Shaheen N. Awan

Toward the development of an objective index of dysphonia severity: A four‐factor acoustic model

During assessment and management of individuals with voice disorders, clinicians routinely attempt to describe or quantify the severity of a patients dysphonia. This investigation used acoustic measures derived from sustained vowel samples to predict dysphonia severity (as determined by auditory‐perceptual ratings), for a diverse set of voice samples obtained from 134 adult females, with and without voice disorders. Stepwise multiple regression analysis on all voice samples, followed by randomized and repeated cross‐validation (random selection of 75% of the original 134 voice sample corpus; 100 iterations) indicated that a four‐variable model comprised of time and spectral‐based acoustic measures was able to strongly predict perceived severity of dysphonia (mean R = .880; mean R2 = .775). A cepstral‐based measure (CPP/EXP ratio) was determined to be the most significant contributor to the prediction of dysphonia severity, though it is clear that the addition of other acoustic measures (pitch sigma; shimmer (dB); and the Discrete Fourier Transformation ratio, a measure of low versus high frequency spectral energy) add substantially to the accurate prediction of severity. The results are interpreted and discussed with respect to the key acoustic characteristics that contributed to the prediction of severity, the value of identifying a subset of time and spectral‐based acoustic measures which appear sensitive to a perceptually diverse set of voices, and the possible use of acoustic models in guiding auditory‐perceptual ratings.

Estimating dysphonia severity in continuous speech: Application of a multi-parameter spectral/cepstral model

The purpose of the study was to identify a sub-set of spectral/cepstral-based analysis methods that would most effectively predict dysphonia severity (as estimated via auditory-perceptual analysis) in samples of continuous speech. Acoustic estimates of dysphonia severity were used as an objective treatment outcomes measure in a set of pre- vs post-treatment speech samples. Pre- and post-treatment continuous speech samples from 104 females with primary muscle tension dysphonia (MTD) were rated by listeners using a 100 point visual analogue scale (VAS) and analysed acoustically with spectral/cepstral-based measures. Stepwise linear regression produced a three-factor model consisting of the cepstral peak prominence (CPP); the mean ratio of low-to-high frequency spectral energy; and the standard deviation of the ratio of low-to-high frequency spectral energy that was strongly correlated with perceived dysphonia severity ratings (R = .85; R2 = .73). Mean differences between predicted vs perceptual ratings for pre- and post-treatment speech samples were < 6 points on the 100 point VAS; mean absolute differences between predicted and perceived ratings were < 16 points on the 100 point VAS (equivalent to within one scale value on commonly used 7-point equal-appearing interval rating scales). A multi-parameter acoustic model consisting of spectral/cepstral-based measures shows considerable promise as an objective measure of dysphonia severity in continuous speech, even across the diverse voice types and severities observed in pre- and post-treatment MTD speech samples.

The aging female voice: acoustic and respiratory data.

The purpose of this study was to extend understanding of the effects of aging on the female voice by obtaining measures of both acoustic and respiratory‐based performance in groups of 18–30, 40–49, 50–59, 60–69, and 70–79‐year‐old subjects. Acoustic measures of speaking fundamental frequency (SFF), pitch sigma, jitter, shimmer, and signal‐to‐noise ratio, as well as respiratory‐based measures of vital capacity (VC), maximum phonation time (MPT), and phonation quotient (PQ) were obtained. Results indicated that the aging groups differed significantly in terms of SFF, pitch sigma, MPT, and VC. In addition, discriminant function analysis was used to classify subjects into age group via a three‐variable model consisting of VC, SFF, and pitch sigma (84% accuracy), and into pre‐ vs. post‐menopausal status via a two‐variable model consisting of VC and pitch sigma (92% accuracy). It appears that declinations in the respiratory and laryngeal mechanisms may occur simultaneously in the aging female.

The effect of anchor modality on the reliability of vocal severity ratings.

