PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of jslhrSubscribeFor AuthorsRelated TopicsJSLHRJSLHR
 
J Speech Lang Hear Res. 2016 October; 59(5): 1025–1034.
PMCID: PMC5345552

Effects of Biofeedback on Control and Generalization of Nasalization in Typical Speakers

Abstract

Purpose

The purpose of this study was to determine the effects of biofeedback on control of nasalization in individuals with typical speech.

Method

Forty-eight individuals with typical speech attempted to increase and decrease vowel nasalization. During training, stimuli consisted of consonant-vowel-consonant (CVC) tokens with the center vowels /a/ or /i/ in either a nasal or nonnasal phonemic context (e.g., /mim/ vs. /bib/), depending on the participant’s training group. Half of the participants had access to augmentative visual feedback during training, which was based on a less-invasive acoustic, accelerometric measure of vowel nasalization—the Horii oral–nasal coupling (HONC) score. During pre- and posttraining assessments, acoustically based nasalance was also measured from the center vowels /a/, /i/, /æ/, and /u/ of CVCs in both nasal and nonnasal contexts.

Results

Linear regressions indicated that both phonemic contexts (nasal or nonnasal) and the presence of augmentative visual feedback during training were significant predictors for changes in nasalance scores from pre- to posttraining.

Conclusions

Participants were able to change the nasalization of their speech following a training period with HONC biofeedback. Future work is necessary to examine the effect of such training in individuals with velopharyngeal dysfunction.

The ability to produce intelligible speech relies on both the accurate auditory perception of speech sounds and somatosensory (i.e., tactile and proprioceptive) information about the state of the vocal tract during their production (Guenther, 1994, 1995; Perkell, 1980; Tourville & Guenther, 2011). This feedback is important for control of the velopharyngeal (VP) port, the connection between the pharynx and the nasal cavity.

In spoken English there are three nasal phonemes, /m/, /n/, and /ŋ/, which require an open VP port. These phonemes are produced with a high degree of nasalization: the passage of acoustic energy through the open VP port and the nasal cavity. Due to coarticulation, vowels that are produced in a nasal context, either preceding or following a nasal phoneme, are also typically produced with an open VP port (Ali, Gallagher, Goldstein, & Daniloff, 1971; Kent, Carney, & Severeid, 1974). In obstruents, conversely, the VP port is closed to allow air pressure to build up in the oral cavity. Nasality is the percept of the nasalization of a voiced speech sound. When oral vowels are produced with an inappropriately open VP port, an excess of acoustic energy emanates from the nasal passages, which is perceived as hypernasality. In speech perceived as hyponasal, the VP port is closed to an inappropriate degree in sounds that are intended to be nasalized, resulting in reduction of the acoustic energy emitted through the nasal cavity.

Augmentative visual feedback may be employed to facilitate speech rehabilitation in VP mislearning, in which inappropriate articulation of the VP mechanism is not caused by a physical or anatomical problem with the VP mechanism (Loney & Bloem, 1987; Peterson-Falzone, Trost-Cardamone, Karnell, & Hardin-Jones, 2006; Trost-Cardamone, 1989) and potentially may be useful for individuals with VP incompetency with neurological origins. At the present time, augmentative visual feedback is often provided in clinical settings via two types of instrumentation: (a) a nasal endoscope, providing a direct visualization of the VP port opening, or (b) a nasometer, providing an indirect measure of VP port status. Although studies demonstrate that this biofeedback can be beneficial in assisting with VP port control in both children and adults (Brunner, Stellzig-Eisenhauer, Pröschel, Verres, & Komposch, 2005; Van Lierde, Claeys, De Bodt, & Van Cauwenberge, 2004; Witzel, Tobe, & Salyer, 1988, 1989; Yamaoka, Matsuya, Miyazaki, Nishio, & Ibuki, 1983; Ysunza, Pamplona, Femat, Mayer, & García-Velasco, 1997), both require invasive or bulky and uncomfortable hardware (Pratt & Hricisak, 1994), limiting their usability in clinical environments.

Although not typically used in clinical settings, vibration of the exterior surface of the nose, transduced by an accelerometer, has also been suggested as a correlate of perceived nasality (Horii, 1980, 1983; Laczi, Sussman, Stathopoulos, & Huber, 2005; Mra, Sussman, & Fenwick, 1998; Redenbaugh & Reich, 1985). The amplitude of this vibration can be either measured in isolation (Stevens, Kalikow, & Willemain, 1975) or compared to the combined oral and nasal acoustic output via a single microphone, referred to as the Horii oral–nasal coupling score, or HONC score (Horii, 1980, 1983). HONC instrumentation is lightweight, doesn’t require bulky hardware, and has low power consumption (Stevens et al., 1975). Furthermore, HONC scores are correlated with experts’ perception of nasality (Horii, 1983), and can effectively differentiate between nasalized and nonnasalized speech tokens (Mra et al., 1998; Thorp, Virnik, & Stepp, 2013; Varghese, Mendoza, Braden, & Stepp, 2014). In addition, previous work has shown that the use of HONC scores to differentiate between nasalized and nonnasalized speech tokens produces results comparable to the use of nasalance (Thorp et al., 2013).

