Assessing Cul-de-Sac Resonance in the Deaf Population and Selecting Appropriate Intervention Techniques
I. Introduction / purpose
The hearing impaired population, specifically the profoundly deaf, have many distinctive, perceptually similar abnormalities in the quality of their voices. Many studies have examined the speech of the deaf population, specifically concerning abnormally high fundamental frequency, lack of respiratory support, breathy voice quality and hypernasal or cul-de-sac resonance. It is the resonant characteristics of the speech of the deaf population that will be the primary focus of this discussion. How do we determine, both perceptually and instrumentally, the various identifying characteristics of each type of resonance? After identifying the distinct type of deviant resonance, how will we then go about remediating these deviances? This paper will discuss a general overview of resonance disorders, the assessment techniques used to identify each type of resonance disorder and a review of recent literature pertinent to these assessment techniques. Following this general discussion, I will attempt to illustrate research-directed assessment and intervention practices for cul-de-sac resonance, possible clinical applications and directions for future research.
II. General Overview
According to Kummer (1996), "resonance is the quality of the voice that results from sound vibrations in the pharynx, oral cavity, and nasal cavity." The relative balance of nasal versus oral sounds produced within these cavities determines the type of voice which is viewed as "normal" or deviant, depending on the type and amount of nasality present. Normal resonance depends on both structural and functional variables; structurally, the resonating cavities must be in tact and free from fistulas or obstructions, the velum must be of appropriate length and functioning adequately for complete velopharyngeal closure and the other articulators (tongue, teeth, lips and jaw) must be in tact and functioning in such a manner as to support the various resonances. Any abnormal structures or functioning of the structures has the potential to cause disturbances and abnormalities in the resonance of an individual’s speech.
The three main types of resonance disorders are hypernasality, hyponasality and cul-de-sac or pharyngeal resonance. Hypernasality is a resonance disorder, primarily due to some type of velopharyngeal inadequacy, which allows an excess amount of resonance in the nasal cavity to occur. This excess nasal resonance may be due to either physical abnormalities of the velum or an inadequacy in its functioning and may or may not be accompanied by audible nasal emissions of air. Hyponasality, or denasality, is the opposite of hypernasality and occurs when there is some type of blockage in the nasal cavity or, in some cases, inadequate opening of the velopharyngeal port. This type of nasality is typical of an individual with a head cold. Cul-de-sac, or pharyngeal resonance, occurs when the acoustic energy gets "trapped" and resonates in the pharyngeal area rather than in the oral or nasal cavities. Perceptually, this is a muffled type of sound and can occur in clients with "very large tonsils and adenoid pad" (Kummer et al., 1993) or in clients with both velopharyngeal incompetence and a nasal blockage. Most commonly, however, cul-de-sac resonance is associated with the speech of the deaf population.
III. Assessment of Resonance Disorders
In order to appropriately diagnose and evaluate a resonance disorder, a perceptual assessment must first be completed. During such an assessment, the clinician should first evaluate whether the resonance is normal or abnormal. According to Kummer & Lee (1996), "resonance can be said to be abnormal if the quality or intelligibility of speech is affected by inappropriate transmission of acoustic energy in the vocal tract." Additional subjective testing should include observation of velopharyngeal sufficiency, intensity of consonants and any nasal emissions during an articulation test, which could indicate velopharyngeal valving problems. In addition, the clinician should observe the client during a connected speech task due to the increased demands that this task places on the velopharyngeal valving system. The trained clinician should be able to distinguish whether the resonance characteristics are indicative of hypernasality, hyponasality, cul-de-sac resonance or a mixture of these. For the inexperienced clinician, however, this may prove to be more difficult and further probing may be necessary. In addition to probing for a resonance disorder by using connected speech tasks with or without nasal consonants, manually closing the nares during speech may give help to identify the problem; if resonance is abnormal, and closing the nares does not change the quality of the resonance, the resonance is either hyponasality or cul-de-sac resonance. To further identify the resonance during the production of nasal consonants, if the productions sound more like the oral cognates of the nasal sounds, hyponasality is indicated.
