Assessment of nasalization in the speech of deaf children

Detection and assessment of hypernasality in repaired cleft palate speech using vocal tract and residual features

The presence of hypernasality in repaired cleft palate (CP) speech is a consequence of velopharyngeal insufficiency. The coupling of the nasal tract with the oral tract adds nasal formant and antiformant pairs in the hypernasal speech spectrum. This addition deviates the spectral and linear prediction (LP) residual characteristics of hypernasal speech compared to normal speech. In this work, the vocal tract constriction feature, peak to side-lobe ratio feature, and spectral moment features augmented by low-order cepstral coefficients are used to capture the spectral and residual deviations for hypernasality detection. The first feature captures the lower-frequencies prominence in speech due to the presence of nasal formants, the second feature captures the undesirable signal components in the residual signal due to the nasal antiformants, and the third feature captures the information about formants and antiformants in the spectrum along with the spectral envelope. The combination of three features gives normal versus hypernasal speech detection accuracies of 87.76%, 91.13%, and 93.70% for /a/, /i/, and /u/ vowels, respectively, and hypernasality severity detection accuracies of 80.13% and 81.25% for /i/ and /u/ vowels, respectively. The speech data are collected from 30 control normal and 30 repaired CP children between the ages of 7 and 12.

Vowel production of Mandarin-speaking hearing aid users with different types of hearing loss

In contrast with previous research focusing on cochlear implants, this study examined the speech performance of hearing aid users with conductive (n = 11), mixed (n = 10), and sensorineural hearing loss (n = 7) and compared it with the speech of hearing control. Speech intelligibility was evaluated by computing the vowel space area defined by the Mandarin Chinese corner vowels /a, u, i/. The acoustic differences between the vowels were assessed using the Euclidean distance. The results revealed that both the conductive and mixed hearing loss groups exhibited a reduced vowel working space, but no significant difference was found between the sensorineural hearing loss and normal hearing groups. An analysis using the Euclidean distance further showed that the compression of vowel space area in conductive hearing loss can be attributed to the substantial lowering of the second formant of /i/. The differences in vowel production between groups are discussed in terms of the occlusion effect and the signal transmission media of various hearing devices.

Causes of Velopharyngeal Incompetence

VPI is a structural or functional aberration of the velum, pharynx, or both that underlies an accompanying speech problem. The onset of VPI may occur anytime during an individual's life but is most prevalent in children between the ages of about 2 through 5 and is typically a sequela of cleft palate. The cleft is usually overt and repaired surgically during infancy but may be a submucous cleft palate in which surgical repair generally occurs later in life. VPI sometimes also accompanies a pharyngeal deficit, particularly a deep pharynx. Other important causes of VPI include various neurologic disorders and hearing loss, although the latter may be qualitatively different from the VPI attributable to other sources. This report is a discussion of these etiologies.

Acoustic characteristics of vowels produced by Greek intelligible speakers with profound hearing impairment II: The influence of stress and context

PURPOSE

The present paper examines the influence of stress and context on selected acoustic characteristics of vowels produced by six adult Greek intelligible speakers with profound hearing impairment and six speakers with normal hearing (three males and three females in each group).

METHOD

F1, F2 and F3 formant frequencies and the duration of vowels are measured in words of the form /(')pVCV/ with V = /i, ɛ, ɐ, ɔ, u/ and C = /p, t, k, s/. Variation in these parameters due to context and stress is reported for the two genders. A co-articulatory measure and three measures that examine the area of the vowel space and differences along the F1 and F2 axes are calculated.

RESULT

The results show a reduction of the vowel space in the unstressed condition for both groups, while vowel duration was found to be significantly longer in this condition for the speakers with hearing impairment. In addition, smaller C-to-V carryover co-articulatory effects were found for the speakers with hearing impairment.

CONCLUSION

Findings are discussed within the framework of perceptual and production constraints in hearing impairment and with reference to current models of co-articulation.