The purpose of this study was to examine the nasalization of vowels in individuals with typical speech before and after a training period with biofeedback based on HONC scores. Although originally proposed as a training modality (Stevens et al., 1975), HONC instrumentation not been explicitly examined as a biofeedback modality to produce changes in nasalization. Adults with typical speech participated in a single training session in which they attempted to increase and decrease the nasalization of consonant-vowel-consonant (CVC) speech tokens. Augmentative visual feedback based on HONC scores was provided to half of the participants during training, whereas the other half of the participants only had access to their typical auditory feedback. During training, half of participants produced CVC speech tokens in a nasal phonemic context (/m/ and /n/), and the other half in a nonnasal phonemic context (/b/ and /d/). We hypothesized that individuals who received augmentative visual feedback during training would exhibit larger changes in their ability to modify their degree of nasalization as a result of training. We also hypothesized that participants would be more effective at changing the nasalization of tokens on which they were explicitly trained.

Method

Participants

Participants were 48 adults (30 women, 18 men) with an average age of 20.2 years (SD = 3.9 years) and no history of speech, language, or hearing disorders. All participants were fluent speakers of American English and were not fluent in any language with contrastive vowel nasalization (e.g., French). Testing occurred over 2 sequential days: Session 1 (approximately 90 min) and Session 2 (approximately 30 min). All participants provided written consent in compliance with the Boston University Institutional Review Board and were compensated for participation.

Experimental Design

Participants completed (a) a preassessment (Pre) of nasalance, (b) a training session with or without biofeedback based on the HONC scores, (c) a postassessment (Post) of nasalance, and (d) a follow-up assessment (Follow-up) of nasalance on the next day (see Figure 1). During all portions of the experiment, participants were instructed to either increase or decrease the nasalization of CVC speech tokens.

Figure 1.
Experimental design with the number of subjects and training group designation indicated. Participants were presented visually with CVC speech tokens. During training half of the participants saw speech tokens in a nasal phonemic context (Nas) and half ...

During the training portion, participants were assigned to one of four groups, with 12 people in each group (see Figure 1). Two groups were trained on stimuli with vowels in a nasal phonemic context (Nas) and two groups were trained on stimuli with vowels in a nonnasal phonemic context (Non). In addition, two groups (one Nas and one Non) only had access to their typical auditory feedback (0vfb), whereas the remaining two groups received augmentative visual feedback (+vfb) in addition to having access to their typical auditory feedback.

Signals Collected

All recordings were conducted in a sound-treated room. Nasalance was collected during assessments with the Nasometer II, model 6450 (Kay Elemetrics Corp, Lincoln Park, NJ) using standard Nasometer II software with a sampling frequency of 11025 Hz. HONC data were collected during training with a headset microphone (Shure model WH20; Shure, Inc., Niles, IL) and BU Series 21771 accelerometer (Knowles Electronic, Itasca, IL). The accelerometer signal was band-pass filtered from 400–1000 Hz and the microphone was filtered from 25–420 Hz, allowing for utilization of the low-frequency portion of the signals (Thorp et al., 2013; Varghese et al., 2014). The microphone was placed 7 cm from the mouth at a 45° angle. The accelerometer was placed on the lateral aspect of the nose, inferior to the nasal bone and superior to the ala (Lippmann, 1981). Both the microphone signal and nasal accelerometer signals were preamplified by a RME Quadmic II (RME, Haimhausen, Germany) and sampled at 44100 Hz using one of two sound cards: a MOTU ultralite mk3 hybrid (model UltraLite3Hy; MOTU, Cambridge, MA) or Komplete Audio 6 (Native Instruments, Los Angeles, CA).

Assessments

Nasalance was measured during the three assessment periods (Pre, Post, and Follow-up). Regardless of training group, during the assessments all participants produced eight CVC speech tokens in a nasal phonemic context and eight tokens in a nonnasal phonemic context. All tokens had a center vowel that was either /a/, /i/, /æ/, or /u/ and were produced in both a congruent manner (e.g., say nonnasal “bob”) and an incongruent manner (e.g., say nasal “bob”), resulting in 32 unique productions. These 32 unique tokens were presented randomly in each of the three blocks for the Pre assessment and the Post assessment. For the Follow-up assessment, the speech tokens were presented in five blocks, with only the last three repetitions of the 32 tokens analyzed.

