The reason for this was to rule out any effect from compensation for nasality due to auditory feedback (Coleman, 1963/1964; Andrews & Rutherford, 1972). To obtain this new test environment, a second tape recorder, a Sharp Educator

PAGE 86

-77#376895 TLR with headphones (Thora-Test SHP-2, 8 ohm), was employed in the experiment. The masking noise consisted of a prerecorded multitalker tape. Twenty speakers were recorded simultaneously on this tape. It was stated that "the recording is unintelligible and the level is stable over time, having a relatively flat spectrum (-2.5 dB) from 200 to 500 Hz" (Holmes, Frank, & Stoker, 1983, p. 88). The tape was obtained from a University of Florida, Department of Speech, audiologist. The multitalker noise tape was presented to the subject binaurally at a sound pressure level of 90 decibels as it was adjusted with a sound-level meter (Davis, 1970). The second test of hypernasality consisted of the perceptual rating of eight paired consonant-vowel-syllables (eve-syllables). The eve-syllable "pin" was paired with another eve-syllable that contained the plosive consonant sound /p/, the vowel sound /i/ followed by one each of the six plosive sounds /p/, /b/, /t/, /d/, /k/, and /g/. There were, thus, six different syllable combinations. The vowel /i/ was selected because, as a high vowel, it has been demonstrated to be the sound where the least velopharyngeal opening was observed, thus indicating a high level of intra oral air pressure (Moll, 1962; Watterson & Emanuel, 1981a). Two of the pairs "pin-pib" and "pin-pip" were repeated, but not in consecutive order. This repetition was included to perform a reliability test among and

PAGE 87

-78between the judges. The six judges rated nasality accord ing to a five-point ordinal rating scale (Appendix E). Procedures of Data Collection The speech samples obtained through the administration of the articulation and the hypernasality tests were tape recorded. There were five different tests altogether--two articulation tests and two nasality tests, of which one nasality test was repeated in a masking noise environment. With the above five tests and the 19 different conditions under which they were tested, a tot a l of 95 test objectives was to be perceptually judged. A test tape was then arranged with the 95 tests occurring in random order. Due to the repetitious nature and the bulk of this test battery, it is not included in its entirety in the disser tation. Judges The recorded speech samples were subject to perceptual judgment. Si x speech-language pathologists served a s judges. They were selected because of their experience in working with resonance and articulation disorders associated with VPI. All but one of the judges were able to evaluate all 95 tests in one sitting. Each judge spent an average time of one hour and 30 minutes for this

PAGE 88

-79evaluation. The range was from one hour and 20 minutes to one hour and 45 minutes. The 95 tests were recorded on three tapes. Tape one contained tests one through 38, tape two tests 39 through 78, and tape three the tests 79 through 95. Each judge was presented with a different sequence of the three tapes, as shown in Appendix F. This was arranged to avoid a tiring influence on the judges' response. Prior to testing, each judge was familiarized with the four tests and read a set of instructions (Appendix G). The judges were told that they could stop the test tape, using a stop button on the tape recorder, should they need to stop for replay or for taking a rest. The judges were not informed of the masking feature in some of the material. Each judge was required to make a total of 589 (31 x 19) articulation test judgments and to rate articu latory proficiency from 76 (4 x 19) connected speech samples (sentences) with a five-point scale. Additionally each judge was asked to make 380 (2 x 10 x 19) judgments on the hypernasality test (190 responses from the unmasked version and 190 responses from the masked version). The rating of hypernasality included judgments of 171 (8 x 19) paired eve-syllables based on a five-point scale. Varia tion in the listening environment was minimized and at all

PAGE 89

-80times judging was performed in a closed room with only the judge and the experimenter present. Instrumentation Three instrumental tests were employed to further assess the effect of palatal openings on speech. These tests represented measurements of nasal airflow, aero dynamic measurements of intraoral air pressure, and dif ferential pressure. The first test, the Nasal Emission Test, supplies the airflow data. The second and third tests were performed by employing the PERCI II instrument which provides the measurements of intraoral air pressure and differential pressure. Nasal airflow. The data on nasal emission of air were collected by employing the Nasal Emission Test as seen in Appendix C. This is a standardized test employing simple and accurate means to measure the presence or absence of nasal air emission. The test is employed frequently at the University of Florida's Cranof ac ial Center and the Department of Communicative Disorders. described in detail in Chapter II. The test was Aerodynamic measurements. Objective instrumental measurements were ne x t applied to measure the effects of palatal openings. The aerodynamic entities that were selected for testing, which are closely related to normal

PAGE 90

-81speech production, were intraoral air pressure and differ ential pressure. Measurement of nasal emission of air was also attempted but the results were inconclusive. They were, therefore, excluded from the study (see section of PERCI below, page 82). The data from the intraoral air pressure and differ ential pressure measurements were gathered to increase the reliability of the perceptual tests. Furthermore, they were collected to determine whether any of these measures could be singly employed for nasality detection. To compute intraoral air pressure and differential pressure an instrument, PERCI II, was employed. The func tion of this instrument was described in Chapter II. To make PERCI II operative, two separate pressure transducers and a pneumotachograph had to be connected to the instru ment via catheters. These catheters had to be of different diameters, dep e nding on what part it would serve. For testing of intraoral air pressure during speech, an oral catheter, two millimeters in diameter, was attached to a Stratham PMSETC pressure transducer, which in turn was connected to the pressure channel on PERCI II. Before testing was begun, the pressure channel of the unit was calibrated with a water manometer, constructed by the investigator. For the differential pressure test, an additional catheter was added to the pressure transducer. The catheter

PAGE 91

-82size was four millimeters in diameter. It was threaded onto a nasal olive that was to be placed in one of the subject's nostrils during testing (Figure 6). It was planned to include an additional test of nasal emission with this instrument. It was thought that such results obtained by instrumental testing would enhance tne results from the perceptual testing of nasal emission. To measure nasal emission, the flow channel of the instrument had to be made operative. Nasal flow is measured with a nasal catheter with an airtight fit into one nostril. The other end of the catheter is connected to a Fleisch pneumo tachograph (pneumotach) Type One. The pneumotach is, in turn, connected to a heater power supply (filament trans former of 6.3 Volt at 1.2 ampere), used to heat the mesh screen inside the pneumotach cylinder. This is necessary to prevent condensation from forming on the screen, which would lead to an erroneous increase in the airflow signal (Micro tronics Corporation, n.d.). The small pressure differential developed across the screen in the pneumotach is then detected by a Stratham PM15ETC pressure transducer and the signal is sent to PERCI II for read-out. Calibration of the flow channel was performed by connecting the pneumotach to a captured air source. A flow gauge (Fischer and Porter Company Precision Bore Flowrator Tube No. FP-1/2-17-G-10/27 sec/sec-air met. @ 14.7 P.S.I.A. & 70) with a scale range of 500 cubic

PAGE 92

-83centimeters per second was connected to the output side of the pneumotach. The calibration of the flow channel was not linear. Deviations were occurring in the lower ranges. The instrument could be calibrated only at the ranges from 500 cubic centimeters through 300 cubic centimeters per second. At the level of 250 cubic centimeters the calibra tion errors were measured at 10 units on PERCI II. At 200 cubic centimeters, 100 cubic centimeters, and at 50 cubic centimeters the errors were 20, 40, and 50 units off, respectively. It was expected that the results could be adjusted to this error in calibration. However, during testing, it was noted that the errors were too grave and erratic. Therefore, due to this internal calibration instability, this measurement of nasal air emission was excluded from the study. Procedures of Data Collection Data from the Nasal Emission Test were collected by the examiner at the time of the original testing. A small paper paddle was held to the subject's nose while he pro duced each of the 10 words from the test. The test was repeated twice for reliability. Any movement of the paper paddle, indicating nasal air flow, was registered. This test should yield a score of zero movement of the paddle to record no nasal air leakage. The data from this test

PAGE 93

-84of nasal emission were collected by the examiner and repeated twice for the three baselines and the 16 experi mental test combinations. The procedures for testing intraoral air pressure and differential pressure called for employing the instrument PERCI II. The two tests were measured during speech for each of the 19 palatal conditions. The test with PERCI II followed the perceptual testing for each baseline and palatal opening. The speech sample employed was one of those suggested in the PERCI manual (Microtronics Corporation, n.d.). The speech sample for both intraoral air pressure and differential pres sure was the word "pop." Five readings were recorded by the investigator. It was explained in the PERCI manual that words and phrases containing the plosive sound /p/ were preferred. When these sounds are produced, the mouth is closed and a stagnant column of air is created. This is desirable because a normal measurement during the baseline tests would rule out any velopharyngeal incompetence, which was especially necessary for the accuracy in the experimental tests with the research prosthesis. These sounds, furthermore, eliminated any problems that could occur from tongue placement. For example, if the place of articulation would occur behind a palatal opening, the true speech effect of that opening would not be accurately documented. By employing plosive sounds the effect of respiratory effort

PAGE 94

-85is also controlled in the presence of palatal openings. In the testing of differential pressure, a ratio of oral and nasal pressure is measured and the respiratory effort is cancelled out, because it affects both pressures equally. It was expected that, with a negligible VPI, the measure ments of intraoral air pressure and differential pressure could be used accurately to measure the influence that defects in the hard palate, such as fistulas, would have on these parameters important for normal speech production. Reliability Testing The conclusion of the reviewed literature on perceptual judgment rating of nasality showed low intraand inter reliability (see Chapter II). Evidence from perceptual rating of articulation had been found to indicate good intrareliability but less reliable interagreement. Other perceptual tests, such as the employment of standardized tests for articulation and hypernasality, had not demon strated these reliability problems. The knowledge of these findings was the prerequisite for the reliability testing, on the hypernasality rating test that was performed in this study. As described above, there were six different paired eve-syllables in the test. Two of these eve-syllable pairs were selected for repeti tion for the purpose of reliability testing, thus increasing

PAGE 95

-86the total number of tested syllable pairs to eight. The repeated syllables were "pin-pip" and "pin-pib." These eve-syllables were nestled into the list of the original test syllables. The judges were not instructed about this procedure. No one of the judges commented on the fact that they had identified the repeated syllable-pairs in the test material. The reliability data were extracted from the total number of responses on the hypernasality rating test and were treated separately. Both intrareliability and interreliability were analyzed with the Spearman rank order correlation procedure. Replication Study The velar opening (D) has been treated as one of the experimental palatal openings. The same hole sizes were drilled and the same series of tests were employed for each of the D-openings. These D-openings were, furthermore, intermixed in the sampling regarding test order and no differences in data collection were considered. Nevertheless, D was originally included for comparison studies with previous results from research on VPI. It was intended that if the D-results compared favorably to previous results on VPI and speech, a common ground could be established for comparing the effects of VPI on speech with those from palatal fistulas.

PAGE 96

-87Data Analysis The data collected from the above-described testing procedures were subjected to statistical analysis. The data were organized to permit comparisons between the different test implementations. The choice of appropriate statistical tests was discussed with a statistics consultant from the Department of Biostatistics at the J. Hillis Miller Health Center. To accurately answer the research questions in Chapter II, the Statistical Analysis System (SAS) was employed (Statistical Analysis System Institute, Inc., 1982a,b). Because of the complicated nature of the data, each of the perceptual tests was computed independently. The aerodynamic test scores of intraoral air pressure and differential pressure were computed together. The program of statistical analyses included means, standard deviations, and anal y sis of variance (Marks, 1982a,b). Duncan's mul tiple range test was employed for a posteriori pairwise comparison among the means (Kirk, 1968). Its purpose is to further identify which means a re different from each oth e r after the analysis of variance has been performed. If significant differences can be identified with this a posteriori test, there is a high degree of confidence in the conclusions. To determine the degree of relationship between the variables Spearman's rank order correlation coefficients were computed.