Hearing impairment and vowel production. A comparison between normally hearing, hearing-aided and cochlear implanted Dutch children

This study investigated the acoustic characteristics of the Belgian Standard Dutch vowels in children with hearing impairment and in children with normal hearing. In a balanced experimental design, the 12 vowels of Belgian Standard Dutch were recorded in three groups of children: a group of children with normal hearing, a group with a conventional hearing aid and a group with a cochlear implant. The formants, the surface area of the vowel space and the acoustic differentiation between the vowels were determined. The analyses revealed that many of the vowels in hearing-impaired children showed a reduction of the formant values. This reduction was particularly significant with respect to F2. The size of the vowel space was significantly smaller in the hearing-impaired children. Finally, a smaller acoustic differentiation between the vowels was observed in children with hearing impairment. The results show that even after 5 years of device use, the acoustic characteristics of the vowels in hearing-assisted children remain significantly different as compared to their NH peers.

Nasalance in Cochlear Implantees

Objectives

Speech intelligibility is severely affected in children with congenital profound hearing loss. Hypernasality is a problem commonly encountered in their speech. Auditory information received from cochlear implants is expected to be far superior to that from hearing aids. Our study aimed at comparing the percentages of nasality in the speech of the cochlear implantees with hearing aid users and also with children with normal hearing.

Methods

Three groups of subjects took part in the study. Groups I and II comprised 12 children each, in the age range of 4-10 years, with prelingual bilateral profound hearing loss, using multichannel cochlear implants and digital hearing aids respectively. Both groups had received at least one year of speech therapy intervention since cochlear implant surgery and hearing aid fitting respectively. The third group consisted of age-matched and sex-matched children with normal hearing. The subjects were asked to say a sentence which consisted of only oral sounds and no nasal sounds ("Buy baby a bib"). The nasalance score as a percentage was calculated.

Results

Statistical analysis revealed that the children using hearing aids showed a high percentage of nasalance in their speech. The cochlear implantees showed a lower percentage of nasalance compared to children using hearing aids, but did not match with their normal hearing peers.

Conclusion

The quality of speech of the cochlear implantees was superior to that of the hearing aid users, but did not match with the normal controls. The study suggests that acoustic variables still exist after cochlear implantation in children, with hearing impairments at deviant levels, which needs attention. Further research needs to be carried out to explore the effect of the age at implantation as a variable in reducing nasality in the speech and attaining normative values in cochlear implantees, and also between unilateral versus bilateral implantees.

Nasalance and nasality in children with cochlear implants and children with hearing aids

OBJECTIVES

In prelingually deaf children, many speech production aspects including resonance, are known to be problematic. This study aimed to investigate nasality and nasalance in two groups of prelingually hearing impaired children, namely deaf children with a cochlear implant (CI) and moderate-to-severely hearing impaired hearing aid (HA) users. The results of both groups are compared with the results of normal hearing children. Besides, the impact of the degree of hearing loss was determined.

METHODOLOGY

36 CI children (mean age: 9;0y), 25 HA children (mean age: 9;1y) and 26 NH children (mean age: 9;3y) were assessed using objective assessment techniques and perceptual evaluations in order to investigate the nasal resonance of the three groups. Ten HA children had thresholds above 70dB (range: 91dB-105dB) and fifteen below 70dB (range: 58dB-68dB). The Nasometer was used for registration of the nasalance values and nasality was perceptually evaluated by two experienced speech therapists using a nominal rating scale (consensus evaluation).

RESULTS

For nasal stimuli, both CI children and HA children showed lower nasalance values in comparison with NH children. The opposite was observed for the oral stimuli. In both hearing impaired groups, cul-de-sac-resonance was observed on a significantly larger scale than in the NH group, and the HA children were judged to be significantly more hypernasal in comparison with NH children.

CONCLUSIONS

Despite the fact that a substantial number of the CI and HA children demonstrate normal (nasal) resonance quality, this aspect of speech production is still at risk for hearing impaired children.