Training Session

The training session was completed with the HONC instrumentation. Throughout the training session, participants were asked to produce four tokens, once in a congruent manner and once in an incongruent manner, for a total of eight unique speech tokens, which were then randomized and repeated in 28 training blocks. Half the participants were trained in a nonnasal phonemic context and the other half were trained in a nasal phonemic context. Participants in the Non (0vfb) and Nas (0vfb) groups were presented with directions regarding what to produce; no augmentative visual feedback was given. All speech tokens had the center vowels /a/ and /i/. The training of only two of the vowels allowed for the evaluation of generalization (Maas et al., 2008) to the untrained vowels during the assessment periods.

For participants in the Nas (+vfb) and Non (+vfb) groups, augmentative visual feedback based on real-time HONC scores was provided during the training session via a custom MATLAB (2013) graphical user interface. A HONC score (see Equation 1; Horii, 1980) was calculated from the vowel of each speech token using MATLAB. HONC scores were calculated over the center 25% of the duration of each production, determined via a 5% threshold above the noise floor of the microphone signal. The root mean square (RMS) of the accelerometer signal, RMS A, was divided by the RMS of the microphone signal, RMS M. To reduce between-speaker variability, a normalization factor, defined as RMS M of the sustained /m/ /RMS A of the sustained /m/, was calculated from the average of six productions of a sustained /m/ recorded prior to the training session (see Eq. 1; Horii, 1980).

HONC=RMSMofthesustained/m/RMSAofthesustained/m/×RMSARMSM
(1)

Prior to each training trial, participants were shown directions on a computer screen (e.g., say nasal “bob”), and an arrow pointing upward for attempts to increase nasalization, or downward for attempts to decrease nasalization. In addition, a threshold line was displayed on the screen. Prior to the start of the training, participants were asked to produce, in their typical speaking voice, eight CVC speech tokens, four in a nasal phonemic context and four in a nonnasal phonemic context, with the center vowels /a/ and /i/, to establish the initial threshold. For intended increases in nasalization, participants were given positive visual feedback if they produced a HONC score above the threshold, whereas for intended decreases in nasalization, participants were given positive visual feedback if they produced a HONC score below the threshold. After each production, a vertical bar corresponding to the HONC score and the threshold were displayed. If the participant was successful, the vertical bar was green; otherwise, the vertical bar was red.

During the first 80 trials, thresholds were adjusted using an adaptive staircase procedure, separately for trials with attempted increases in nasalization and for trials with attempted decreases in nasalization. If the participant was successful for two successive trials with the same attempted changes in nasalization, the threshold was moved to make the next trial more challenging; that is, moved up for attempted increases in nasalization and moved down for attempted decreases in nasalization. If the participant continued to be successful, the threshold would continue moving after every subsequent trial until the participant was unable to reach the threshold. When the participant was unable to reach the threshold, the threshold was moved to make the next trial easier. The amount the threshold moved, the step size, when the participant was successful was initially set to 0.20 and systematically decreased throughout the initial part of training until it reached 0.02. The step size for unsuccessful trials was set to one-fifth of the successful step size. These step sizes were chosen empirically based on pilot studies and resulted in stabilization of participants’ thresholds. After the first 80 trials, the thresholds were no longer modified adaptively. At that point, the average HONC score of the last five reversals (a change in the direction of the threshold) was calculated and the threshold was maintained at that level for the remainder of the trials. This threshold was chosen such that the participant was able to produce the correct nasalization the majority of the time, but not on every trial, in order to provide a constant challenge (Guadagnoli & Lee, 2004).

Thus, the augmentative visual feedback in this experiment utilized “knowledge of results” as well as “knowledge of performance” (Gentile, 1972; Salmoni, Schmidt, & Walter, 1984). Participants were provided with feedback in the form of the vertical bar color providing them knowledge of results; that is, whether they were successful or not. In addition, as the vertical bar height corresponded to the HONC score of their production, participants were aware of the how far their productions were from their target (the threshold line), providing them with the knowledge of performance. Further, this presentation of augmentative visual feedback was faded (reduced in frequency) throughout the training portion (Winstein & Schmidt, 1990; Wulf & Schmidt, 1989). The fading of feedback was intended to mitigate the possibility of individuals relying solely on the augmentative (visual) feedback, rather than their own intrinsic (auditory) feedback (Salmoni et al., 1984; Winstein & Schmidt, 1990). In the Nas (+vfb) and Non (+vfb) groups the first 80 trials had 100% augmented visual feedback, the feedback was faded to 50% for the next 48 trials, 25% of the following 48 trials, and 0% of the last 48 trials in the training. The presentations of the trials in which the visual feedback was not shown were identical to the trials seen by the Nas (0vfb) and Non (0vfb) groups throughout training.