PAGE 97

CHAPTER IV RESULTS This study represents an attempt to start defining and quantifying the influence of palatal fistulas on speech. Specific sizes and locations were obtained by systematic ally manipulating artificial openings in a removable palatal prosthesis. Four locations along the palatal midline were utilized and four artificial fistula sizes were used. These ranged from 5 mrn 2 to 30 mm 2 . To address the questions posed in Chapter II, data were gathe red from several tests of speech and from measurements of the speech cor relates of nasal air emission , intraoral air pressure, and differential pressure. There were seven types of data obtained from these tests and measures. Five of these data sets were collected from two perceptual tests of articula tion, two perceptual tests of hypernasality, and one set of data originating from an instrumental pass-fail test of nasal air emission . The last two data sets were obtained from instrumental test measurements of intraoral air pressure and differential pressure. The purpose of this study was to quantify and define palatal openings by their influence on speech through descriptive statis ti cs , rather than to ana lyz e speech as -88

PAGE 98

-89it was affected by apalatal openings. The objectives were to define the critical size/location of the different open ings, whetherspeech was negatively influenced in any way, and to quantify each opening size/position by the influence it exerted on speech behavior. Therefore, each separate test was analyzed relative to error production on the perceptual listening tests. The instrumental tests were analyzed relative to the necessary intraoral air pressure and differential pressure for maintaining normal speech production and to quantitatively correlate each size/location to its deviation. The presence of any nasal emission of air was recorded, and the data from two readings were analyzed. Perceptual Measurements There were four perceptual tests in the test battery judged by six judges. The tests employed were (1) the Bzoc h Error Pattern Screening Articulation Test, (2) the Articu lation Rating Test, (3) the Cul-de-sac Hypernasality Test, and (4) the Hypernasality Rating Test. The Cul-de-sac Hypernasality Test was used twice, the second time with auditory masking, thus increasing to five the actual number of evaluated perceptual tests. Each of these five tests was assessed with the three baselines (BN, BR, and PER) and the 16 experimental openings

PAGE 99

-90(A, B, C, and D locations and the sizes 5 20 mm 2 , and 30 mm 2 ) in random order. Articulation Test 2 mm , 10 2 mm , The response variables from the BEPSAT we re determined by assessing all test sounds contained in each of the 31 words. The number of correct articulations was extracted and used in the data analysis as the cutoff score that was to id en tif y complete articulatory proficiency versus non proficiency. No differentiation of error type was deter mined. The means and standard deviations of correct articulation among the six judges for all 19 test conditions are shown in Table 1 and Table 2. Figure 10 and Figure 11 illustrate these results graphically by size and position. The analysis of va riance indicated significant dif ferences between the means (F = 2.12, df = 15, p < 0.01). Further analysis of the ind epende nt variables, position and size, did not show significant differences between the positions (F = 0.15, df = 63, p = 0.92). The differences between the sizes were, however, significant (F = 8.69, df = 3, p < 0.01). On a size-by-size a nalysis, the a posteriori Duncan Multiple Range T es t was employed to identify furth e r differences among t he sizes. It indicat ed that the 5 mm 2 openings were not si gnificant ly diff e rent from the baseline measurements. However, all larger

PAGE 100

-91Table 1. Means of normal articulation scores among the six judges (Articulation Test; n = 31). Size (mm2) Position 5 10 20 A 28 23 15 B 29 26 18 C 27 24 21 D 26 25 20 Baselines BN BR PBR 27 29 30 30 23 20 20 21 Table 2. Standard deviations of the normal articulation scores among the six judges (Articulation Test; n = 31). Size (mn/) Position 5 10 20 30 A 3. 5 7 6.83 8.95 6.24 B 2. 7 3 4 . 11 9 . 83 7.90 C 4.49 6.25 7.03 6.33 D 3.31 5.89 7.17 9 .21 Baselines BN BR PBR 3.88 3.09 0. 81

PAGE 101

31 31 I 20.... 20 1010 -2 10 2 5 mm mm I \0 [\.) A B C D A B C D I 31 31 -20 -~-20 -10 _.._ 10 --ti) (1J 2 mm2 20 mm 30 0 C) U) A B C D A B C D Figure 10. Mean number of correct articulation scores among the six judges for each size (Articulation Test; n = 31)

PAGE 102

Ul Q) H 0 0 Cl) Ul Q) H 0 0 Cl) 31 20 10 0 31 20 10 0 --Baselines-BN BR --Baselines-BN BR -93PBR 5 PBR 5 10 Position A 20 2 30 mm Position B 10 20 30 mm 2 Figure 11. Mean number of correct a rticul a tion scores among the si x judges f or each position (Articulation Test; n = 31).

PAGE 103

(/J Q) 0 u U) (/J Q) 0 u U) 31 20 10 0 31 20 10 0 --Baselines-BN BR --B aselines -BN BR Figure 11. continued -94PBR 5 10 PBR 5 10 Position C 20 Position D 20 2 30 mm 2 30 mm

PAGE 104

-95openings were significantly different from the baselines. These significant differences did occur when the 10 rnm 2 openings were introduced. The Duncan a posteriori test showed no differences among positions. A Spearman rank order correlation matrix indicates that size is influencing the response variable signifi cantly (r = -0.42, p < 0.01) as can be seen in Table 3. The negative correlation indicates that there are fewer correct articulation responses with increased opening size. Regarding position, no correlation was obtained (r = 0.02, p = 0.84). Articulation Rating The judgments of the presence or absence of any articu lation errors on the four sentences in the Articulation Rating Test determined how the pa lat a l openings, of dif ferent sizes/locations, would influ e nce speech. The number of judgments of no articulation errors was ext racted for data analysis. The means and standard deviations of no nasality judgments among the s i x judg e s fo r al l the 19 test condi tions are demonstrated in T a bl e 4 and Table 5. A graphical representation of no nasality judgments by size and position is given in Figure 12 and Figure 13. An analysis of variance indicated significant differ ences for the dependent variable , no nasality (F = 6.72,

PAGE 105

-9 6 Table 3. Spearman's rank order correlation coefficients for the corr e ct judgments from the Arti c ulation Test. Position Size Correct 0.02007 0. 8 461 -0.42506 0.0001

PAGE 106

-97Table 4. Mean number of normal articulation scores among the six judges (Articulation Rating; n = 4). Size 2 (mm ) Position 5 10 20 30 A 3.16 1. 50 0.83 1.16 B 3.16 1. 50 0.83 0.50 C 3.66 1. 66 1. 00 1.00 D 2.66 2.00 1. 50 1. 50 Baselines BN BR PBR 3.16 4.00 3.33 Table 5. Standard deviations of normal articulation scores among the six judges ( Articulation Rating; n = 4). Size (mm2) Position 5 10 20 30 A 1. 60 1. 37 1. 6 7 1. 83 B 1. 98 1. 76 1. 60 0.54 C 0.51 0.96 0.67 1. 6 7 D 1. 36 1. 78 1. 76 1. 51 Baselines BN BR PBR 0.98 0.00 1. 21

PAGE 107

Ul G) H 0 4 0 4 c)5 0 A B C 2 5 mm D 20 2 mm 4 0 4 0 10 A B C D 30 2 mm 2 mm .__ _____________ _.._ __ ___ _ __._ ___ _ _ __._ ___ ~--~ ----A B C D A B C D Figure 12. Mean number of normal articul a t i on scores among six judges for each siz e (Ar t icul a t io n R atin g ; n = 4). I I.O co I

PAGE 108

en G) 4 --Baselines-u U) 0 en G) 0 u U) 4 BN BR --Baselines--99Position A PBR 5 1 0 20 30 mm 2 Position B 0 ~-----__._ ____ ...._ _ _.__ __ _____._ ______ _..__ 2 BN BR PBR 5 10 20 30 mm Figure 13. Mean number of normal articulation scores among the six judges for each position (Articulation Ra ting; n = 4) .

PAGE 109

(/) Q) ),.. 0 (.) 4 CJ) 0 (/) Q) ),.. 0 (.) CJ) 4 --Baselines-BN BR --B a selin e s--100Position C \ PBR 5 10 20 Position D 0 '----------~-~--.__ __ ___.. ___ -l..__ 30 mm 2 BN BR PBR 5 10 20 Figur e 13. continued

PAGE 110

-101df = 15, p <0.01). The differences among the four positions were not significant (F = 1.06, df = 3, p = 0.37). How ever, the size variable indicated significant differences (F = 27.24, df = 3, p < 0.01). The Duncan Multiple Range Test indicated significant differences between the 5 mm 2 and the 20 mm 2 openings. The 5 mm 2 opening did not differ significantly from the baseline measures (zero opening) or the 10 mm 2 opening, and no significant difference was found between the 20 mm 2 and the 30 mm 2 openings. The Spearman correlation coefficients show negative correlational associations (r = -0.62, p < 0.01) between size and the response variable. No relationship (r = 0.01, p = 0.85) was found between position and the response variable (Table 6). Cul-de-sac Hypernasality Test The 10 words from the standardized hypernasality test were assessed for normal or hypernasal voice quality by the six judges. The response variables, the words produced with hypernasal quality, were added and statistically analyzed for each palatal opening size/location and for each of the six judges. There were two versions of this test. In one version the subject was exposed to auditory masking with loud noise. In the second version no auditory masking noise was employed. The two versions of this test were treated separately.