Effects of spectral content on Horii Oral-Nasal Coupling scores in children

A miniature accelerometer and microphone can be used to obtain Horii Oral-Nasal Coupling (HONC) scores to objectively measure nasalization of speech. While this instrumentation compares favorably in terms of size and cost relative to other objective measures of nasality, the metric has not been well characterized in children. Furthermore, the measure is known to be affected by vowel loading, as speech loaded with "high" vowels is consistently scored as more nasal than speech loaded with "low" vowels. Filtering the signals used in computation of the HONC score to better isolate the correlates of nasalization has been shown to reduce vowel-related effects on the metric, but the efficacy of filtering has thus far only been explored in adults. Here, HONC scores for running speech and the vowel portions of consonant-vowel-consonant tokens were calculated for the speech of 26 children, aged 4-9 yrs. Scores were computed using the broadband accelerometer and speech signals, as well as using filtered, low-frequency versions of these signals. HONC scores obtained using both broadband and filtered signals resulted in well-separated scores for nasal and non-nasal speech. HONC scores computed using filtered signals were found to exhibit less within-participant variability.

The influence of speaking rate on nasality in the speech of hearing-impaired individuals

PURPOSE

The purpose of this study was to determine whether deliberate increases in speaking rate would serve to decrease the amount of nasality in the speech of severely hearing-impaired individuals.

METHOD

The participants were 11 severely to profoundly hearing-impaired students, ranging in age from 12 to 19 years (M = 16 years). Each participant provided a baseline speech sample (R1) followed by 3 training sessions during which participants were trained to increase their speaking rate. Following the training sessions, a second speech sample was obtained (R2). Acoustic and perceptual analyses of the speech samples obtained at R1 and R2 were undertaken. The acoustic analysis focused on changes in first (F(1)) and second (F(2)) formant frequency and formant bandwidths. The perceptual analysis involved listener ratings of the speech samples (at R1 and R2) for perceived nasality.

RESULTS

Findings indicated a significant increase in speaking rate at R2. In addition, significantly narrower F(2) bandwidth and lower perceptual rating scores of nasality were obtained at R2 across all participants, suggesting a decrease in nasality as speaking rate increases.

CONCLUSION

The nasality demonstrated by hearing-impaired individuals is amenable to change when speaking rate is increased. The influences of speaking rate changes on the perception and production of nasality in hearing-impaired individuals are discussed.

Effect of Cochlear Implantation on Nasality in Children

Hypernasality is a commonly perceived characteristic of speech in deaf adults and children, but the mechanism of this abnormal nasal resonance is poorly understood. The impact of cochlear implantation on nasalance measures in children with severe auditory deprivation has not been previously reported. We conducted a study of nasality in 6 deaf children who had undergone cochlear implantation. Voice recordings were obtained before surgery and 6 months after activation of the implants. The MacKay-Kummer SNAP Test—which consists of a syllable-repetition subtest and a picture-cued subtest—was used to obtain nasalance scores for oral (bilabial, alveolar, velar, and sibilant) and nasal phonemes. Before cochlear implantation, mean nasalance scores were significantly higher than normal during the production of oral phonemes for both subtests (p ≤ 0.05). Six months after activation, the nasalance measures for all components of the syllable-repetition subtest had been restored to within 1 standard deviation of normal. For all oral phonemes of the picture-cued subtest, the elevated nasalance scores were consistently lower after cochlear implant activation, although the difference was statistically significant only for velar tasks. Nasalance scores for nasal phonemes were within 1 standard deviation of normal both before and after implant activation. Our study showed that cochlear implantation partially corrects elevated nasalance measures. Disturbances in nasal resonance may be caused in part by the inability of deaf speakers to monitor velopharyngeal valving with auditory feedback. The trend toward improved nasalance scores after implantation highlights the role of auditory feedback in monitoring velopharyngeal function. Visual biofeedback may be required to further normalize hypernasal speech in profoundly deaf children.