Data Analyses

Analyses for this study focused on the clinically relevant tasks, the two incongruent conditions, and the congruent condition for nonnasal tokens. Therefore, nasalance scores were split into three different subsets for analyses, determined by the token type and the intended nasalization direction (see Table 1). Nasalance scores were calculated from the middle portion of the vowel in the CVC token with the Nasometer II software. The Nasometer calculates nasalance (%), which is the ratio of nasal acoustic energy to nasal-plus-oral acoustic energy, multiplied by 100 (Fletcher, 1972). Average changes in nasalance scores from Pre assessment to both the Post and Follow-up assessments were calculated to evaluate the magnitude of change in nasalance scores. On the basis of the larger nasalance changes noted in the incongruent (I) condition, additional analyses were conducted on the I-nas and I-non subsets. Statistical analyses were performed using Minitab Statistical Software (Minitab, 2012). Given the exploratory nature of the study, adjustments were not made for multiple testing. Two linear regressions, one for the I-non subset and one for the I-nas subset examined whether the dependent variables for change in nasalance scores from the Pre to the Post assessments were influenced by augmentative visual feedback (+vfb or 0vfb), phonemic context (Nas or Non), or the interaction between augmentative visual feedback and phonemic context during training. Two additional linear regressions, one for I-non and one for I-nas, were similarly performed to examine whether the changes in nasalance scores from the Pre to Follow-up assessments were influenced by augmentative visual feedback, phonemic context, or the interaction between augmentative visual feedback and phonemic context during training.

Table 1.
Three subsets utilized for statistical analyses.

To examine whether differences in the nasalance scores in the incongruent subsets of vowels were contingent on whether the vowel was trained (/a/ or /i/) or untrained (/æ/ or /u/), four paired-samples t tests were performed. Two paired-samples t tests compared the differences in nasalance (trained vs. untrained vowels) at the Pre assessment compared to the differences in nasalance at the Post assessment. Two additional paired-samples t tests compared the differences in nasalance at the Pre assessment compared to the Follow-up assessment.

Results

Participants demonstrated the largest changes in nasalance scores in the incongruent productions when they received augmentative visual feedback during training and were trained in the specific phonemic context that was being assessed. These changes were maintained to the Follow-up assessment the next day. Although specificity of the phonemic context of trained stimuli was a significant predictor of training effects, there was no clear evidence for such specificity based on the vowels of stimuli used during training.

Descriptive Changes in Nasalance

During the congruent subset C-non there was minimal change in the nasalance values (see Figure 2). In contrast, participants were able to produce intended changes in nasalization in both incongruent subsets (I-non and I-nas) regardless of the phonemic context in which they were trained (see Figure 2). However, although participants were able to produce intended changes in nasalization in all subsets, participants who had training specificity and received augmentative visual feedback had the largest changes in nasalization. For the I-non productions, individuals who received augmentative visual feedback and were trained on nonnasal tokens (Non, +vfb) had the largest increases in nasalance from Pre assessment to Post assessment and from Pre assessment to Follow-up assessment. For the I-nas productions, individuals who received augmentative visual feedback and were trained on nasal tokens (Nas, +vfb) had the largest decreases in nasalance from the Pre assessment to the Post assessment, and from the Pre assessment to the Follow-up assessment.

Figure 2.
Representation for the magnitude change in nasalance in the three subsets. Shaded squares indicate a change in the correct direction for that subset.

Incongruent Manners of Production

Augmentative visual feedback, phonemic context, and the interaction of Augmentative Visual Feedback × Phonemic Context during training were examined as predictors for the change in nasalance scores from the Pre assessment to the Post assessment, and the Pre assessment to the Follow-up assessment within each subset.

Nonnasal Tokens Produced With Increased Nasalance (Subset I-non)

Neither augmentative visual feedback, nor phonemic context, nor the interaction of Augmentative Visual Feedback × Phonemic Context during training were significant predictors for the changes in nasalance scores from the Pre to the Post assessments in the I-non subset (PostPre; see Table 2). However, the interaction of Augmentative Visual Feedback × Phonemic Context significantly predicted the change in nasalance from the Pre to the Follow-up assessments (interaction effect; β =15.68, p = 0.04; see Figure 3). Within the two groups that received augmentative visual feedback, individuals who were trained on nonnasal tokens (Non, +vfb), and therefore were trained in the phonemic context specific for this particular subset, had a larger increase in their nasalance between the Pre and the Follow-up assessments than individuals who were trained in a nasal phonemic context (Nas, +vfb). In the groups that did not receive augmentative visual feedback, individuals who were trained in a nasal phonemic context (Nas, 0vfb) had a greater increase in their nasalance than individuals who were trained in a nonnasal phonemic context (Non, 0vfb) from the Pre to the Follow-up assessments.