PAGE 111

Table 6. -102Spearman's rank order correlation coefficients for the judgment, no errors from the Articulation Rating Test. Position Size No errors 0.07820 0.4489 -0.45755 0.0001

PAGE 112

-103The means and standard deviations for both the unmasked and the masked versions are presented in Table 7 and Table 8, with graphical representation in Figure 14 and Figure 15. It can be interpreted from these tables and figures that the ratio between the correlation coefficients equals 0.39, meaning that there is a 39 percent higher error probability occurring in the masked version of the test. Three excep tions are noted for the palatal openings A20, A30, and on D30. These data also indicate smaller standard deviations occurring with the masked version of the test. It should be noted that the baseline testing also displayed similar differences. The Spearman correlation coefficients show higher association between the masked response variable and size than does the unmasked response variable. Size has a significant correlation coefficient with the masked variable (r = 0.28, p < 0.01), but is insignificant with the unmasked version. Position does not show any significant correlation with either of the two versions of this test (Table 9) . Unmasked version. The analysis of va riance procedure from the unmasked response variable indicated highly sig nificant differences among the means (F = 4.32, df = 15, p < 0.01). The sources of variation come from position and size and there is a significant interaction between these two variables (F = 6.03, df = 9, p < 0.01). The main

PAGE 113

Table 7. Position AUM AM BUM BM CUM CM DOM DM Mean (M) (n = UM M -104scores of the unmasked (UM) and the versions from the Hypernasality Test 10) . 5 3.16 4.50 6.33 7.83 4.50 6. 16 2.66 4.33 BN 2.00 2.50 Size 10 3.00 4.00 4.16 4.66 1. 83 4.83 2.66 3.00 Baselines BR 8.50 5.33 (mm 2) 20 4.16 2.16 0.66 2.50 5.00 5.33 0.50 1.16 PBR 6.33 6.00 masked 30 3.00 2.83 1. 83 4.00 3.50 3.50 6. 16 4. 16

PAGE 114

Table 8. Position AUM AM BUM BM CUM CM DUM DM -105Standard deviations of the unmasked (UM) masked (M) versions of the Hypernasality (n = 10) . Size (rnrn2) 5 10 20 2.13 1. 54 2.71 2.88 2.44 1. 72 1. 75 2.13 0.81 1. 32 2.65 0.87 2.25 1. 32 2.00 2.78 3.31 2.65 2. 16 1. 50 0.54 3.26 2.82 0.75 Baselines UM M BN 1. 89 2.07 BR 1. 97 1. 75 PER 2.58 2.60 and Test 30 2.96 1.83 1. 4 7 2. 19 2.25 1. 97 2.71 1. 94

PAGE 115

5 2 10 I 10 mm mm 10 J ... ' ....... 5 ' ----------.... ~-0 0 A B C D A B C D 2 20 mm 30 mm 10 10 t / .. -------/--4 A B C D A B C D Figure 14. Mean scores on the hypernasality test among the six judges for each size (Hypernasality Test; n; 10). 2 2 Unmasked Masked I I-' 0 O'I I

PAGE 116

U) Q) 0 () Cf) U) Q) 0 () Cf) 10 5 0 10 5 0 -1 0 7B ase li nes -,,... _., / / / BN BR PBR -Baselines -___ .,,,., .... / / / BN BR P B R ..... 5 ,,,.., / ' 5 ' ' ' P o s i ti o n A .... , ' ... ... ..... 10 20 Posi ti on B ' ' 1 0 ' ' 2 0 30 2 mm 30 mm 2 Unm asked M a sked F i g ur e 15. Mea n sco r es o n t he hype rn asa lit y t es t a m o n g th e si x j u dges for ea ch p o s it io n ( Hype rna sa lit y T e st ; n = 10) .

PAGE 117

-108Position C --Baselines-10 ' _ ... 5 ' ' -' ' ' " UJ QJ >-l 0 u CJ) 0 BN BR PBR 5 10 20 3 2 mm Position D --Baselines-10 ---5 r ", / / ,.__ ,, UJ , ' (1) '• / / :....i ' / ' 0 ' u ' ,'. ___,, CJ) 0 BN BR PBR 5 10 20 30 mm2 Unmasked Masked ----Figure 15. continued

PAGE 118

109 Table 9 . Spearman ' s ra n k order corre l atio n coeff i c i en t s for the judgment of hypernasali t y from t he Hypernasality Test. U nmasked Masked Position 0 . 0398 0 . 03352 0 . 6989 0 . 7457 Size 0 . 1140 0 . 2897 0 0 . 2687 0 . 0 0 4 Unmasked 1. 0000 0 .5 6381 0 . 0000 0 . 000 1

PAGE 119

-110effects of position revealed no significant differences (F = 0.52, df = 3, p = 0.67) and the main effects of size were also nonsignificant (F = 3.02, df = 3, p = 0.03). In order to identify where these differences occur, the Duncan multiple comparison procedure was employed . The results show no significant differences among the diffe rent palatal fistula openings. The sma ll es t size opening of 5 mm 2 was not statistically different from the baseline measures at the 0.05 leve l of significance. The Duncan Multiple Range Test demonstrated no significant differences between the sizes 10 mm 2 and 20 mm 2 , b ut these differed from the 5 mm 2 opening. In this test there were, furthermore, no differences between the sizes 5 mm 2 and 30 mm 2 Masked vers ion. The masked version of this test indi cates significant variables among the means (F = 2.75, df = 15, p < 0.01). Th e re was no eviden ce ior a significant interaction between position and size (p = 0.27). In the test of main effects, position (F = 3.61, df = 3, p < 0.01) and siz e (F = 6.39, df = 3, p < 0.01) both showed significant differences. The Duncan multiple comparison test demonstrates that the differences occurred between the experimental positions grouped in C and B versus the group ing A and D. Analyzing the differences between th e sizes, the Duncan comparison test demonstrates that ther e are no differences between the baselines and the 5 mm 2 opening, and that the 5 mm 2 ope ning is s t a tistically different

PAGE 120

-111from the 10 mm 2 and th e 20 mm 2 openings at the 0.05 significance level. Hyp e rn a sality Rating Eight eve-syllable pairs were rated perceptually for hypernasal voice quality. All the no nasality judgm e nts were extracted since this measure represents the cutoff scor e which is employed to a id to the definition and quantification of palatal openings. ana lyz ed . Th ese data were statistically Means and standard deviations of no nasality for the six judges are represented numericall y in Table 10 and T able 11. Figure 16 and Figure 17 demonstrate the da t a graph icall y . The analysis of variance from the response va ri a ble, no nasality, demonstra t ed highly significant d iff e rences among th e means (F = 2.31, df = 15, p < 0.01). Insignificant differences of the main effect , pos iti o n, was shown (F = 0.35, df = 3, p = 0.78). However, highly significant differences we r e demonstra t ed by the main effect, size (F = 10.28, df = 3, p < 0.01). A size-by-size comp a rison with the Dun can ( p = 0.05) indic a t ed no significant differences be t ween the baselines and th e 5 rnrn 2 opening. However, there were significant differences between the 5 mm 2 and all of th e larg e r openings . Th e opening sizes,

PAGE 121

-112Table 10. Means of normal resonance scores among the six judges (H ypernasality Rating; n = 8) . Size (mm2) Position 5 10 20 30 A 5.66 3.16 2.16 1.50 B 7.83 5.33 1. 66 1. 66 C 7.33 7.50 0.16 1. 66 D 4.66 6.00 1. 83 3.16 Baselines BN BR PBR 7 . 66 6.83 7.83 Table 11. St a ndard dev i ations of resonance scores among th e si x judges (H ypernasality Rating; n = 8 ). Size (mm 2 ) Position 5 10 20 30 A 1. 96 3.71 3.48 2.34 B 0.40 3 .07 2.65 2.87 C 1. 21 1. 22 0.40 2.58 D 3.14 1. 54 3.25 1. 22 Baselin e s BN BR PBR 0.51 2 . 85 0.40

PAGE 122

Ul (]) H 0 u Cl) 8 5 0 8 5 0 A A B C D 2 20 mm B C D 8 10 mm 2 5 0 A B C D Position 8 30 2 mm 5 0 A B C D Pos1 ion Figure 16. Mean number of normal resonance scores among the six judges for eac h size (H ypernasality Rating; n = 8 ). f--' f--' w I

PAGE 123

rJJ Q) H 0 u U) rJJ Q) H 0 u (I) 8 5 0 8 5 0 --B aselines -BN BR --B ase lin es -BN BR -114PBR 5 10 PBR 5 10 Position A 20 2 30 mm Position B 20 Figure 17. Mean number of normal resonance scores among the six judges for each position (Hypernasality Ra ting; n = 8) .

PAGE 124

U) (1) H 0 0 CJ) U) (1) H 0 0 8 5 0 8 5 CJ) 0 --Baselines-BN BR --Baselines-BN BR Figure 17. continued -115PBR 5 \ PBR 5 Position C 10 20 Position D 10 20 30 mm 2 2 30 mm

PAGE 125

-1162 10 mm , 20 rnrn 2 , and 30 2 mm ' were not si g nificantly different from each other. The relationship between the response variable, no nasality, and size is represented by a negative correlation of -0.45 (p < 0.01), indicating that increased opening size will negatively influence speech production . There is no relationship between position and the response variable (r = 0.78, p = 0.44) (Table 12). Instrumental Measurements Nasal Emission T es t The first instrumental evaluation tool to be used was the Nasal Emission Test. To assess nasal emission, 10 responses were ob t a in ed twice. The repeated response data were exa ctl y eq ual to the data obtained during the original test experiment. The results are presented in Table 13. The results from this test can be analyzed best by simply describing that all three baseline tests demonstrated no nasal emission. Howe ver , as soon as the smallest opening 2 of 5 mm was introduced,nasal air emission was measured throughout. When the subject produced the words with the 2 . h . 5 mm openings, e was able , at the A and B positions, to prevent air leakage from th e nose for 1 trial and 3 trials, respectively, out of a total of 10. This also occurred at the B position with the 10 rnrn 2 opening where he could

PAGE 126

-117Table 12. Spearman's rank order correlation coefficients for the judgment, no nasality, from the Hypernasality Rating. Position Size No nasality 0.01872 0.8571 -0.62740 0.0001

PAGE 127

Table 13. Mean errors of nasal air emission (NEM) (Nasal Emission Test; n = 10). Position Size NEM Position Siz e NEM A 5 9 B 5 7 A 10 10 B 10 8 A 20 10 B 20 10 A 30 10 B 30 10 C 5 10 D 5 10 C 10 10 D 10 10 C 20 10 D 20 10 C 30 10 D 30 10 f-' f-' CXJ I Baselines BN BR PBR 0 0 0

PAGE 128

-119prevent nasal airflow on two trials (Figure 18 and Fig ure 19) . The relationship between the nasal airflow variable and size versus position is indicated by a Spearman rank order correlation coefficient. Table 14 shows a significant association {p < 0.01) between both variables, size and position, to the response variable. Aerodynamic Measurements There were two aerodynamic test measurements. They were the measures of intraoral air pressure and differen tial pressure. The instrument PERCI II was employed for the collection of data. The data were obtained from repeated measurements of five repetitions of a tested word "pop. II These two aerodynamic measurements were tested for all 19 test conditions and in the same order as the per ceptual testing was performed. Intraoral air pressure. Means and standard deviations of the values of intraoral air pressure were computed and are represented in Table 15 and Table 16. A graphical representation of intraoral air pressure by size and posi tion is given in Figure 20 and Figure 21. An analysis of variance of intraoral air pressure for the word "pop" shows highly significant differences among the 16 openings (F = 3.67, df = 15, p < 0.01). It is

PAGE 129

10 5 0 A 10 Ul 5 (I) H 0 u (/) 0 A B C B C 2 5 mm D 20 mm 2 D 10 mrn 2 10 5 0 A B C D 2 30 mm 10 5 0 A B C D Figure 18. Mean errors on the nasal emission test for each size (n = 10). I I-N 0 I

PAGE 130

Ul (!) H 0 0 Cf) Ul (1) H 0 0 Cf) 10 5 0 10 5 0 --B ase lines-BN BR -Baselines -BN BR -121Position A PBR 5 10 20 30 mm Position B PBR 5 10 20 30 mm 2 Figure 19. Mean e rrors on the Nasal Emission Test for eac h position (n = 10).