The purpose of this study was (1) to confirm if anchors (ie, perceptual references) and training affect the inter- and intrarater reliability of perceptual analysis of various voice types and severities compared to receiving training alone, and (2) to determine whether the modality in which the anchor is presented affects rater reliability. In this study, modality refers to whether the anchor is presented auditorily, visually via a written definition (a textual anchor), or a combination of both anchor types. A randomized multigroup comparison was performed. Forty inexperienced judges were selected to rate 36 sustained vowel voice samples of various voice types (ie, normal, breathy, hoarse, and rough) in terms of perceived vocal severity using four different methods (No Anchor, Textual Anchor, Auditory Anchor, and Combined Textual/Auditory Anchors). Subjects were randomly assigned to one of the four conditions. Before the rating task, all subject groups received a brief training sessions (15-20 minutes in duration) in which voice quality type and severity definitions were provided and representative voice samples were listened to. A computer program was developed to present anchors in the form of an auditory sample, written definition, or both. A no anchor condition was also presented. Results indicated that the combination of training and anchors significantly improves the interrater reliability of perceptual voice ratings. In addition, the use of auditory anchors resulted in 95% confidence intervals that were significantly smaller for rating mild voice disorders, and both breathy and hoarse voice qualities. Textual anchors did appear to show some improvement over training alone (ie, no anchors), but were generally not as strong as the use of auditory anchors. However, the combination of textual and auditory anchors resulted in the greatest degree of interrater reliability as assessed via mean correlations. The ratings produced by the Auditory and Combined Anchor groups were also observed to have significantly smaller 95% confidence intervals for the rating of mildly dysphonic voice than the No Anchor or Textual Anchor groups, indicating greater consistency and increased precision of ratings for this level of severity. Anchors were not only useful in improving measures of reliability among judges, but were also able to do so without significantly increasing the amount of time that the judges had to spend on the rating tasks. No significant effect of anchors was observed for measures of intrarater reliability. The use of perceptual anchors (in particular, auditory anchors), in conjunction with training and computerized rating procedures, may provide the consistent methodology necessary to improve the reliability of perceptual judgments. Although textual anchors were not as effective as auditory anchors, the combination of auditory and textual information may result in overall improvements in interrater reliability.

Toward Validation of the Cepstral Spectral Index of Dysphonia (CSID) as an Objective Treatment Outcomes Measure

OBJECTIVES To examine the validity of the Cepstral Spectral Index of Dysphonia (CSID) as an objective treatment outcomes measure of dysphonia severity. METHOD Pre- and posttreatment samples of sustained vowel and connected speech productions were elicited from 112 patients across six diagnostic categories: unilateral vocal fold paralysis, adductor spasmodic dysphonia, primary muscle tension dysphonia, benign vocal fold lesions, presbylaryngis, and mutational falsetto. Listener ratings of severity in connected speech were compared with a three-factor CSID model consisting of the cepstral peak prominence (CPP), the low-to-high spectral energy ratio, and its standard deviation. Two additional variables, the CPP standard deviation and gender, were included in the five-factor CSID model to estimate severity of vowels. RESULTS CSID-estimated severity for sustained vowels and connected speech was strongly associated with listener ratings pretreatment, posttreatment, and change observed pre- to posttreatment. Spectrum effects were examined, and severity of dysphonia did not influence the relationship between listener perceived severity and CSID-estimated severity. CONCLUSIONS The results confirm a robust relationship between listener perceived and CSID-derived dysphonia severity estimates in sustained vowels and connected speech across diverse diagnoses and severity levels and support the clinical utility of the CSID as an objective treatment outcomes measure.

Nasalance distance and ratio: two new measures.