Table 2.
Results from linear regression examining the change in nasalance from the Pre to the Post assessments and from the Pre to the Follow-up assessments for subset I-non.
Figure 3.
Mean and 95% confidence intervals for the change in nasalance for nasal tokens produced with attempted decreases in nasalization (subset I-nas). Left, scores from the Pre to the Post assessments; right, scores from the Pre to the Follow-up assessments. ...

Nasal Tokens Produced With Decreased Nasalance (Subset I-nas)

A linear regression model for the change in nasalance scores from the Pre to the Post assessments (PostPre) revealed that augmentative visual feedback and phonemic context were not significant predictors for the changes in nasalance scores in the subset I-nas. However there was a trend towards significance in the interaction of Augmentative Visual Feedback × Phonemic Context (interaction effect; β = −11.37, p = 0.07; see Figure 4). Within the groups that received augmentative visual feedback, individuals who were trained on nasal tokens (Nas, +vfb), and therefore were trained in the phonemic context specific for this particular subset, trended towards a larger decrease in their nasalance scores than individuals who were trained in a nonnasal phonemic context (Non, +vfb). In the group that did not receive visual feedback, individuals who were trained on nasal tokens (Nas, 0vfb) had close to no change in nasalance scores at the Post assessment, whereas individuals who were trained on nonnasal tokens (Non, 0vfb) had a small decrease in nasalance at the Post assessment. Similar results were seen for the changes in nasalance scores from the Pre assessments to the Follow-up assessment, which also showed a trend toward an interaction effect (β = −13.88, p = .07; see Figure 4). As independent factors, augmentative visual feedback and phonemic context were not significant predictors for the changes in nasalance scores from the Pre to the Follow-up assessments (see Table 3).

Figure 4.
Mean and 95% confidence intervals for the change in nasalance for nasal tokens produced with attempted decreases in nasalization (subset I-nas). Left, scores from the Pre to the Post assessments; right, scores from the Pre to the Follow-up assessments. ...
Table 3.
Results from linear regression examining the change in nasalance from the Pre to the Post assessments and from the Pre the Follow-up assessments for subset I-nas.

Effects of Vowel Training Specificity

Differences between nasalance scores for trained vowels (/i/ and /a/) and nasalance scores for untrained vowels (/æ/ and /u/) were compared at each assessment period for each incongruent subset. For subset I-non the difference in nasalance between trained and untrained vowels during the Pre assessment was not significantly different than the difference in nasalance between trained and untrained vowels during either Post assessment or the Follow-up assessment (both p > .05). For the subset I-nas, the difference in nasalance between trained and untrained vowels during the Pre assessment was not significantly different than the difference in nasalance during the Post assessment, t(48) = 0.87, p = .39. However, trained vowels had a 9.9 point higher nasalance score than untrained vowels on the Pre assessment compared to a 7.8 point higher score during the Follow-up assessment, t(48) = 2.74, p = .01.

Discussion

Overall, this study demonstrated the benefit of using the HONC instrumentation for training control of nasalization. Results from this study indicated that both specificity of phonemic context and presence of augmentative visual feedback during training were meaningful components of a training paradigm aimed at changing the level of nasalization (measured via nasalance).

Congruent Productions

Individuals with typical speech were not able to significantly change the nasalization of congruent productions. This is likely due to the anatomical constraints involved in speech production, making it difficult for individuals with typical speech to make large changes during these productions.

Incongruent Productions

Participants were more successful at changing nasalance scores of the incongruent productions. Many participants were able to change their nasalance scores to values that are considered clinically significant and would be cause for further evaluation in a clinical environment. In particular, for an oral passage (e.g., zoo passage; Fletcher, 1972), a nasalance score of 32 has been proposed as a critical threshold for diagnosing VP impairments (Dalston, Warren, & Dalston, 1991c). Participants in the Non (+vfb) group were able to produce nasalance scores above 32 during the Post and Follow-up assessments (range = 32.1–35.1) when they were asked to produce nonnasal words with increased nasalization (subset I-nas). Dalston et al. (1991a, 1991b) also noted that when examining nasal-loaded sentences, production of nasalance scores that were under 50 resulted in a hyponasal quality. In subset I-nas, nasalance scores in the Post assessment for the Non (0vfb) and Nas (+vfb) groups, as well as nasalance scores from the Follow-up assessments in the Non (0vfb), Nas (0vfb), and the Nas (+vfb) groups were below a nasalance score of 50 (range = 34.3–49.0). The lowest nasalance scores were in the Nas (+vfb) group, with an average nasalance score of 38.6 at the Post assessment and 34.3 at the Follow-up assessment. Therefore, participants with typical speech in this study were able to produce speech with nasalance scores that approximated individuals with VP dysfunction. However, it should be noted that these cutoff values were developed for running speech, unlike the vowel stimuli in the current study.