Although the perceptual assessment may seem to be relatively straightforward and conclusive, a variety of research indicates that there are many additional variables that may affect perception of abnormal resonance. Colton & Cooker (1968), Fletcher & Higgins (1980) and others have shown that, for normal hearing individuals, a slower rate of speech has a direct effect on others’ perceptions of increased nasality. Since deaf individuals tend, on average, to have slower rates of speech, this may be perceived as a higher degree of nasality. Other, more specific, features that may affect the perception of resonance are, "reduction in the intensity of the 1st formant, one or more antiresonances within the spectrum, reinforced harmonics at resonances at which energy is not expected and a change or shift in the center frequencies of the formants" (Fletcher & Daly, 1976). In other words, it is sometimes very difficult to perceptually differentiate resonance abnormalities due to the inherent influence of other variables such as pitch, intensity and rate of speech. In addition, there are some normal, dialectical, variations of nasality that are considered, "acceptable for normal speech" (LaPine, Stewart & Tatchell, 1991). Because of these differences, instrumental assessments are warranted as part of a complete evaluation.
An instrumental assessment should be conducted following the perceptual assessment to further identify the resonance abnormality. One of the most common instruments currently used to assess resonance abnormalities is the nasometer (Kay Elemetrics, Pine Brook, NJ). The nasometer is, "a computer-based instrument that is designed to be used with either an IBM-compatible or Apple personal computer and consists of a headset that has directional microphones for the nose and mouth" (Kummer & Lee, 1996). These microphones detect the acoustic energy that is emitted from the oral and nasal cavities and then compute a ratio of nasal acoustic energy to total acoustic energy to arrive at an average "nasalance" score. Although this score can be useful in determining the resonance and nasality, the clinician must be knowlegeable in interpreting the score; "a combination of hyponasality and nasal emission can affect the nasalance score to a significant degree" (Dalston, Warren & Dalston, 1991). Thus, excess nasal emissions may incorrectly identify a client as being hypernasal when, in fact, they are hyponasal or have cul-de-sac resonance with some nasal emissions.
The Key Elemetrics nasometer is actually the third generation of an instrument which uses the concept of TONAR (The Oral Nasal Acoustic Ratio). The original concept of nasalance (and TONAR) was first described by Fletcher (1970) who wanted to devise an instrument that would, "quantify perceived vocal nasality" (LaPine, Stewart & Tatchell, 1991). Fletcher extended the work of other researchers (such as Counihan and Pearce-1965) who used devices such as the U-tube, manometers, oral manometers and flowmeters to assess oral and nasal air pressure. In 1972, Fletcher developed the "Zoo Passage," a phonetically balance passage with no nasal consonants and no meaningful context, to be used with TONAR II. Other instrumentation that has been used to assess abnormal resonance (or nasalance) include: the ZIPPO (Zemlin Index of Palatopharyngeal Opening; Zemlin & Fant, 1972), spectral analysis, oral and nasal sound pressure, accelerometry (Stevens, Kalikow, and Willemain, 1975), and HONC (The Horri Oral Nasal Coupling index; Horri, 1980).
In addition to perceptual and instrumental assessments of resonance, an intraoral examination should always be included in an evaluation. Through this type of examination, a clinician can identify the presence of palatal fistulas, the relative length of the velum and velar mobility. However, because velopharyngeal closure occurs above the level of the oral cavity, complete functioning of the velum cannot (and should not) be assessed from this view.
IV. Review of Literature
The assessment of cul-de-sac resonance is comparatively more difficult to assess than either hypernasality or hyponasality. Cul-de-sac (or pharyngeal) resonance is not necessarily identifiable by nasality or denasality alone but rather consists of unique acoustic characteristics. The use of the nasometer therefore, will primarily indicate the absence of other resonance qualities (hypernasality or hyponasality) rather than indicating the presence of cul-de-sac resonance. The following research articles address some aspects of cul-de-sac (or pharyngeal) resonance and offer some evidence for its identification and remediation.