PAGE 131

10 Ul 5 Q) H 0 u CJ) 0 10 Ul Q) 5 H 0 u CJ) 0 Figure 19. --Baselines-BN BR --Baselines-BN BR continued -122Position C PBR 5 10 20 30 mm 2 Position D PBR 5 10 20 30 mm 2

PAGE 132

-123Table 14. Spearman's rank order correlation coefficients for the values of intraoral air pressure (IPOP) and differential pressure (DPOP) and nasal airflow (NEMT). Position Size Instrumental Measures IPOP 0.6222 0.5835 -0.05326 0.6389 DPOP 0.25981 0.0199 0.10944 0.3338 NEMT 0.33799 0.0022 0.49810 0.0001

PAGE 133

-124Table 15. Means and standard deviations of intraoral air pressure ( IPOP) and differential pressure (DPOP) when the word "pop" was pronounced. IPOP DPOP Position Size X S.D. X S.D. A 5 2.64 0.78 l. 66 0.59 A 10 2.35 0. 84 l. 36 0.35 A 20 2.08 0.23 1. 14 0.22 A 30 2.30 0.48 l. 39 0.29 B 5 2.94 0.76 1. 49 0.20 B 10 2.23 0.65 1. 61 0.28 B 20 2.27 0.45 1. 10 0.22 B 30 2.33 0.29 1. 19 0.35 C 5 3.32 1.16 1. 28 0.12 C 10 4.13 0.90 1. 98 0.98 C 20 2.80 0.92 4.47 0.75 C 30 2.98 0.23 l. 6 7 0.5 4 D 5 1.92 0.60 l. 64 0.56 D 10 2.32 0.64 l. 48 0.27 D 20 2.04 0.16 1. 78 0.45 D 30 2.51 0.26 1. 63 0.30 BN 0 6.19 1. 43 3. 59 0.32 BR 0 6.84 0.60 5.05 1.04 PBR 0 5.26 0.85 1. 43 0.70

PAGE 134

Table 1 6 . Intraoral air pressure and differential pressure measurements for each fistula size and location. Positions BN BR PBR A B C D Size 2 ( mm ) Intraoral Air Pressure ( cm H2O ) 2.64 2.94 3 . 32 1. 92 5 6 .1 9 6 . 84 5 . 26 2 . 35 2.23 4 . 13 2 . 32 10 2 . 08 2 . 27 2 . 80 2 . 04 20 2 . 30 2 . 33 2 . 98 2.51 30 Differential Pressure ( cm H2O ) 1. 66 1. 49 1. 28 1. 64 5 3. 59 5 . 05 1. 43 1. 36 1. 61 1. 98 1. 48 10 1. 14 1. 10 4 . 47 1. 78 20 1. 39 1. 91 1. 67 1. 63 30 I I-' N lJl I

PAGE 135

5 mm 2 10 mm 2 7 t t Q) t)"\ i:: cu H t)"\ / i:: cu H r-1 cu 0 s N H ::r: 0 z s 3 CJ + r-1 cu I s / I-' H N 0 / O'I z / I + Q) H ::l en en Q) H 0-1 0 A B C D A B C D Position Figure 20. Intraoral air pressur e variation as related to fistula position for each size.

PAGE 136

2 I 2 t 20 mm 1 ru 30 mm 7 / / / / (1) tJ, tJ, rd rd / H H / / r-1 r-1 B / rd 0 E: / I N H / f-' 0 0 N z z --.J E: 3 I u + + (1) H ::J Ul Ul (1) H p.. 0 A B C D A B C D Position Figure 20. continued

PAGE 137

0 N ::r:: 8 0 (jJ H ::1 UJ UJ (jJ H 0. 7 3 0 Figure 21. -128Position A = Position B = Position C = 0 Position D = D 5 10 20 30 mm 2 Intraoral air pressure variation as related to fistula size for each position.

PAGE 138

-129interesting to note that the main effect, position, was a significant variable in this test (F = 11.61, df = 3, p < 0.01). Size, however, was insignificant (F = 1.98, df = 3, p = 0.13) as well as any interaction between the two variables (F = 1.58, df = 9, p = 0.14). A Duncan a posteriori comparison test shows that the significant differences occur between the baselines and the experimental openings. By analyzing the individual experimental openings, A20, D20, and DOS proved to be the most different at the 0.05 level of significance. A Spearman rank order correlation coefficient of 0.06, p = 0.58, indicates no correlation between position and the intraoral pressure when the word "pop" was produced. Neither was a significant correlation obtained between the variables, size and the word "pop" (Table 14). Differential pr e ssure. Means and standard deviations of the differential pressure measur e ments were computed and are shown in Tables 15 and 16. Figure 22 and Figure 23 demonstrate graphically the differential pressure variation as related to position and size. An analysis of variance of differential pressure for the word "pop" shows highly significant differences among the 16 openings (F = 13.86, df = 15, p < 0.01). The inter action between the v a riables, position and size, was significant (F = 15.14, df = 9, p < 0.01) as well as the main effects of position (F = 16.98, df = 3, p < 0.01) and

PAGE 139

0 N ::r:: 8 u Q) H ::J U1 U1 Q) H 0. 5 2 10 2 t mm mm t 7 / Q) Q) O' O' i::: i::: ro ro H H rl rl ro ro 8 8 H / H 0 0 z z 3 t + 0 A B C D A B C D osi ti on Figure 22. Differential pressure variation as related to fistula position for each size. I I-' w 0 I

PAGE 140

20 2 I t ru / 30 2 t mm mm 7 / / Q) / ty, I t,, / j i / / , / / ; / 0 / N e / I ::r:: 0 / I-' 0 f=: z z / w I-' u 3 I f t Q) H ::l UJ UJ Q) H p.. 0 A B C D A B C D Pos iti o n Fi g ur e 22. continu e d

PAGE 141

-132Position A = 7 Position B = ,.. Position C = 0 Position D = D 3 0 5 10 20 30 mrn 2 Figur e 23 . Di fferen ti al press ur e variation as related to fistula size for each position .

PAGE 142

-133of size (F = 6.89, df = 3, p < 0.01). The Duncan multiple comparison test indicates significant differences between the two baselines, BN and BR, and the experimental open ings, except for the opening C20. The experimental opening C20, however, is not significantly different from the base line, BR. The baseline, PBR, is found not to be signifi cantly different from the experimental openings. Among the experimental openings, the only significant differences are to be found between Cl0 and B20. The results from the Spearman correlation indicate a significant relationship between position and differential pressure when the word "pop" was produced (r = 0.26, p = 0.01). The variable, size, however, does not correlate significantly with differential pressure when the word "pop" is pronounced (r = 0.10, p = 0.33) (Table 14). Reliability It was necessary to perform reliability testing on an extracted portion from some of the above testing material. It was important to perform this reliability testing due to poor reliability results reported in the literature (see Chapter II and Chapter III). It has been reported that, in the judgment of hypernasality, both intraand int e rrater ~eliability are low. Reliability testing was performed on material from the Hypernasality Rating Test.

PAGE 143

-134The results form two paired nonsense syllables "pin-pip" and "pin-pib" were extracted from the test materkal, as they had been repeated twice. Intr a rater Reliability The testing of intrarater reliability was computed with the Spearman's rank-order correlation coefficient. The Spearman's rank-order correlation coefficients, suitable for ranked data and ordinal scales, were statistically analyzed and are represented in Table 17. From the statistical analysis, the correlation coefficients are mainly in the .90, .80, or .70 ranges with significance levels of p < 0.01. This occurred for all values except for judge number one that exhibits a 0.36 correlation coefficient at a 0.02 level of significance. Interrater Reliability The agreement among the judges was computed statistically employing Spearman's rank-order correlation coefficients. As Table 18 indicates, the interrater reliability factor was satisfactory. Interrater reliability among the judges was significant at the 0.0001 level of significance for all judgments except for the judgments made by judge number two. The correlation coefficients for judges one, three, four, five, and six were all in the midand the higher .70

PAGE 144

Table 17. Spearman's rank order correlation coefficients for the intr arater r e liJbility ( se lf-agreement) t es t. First Response Second Respons e Fl F2 F3 F4 FS F6 S l 0 . 36758 0.0232 S2 0.79292 0.0001 SJ 0.92942 0.0001 S4 0.73983 0.0001 ss 0.92214 0.0001 S6 0.83442 0.0001 I f-1 w V I I

PAGE 145

Table 18. Spearman's rank order correlation coefficients for the interrater reliability test (agreement among the si x judg es ). Jl J2 J3 J4 JS J6 Jl 1.00000 0.36137 0.82155 0 .74797 0.79525 0.78265 0.0000 0.0258 0.0001 0.0001 0.0001 0.0001 J2 0.36137 1. 00000 0.3 3 658 0.20807 0.34132 0.31082 0.0258 0.0000 0.0388 o. 2100 0.0360 0.0575 J3 0.82155 0.33658 1.00000 0.79272 0.88117 0.88582 0.0001 0.0388 0.0000 0.0001 0.0001 0.0001 J4 0.74797 0.20807 0.79272 1. 00000 0.78778 0.75618 0.0001 0.2100 0.0001 0.0000 0.0001 0.0001 JS 0.79525 0.34132 0.88117 0.78776 1. 00000 0.87514 0.0001 0.0360 0.0001 0.0001 0.0000 0.0001 J6 0.78265 0.31082 0.88582 0.75618 0.87514 1. 00000 0.0001 0.0575 0.0001 0.0001 0.0001 0.0000 I f-' w (J'\ I

PAGE 146

-137and .80 ranges. The correlation coefficients for judge number two, however, were in the .20 a nd the .30 ranges with significanc e levels of only 0.02, 0.03, and 0.05.

PAGE 147

CHAPTER V DISCUSSION AND CONCLUSIONS The main purpose of this study has been to initiate quantitative definitions of palatal fistulas by assessing their influence on speech and resonance. Further objec tives were to establish the critical size and location of these openings and to determine when and how the fistula is detrimental to normal speech production. Four palatal openings were artificially introduced through a palatal research prosthesis and defined as to location and size. All four openings simulated four different fistula sizes ( 2 2 2 2 30 rnrn2). 5 mm, 10 mm , 0 mm , and Each of th e se three openings (A, B, and C) act e d a s a hard p al a t e fistula and the fourth opening (D) was selected for comparison with VPI (Figure 9 ) . It was intended that conclusions from previous reports on VPI would s e rve a s v alidit y comparisons for the D-opening. In other words, if the dat a , collect e d from measurements with the D-op e nings, compared well with data on VPI, this new data would constitut e a v alid basis for the validity of the exp e rimental model a pplied in this research. Four questions were asked in Chapter II and several testing tools were employed to a nswer them. The results -138

PAGE 148

-139from the listener judgments and the instrumental measurements were reported in Chapt e r IV. It is th e purpose of this chapter to discuss the implications of these results. Discussion Clinical expe rienc e and qualitative descrip tion s have implied that speech is negatively affected by the increasing size of a palatal fistula. Th e significance of location has not been clearly estab lished, although qualitative statements have d ir ected a tt en tion to the fa ct that pos teriorly located openings introduce more speech p roblems than openings located ante ri orly . Hard Palate Fistulas Related to Velopharyngeal Insufficiency and Speech Many studies have addressed VPI and its influence on speech. The results from such studies have concluded that speech, in terms of nasality and a rticul at ion, is nega ti ve ly affected by openings in the ve lop ha r yngea l port. This has been demonstrated from perceptual judgment testing. Instrumental methods, such as spectrography and aerodynam ic measurements, have substantiated such claims by confirming that chan ges in spectrographic data and intraoral a ir pressure can be registered with even smaller VPis than when employing perceptual tests.