Although previous research has indicated that knowledge of results feedback is initially beneficial (Guadagnoli & Lee, 2004), a continuation of feedback may have a detrimental impact on learning (Salmoni et al., 1984; Winstein & Schmidt, 1990). Individuals can begin to rely on the augmentative feedback instead of their own intrinsic feedback (Salmoni et al., 1984; Winstein & Schmidt, 1990). One study found a negative relationship between how frequently participants were presented with feedback and how successful they were at nasalizing a sustained /i/ during a 10-min training period (Steinhauer & Grayhack, 2000). However, it is possible that this negative relationship was either due to too short a training period, or too easy a task to allow for further improvements. The current study did not demonstrate a negative effect of feedback in the incongruent productions. One possible explanation for this observation is that the fading of augmentative visual feedback during training may have prevented the overreliance on external feedback as opposed to internal feedback.

The strongest predictor for changes in nasalance for the incongruent productions was the interaction between augmentative visual feedback and phonemic context. For subset I-non, within the groups that received augmentative visual feedback, participants who were trained in a nonnasal phonemic context were more effective at increasing their nasalization from the Pre to the Follow-up assessments relative to individuals who trained in a nasal phonemic context. The largest changes in nasalance scores for individuals who received augmentative visual feedback from the Pre to the Post assessment in subset I-nas were individuals who were trained in a nasal phonemic context relative to individuals who were trained in a nonnasal phonemic context. Participants in subset I-nas not only retained these changes in nasalance scores, but the scores further decreased when nasalance was evaluated during the Follow-up assessment. Although participants who were provided with augmentative visual feedback and phonemic context during training were effectively able to change their nasalizations, they displayed increased variability in the magnitude of these changes relative to individuals who did not receive both augmentative visual feedback and phonemic context during training. Overall, results from these incongruent productions indicated that it was not solely the presence of augmentative visual feedback or phonemic context during training that had an influence on changes in nasalization, but rather the interaction between these factors.

Although the results indicated the importance of the specificity of phonemic context (nasal or nonnasal) during training, there was less clear evidence of a vowel specificity effect during training in the incongruent productions. In all of the assessment periods, trained vowels (/a/ and /i/) always had higher nasalance scores relative to the untrained vowels (/æ/ and /u/). This was likely due to the intrinsic nasalance values of the vowels (Awan, Omlor, & Watts, 2011; Gildersleeve-Neumann & Dalston, 2001); however, this did not affect how successful participants were at changing the degree of nasalization of stimuli.

Limitations and Future Directions

This study demonstrated the effectiveness of utilizing the less-invasive HONC instrumentation during a training paradigm to change nasalization in adults with typical speech, suggesting both research significance and clinical potential. This training paradigm was for a relatively short time period, and effects were seen immediately after the training period as well as the next day. Although this carryover is extremely important for clinical use and effectiveness of the training paradigm, future investigations of carryover should be performed.

This study presented a proof of concept for the use of the HONC instrumentation as a training tool, but it did not explore the potential effects of structural differences that may occur in individuals with VP dysfunction (Kummer, 2008). Future work will examine this paradigm in individuals with VP dysfunction to determine the effectiveness in changing nasalization in clinical populations. Furthermore, this biofeedback was only aimed at addressing differences in resonance in vowels. Other errors, such as weak production of pressure consonants, may not be as effectively changed with this paradigm. Last, although this study showed significant changes in nasalance, it was beyond the scope of the current project to assess the perception of nasality. The presence of hypernasality in speech has been shown to be received negatively by both children (Blood & Hyman, 1977) and adults (Lallh & Rochet, 2000; McKinnon, Hess, & Landry, 1986). Therefore, understanding the potential changes in the perception of nasality is an essential next step.

Conclusions

This study demonstrated the effectiveness of utilizing HONC instrumentation in a training paradigm to change the nasalization of speech production in adults with typical speech. Participants demonstrated the largest changes in nasalance in the incongruent productions when they received augmentative visual feedback during training and were trained in the phonemic context that was being assessed. These changes were maintained to the Follow-up assessment the next day. Although specificity of phonemic context during training was important, there was not a clear effect of specificity of the vowel used during training. Future work will examine the use of HONC scores in training nasalization changes in individuals with VP dysfunction.

Acknowledgments

This work was supported in part by the Deborah Munroe Noonan Memorial Research Fund (awarded to Cara Stepp), National Institute on Deafness and Other Communication Disorders Grant DC012651 (awarded to Cara Stepp), and a grant from the Dudley Allen Sargent Research Fund, Boston University, (awarded to Elizabeth Heller Murray). The authors would also like to acknowledge Kate Girouard and Ali Martinson for their help with data analysis.