- Normal and reduced phonological space: the production of English vowels by deaf adolescents" (Monsen, 1976)
The purpose of this study was to spectrographically measure the formant frequencies of three vowels (/i/, /a/, / /) in the speech of 36 deaf and 4 normally-hearing adolescents in order to determine the phonological space defined by the maximum and minimum values of the first and second formants of each vowel. It is hypothesized that this phonological space is reduced for many deaf subjects. The author suggests that the reasons for this reduction are due to abnormal auditory and visual feedback available to the deaf child. The subjects were asked to read simple sentences such as "I like ice cream" or "He plays baseball." These sentences were recorded and played back for 35 naïve and 15 experienced listeners who were asked to transcribe in normal English orthography what each subject said. A percentage score reflecting general speech intelligibility of each child was then obtained by averaging the scores of the 50 listeners. Fifteen words containing the vowels /i/, /a/, & / / were then selected from the recorded sentences and measured on the Voice Identification spectrograph (series 700) for specific F1 and F2 formant analysis. Five examples of each vowel were then plotted on an F1-F2 graph. Results indicated that, for the deaf subjects, the second formant tended to "float around" 1800 Hz, regardless of which vowel was being articulated. In addition, the phonological space used by the deaf subjects was greatly reduced in comparison with the space used by the normal hearing subjects for all vowels. Two possible reasons given by the author for the reduction of this phonological space between the vowels are: first, that since the first formant is perceptually more salient to a deaf child than the second formant (which is typically in the higher frequencies between 1000-2500Hz) the deaf child does not naturally produce many variations in F2 and secondly, because there is relatively low visibility of the articulatory movements of the tongue, the formants affected by these movements have less variability. Thus, because the tongue is, "primarily responsible for producing maximally high and low frequency values of the second formant," these formants tend to be "neutralized" at some central frequency between 1000 & 2500 Hz.
- "Cephalometric and cineradiographic study of deviant resonance in hearing-impaired speakers" (Subtelny, Walter, Whitehead & Subtelny, 1989)
The purpose of this study was to investigate the physiological correlates of deviant pharyngeal resonance as manifested in hearing-impaired adult speakers with near normal respiratory and phonatory function. The techniques of cephalometric roentgenography and cineradiography were used because, according to the author, they are "particularly well suited to study oral and pharyngeal relationships during static and dynamic speech production." The subjects used in this study were four hearing-impaired (HI), adult speakers with pharyngeal voice quality who had been diagnosed as having "prelingual, severe, bilateral sensorineural hearing losses." All subjects were fitted with hearing aids before 4 years of age and, at the time of the study, wore their aids "all or most of the time." Ten adult females with normal hearing were used to establish the cineradiographic normal reference. Speech samples that were elicited and recorded included the sustained vowels /i/, /u/ and /a/ and the sentence "Peter Piper picked a peck," which provided multiple articulations of the stop /p/ in all positions. The author notes that the analysis was designed to measure the oral cavity dimensions and horizontal pharyngeal dimensions at four anatomical sites because, "retraction of the tongue has been hypothesized as the physiological basis for pharyngeal resonance in hearing-impaired speakers (Boone, 1977)."