Acoustic analysis and detection of hypernasality using a group delay function

In this paper, we describe a group delay-based signal processing technique for the analysis and detection of hypernasal speech. Our preliminary acoustic analysis on nasalized vowels shows that, even though additional resonances are introduced at various frequency locations, the introduction of a new resonance in the low-frequency region (around 250 Hz) is found to be consistent. This observation is further confirmed by a perceptual analysis carried out on vowel sounds that are modified by introducing different nasal resonances, and an acoustic analysis on hypernasal speech. Based on this, for subsequent experiments the focus is given only to the low-frequency region. The additive property of the group delay function can be exploited to resolve two closely spaced formants. However, when the formants are very close with considerably wider bandwidths as in hypernasal speech, the group delay function also fails to resolve. To overcome this, we suggest a band-limited approach to estimate the locations of the formants. Using the band-limited group delay spectrum, we define a new acoustic measure for the detection of hypernasality. Experiments are carried out on the phonemes /a/, /i/, and /u/ uttered by 33 hypernasal speakers and 30 normal speakers. Using the group delay-based acoustic measure, the performance on a hypernasality detection task is found to be 100% for /a/, 88.78% for /i/ and 86.66% for /u/. The effectiveness of this acoustic measure is further cross-verified on a speech data collected in an entirely different recording environment.

Comparison of the overall intelligibility, articulation, resonance, and voice characteristics between children using cochlear implants and those using bilateral hearing aids: a pilot study

The purpose of this study was to determine and to compare the overall intelligibility, articulation, resonance, and voice characteristics in children using cochlear implants (CI) and children using conventional hearing aids (HA). Nine prelingually deaf children using CI and six children with a prelingual severe hearing loss using HA, were selected to participate. Objective (DSI, nasalance scores) as well as subjective assessment techniques (perceptual evaluations) were used. Both the CI and HA children demonstrated normal vocal quality and resonance but showed the presence of articulation disorders. In the CI children, intelligibility was significantly better compared to the HA children. Significantly more phonetic and phonological disorders were present in the HA children. The results of this study show a poorer intelligibility of the HA children in comparison with the CI children which is probably due to the occurrence of significantly more phonetic and phonological disorders. Future detailed analysis in a larger sample of CI and HA children may help further clarify the issue of speech and voice characteristics and may demonstrate an important prognostic value.

The influence of dentures on nasalance values in speech

OBJECTIVE

The purpose of this study was to investigate the effects of maxillary dentures on nasalance values in normal elderly individuals.

DESIGN

A three-factor within-subjects experimental design was employed to analyze the nasalance scores obtained.

PARTICIPANTS

Twenty English-speaking women aged 61 to 81 years, who wore complete maxillary dentures, participated in the study.

INTERVENTION

A Nasometer was used to collect nasalance data. Participants read three standard passages aloud, three times each, in two conditions: with and without their maxillary dentures.

RESULTS

Results revealed that nasalance values were significantly lower with the maxillary dentures removed (p<.01), although the difference in nasalance between the two denture conditions averaged no more than 2%. Nasalance scores obtained both with and without maxillary dentures fell within one standard deviation of the mean for a reference group of comparably aged English-speaking women.

CONCLUSIONS

The results suggest that the existing collection of normal reference data for nasalance may be safely used for elderly individuals regardless of their status with respect to natural or prosthetic maxillary dentition.

The effect of auditory feedback on the control of oral-nasal balance by pediatric cochlear implant users

OBJECTIVE

This study examined the control of oral-nasal balance by pediatric cochlear implant (CI) users, with and without auditory feedback.

DESIGN

Five CI users read lists of sentences in two conditions: with their devices on and with their devices off. Their oral-nasal balance (ratio of energy radiated from the oral and nasal cavities) was measured in both conditions and compared with values obtained from children with normal hearing.

RESULTS

CI users showed different patterns of abnormal oral-nasal balance with their devices off, but they generally achieved values that were closer to normal when their devices were on.

CONCLUSIONS

The results suggest that children with CIs use the auditory signal provided by their device to improve their control of nasalization. It is also possible that at least part of the changes in oral-nasal balance were driven by changes in related articulatory parameters.