PAGE 149

-140It was demonstrated in the present work that hard palate fistulas representing an opening between the oral and the nasal cavities also impair speech. With fistulas, as with VPI, both articulation and resonance quality are negatively influenced. In determining the negative effect from hard palate fistulas both perceptual judgment and instrumental tests were employed. The sensitivity of the instrumental tests, noted in studies on VPI, was confirmed in this study. General agreement among authors, who employed several perceptual measurements in their studies, implies that speech impairments of nasality and articulation are introduced when there is a velopharyngeal opening between 2 2 2 10 mm and 20 mm. With openings of approximately 20 mm ' hypernasal speech was always r eg ister ed (Isshiki et al., 1968; Shprintzen et a l., 1977; Warr e n, 1 964a ,b). Employing instrumental tests, such as spectrography or aerodynamic measurements, smaller changes of nas al it y could be detected than with perceptual tests. This was demonstrated by Bjork (1961) and by Warren et al. (1985). Bjork (1961) stated that small acoustic changes were registered with the spectrograph at velopharyngeal openings of approximately 10 mm 2 . Warren et al. (1985) reported velopharyngeal openings from 10 mm 2 to 19 mm 2 as having borderline inadequacy. However, they reported further that an opening as small as 5 mm 2 can introduce mild inconsistent hypernasality. It was determined, in this present study, that

PAGE 150

-141all four simulated fistulas, including the three hard palate fistulas and the velarly positioned one, compared well to previous research on velopharyngeal insufficiency. Palatal openings as small as 5 mm 2 showed decreased intra oral air pressure and differential pressure and evidenced nasal emission of air. These results were obtained fro m the three instrumental tests. Two of the perceptual listen ing tests indicated that speech began to be distorted at the 10 mm 2 openings, wh ile th e other thr ee percep tu al tests 2 signaled speech deterioration at openings of 20 mm. In summarizing the findings, from this study and f rom previous research, it has been demonstrated that, regardless of th e testing method, both hard palate fistulas and ve lopharyng ea l insuffici ency cause a mild speech diso rd e r to oc cur with 5 mm 2 openings. Consistent errors of speech were intro duced with hard palate fistulas and velopharyngeal i nsuf ficiency between 10 mm 2 and 20 mm 2 . Speech was always associated with hypernasality a nd misarticulations when the openings were 20 mm 2 or 30 mm 2 . Effects of Fistula Size on Speech The findings indic a te that th e size factor was mor e important than the factor of position. W hen the size of a palatal fistula was incr ea s ed , speech was negatively influenced. This could be m eas ured throughout all of the

PAGE 151

-142eight different test measurements, obtained through the five subjective perceptual measurements and the three objective instrumental measures. The findings from the five different tests of articulation and resonance can be collectively described b e cause of their similarities (Figures 11, 13, 15, and 1 7) . The r e sults of the three instrumental tests of nasal airflow, intraoral air pressure, and differential pressure support this indication (Figures 19, 21, 23). It was observed that the small e st opening, 5 2 mm ' always caused a lowering of the score, indicating both articulatory inadequacy and introduction of hypernasal voice quality. However, this change was small and the differences were not significant. The articulatory errors were charac terized as inconsist e nt imprecisions or distortions. Hyper nasality wa s regist ered inconsist en tly as slightly nasal. The 10 mm 2 opening g e nerally represented a larg e r increment of errors for all five tests. In spite of this seemingly 2 large increase of errors, fistulas with 10 mm openings also failed to demonstrate significance on all but two of the perceptual tests. The Articulation Rating and the masked version of the Hypernasalit y Test indicated sig nificant differences of the speech production caused by 2 10 mm palatal op e nings. Openings as large as 20 mm 2 caused a significant increase in errors for all perceptual tests. When asses ing the effects of the 30 mm 2 opening, a nonlinearity was

PAGE 152

-143observed in the form of an apparent leveling off in the error count. This phenomenon was found with only a few exceptions in all of the perceptual studies and was con firmed by the instrumental test results. Whether this is actually a leveling off that would extend through even larger size openings cannot be confirmed since sizes larger than 30 mrn 2 were not tested. Previously reported similar incidents involving velopharyngeal insufficiency indicated a le ve ling off of fricative intelligibility and nasality at openings ranging from 3.5 millimeters to 7.0 millimeters (Subtelny e t al., 1961). For comparison with the measures in the prese n t study, a 3.5-millimeter diameter opening 2 corresponds to 9.6 mm, and a 7.0-millimeter opening equals 2 38.5 mm. Other nonlinear relationships were evidenced between VPI and intr ao r a l pressure (Warren, 1964a; Warren & Devereux, 1966; Warren et al., 1981). This nonlinearity 2 was found to begin at sizes larger than 20 mm. However, when the openings exceeded 7.5 millimeters (44.2 mm 2 ), intelligibility and nasality continued to deteriorate (Subtelny et al., 1961). Evidence from the different test measurements in the 2 present work clearly show that 10 mm palatal fistulas start to affect speech negatively. When th e openings reach 20 mrn 2 , speech is signi f icantly deteriorated both in terms of articulation and nasal resonance. When analyzing the three instrumental tests, separate from the perceptual

PAGE 153

-144tests, a significant number of errors was found with the introduction of a 5 mm 2 fistula. Thus it can be concluded that the instrumental tests show higher sensitivity in detecting the negative influences aff e cting speech than do the perceptual listening tests. Critical Fistula Size By specifically estimating the importance of size, the response data were pooled from all tests (Table 19). It can be concluded from this table that an opening as small as 5 mm 2 introduces negative effects that interfere with normal speech production. This measure is statistically valid in all the thre e instrumental tests. However, the five perceptual tests indicate the critical size as lying between 10 mm 2 and 20 mm 2 . The sensitivity of the instru mental measur e ments proved to be superior to the perceptual tests. The lack of sensitivity of the perceptual tests can be attributed to the fact that they had to rely on human listening skills for detection of any speech defects. The apparent sensitivity of the instrument a l tests was con firmed when compared to the perceptual t es ts. Effects of Fistula Position on Speech The significance of fistula position was not conclu sive in all tests. However, in four out of the eight tests,

PAGE 154

-145Table 19. Critical size (mm 2 } of palatal fistulas for each of the eight measurements of speech. Test Critical AT ARATE HTUM HTM HRATE NEMT Po L'ip AT= Articulation Test ARATE = Articulation Rating Test HTUM = Hypernasality Test: Unmasked HTM = Hypernasality Test: Masked HRATE = Hypernasality Rating Test NEMT = Nasal Emission Test Po= Intraoral air pressure L'i p = Differential pressure 20 10 20 10 20 5 5 5 size (mm2}

PAGE 155

-146fistula location was demonstrated to be significant. Table 20 shows these results. The tests that displayed most sensitivity in conducting the size analysis were the three instrumental tests (nasal airflow, intraoral pressure, and differential pressure) and the masked version of the Hyp e rnasalit y Test. These same tests were the ones that identified the significance of fistula position in this quantification attempt. The a posteriori tests identify those locations that are not significantly different from each other. The y are marked in Table 20 by underlining. By examining the significant results of fistula posi tion, the following can be interpr e ted. The velar posi tion, D, appears in three of th e four tests as being the position exhibiting the most speech errors. The an terior position, A , appears once as the position with the most errors. Further exam ination of any trends from the remain ing tests supports the observation that fistulas occurring at the D position introduce more speech disorders than do fistulas at other locations. Fistulas at position A are second and position B (the mid-opening in the hard palate) are third. At no occasion was the posterior position, C, registered as the opening to score the most errors. Thus, it can be concluded that, although fistula position does not show significanc e through all tests, it does so 50 percent of the time.

PAGE 156

147 Tab l e 2 0. The infl u ence of fi s t u la posit i on of speech . Signif i cance C r itica l T est p < 0 .01 f i s t u l a pos it ion A T No B ,C, D , A ARA T E No D ,C, A , B H TU M No C, A , B , D H T M Yes C , B , A , D HRATE No C, B , D , A N E M T Yes ~, ~ , C , D Po Yes ~ , B,A , D 6 p Yes ~, D , B , A A T= Ar t iculation T est ARA T E = Articulat i on Rat i ng T e s t HT UM = H ypernasal i ty T es t: Un m asked H TM = H ypernasality Test : Masked H RA TE = Hypernasali t y Ra ti ng Test NEM T = Nasal Emission Test P o = Intr aora l air pressu r e 6p = Di ffe r ent i al press u re Wors t ra ti ng A B D D A D D A U n de r lined = Positions that are not significantly different from each o t her ; the listing shows a trend if no significance is indicated.

PAGE 157

-148With the assumption that the fistula position D (drille d in the velar part of the prosthesis) can be compared to VPI, it appears that VPI causes more speech inadequacies than any hard palate fistula. Furthermore, among the three hard palatal fistulas (A, B, and C) it can be implied that a fistula located anteriorly (A) would cause more speech problems than fistulas in the middle of the hard palate (B). A fistula at the posterior position (C) does not result in any speech disturbances. This does not mean that a fistula at position C does not have any negative influences on speech. It merely indicates that a fistula at the posi tion C does not in itself cause any significant speech disorders, as would be the case if the fistula were found at A or at the velar position, D. When the size factor is taken into account, it will a lways be the main effect according to which the speech defect can be related. Previous qualitative reports have indicated that posteriorly positioned fistulas were causing more speech disorders than those located anteriorly. From this study, it can be concluded that, before we can make such a judg ment, the exact location of the fistula should be known. If the fistula occurs in the soft palate, the above state ment of position is true. However, relative to hard palate fistulas, it does not appear to be true. With the results from this work, it is evident that there is an important need to specify the exact location of a fistula in order

PAGE 158

-149to attempt any inference of its possible influence on speech production. Auditory Masking The Hypernasality Test was adm inist e red twice . The second time the testing was performed, the test subject was exposed to auditory masking noise . This auditory masking experiment was carried out to evaluate a possible influence of speech loudn ess on nasality. Clinical experience has shown that speakers with in adeq u a t e velopharyngeal closure present with a reduced speech loudn e ss level . This was also discussed by Bernthal a nd Beukelman (1977). The reduc tion of the loudness level has been expla in ed as an attempt of the subject to compensate for nasality through auditory feedback. Furthermore, a damping of t he a coustic speech signal may occur because of the increased oronasal coupling, which would cause the overall int ens it y level to become reduced. In this study, auditory masking was employed in order to prevent any audi tor y feedback and thereby avoid any com pensation attempts that might take place during testing. Due to the high noise level of 90 decibels , this experiment had to be minimized to comprise only one test per baseline and experimental opening . Figure 14 and Figure 15 show the relationship between th e two tests. The correlation

PAGE 159

-150study indicat ed that the masked version was significantly more sensitiv e in detecting more nasality errors when the factor of siz e was manipulated than was found in the unmasked version. B y employing auditory masking more e rrors of nasality were ev idenced and these erro rs could be 2 measured with pala tal fistulas being as small as 10 mm. The unmasked ve rsion of this test did not register such speech disorders until th e fistula size was 20 rnrn 2 large. The factor of posit i on was not a si g nificant variable on the two versions of the Hyp e rnasality Test. There was, however, a difference between the unmasked and the masked versions, indicating higher sensitivity when auditory maskin g was employed. Reliability It has been indicated from previous research that bot h intraand interrater reliability have been poor for perceptual jud gme nt testing of nasality. From this work, intra r a ter reliability (s e lfag r ee ment) of each judge was fo und to be high for all but judge number one (T ab l e 1 7 ). Interr ater reliability, or the agreement among the six judges, wa s also high for all but judg e number two. It appears that these generally high measures of reliability may b e attributed to the speech sa mpl e which consisted of paired eve-syllables e mploying plosive sounds connected by the vowel /i/ requirin g a high l eve l of intr ao r a l air pressure.