Results of this study were somewhat surprising; the normal hearing speakers displayed the expected differences in tongue positions for each of the vowels produced, however, the HI speakers, displayed some consistent variations in their tongue positions. Because the speech and voice characteristics of HI speakers are extremely variable, the consistencies in their tongue movements were not expected. In general, the HI speakers with pharyngeal resonance showed, "less than normal shifting of the tongue within the horizontal plane of the oral cavity and greater than normal variation in lip, mandible, velum, hyoid and epiglottis positions." Specific "aberrant" movements identified were tongue placement and retraction, elevated hyoid (with a large vertical distance between the hyoid and the laryngeal sinus) and retracted tongue root (associated with deflection of epiglottis toward the pharyngeal wall). The limited tongue movements reflect a generally "neutralized" tongue position in those HI speakers with pharyngeal resonance. Because of the relationship between tongue position and the frequency location of the second formant, previous inferences concerning tongue movement have been based upon studies of formant structure. This finding of consistent "aberrant" tongue movement supports the concept of "neutralization" put forth by Monsen in 1976. In his studies, he found that, in the speech of the hearing impaired, the second formant of all vowels, tended to center around 1800 Hz, regardless of the vowel that was intended to be spoken. Although the tongue positions in this study appeared to be in this "neutral" position, the vowels were not so distorted that they could not be identified. The author hypothesizes that this may be due to the age (early 20’s) of the subjects and years of aural/oral training and learning compensatory strategies.
The most significant finding of this study, however, involves the deviation of the tongue root in the lower pharynx of the HI speakers. The consistency of this finding among the speakers supports Boone’s (1977) hypothesis regarding the cause of pharyngeal resonance as being a retracted tongue resulting in the lowering of the second formant frequency. Increasing the pharyngeal cavity length influences the total vocal tract length and, thus, "lowers all formant frequencies, especially those that can be regarded as a back cavity resonance" (Lindblom and Sundberg, 1971). Although the F2 formant is affected directly by the limited tongue height in the HI speaker, all other formants are also affected by the change in cavity length. Thus, it would be very difficult to identify cul-de-sac resonance solely on the basis of lowered F2 without other acoustic characteristics present.
- "Spectral study of deviant resonance in the speech of women who are deaf"
(Subtelny, Whitehead & Samar, 1992)
The purpose of this study was to extend the hypothesis established at the conclusion of the 1989 study by Subtelny, Walter, Whitehead & Subtelny; the authors pursued the, "hypothesized acoustic correlate of the deviant resonating relationships of the recorded vowels from the previous radiographic study" through spectral analysis of the recorded vowels. Through this spectral analysis, the authors hoped to determine three things: First, is the 2nd formant frequency specifically lowered in deaf speakers when pharyngeal resonance is perceptually apparent during production of vowels /i/, /u/ and /a/; second, does the overall formant structure of /i/, /u/ and /a/ produced by the HI speakers differ from that of the same vowels produced by the normally hearing women; and lastly, are the commonly described relationships between 2nd formant frequency and horizontal tongue position and 1st formant frequency and vertical tongue position demonstrated in women who are hearing impaired and have pharyngeal resonance.
The audio recordings from the previous study (which had been recorded onto a
Nagra IV-S tape recorder) were low-pass filtered at 5 kHz (Rockland system 816; nominal rejection rate of 40 dB/octave) and then digitized with 12-bit precision at 10kHz with a Masscomp laboratory computer. Using the Interactive Laboratory System (1986) (ILS) formant frequency extraction routine, which analyzes spectral characteristics using the LPC method, the 1st, 2nd and 3rd formants were calculated and averaged over the 900-msec sample.
Results from this study indicated that there were significant differences (p<.01) between the normal hearing and the HI groups with respect to F1, F2 and F3 on each of the three test vowels. However, the general relationships between the frequency of F2 and horizontal tongue position and the frequency of F1 and vertical tongue position were maintained in the three vowels produced by the HI women. Thus, although the "phonological space" among the three vowels was drastically decreased in the productions of the HI women, the general relationships among the formants of all three vowels was preserved. This indicates that, although the restricted tongue movement of the HI women caused a "clustering" of the 2nd formant frequency in the 1500-2100 Hz frequency range, the preservation of the general formant relationships of each vowel by the HI speakers allowed for acoustically distinguishable (although somewhat neutral) vowel productions.
This finding would support the need for thorough perceptual analysis prior to any instrumental analysis so that the successful production of the vowels (using compensatory strategies) is not overshadowed by apparent spectral differences.