Acoustic correlates of English and French nasalized vowels

Acoustic analysis of nasalized vowels in the frequency domain indicates the presence of extra peaks: one between the first two formants with amplitude P1 and one at lower frequencies, often below the first formant, with amplitude P0. The first-formant amplitude A1 is also reduced relative to its amplitude for an oral vowel. These acoustic characteristics can be explained by speech production theory. The objective of this study was to determine the values for the acoustic correlates A1-P1 and A1-P0 (dB) for quantifying nasalization. They were tested as measures of nasalization by comparing vowels between nasal consonants and those between stop consonants for English speakers. Also, portions of nasal vowels following a stop consonant were compared for speakers of French, which makes a linguistic distinction between oral and nasal vowels. In the analysis of English, the mean difference of A1-P1 measured in oral vowels and nasalized vowels had a range of 10 dB-15 dB; the difference of A1-P0 had a range of 6 dB-8 dB. In the study of French, the difference of A1-P1 measured between the least-nasalized portion and the most-nasalized portion of the vowel had a range of 9 dB-12 dB; for A1-P0, the difference ranged between 3 dB and 9 dB. In order to obtain an absolute acoustic measure of nasalization that was independent of vowel type, normalized parameters were calculated by adjusting for the influence of the vowel formant frequencies.

HONC measures in men and women: validity and variability

Horii's (1980) Oral Nasal Coupling index (HONC) was used in measures of 20 women and 20 men with normal speech to validate procedures used for assessment of disordered nasal resonance. Subjects produced sustained vowels [i] and [a], repeated single word productions of "baby" and "mamie" and nonnasal and nasal sentences. Results showed that the correction factors generated during [m] calibration procedures differed significantly between women and men, but not over time within the single measurement session. Differences were also found for the decibel levels produced in the voice channel during the [m] calibration procedure: the women used higher dB levels, particularly towards the end of the session, than the men. In addition, dB levels differed over time. Differences of 13 dB (HONC) were found to separate nonnasal from nasal sentences supporting the validity of the HONC measure. Smaller differences were found between sustained vowel and repeated single words and nasal sentences. Greater variability found for vowel productions also suggests that the sentence stimuli may be more effective for demonstrating hyper- or hyponasality.

Nasality and velopharyngeal function in five hearing impaired adults

Cineradiographic observations of velopharyngeal functioning and listener judgments of hypernasality were compared for five hearing impaired adults. All of the hearing impaired speakers were perceived to have speech characterized by excessive nasality, while only two of the subjects exhibited any velopharyngeal opening. The relationships between the perception of excessive nasality and variables other than the physiology of the velopharyngeal mechanism are discussed.

Speech production in hearing-impaired children

Investigations in recent years have indicated that only about 20% of the speech output of the deaf is understood by the "person-on-the-street." This lack of intelligibility has been associated with some frequently occurring segmental and suprasegmental errors. Errors of voicing and omission of consonants as well as vowel substitutions are common articulatory problems. Suprasegmental errors are associated with problems of timing, intonation, and voice quality. Yet, in spite of the relatively consistent findings regarding these frequently occurring errors, the correlation between the errors and overall speech intelligibility is less clear. The interrelationship between error types and the frequency of occurrence of a particular error type must be taken into consideration when interpreting the true effect of the error type on speech intelligibility. It is only when this is determined that one can begin to predict how much improvement to expect from correction of specific errors.

A radiographic investigation of velar positioning in hearing-impaired young adults

Listener judgments of the presence and severity of hypernasality were made of speech samples from 25 hearing-impaired young adults. These judgments were compared to the results of manometric testing, cephalometric analyses of velar positioning, and analyses of nasal air emission, nasal rustle, and glottal stops. Only two examples of velopharyngeal opening were observed in the radiographic analysis. Also, those subjects judged to be hypernasal failed to exhibit behavior on the other clinical measurements, which was consistent with expectations based on observations of hypernasality in the oral-facial cleft-palate population. The results of this study illustrate the need for continued research into the anatomical and physiological attributes of the perceived hypernasality in the hearing impaired. Also, the remedial consequences of these findings are discussed.