PAGE 160

-151Because of high intraand interrater reliability, it appears that the conclusions from this study should be similarly reliable. Conclusions The primary objective of this work was to conduct basic research to quantitatively describe hard palate fistulas and their influ e nce on the speech variables, articulation, and resonance quality. The speech correlates of nasal air flow , intraoral air pressure, and differential pressure also were included in the test battery to increase the validity of the results. This quantitative information will augment the facilitation of t he choice of treatment for hard and soft palate fistulas. One of the main goals for the speaker and, thus, in the treatment of patients with a palatal disorder, is to establish ac cept a bl e speech as it is perceived by fellow communicators . If a critical size opening can be determined by e mploying perceptual listening tests, this seems to be th e most import ant means of assess ing spe e ch as it is functionally effec tive, even though errors may be detected are employed. sooner when instrument a l tests A close relationship was found to exist between the results from the simulated hard palate fistulas and the soft palate fistula as well as with results on VPI from previous research studies. The r es ults from this study

PAGE 161

-152show that palatal fistulas affect speech differently depending on their size and position. It was demonstrated that a mild speech disorder is created with a palatal fistula as small as 5 mm 2 . Trends from the subjective judgment measure ments demonstrated this feature, which was determined to be significant by the objective instrumental tests. Speechlanguage pathologists are, however, mainly interested in the definition of palatal fistulas according to the effect that they exert on speech. The results from the perceptual tests, in this study, support the clinical qualitative findings that increasing fistula size causes an increasing amount of speech disorders. It was, however, determined that fistula sizes between 10 mm 2 and 20 mm 2 negatively influenced both articulation and resonance quality. of 20 mm 2 and 30 mm 2 increased the severity of the speech disorder. Sizes A leveling off was found to occur with openings of 2 30 mm. It is, however, not certain that such a leveling off would continue with increasing sizes. Similar phenomena of a leveling off were observed with velopharyngeal insuf ficiencies. However, as size continued to increase, speech continued to deteriorate. The effect of position of hard palatal fistulas was negligible. Although velar-positioned fistulas were observed to cause more speech problems, the anterior hard palatal position was observed to initiate more speech

PAGE 162

-153disorders than the other positions among the hard palatal fistulas. The conclusion is, therefore, that palatal open ings located posteriorly in the hard palate tend to cause the least disorders of articulation or nasality. The results from this study appear explicit. Some unexplained results, however, did occur. These do not form a trend, but appear solely to be caused by isolated errors of measurement (Figure 11 and Figure 15). The possible error sources may be attributed to complications in the nasality testing. It has been suggested also that varia tions in configuration of the oral tract may occur. One hypothetical exp lanation would be that measurement dif ferences could occur from attempts to compensate, that is, to try and self-correct by changing the oral configuration. Such chang es of configuration would typically involve factors such as position of the ton gue , the size of the mouth opening, and the size and shape of the ve lopharyngeal opening. It should also be mention e d that the subject had worn the research pros thesis for 10 days . Although he was perceptuall y judged to have mastered normal speech, it is uncertain whether his speaking behavior could have been improved further if the p r act icing time had been longer. The subject also expressed sli ght dissatisfaction with the research prosthesis. With the basic quantitative def initions obtained from this study, it seems only to be the beginning of a developing

PAGE 163

-154research on hard palatal fistulas. Further research is strongly advocated. Specific interests would be to investi gate further importance of fistula position. One question would be whether smaller decrements between the positions would influence speech production. Furthermore, it would be interesting to investigate whether fistulas located away from the midline would signal any differences. Fistula shape may be another important factor. It is felt that quantitative information of fistula shape would contribute to the understanding of palatal fistulas and their influence on speech. Finally, comparisons of these simulated palatal fistulas to pathological oronasal fistulas in a control study would enhance the data from this model. Following such a study, guidelines could be established that would contribute to the treatment of palatal fistulas.

PAGE 164

V) ..J w "' u w 0 3: ..J w > w ..J a ..J 2 ei a: i= " a: 1 2 l 210 1 0 0 f~ 10 . .,, 20 IQ l0 60 70 80 0 j 1 0 0 II 0 I 2 0 +-+APPENDIX A AUDIOMETRIC EVALUATION I Evaluator: ________ _ _ Date : __________ Audiom•ter : FREQUENCY IN II ERTZ l00 l000 2000 7 50 1 )00 3000 I I I /I'\ I ,,. " I . f, -,,._ ..... " fJ~ ~ ;,----.. I ,i,, " , I i '-/ r'\ I I I ' V I I I ' I I I ' I I I I I t I I I . I I I I ' I I ' I ' I ' ' I ' I I I I I I I I I I ' ' ' t I I I I ' I I I I ' I I I : I I . , I I I I I I I I I I I ' I I ' I I I . I I I I I I I I I I I I I -+ -t-t6000 I I l -+ I I I I I I I I I I , I I I I ' I ' I I I I I ' I CO DE : ROOD AIR CONDUCTION MASK EO AIR ,er.,. -'< tiJ oo R i~h• ( R rJ) Ri~h, (Re d) )(--: Lrh (Blue} o oLdt {Blud DON E CON DUCTIO N MASKED DONE ,_. R i~ht (Red) ,__. Ld, (Blur) Ot her : Tcu Quiet ___ _ Mod. Qui., __ Noisy ___ _ P T SRT A..-c . R L Ri~ht (Red) -.4 ,oLeft (Blue ) T ~st Rdi•biH{y ~ --fair ___ _ P oo r ___ _ Di•c. SL I E """ IT"' R L -+-:c 1c 1;+-j ,._ ____________ E_ . F_ F_E_ _ _T_I V _ E _ . _M_A_ S _ K _I_ N _G ______ ~B~C~L_ IR _ R __ ~. ----------~---~ REMAR KS: II<. R L -155

PAGE 165

APPENDIX B ORDER OF THE EXPERIMENTAL TEST CONDITIONS 1. 0 Baseline (BN) 2. 0 Baseline (BR) 3. C 10 4. C 20 5. D 10 6. C 30 7. A 10 8. B 10 9. D 20 10. A 20 11. B 20 12. B 30 13. A 30 14. D 5 15. A 5 16. B 5 17. D 30 18. C 5 19. 0 Post-Baseline (PBR) -156

PAGE 166

APPENDIX C CRANIOFACIAL CENTER SPEECH TESTS The Bzoch Error Pattern Screening Articulation Test (BEPSAT) C = Correct I= Indistinct due to nasal emission D = Distortion SS = Simple substitution GS= Gross substitution O = Omission /p/ /bl /ti /di /k/ lg/ I f I /vi I -e I I ,r I /s/ /z/ I I !) I /wl Ill /y/ I r I /m/ / n/ I) I /sp/ / s tr/ /st/ /SK/ /sm/ I b I/ /kl/ /br I PLOSIVES C T D SS GS 0 aPPle baBy rnounTain canDy chiCKen waGon FRICATIVES elePHant shoVel too Tl l brush feaTHer bicyCLe sciSSors diSHes televiSion sandWich I I b:illons I I onlOns ! i allllow I I -NASALS haMMer baNana haNGer 11 I 11 11 BLENDS SPider STRawberries STar SKirt i SMoke LJ_, _. Block I Clown -------B 1100111 -157

PAGE 167

-158Hypernasality Test Circle words or vowels on which a shift in tone quality occurred when nares were clos ed (i.e., indicated vela pharyngeal insufficiency); ask patient first to repeat word lou dly, then to say it again while nares are pinched . A. beet bought bit boat bait boot bet but bat Bert For infants or speechless try simpler test of alternately pinching and opening nares as patient utters prolonged vowels /i/ as in see and /u/ as in new; circle those that shift. B. /i/ /u/ /i/ /u/ /i/ /u/ /i/ / u/ /i/ /u/ Hyp e rnasality Score .... / .... Nasal Emission T e st Circle words or syllables utt e r ed that ev idenced nasal air flow (i.e., indicated velopharyngeal insufficiency); air flow checked by FLA II ( ) , Paddle ( ) , Bubble ( ) , other A. people baby paper Bobby puppy bubble pepper B.B. piper bye -b ye B. For infants test /pi/ syll a bl e ten times: p p p p p p p p p p N asa l Emission Score .... / .... Used by permission, K.R. Bozch.

PAGE 168

APP E N D IX D AR T I C ULA T ION RATING TEST Ordina l Scale for Judging Articulation 1 . No articulation errors and comple t ely intel l ig i ble . 2. Makes a few art ic ulation errors such as imprecise and distorted so u nd s d u e t o nasal emission ; speech is s till within normal limits . 3 . Makes consistent ar t ic u lat i on errors such as substi tut i ons and omissions which reduce overal l intelli gib i lity . 4. Makes consistent art i culation errors such as above including gross subst i tution errors (glottal stops and pharyngeal fricatives) which reduce overall intelli gibility grossly. 5. Unintelligibl e Circle the number wh i ch bes t corresponds to your judgment of articulation 1. In the evening Connie Watches TV with me 1 2 3 4 2 . Baby looks as mother washes the dish e s 1 2 3 4 3. The soldier shot an arrow at a grasshopper 1 2 3 4 4 . The wicked cat got in the buggy l 2 3 4 1595 s s 5

PAGE 169

APPENDIX E HYPERNASALITY RATING TEST Ordinal Scale for Hypernasality Rating 1. none 2. mild 3. moderate 4. severe 5. unintelligible pin pib pin pit pin pip pin pik pin pid pin pib pin pig pin pip No detectable hypernasality. Sliqhtl v nasal but still acceptable as within normal variation; only noticeable to a trained ear. Hypernasal resonance quality is evi dent and should be apparent to a layman . Very nasal; adversely reduces the intelligibility for the total speech pattern. Unintelligible. Circle the number which best corresponds to your judgment of nasality 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 160

PAGE 170

APPENDIX F SEQUENCE OF THE THREE TEST TAPES PRESENTED FOR PERCEPTUAL JUDGMENT OF THE 95 TESTS (FIVE SPEECH TESTS AT 1 9 FISTULA CONDITIONS) J ud g e Test ord e r 1 1 2 3 2 1 3 2 3 2 1 3 4 2 3 1 5 3 1 2 6 3 2 1 -161

PAGE 171

Instructions APPENDIX G INSTRUCTIONS FOR THE PERCEPTUAL JUDGMENT PERFORMANCE First, I would like you to familiarize yourself with the two tests, the articulation test and the hypernasality test. In the articulation test you are to evaluate the capitalized sounds only. They are indicated on the test form. In the hypernasality test you are to listen for a shift of tone between the two presentations of each of the 10 words. Do not mark any misarticulation that you may hear. The next two tests use one rating scale each for makin g the judgments of articulation and hypernasality. Look at the rating scales closely and refrain from judging any nasality in the articulation rating test or from judging any misarticulations in the hypernasality test. There will be 95 tests on these three tapes which correspond to approximately 90 minutes of listening time. The tests will be presented in random order. Each test to be judged is numbered on the tape and this number corres ponds to the number at the top left corner of each sheet. There is a short pause (approximately two seconds) between -162

PAGE 172

-163each test word, word pairs, or test sentences for you to make a decision and to mark it. There is a pause button available to you on the tape recorder, and you may use it if you feel it is necessary, but try to make your judgments without using it. If you need to stop for some reason, you are allowed to do so.