- "A preliminary investigation concerning the use of nasometry in identifying patients with
hyponasality and/or airway impairment" (Dalston, Warren & Dalston, 1991)
The purpose of this study was to determine the extent to which acoustic measurements of speech using a Kay Elemetrics nasometer corresponded with clinical judgments of hyponasality and aerodynamic measurements of nasal cross-sectional area. The subjects included 76 individuals who were clinically determined to have various degrees of nasal airway impairment. Results of this study indicated that the nasalance scores obtained by the nasometer were "somewhat accurate" in identifying patients without denasality in their speech (34 of 43), but that a large number of hyponasal subjects (17 of 33) obtained high nasalance scores. Interestingly, when patients with discernible nasal emissions were eliminated from the analysis, the nasometry scores became much more accurate; "Among the 30 patients who manifested no nasal emission, the nasometry scores could be used successfully to identify all subjects with hyponasality. In fact, these scores correctly categorized 90% of all the subjects (27 of 30)." This over-generalization of hypernasality due to the presence of nasal emissions is important also in the analysis of cul-de-sac resonance. If the clinician depends solely on the results of the instrumentation, a client with cul-de-sac resonance AND nasal emissions may be incorrectly identified as being hypernasal.
V. Research-supported directions of assessment and intervention for cul-de-sac resonance
An accurate and complete evaluation of an individual suspected of having any resonance disorder should always involve a perceptual assessment. Because this part of the evaluation is important in determining whether or not the speech produced is abnormal in some way, this assessment must be done prior to any instrumental analysis.
Nasometry is one instrumental assessment technique that may be used to detect overall nasalance in the speech. When cul-de-sac resonance is suspected, however, the nasometer serves to "rule out" hyper- and hyponasality in an effort to more clearly define the pharyngeal resonance. The nasometer can be used reliably to detect nasalance with one qualification, the presence of nasal emissions. Dalston et al. (1991) expressed these concerns at the conclusion of their study when they wrote, "We feel that Nasometer measurements can be used with considerable confidence in corroborating clinical impressions of hyponasality… (however) this is true only for those individuals who do not manifest clinically discernible nasal emission" (Dalston et al., 1991). In any individual exhibiting nasal emissions, clinicians must be wary of possible skewed nasalance scores in the direction of hypernasality.
Spectral analysis of the first and second formants in deaf speakers has clearly illustrated the decreased phonological space used to articulate maximally different vowel sounds. Angelocci, Kopp & Holbrook (1964) found that in a F1-F2 frequency plot of the three extreme vowels, the productions of the deaf subjects overlapped extensively with each other. These vowels were described as, in general, being more "centralized" or "schwa-like" than normal. Angelocci, et al. concluded that the deaf subjects did not have, "clearly defined articulatory vowel target areas." Although this research is important in that it provides the instrumental evidence for the acoustic perception of neutralized vowel sounds, it is also important to note that many deaf speakers maintain the intelligibility of distinct vowels even with reduced phonological space. According to Monsen (1976), "the fact that the phonological space of vowel articulation may be reduced for some deaf speakers does not necessarily mean that the vowels are ill-defined phonemically." However, Monsen does concede that the "probability of overlap of vowel targets is greatly increased" with this reduced phonological space. Clinically, the spectral analysis of "extreme" vowels and the resultant measurement of phonological space may be an effective technique for pre- and post-baseline measurements in illustrating the acquisition of additional phonological space between vowels.
Current research supports Boone’s (1977) hypothesis that excessive pharyngeal or cul-de-sac resonance is (at least in part) due to incorrect tongue placement in the oral cavity. Since the retraction of the tongue causes the sound to be "trapped" in the pharyngeal area, it is necessary for the clinician working with such a client to encourage and demonstrate a more anterior tongue position during speech. When working with cul-de-sac resonance disorders, according to Subtelny et al. (1989), "Speech-Language services should promote and habituate an elevated and fronted carriage during speech." Since these aberrant tongue positions affect not only resonance, but also the pitch and intensity of speech, the Visipitch (Kay Elemetrics) or other similar visual feedback device may be used therapeutically; providing visual feedback to alter the pitch or the intensity of the speaking voice will, indirectly, promote more forward tongue movements.