Funding Statement

This work was supported in part by the Deborah Munroe Noonan Memorial Research Fund (awarded to Cara Stepp), National Institute on Deafness and Other Communication Disorders Grant DC012651 (awarded to Cara Stepp), and a grant from the Dudley Allen Sargent Research Fund, Boston University, (awarded to Elizabeth Heller Murray).

References

  • Ali L., Gallagher T., Goldstein J., & Daniloff R. (1971). Perception of coarticulated nasality. The Journal of the Acoustical Society of America, 49(2B), 538–540. [PubMed]
  • Awan S. N., Omlor K., & Watts C. R. (2011). Effects of computer system and vowel loading on measures of nasalance. Journal of Speech, Language, and Hearing Research, 54(5), 1284–1294. [PubMed]
  • Blood G. W., & Hyman M. (1977). Children’s perception of nasal resonance. Journal of Speech and Hearing Disorders, 42(3), 446–448. [PubMed]
  • Brunner M., Stellzig-Eisenhauer A., Pröschel U., Verres R., & Komposch G. (2005). The effect of nasopharyngoscopic biofeedback in patients with cleft palate and velopharyngeal dysfunction. The Cleft Palate–Craniofacial Journal, 42(6), 649–657. [PubMed]
  • Dalston R. M., Warren D. W., & Dalston E. T. (1991a). The identification of nasal obstruction through clinical judgments of hyponasality and nasometric assessment of speech acoustics. American Journal of Orthodontics and Dentofacial Orthopedics, 100(1), 59–65. [PubMed]
  • Dalston R. M., Warren D. W., & Dalston E. T. (1991b). A preliminary investigation concerning the use of nasometry in identifying patients with hyponasality and/or nasal airway impairment. Journal of Speech, Language, and Hearing Research, 34(1), 11–18. [PubMed]
  • Dalston R. M., Warren D. W., & Dalston E. T. (1991c). Use of nasometry as a diagnostic tool for identifying patients with velopharyngeal impairment. The Cleft Palate–Craniofacial Journal, 28(2), 184–189. [PubMed]
  • Fletcher S. G. (1972). Contingencies for bioelectric modifications of nasality. Journal of Speech and Hearing Disorders, 37, 329–346. [PubMed]
  • Gentile A. M. (1972). A working model of skill acquisition with application to teaching. Quest, 17(1), 3–23.
  • Gildersleeve-Neumann C. E., & Dalston R. M. (2001). Nasalance scores in noncleft individuals: Why not zero? The Cleft Palate–Craniofacial Journal, 38(2), 106–111. [PubMed]
  • Guadagnoli M. A., & Lee T. D. (2004). Challenge point: A framework for conceptualizing the effects of various practice conditions in motor learning. Journal of Motor Behavior, 36(2), 212–224. [PubMed]
  • Guenther F. H. (1994). A neural network model of speech acquisition and motor equivalent speech production. Biological Cybernetics, 72(1), 43–53. [PubMed]
  • Guenther F. H. (1995). Speech sound acquisition, coarticulation, and rate effects in a neural network model of speech production. Psychological Review, 102(3), 594–621. [PubMed]
  • Horii Y. (1980). An accelerometric approach to nasality measurement: A preliminary report. The Cleft Palate Journal, 17(3), 254–261. [PubMed]
  • Horii Y. (1983). An accelerometric measure as a physical correlate of perceived hypernasality in speech. Journal of Speech and Hearing Disorders, 26(3), 476–480. [PubMed]
  • Kent R. D., Carney P. J., & Severeid L. R. (1974). Velar movement and timing: Evaluation of a model for binary control. Journal of Speech and Hearing Disorders, 17(3), 470–488. [PubMed]
  • Kummer A. W. (2008). Cleft palate and craniofacial anomalies (2nd ed.). Clifton Park, NY: Thomas Delmar Learning.
  • Laczi E., Sussman J. E., Stathopoulos E. T., & Huber J. (2005). Perceptual evaluation of hypernasality compared to HONC measures: The role of experience. The Cleft Palate–Craniofacial Journal, 42(2), 202–211. [PubMed]
  • Lallh A. K., & Rochet A. P. (2000). The effect of information on listeners’ attitudes toward speakers with voice or resonance disorders. Journal of Speech, Language, and Hearing Research, 43(3), 782–795. [PubMed]
  • Lippmann R. (1981). Detecting nasalization using a low-cost miniature accelerometer. Journal of Speech, Language and Hearing Research, 24(3), 314–317. [PubMed]
  • Loney R., & Bloem T. (1987). Velopharyngeal dysfunction: Recommendations for use of nomenclature. The Cleft Palate Journal, 24(4), 334–335. [PubMed]
  • Maas E., Robin D. A., Hula S. N. A., Freedman S. E., Wulf G., Ballard K. J., & Schmidt R. A. (2008). Principles of motor learning in treatment of motor speech disorders. American Journal of Speech-Language Pathology, 17(3), 277–298. [PubMed]