OBJECTIVES Mean nasalance in speakers with perceptually normal nasal resonance can differ in magnitude considerably. In addition, categorizations of speech based on nasalance scores may not agree with perceptual judgments. To overcome this limitation, we evaluated two new simple measures derived from mean nasalance data: the nasalance distance (range between maximum and minimum nasalance) and the nasalance ratio (minimum nasalance divided by maximum nasalance). SETTING Department of Oral and Maxillofacial Surgery, University of Technology, Munich, Germany. SUBJECTS The sample consisted of 133 cleft lip and palate patients with normal nasal resonance or varying degrees of hypernasality. PROCEDURES Oral and nasal acoustic measurements were made using the NasalView system. Nasalance distance and nasalance ratio were calculated for five non-nasal and three nasal sentences from the modified Heidelberg Rhinophonia Assessment Form. RESULTS Optimum cutoffs were derived from receiver-operating characteristics. Results for the sentence stimuli ranged from 64.4% to 89.6% sensitivity and from 91.2% to 94.1% specificity. When the analysis was limited to only one nonnasal and one nasal sentence, results ranged from 79.7% to 87.5% sensitivity and from 88.2% to 97.1% specificity. CONCLUSIONS We conclude that the two new measurements are valuable in routine clinical examinations. Nasalance distance and ratio derived from sentence stimuli are two useful and easily applicable measures that can be used to supplement the nasalance mean value.

Exploring the Relationship Between Spectral and Cepstral Measures of Voice and the Voice Handicap Index (VHI)

OBJECTIVES The purpose of this study was to examine the strength of relationship between impairment-level acoustic measures derived from spectral- and cepstral-based analyses (including the cepstral peak prominence [CPP]; ratios of low vs high frequency spectral energy; and the respective standard deviations [SDs] for these measures) and a disablement measure (the total Voice Handicap Index [VHI] score) in a large and diverse group of voice-disordered and control subjects. The relationship between total VHI and the Cepstral Spectral Index of Dysphonia (CSID-a multivariate estimate of dysphonia severity) was also examined. METHODS Subjects were 332 adults (116 males and 216 females) comprised of voice-disordered subjects who presented to a physician with a voice-related complaint (n=258) and a group of nonvoice-disordered control subjects (n=74). A VHI 30-item score and speech/voice samples including the second and third sentences of The Rainbow Passage and productions of the sustained vowel /ɑ/ were obtained for each subject. Sentence and sustained vowel samples were analyzed using the Analysis of Dysphonia in Speech and Voice (ADSV) program (ADSV model 5109 v.3.4.2; KayPENTAX, Montvale, NJ). RESULTS Across all subjects, low-to-moderate strength Spearman rho (rs) correlations were observed between the total VHI and the CPP and the CSID in both speech and vowel contexts and for the CPP SD from continuous speech (rss ranging from -0.45 to -0.49 for VHI vs CPP; 0.47 for VHI vs CSID; -0.44 for VHI vs CPP SD). Several other measures obtained from spectral or cepstral analyses also were observed to correlate with total VHI, although increased variability in the strength, direction, and overall significance of these other variables was observed depending on gender and elicited context. CONCLUSIONS Voice-related disablement occurs within a context. In contrast, impairment-level measures of phonatory function (like the spectral and cepstral measures included in this study) are by nature decontextualized and appear to correlate low-to-moderately with quality of life measures like the VHI. Therefore, spectral and cepstral acoustic measures and the VHI should be viewed as providing relatively unique, meaningful, and complementary information.

Tracking voice change after thyroidectomy: application of spectral/cepstral analyses

This study evaluates the utility of perioperative spectral and cepstral acoustic analyses to monitor voice change after thyroidectomy. Perceptual and acoustic analyses were conducted on speech samples (sustained vowel /ɑ/ and CAPE-V sentences) provided by 70 participants (36 women and 34 men) at four study time points: prior to thyroid surgery and 2 weeks, 3 months and 6 months after thyroidectomy. Repeated measures analyses of variance focused on the relative amplitude of the dominant harmonic in the voice signal (cepstral peak prominence, CPP), the ratio of low-to-high spectral energy, and their respective standard deviations (SD). Data were also examined for relationships between acoustic measures and perceptual ratings of overall severity of voice quality. Results showed that perceived overall severity and the acoustic measures of the CPP and its SD (CPPsd) computed from sentence productions were significantly reduced at 2-week post-thyroidectomy for 20 patients (29% of the sample) who had self-reported post-operative voice change. For this same group of patients, the CPP and CPPsd computed from sentence productions improved significantly from 2-weeks post-thyroidectomy to 6-months post-surgery. CPP and CPPsd also correlated well with perceived overall severity (r = −0.68 and −0.79, respectively). Measures of CPP from sustained vowel productions were not as effective as those from sentence productions in reflecting voice deterioration in the post-thyroidectomy patients at the 2-week post-surgery time period, were weaker correlates with perceived overall severity, and were not as effective in discriminating negative voice outcome (NegVO) from normal voice outcome (NormVO) patients as compared to the results from the sentence-level stimuli. Results indicate that spectral/cepstral analysis methods can be used with continuous speech samples to provide important objective data to document the effects of dysphonia in a post-thyroidectomy patient sample. When used in conjunction with patients self-report and other general measures of vocal dysfunction, the acoustic measures employed in this study contribute to a complete profile of the patients vocal condition.