Although much research has been completed regarding resonance disorders and, in particular, cul-de-sac resonance, there is still much work to be done. There is a need for additional research in the areas of both instrumental assessment and intervention of cul-de-sac resonance. However, even with the advent of more advanced techniques, the clinician must ensure that he uses a variety of perceptual and instrumental measures to correctly identify and remediate the resonance disorder.
Angelocci, A.A., Kopp, G.A. & Holbrook, A. (1964). The vowel formants of deaf and normal-
hearing eleven to fourteen-year-old boys. Journal of Speech and Hearing Disorders, 29(2),
Baken, R.J. (1987). Clinical measurements of speech and voice. Boston, MA: College-Hill Press.
Calvert, D. (1962). Deaf voice quality: A preliminary investigation. Volta Review, 64, 402-405.
Colton, R.H. & Cooker, H.S. (1968). Perceived nasality in the speech of the deaf. Journal of
Speech and Hearing Research, 11, 553-559.
Dalston, R.M., Warren, D.W. & Dalston, E.T. (1991). A preliminary investigation concerning the
use of nasometry in identifying patients with hyponasality and/or nasal airway impairment.
Journal of Speech and Hearing Research, 34, 11-18.
Fletcher, S.G. (1973). Perceptual skills in clinical management of nasality. Folia Phoniatrica, 25,
Fletcher, S.G. & Daly, D.A. (1976). Nasalance in utterances of hearing-impaired speakers.
Journal of Communication Disorders, 9, 63-73.
Fletcher, S.G. & Higgins, J.M. (1980). Performance of children with severe to profound auditory
impairment in instrumentally guided reduction of nasal resonance. Journal of Speech and
Hearing Disorders, 45, 181-194.
Fletcher, S.G., Mahfuzh, F. & Hendarmin, H. (1999). Nasalance in the speech of children with
normal hearing and children with hearing loss. American Journal of Speech-Language
Pathology, 8, 241-248.
Haapanen, M.L. (1991). A simple clinical method of evaluating perceived hypernasality. Folia
Phoniatrica, 43, 122-132.
Haapanen, M.L., Liu, L., Hiltunen, T., Leinonen, L. & Karhunen, J. (1996). Cul-de-sac
hypernasality test with pattern recognition of LPC indices. Folia Phoniatrica, 48, 35-43.
Horii, Y. (1983). An accelerometric measure as a physical correlate of perceived hypernasality in
speech. Journal of Speech and Hearing Research, 26, 476-480.
Kent, R.D. (1979). Isovowel lines for the evaluation of vowel formant structure in speech
disorders. Journal of Speech and Hearing Disorders, 44, 504-512.
Kummer, A.W. & Lee, L. (1996). Evaluation and treatment of resonance disorders. Language,
Speech, and Hearing Services in Schools, 27, 271-281.
LaPine, P.R., Stewart, M.G. & Tatchell, J. (1991). Application of nasometry to speech samples of
hearing-impaired children. Perceptual and Motor Skills, 73, 467-475.
Monsen, R.B. (1976). Normal and reduced phonological space: the production of English vowels
by deaf adolescents. Journal of Phonetics, 4, 189-198.
Subtelny, J., Walter, L., Whitehead, R. & Subtelny, J.D. (1989). Cephalometric and
cineradiographic study of deviant resonance in hearing-impaired speakers. Journal of Speech
and Hearing Disorders, 54, 249-263.
Subtelny, J., Whitehead, R. & Orlando, N. (1980). Description and evaluation of an instructional
program to improve speech and voice diagnosis of the hearing impaired. The Volta Review, 82,
Subtelny, J., Whitehead, R. & Samar, V. (1992). Spectral study of deviant resonance in the
speech of women who are deaf. Journal of Speech and Hearing Research, 35, 574-579.