  • MATLAB. (2013). MATLAB R2013a (Version 8.1.0.604) [Computer software]. Natick, MA: MathWorks.
  • McKinnon S. L., Hess C. W., & Landry R. G. (1986). Reactions of college students to speech disorders. Journal of Communication Disorders, 19(1), 75–82. [PubMed]
  • Minitab. (2012). Minitab Statistical Software (Version 16.2.3). State College, PA: Minitab.
  • Mra Z., Sussman J. E., & Fenwick J. (1998). HONC measures in 4- to 6-year-old children. The Cleft Palate–Craniofacial Journal, 35(5), 408–414. [PubMed]
  • Perkell J. S. (1980). Phonetic features and the physiology of speech production. In Butterworth B., editor. (Ed.), Language production: Vol. 1 (pp. 337–372). London, England: Academic Press.
  • Peterson-Falzone S., Trost-Cardamone J. E., Karnell M., & Hardin-Jones M. (2006). The clinician's guide to treating cleft palate speech. St. Louis, MO: Mosby.
  • Pratt S. R., & Hricisak I. (1994). Commercially available computer-based speech feedback systems. Journal of the Academy of Rehabilitative Audiology, 27, 89–106.
  • Redenbaugh M. A., & Reich A. R. (1985). Correspondence between an accelerometric nasal/voice amplitude ratio and listeners’ direct magnitude estimations of hypernasality. Journal of Speech and Hearing Disorders, 28(2), 273–281. [PubMed]
  • Salmoni A. W., Schmidt R. A., & Walter C. B. (1984). Knowledge of results and motor learning: A review and critical reappraisal. Psychological Bulletin, 95(3), 355–386. [PubMed]
  • Steinhauer K., & Grayhack J. P. (2000). The role of knowledge of results in performance and learning of a voice motor task. Journal of Voice, 14(2), 137–145. [PubMed]
  • Stevens K., Kalikow D., & Willemain T. (1975). A miniature accelerometer for detecting glottal waveforms and nasalization. Journal of Speech and Hearing Disorders, 18(3), 594–599. [PubMed]
  • Thorp E. B., Virnik B. T., & Stepp C. E. (2013). Comparison of nasal acceleration and nasalance across vowels. Journal of Speech, Language, and Hearing Research, 56(5), 1476–1484. [PubMed]
  • Tourville J. A., & Guenther F. H. (2011). The DIVA model: A neural theory of speech acquisition and production. Language and Cognitive Processes, 26(7), 952–981. [PMC free article] [PubMed]
  • Trost-Cardamone J. E. (1989). Coming to terms with VPI: A response to Loney and Bloem. The Cleft Palate Journal, 26(1), 68–70. [PubMed]
  • Van Lierde K. M., Claeys S., De Bodt M., & Van Cauwenberge P. (2004). Outcome of laryngeal and velopharyngeal biofeedback treatment in children and young adults: A pilot study. Journal of Voice, 18(1), 97–106. [PubMed]
  • Varghese L. A., Mendoza J. O., Braden M. N., & Stepp C. E. (2014). Effects of spectral content on Horii Oral-Nasal Coupling scores in children. The Journal of the Acoustical Society of America, 136(3), 1295–1306. [PubMed]
  • Winstein C. J., & Schmidt R. A. (1990). Reduced frequency of knowledge of results enhances motor skill learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 16(4), 677–691.
  • Witzel M. A., Tobe J., & Salyer K. (1988). The use of nasopharyngoscopy biofeedback therapy in the correction of inconsistent velopharyngeal closure. International Journal of Pediatric Otorhinolaryngology, 15(2), 137–142. [PubMed]
  • Witzel M. A., Tobe J., & Salyer K. E. (1989). The use of videonasopharyngoscopy for biofeedback therapy in adults after pharyngeal flap surgery. The Cleft Palate Journal, 26(2), 129–134. [PubMed]
  • Wulf G., & Schmidt R. A. (1989). The learning of generalized motor programs: Reducing the relative frequency of knowledge of results enhances memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 15(4), 748–757.
  • Yamaoka M., Matsuya T., Miyazaki T., Nishio J., & Ibuki K. (1983). Visual training for velopharyngeal closure in cleft palate patients: A fibrescopic procedure (Preliminary report). Journal of Maxillofacial Surgery, 11, 191–193. [PubMed]
  • Ysunza A., Pamplona M., Femat T., Mayer I., & García-Velasco M. (1997). Videonasopharyngoscopy as an instrument for visual biofeedback during speech in cleft palate patients. International Journal of Pediatric Otorhinolaryngology, 41(3), 291–298. [PubMed]

Articles from Journal of Speech, Language, and Hearing Research : JSLHR are provided here courtesy of American Speech-Language-Hearing Association