PAGE 173

BIBLIOGRAPHY ibyholm, F.E., Brochgrevink, H., H.C., & Eskeland, G. Palatal fistula following cleft palate surgery. Scandinavian Journal of Plastic and Reconstructive Surgery, 1979, g, 295-300. Adisman, I.K. Cleft palate prosthetics. In W.C. Grabb, s.w. Rosenstein, & K.R. Bzoch (Eds.), Cleft lip and palate: Surgical, dental, and speech aspects. Boston: Little, Brown and Company, 1971. Andrews, J.R., & Rutherford, D. Contribution of nasally emitted sound to the perception of hypernasality of vowels. Cleft Palate Journal, 1972, 2, 147-156. Aram, A., & Subtelny, J.D. Velopharyngeal function and cleft palate prostheses. Journal of Prosthetic Den tistry, 1959, 2, 149-158. Bateman, H.E. A clinical approach to speech anatomy and physiology. Sprinfield, Ill.: C.C. Thomas, 1977. B e nson, D. Roentgenographic cephalometric study of palato pharyngeal closure of normal adults during vowel phonation. Cleft Palate Journal, 1972, 2, 43-50. Bernthal, J.E., & Beukelman, D.R. The effect of chang e s in velopharynge al orifice area on voca l intensity. Cleft Palate Journal, 1977, 1..., 63-77. Bjork, L. V e loph a ryngeal f unction in connected speech: Studies using tomography and cineradiography syn chronized with speech spectrography. Acta Radiologica: Supplement 202, 1961. Bless, D.M., Ewanowski, S.J., & Dibbell, D.G. A technique for temporary obturation of fistulae--A clinical note. Cleft P a late Journal, 1980, l2_, 297-300. Bloch, P.J. Clinical evaluation for the cleft palate t ea m setting. In K.R. Bzoch (Ed.), Communicative disorders of cleft lip and palate (2nd ed .). Boston: Little, Brown and Company, 1979. -164

PAGE 174

-165Bradford, L.J., Brooks, A.R., & Shelton, R.L. Clinical judgment of hypernasality in cleft palate children. Cleft Palate Journal, 1964, l, 329-335. Bradley, D.P. Congenital and acquired paletopharyngeal insufficiency. In K.R. Bzoch (Ed.), Communicative disorders related to cleft lip and palate (2nd ed.). Boston: Little, Brown and Company, 1979. Brescia, N.J. Anatomy of the lip and palate. In W.C. Grabb, S.W. Rosenst e in, & K.R. Bzoch (Eds.), Cleft lip and palate: Surgical, dental, and speech aspects. Boston: Little, Brown and Company, 1971. Brown, W.S., & McGlone, R.E. Constancy of intraoral air pressure. Folia Phoniatrica, 1969, ~' 332-339. Bzoch, K.R. Articulation proficiency and error patterns of preschool cleft palate and normal children. Cleft Palate Journal, 1965, 1, 340-349. Bzoch, K.R. Variations in velopharyngeal valving: A factor of vowel changes. Cleft Palate Journal, 1968, 2, 211-218. Bzoch, K.R. Measurement and assessment of categorical aspects of cleft palate speech. In K.R. Bzoch (Ed.), Communicative disorders related to cleft lip and palate (2nd ed.). Boston: Little, Brown and Company, 1979. Bzoch, K.R., Kemker, F.J., & Di xon -Wood, V.L. The preven tion of communicative disorders in cleft palate infants. In N.J. Lass (Ed.), Speech and language: Advances in research and practice (Volume 10). New York: Academic Press, Inc., 1984. Bzoch, K.R., & Williams, W.N. Introduction, rationale, principles, and related basic embryology and anatomy. In K.R. Bzoch (Ed.), Communicative disorders related to cleft lip and palate (2nd ed.). Boston: Little, Brown and Company, 1979. Coleman, Jr., R.O. The effect of changes in width of velopharyngeal apertur e on acoustic and perceptual properties of nasalized vowels (Doctoral dissertation, Northwestern University, 1963). Dissertation A bstracts, 1964, 25, 688-689. (University Microfilms No. 64-5826) Cosman, B., & Falk, A.S. D e layed hard p a lat e repair and speech deficiencies: A cautionary report. Cleft Palate Journal, 1980, 1:...2., 27-33.

PAGE 175

-166Counihan, D.T. Oral and nasal airflow a nd air pressure measures. In K.R. Bzoch (Ed.), Communicative disorders related to cl eft lip and palate (2nd ed.). Boston: Little, Brown and Compan y , 1979. Counihan, D.T., & Cullinan, W.L. Reliability and dispersion of nasality ratings. Cleft Pal a t e Journal, 1970, 2, 261-270. Crelin, E.S. Development of the upp e r respiratory system. Clinical S vmpos ia, Ciba, 1976, ~' 3-30. Curtis, J.F. The acoustics of nasalized speech. Cleft Palate Journal, 1970, 2, 380-396. Dalston, R.M. activity. Photodetector assessmen t of velopharyngeal Cl eft Palate Journal, 1982, .1:.2_, 1-8. Daniloff, R., Schuckers, G., & F e th, L. The physiology of speech and hearing. Engl ewo od Cliffs, N.J.: Prentice Hall, 1980. D avis , H. Acoustics and psycho-acoustics. In H. Davis & S.R. Silverman (Eds.), He aring and deafness (3rd ed.). New York: Holt, Rinehart and Winston, 1970. Dickson, D.R. Normal and cleft palate anatomy. Cleft Palate Journal, 1972, ~' 280-293. Drane, J.B. Prosthetic considerations in oral ab lativ e surgery . AS H A Reports No . 8 , 1 973 , 39-41. DuBru hl , E.L. Sicher's oral anatomy (7th ed.). St. Louis: C.V. Mosby Comp a n y , 1980. Edgerton, M.T., Sad ove , M., Compton, M ., Bull, G., Blomain, E., McDonald, W., & Bralley, R. Nasa l vibratio n a nalysi s : A noninvasive objective technique to evalu ate the speech of pa ti ents w it h palatopharyngeal disord e rs. Plastic a nd Reconstructive Surgery, 1981, ~' 153-157. Erickson, B. A comparison st u dy between the Iowa Pressure Articulation T est and the Bzoch Error Pattern Diag nostic Screening Articulation Test. Paper presented at the Florida Cleft Palate Association Convention, January 1984. Fletcher, S.G. Th eo ry and instrumentation for quantitative measur e ment of nasality. Cleft Palate Journal, 1970, 2, 601-609.

PAGE 176

-167Fletcher, S.G., & Bishop, M.E. Measurement of nasality with Tonar. Cleft Palate Journal, 1970, 2, 610-621. Fletcher, S.G., Sooudi, I., & Frost, S.D. Quantitative and graphic analysis of prosthetic treatment for "nasalance" in speech. Journal of Prosthetic Dentistry, 1974, 11_, 284-291. Fri tzell, B. 11 Tala i nasan" --ett 2-~r sma terial. Lakartid ningen, 1973, J...Q_, 2536-2538. Garber, S.R., Speidel, T.M., Siegel, G.M., Miller, E., & Glass, L. The effects of presentation of noise and dental appliances on speech. Journal of Speech and Hearinq Research, 1980, ~, 838-852. Gordon, N.C., & Brown, S.L. Closure of oronasoantral defects: Report of case. Journal of Oral Surgery, 1980, ~. 600-605. Gorlin, R., & Gorlin, S.G. Hydraulic formula for calcula tion of the area of the stenolic mitral valve, other cardiac valves, and central circulatory shunts. American Heart Journal, 1951, ..!!_, 1-29. Grabb, w.c., Rosenstein, S.W., & Bzoch, K.R. (Eds.). Cleft lip and palat e : Surgical, dental and speech aspects. Boston: Little, Brown and Company, 1971. Graber, T.M., Bzoch, K.R., & Aoba, T. A functional study of the palatal and pharyngeal structures. The Angle Orthodontist, 1959, 22_, 30-40. Hadding, K., & Petersson, L. Experimentell fonetik. Lund: Broderna Ekstrands Tryckeri, AB, 1970. Hardy, J.C. Airflow and air pressure studies. ASHA Reports, No. 1, 1965, 141-152. Harris, R. Summary of a conference on cleft lip and cleft palate. Journal of American Dental Association, 1980, 100, 396-398. Henderson, H.P. The "tadpole flap": An advancement island flap for the closure of anterior palatal fistulae. British Journal of Plastic Surgery, 1982, }2, 163166. Henningsson, G. 2:an. Rapport ran norsk logopedlags sornmerkurs 1982. Svensk Tidskrift for Foniatri och Logopedi, 1983, No. 1, 2-6.

PAGE 177

-168Holmes, A.E., Frank, T., & Stoker, R.G. Telephone listening ability in a noisy background. Ear and Hearing, 1983, i, 88-90. Horii, Y. An accelerometric approach to nasality measure ment: A preliminary report. Cleft Palate Journal, 1980, 1:..2., 254-261. House, A.S., & Stevens, K.N. Analog studies of the nasalita tion of vowels. Journal of Speech and Hearing Dis orders, 1956, ~' 218-232. Hutters, B. Naeseflow til vurdering at velofaryngeal insufficiens--nogle forelbige iagttagelser. Scan dinavian Journal of Logopedics and Phoniatrics---;-[982, ]__, 135-146. Isshiki, N., Honjow, J., & Morimoto, M. pharyngeal incompetence upon speech. Journal, 1968, 1, 297-310. Effects of vela Cleft Palate James, R.B. Surgical closure of large oroantral fistulas using a palatal island flap. Journal of Oral Surgery, 1980, 2_, 591-594. Kirk, R.E. Experimental design: Procedures for the behavioral sciences. Belmont, Ca.: Brooks/Cole Publishing Company, 1968. Krause, C.J., Tharp, R.F., & Morris, H.L. A comparative study of results of the von Langenbeck and the V-Y pushback palatoplasties. Cleft Palate Journal, 1976, .!l., 11-19. Kuehn, D.P., & Dolan, K.D. A tomographic technique of assessing lateral pharyngeal wall displacement. Cleft Palate Journal, 1975, g, 200-209. Liebman, E.S. Some speech effects produced by experimental variation of oral-nasal cavity coupling (Doctoral dissertation, Northwestern University, 1964). Disser tation Abstracts, 1964, 25, 3743. (University Microfilms No. 64-12,309 Lindsay, W.K. Von Langenbeck palatorrhaphy. In W.C. Grabb, S.W. Rosenstein, & K.R. Bzoch (Eds.), Cleft lip and palate: Surgical, dental and speech aspects. Boston: Little, Brown and Company, 1971. Lubker, J.F. Aerodynamic and ultrasonic assessment tech niques in speech-dentofacial research. ASHA Reports No. 5, 1970, 207-223.