A comparison of trained and untrained vocalists on the Dysphonia Severity Index.

The purposes of this study were (1) to compare trained and untrained singers on the Dysphonia Severity Index (DSI) and its component measures, and (2) to contribute to normative DSI data for trained singers. This study included 36 untrained participants (15 males and 21 females) and 30 participants (15 males and 15 females) with singing experience between the ages of 18 and 30 years. Measures of maximum phonation time (MPT), highest phonational frequency, lowest intensity, and jitter were obtained for each subject and incorporated into the previously published multivariate DSI formula. Results indicated that vocally trained subjects have significantly higher DSI scores than untrained subjects (mean DSI: 6.48 vs 4.00, respectively), with significant differences observed between trained and untrained groups for three of the four components of the DSI (F(0) high; I low; jitter). The findings of this study are consistent with previous reports that indicate significant increases in the DSI with vocal training, and with various studies that have observed increased vocal capability in trained singers versus their untrained counterparts. The results of this study indicate that alternative normative expectations for the DSI may need to be taken into account when using the DSI with patients who have participated in directed vocal training, such as choral participation and voice/singing lessons.

Test-Retest Reliability for Aerodynamic Measures of Voice

OBJECTIVES/HYPOTHESIS The purpose of this study was to investigate the intrasubject reliability of aerodynamic characteristics of the voice within typical/normal speakers across testing sessions using the Phonatory Aerodynamic System (PAS 6600; KayPENTAX, Montvale, NJ). METHODS Participants were 60 healthy young adults (30 males and 30 females) between the ages 18 and 31 years with perceptually typical voice. Participants were tested using the PAS 6600 (Phonatory Aerodynamic System) on two separate days with approximately 1 week between each session at approximately the same time of day. Four PAS protocols were conducted (vital capacity, maximum sustained phonation, comfortable sustained phonation, and voicing efficiency) and measures of expiratory volume, maximum phonation time, mean expiratory airflow (during vowel production) and target airflow (obtained via syllable repetition), peak air pressure, aerodynamic power, aerodynamic resistance, and aerodynamic efficiency were obtained during each testing session. Associated acoustic measures of vocal intensity and frequency were also collected. All phonations were elicited at comfortable pitch and loudness. RESULTS All aerodynamic and associated variables evaluated in this study showed useable test-retest reliability (ie, intraclass correlation coefficients [ICCs] ≥ 0.60). A high degree of mean test-retest reliability was found across all subjects for aerodynamic and associated acoustic measurements of vital capacity, maximum sustained phonation, glottal resistance, and vocal intensity (all with ICCs > 0.75). Although strong ICCs were observed for measures of glottal power and mean expiratory airflow in males, weaker overall results for these measures (ICC range: 0.60-0.67) were observed in females subjects and sizable coefficients of variation were observed for measures of power, resistance, and efficiency in both men and women. Differences in degree of reliability from measure to measure were revealed in greater detail using methods such as ICCs and coefficients of variation than with means comparison testing. The results of this study also show that reliable aerodynamic and associated measures may be elicited using comfortable pitch and loudness. CONCLUSIONS The 1-week test-retest reliability for the majority of aerodynamic and associated acoustic measures assessed in this study is considered good-to-excellent. Clinicians and researchers using aerodynamic and associated measures should be aware of possible significant gender effects that influence both normative expectations as well as the standard error of measurement (ie, typical error) and estimates of minimum difference that may be used to differentiate typical from disordered voice.