PAGE 178

-169Lubker, J.F., & Moll, K.L. Simultaneous oral-nasal airflow measurements and cinefluorographic observations during speech production. Cleft Palate Journal, 1965, 1, 257-272. Marks, R.G. Designing a research project: The basics of biomedical research methodology. Belmont, Ca.: Life time Learning Publications, 1982. (a) Marks, R.G. Anal y zing research data: The basics of bio medical research methodology. Belmont, Ca.: Lifetime Learning Publications, 1982. (b) McKerns, D. Mc. Velopharyngeal mechanism: The factor of sex (Doctoral dissertation, University of Florida, 1968). Dissertation Abstrac t s International, 1969, 20B, 427B-428B. (University Microfilms No. 69-10,922) McWilliams, B.J., Glaser, E.R., Philips, B.J., Lawrence, C., Lavorato, A.S., Beery, Q., & Skolnick, M.L. A comparative study of four methods of evaluating velo pharyngeal adequacy. Plastic and Reconstructive Surgery, 1981, ~' 1-10. Microtronics Corporation. n.d. PERCI. Carrboro, N.C.: Author, Millard, D.R., Jr. Cleft craft: The evolution of its surgery: III. Alveolar and palatal deformities. Boston: Little, Brown and Company, 1980. Moll, K.L. Cinefluorographic techniques in speech research. Journal of Speech and Hearing Research, 1960, l, 227-241. Moll, K.L. Velopharyngeal closure on vowels. Journal of Speech and Hearing Rese a rch, 1962, 2, 30-37. Moll, K.L. Cineradiography in research and clinical studies of the velopharyngeal mechanism. Cl e ft Palate Journal, 1964, l, 391-397. Moll, K.L. A cin e fluorographic study of velopharyngeal function in normals during various activities. Cleft Palate Journal, 1965, 1, 112-122. Morley, M.E. Cleft palate a nd speech (5th ed.). London: E & S Livingstone Ltd., 1962. Morris, H.L. The oral manometer as a diagnostic tool in clinical speech pathology. Journal of Sp e ech and Hearing Disorders, 1966, ll_, 362-369.

PAGE 179

-170Morris, H.L. Etiological bases for speech problems. In D. Spriestersbach & D. Sherman (Eds.), Cleft palate and communication. New York: Academic Press, 1968. Musgrave, R.H., & Bremner, C.J. Complications of cleft palate surgery. Plastic and Reconstructive Surgery, 1960, ~, 180-189. Nylen, B.D. Cleft palate and speech. Acta Radiologica: Supplement 203, 1961. O'Neal, R.M. Oronasal fistulas. In W.C. Grabb, S.W. Rosenstein, & K.R. Bzoch (Eds.), Cleft lip and palate: Surgical, dental, and speech aspects. Boston: Little, Brown and Company, 1971. Palmer, C.R., Hamlen, M., Ross, R.B., & Lindsay, W.K. Cleft palate repair: Comparison of the results of two surgical techniques. The Canadian Journal of Surgery, 1969, g, 32-39. Philips, B.J.W., & Bzoch, K.R. Reliability of judgments of articulation of cleft palate speakers. Cleft Palate Journal, 1969, , 24-34. Proctor, B. Bone graft closure of large or persistent oroma x illary fistula. Laryngoscope, 1969, J..J..., 822826. Ramig, L. A . Effects of e x aminer expectancy on speech ra ti n gs of individuals with cleft lip and/or palate. Cl e ft Pal a te Journal , 1982, ~, 270-274. Reich, A.R., & Redenbaugh, M.A. Relation between nasal/ voice accelerometric values and interval estimates of hypernasality. Cleft Palat e Journal, 1985, ~, 237-245. Reid, D.A.C. Fistul a e in the h a rd palate f ollowing cleft palate surgery. British Journ a l o f Pl a stic Surqery, 1962, 1:.2, 377-384. Reisberg, D.J., Gold, H.O., & Dorf, D.S. A technique for obturating pal a tal fistulas. Cleft Palate Journal, 1985, 1l, 286-2 8 9. Ross, R.B., and Johnston, M.C. Cleft lip a nd palate. Baltimore: The Williams and Wilkins Company, 1972.

PAGE 180

-171Schwartz, Martin F. Acoustic measures of nasalization and nasality. In K.R. Bzoch (Ed.), Communicati ve disorders r ela ted to cleft lip and palate {2nd ed.). Boston: Little, Brown and Company, 1 979. Shaw, N ., & Gilbert, H.R. Research note: A respirometric t e chnique for evaluating velopharyngeal closure in children. Journal of Speech and Hearing Research, 1982, _e, 476-480. Sh e lton, R.L., & Blank, J.L. Oronasal f istulas, intr ao ral a ir pressure, a nd nasal air flow during speech. Cl eft Palate Journal, 1984, Q, 91-99. Shelton, R.L., Brooks, A.R., & Youngstrom, K.A. Articula tion and patterns of palatopharyngeal closure. Journal of Speech and Hearing Disorders, 1964, ~, 390-408. Sh er m an , D. Correl a tion between defective articulation and nasal it y in cleft palate speech. Cleft Palate Journal, 1970, 2, 626-629. Shprintzen, R.J., Croft, C., Lewin, M.L., & Rakoff, S. The relationship of perceived hypernasality to velo pharyngea l gap size during spe ech. Cleft Palate Journal , 1977, 1., 353. (Abstr ac t) Shprintzen, R.J., McCall, G.N., Skolnick, M.L., & L enc ione, R . M . Selective movement of the lateral aspects of the pha r yngea l walls during ve lopharyn geal closure f or speech , blowing, a nd whistling in no rmals. Cleft Pa l ate Journal, 1 97 5, 11_, 51-58. Skolnick, M.L. V id eofluoroscop ic examination of the ve lo pharyngeal portal during phonation in lateral and base projections--A new technique for studying the mechanics of closure. Cl eft Palate Journal, 1970, 2, 803-816. Skolnick, M.L., Glaser, E.R., & McWilliams, B.J. The use and limitati ons of the barium pharyngogram in the detection of velopharyngeal insufficiency. Radiology, 1980, 135, 301-304. Spriestersbach, D.C., & Powers , G . R . Articu lation skills, velopharyngeal c l os ur e , and oral breath pressure of children with cleft palates. Journal of Speech and Hearing Research, 1959, ~, 318-325. Statistical Analysis System Institut e , Inc. SAS user's -----guide: Basics, 1 982 ed ition. C a ry, N.C.: Author, 1982. (a)

PAGE 181

-172Statistical Analysis System Institute, Inc. SAS user's guide: Statistics, 1982 edition. Cary, N.C.: Author, 1982. (b) Stevens, K.N., Kalikow, D.N., & Willemain, T.R. A miniature accelerometer for detecting glottal waveforms and nasalization. Journal of Speech and Hearing Research, 1975, 1:..._, 594-599. Subtelny, J.D., Koepp-Baker, H., & Subtelny, J.D. Palatal function and cleft palate speech. Journal of Speech and Hearing Disorders, 1961, ~' 213-224. Van Demark, D., Bzoch, K., Daly, D., Fletcher, S., Mcwilliams, B.J., Pannbacker, M., & Weinberg, B. Methods of assess ing speech in relation to velopharyngeal function. Cleft Palate Journal, 1985, ~' 281-285. Warren, D.W. Velopharyngeal orifice size and upper pharyn geal pressure-flow patterns in normal speech. Plastic and Reconstructive Surgery, 1964, l_l, 148-162. (a) Warren, D.W. Velopharyngeal orifice size and upper pharyn geal pressure-flow patterns in cleft palate speech: A preliminary study. Plastic and Reconstructive Surgery, 1964, 34, 15-26. (b) Warren, D.W. 26-33. Instrumentation. ASHA Reports No. 9, 1973, Warren, D.W. PERCI: A method for rating palatal efficienc y. Cleft Palate Journal, 1979, .l._, 278-285. Warren, D.W., Dalston, R.M., Trier, W.C., & Holder, M.B. A pressure-flow technique for quantifying temporal patterns of palataopharyngeal closure. Cleft Palate Journ a l, 1985, ~, 11-19. Warren, D.W., & De ve r eux , J.L. An ana log study of cleft palate speech. Cleft Palate J o urnal, 1966, }, 103114. Warren, D.W., & DuBois, A.B. A p r essure -flow technique for measuring ve lophar yngea l orifice area during con tinuous speech. Cleft Pala te Journal, 1964, l, 52-71. Warren, D.W., Hall, D.J., & Davis, J. Oral port constric tion and pressure-air-flow relationships during sibilant productions. Folia Phoniatrica, 1981, _ll, 380-394.

PAGE 182

-173Warren, D.W., & Hoffman, F.A. A cineradiographic study of velopharyngeal closure. Plastic and Reconstructive Surgery, 1961, 2._, 656-669. Watterson, T., & Emanuel, F. Effects of oral-nasal coupling on wh ispered vowel spectra. Cleft Palate Journal, 1981, ~' 24-38. (a) Watterson, T., & Emanuel F. Observed effects of velo pharyngeal orifice size on vowel identification and vowel nasality. Cleft Palate Journal, 1981, .l_, 271-278. (b) Wharton, P.W. An autobiographical study: aspects of cleft palate habilitation. University of Florida, 1978. Psychosocial Master's thesis, Williams, W.N. Radiographic assessment of the structure and function of the speech mechanism. Unpublished manual, Department of Oral Biology, College of Dentistr y , University of Florida, 1986. Witzell, M.A., Clarke, J.A., Lindsay, W.K., & Thomson, H.G. Comparison of results of pushback or von Langen beck repair of isolated cleft of the hard and soft palate. Plastic Reconstructi ve Surgery, 1979, __i, 347-352. Zemlin, W.R . Spe ech and hearing science: physiology. Englewood Cliffs, N.J.: 1968. Ana tomy and Prentice-Hall,

PAGE 183

BIOGRAPHICAL SKETCH The author was born in Stockholm, Sweden, on June 4, 1941. She had the opportunity to study at universities in Austria and West-Germany before graduating in 1972 with a master's degree in liberal arts from the University of Stockholm, Sweden. She worked briefly in city planning in Stockholm. The author moved to Florida, USA, in 1973 and entered the University of South Florida, Tampa, where she graduated in 1975 with a Master of Sci ence degree in speech pathology. In the years 1975 to 1977 she worked as a speech language pathologist at th e rehabilitation clinic and also in private consulting in Sarasota, Florida. In 1977 she returned for doctoral studies in speech pathology to the University of Florida, Gainesville. worked in private practice in Sweden. -174Intermittently, she

PAGE 184

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. William N. Williams, Chairman Associate Professor of Speech I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Thomas B. Abbott Professor of Speech I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Professor of Communicative Disorders

PAGE 185

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. G. Paul Moore Professor Emeritus of Sp ee ch I certify that I have read this study and that in my opinion it c o nforms to acceptable standards of scholarly pres e ntation and is fully adequate, in scope and quality, as a dissertation for the d e gree of Doctor of Philosophy. 1' i, / , , , " .. / " : _ , / Li' ... ~ --_. Nikzad S. Jav id Professor of Removable Prosthodontics I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Pa ul Vv . Wharton Dir ecto r of Development, Hope Haven Clinic, Jacksonville, Florida

PAGE 186

This dissertation was submitted to the Graduate Faculty of the Department of Speech in the College of Liberal Arts and Sciences and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August 1986 Dean, Graduate School

PAGE 187

111 1 111 1 111 HI1!ij~'rn1~~ ~m11 1 m111111 1 111 1 3 1262 08285 310 1


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ESGNATCTJ_8X5BQJ INGEST_TIME 2011-07-01T14:25:55Z PACKAGE UF00102777_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES