Assessment Of Voice Disorders Research Articles

Assessment of Voice Disorders Using Machine Learning and Vocal Analysis of Voice Samples Recorded through Smartphones

Background: The integration of edge computing into smart healthcare systems requires the development of computationally efficient models and methodologies for monitoring and detecting patients’ healthcare statuses. In this context, mobile devices, such as smartphones, are increasingly employed for the purpose of aiding diagnosis, treatment, and monitoring. Notably, smartphones are widely pervasive and readily accessible to a significant portion of the population. These devices empower individuals to conveniently record and submit voice samples, thereby potentially facilitating the early detection of vocal irregularities or changes. This research focuses on the creation of diverse machine learning frameworks based on vocal samples captured by smartphones to distinguish between pathological and healthy voices. Methods: The investigation leverages the publicly available VOICED dataset, comprising 58 healthy voice samples and 150 samples from voices exhibiting pathological conditions, and machine learning techniques for the classification of healthy and diseased patients through the employment of Mel-frequency cepstral coefficients. Results: Through cross-validated two-class classification, the fine k-nearest neighbor exhibited the highest performance, achieving an accuracy rate of 98.3% in identifying healthy and pathological voices. Conclusions: This study holds promise for enabling smartphones to effectively identify vocal disorders, offering a multitude of advantages for both individuals and healthcare systems, encompassing heightened accessibility, early detection, and continuous monitoring.

Automated Acoustic Evaluation of Voice Disorders: A Comprehensive Study on Parameter Analysis Using ANN

Voice analysis serves as a crucial tool in diagnosing voice abnormalities, offering a non-invasive alternative to intrusive procedures. This research digs into a comprehensive study of voice disorder assessment methodologies, focusing on acoustic analysis and classification. The study employs various parameters such as Jitter, Shimmer, and Harmonic-to-Noise Ratio (HNR) alongside an Artificial Neural Network (ANN) classifier. Utilizing the Saarbruecken Voice Database the research aims to distinguish between healthy and dysphonic voices across genders. Principal Component Analysis (PCA) aids in feature selection, enhancing model accuracy. The results exhibit distinct precision levels in male and female groups, showcasing the effectiveness of specific parameters in classification.

Vocal Tract Acoustic Measurements for Detection of Pathological Voice Disorders

Human voice is an important signal that, if analyzed cautiously, will reveal various disorders and diseases of the human body. Vocal profiling is emerging as a new technique for assessment of voice disorders. This work has focused mainly on vocal tract acoustic measurements to detect voice disorders, especially dysphonia. Dysphonia is a communication disorder that, if not detected at an early stage, may lead to serious complications like laryngeal cancer thus affecting health and quality of life. The voice recordings of 52 subjects (26 dysphonic and 26 healthy) were taken from Saarbrucken Voice Database, 28 live samples (14 dysphonic and 14 healthy) were recorded using CSL 4500 tool and 169 subjects (111 dysphonic and 58 healthy) were taken from VOICED Database. These voice samples were used to extract five acoustic parameters: fundamental frequency, jitter, shimmer, pitch, and formants. The results obtained had an accuracy of around 85% which confirms the potentiality of these fundamental parameters in the successful detection of dysphonia as well as other voice disorders.

Clinical Utility and Validation of the Acoustic Voice Quality and Acoustic Breathiness Indexes for Voice Disorder Assessment in English Speakers.

While several acoustic voice metrics are available for clinical voice assessment, there remains a significant need for reliable and ecologically valid tools. The Acoustic Voice Quality Index version 03.01 (AVQI-3) and Acoustic Breathiness Index (ABI) hold potential due to their comprehensive assessment approach, incorporating diverse voice aspects. However, these tools still need to be validated in English-speaking populations. This study assessed the discriminatory accuracy and validity of AVQI-3 and ABI in 197 participants, including 148 with voice disorders. Voice samples were collected, followed by AVQI-3 and ABI calculations. Additionally, auditory-perceptual assessments were conducted by a panel of speech-language pathologists. AVQI-3 and ABI effectively identified disordered voice quality, evidenced by high accuracy (AUCs: 0.84, 0.89), sensitivity, and specificity (thresholds: AVQI-3 = 1.17, ABI = 2.35). Strong positive correlations were observed with subjective voice quality assessments (rs = 0.72, rs = 0.77, p < 0.001). The study highlights AVQI-3 and ABI as promising instruments for clinically assessing voice disorders in U.S. English speakers, underscoring their utility in clinical practice and voice research.

Nyquist Plot Parametrization for Quantitative Analysis of Vibration of the Vocal Folds

ObjectivesThe Nyquist plot provides a graphical representation of the glottal cycles as elliptical trajectories in a 2D plane. This study proposes a methodology to parameterize the Nyquist plot with application to support the quantitative analysis of voice disorders. MethodsWe considered high-speed videoendoscopy recordings of 33 functional dysphonia (FD) patients and 33 normophonic controls (NC). Quantitative analysis was performed by computing four shape-based parameters from the Nyquist plot: Variability, Size (Perimeter and Area), and Consistency. Additionally, we performed automatic classification using a linear support vector machine and feature importance analysis by combining the proposed features with state-of-the-art glottal area waveform (GAW) parameters. ResultsWe found that the inter-cycle variability was significantly higher in FD patients compared to NC. We achieved a classification accuracy of 83% when the top 30 most important features were used. Furthermore, the proposed Nyquist plot features were ranked in the top 12 most important features. ConclusionsThe Nyquist plot provides complementary information for subjective and objective assessment of voice disorders. On the one hand, with visual inspection it is possible to observe intra- and inter-glottal cycle irregularities during sustained phonation. On the other hand, shaped-based parameters allow quantifying such irregularities and provide complementary information to state-of-the-art GAW parameters.

A Prospective Observational Study to Determine the Added Clinical Value of Videokymography to Videostroboscopy in Patients with Change in Voice.

Videolaryngostroboscopy (VLS) is considered gold standard method for assessing voice disorders. But patients with irregular waveform of vocal folds cannot benefit from the VLS. Videokymography [VKG] is a single line real time, high speed imaging technique. It detects voice disorders based on vocal fold vibration characteristics whether the vibrations are regular or irregular. There is no standard clinical protocol or evidence on the clinical relevance of VKG for functional assessment of voice disorders. Since mechanism of voice production depends on vibration characteristics, VKG imaging leads to new possibilities for diagnosis, objective documentation and monitoring of vocal fold behavior in clinical practice in case of voice disorders. This study aims to evaluate clinical value of VKG in addition to VLS as a complementary tool for the assessment of voice disorder.

Similarities and differences across patient completed voice case history questionnaires – A scoping review

Case history assessments are ubiquitously performed across various health professions for diagnostic purposes. Questionnaires are considered a valuable tool within this process. There is currently no standardized tool available to collect such information in the assessment of voice disorders. Conflicting advice from peak bodies and research evidence makes the process heterogenous, haphazard and difficult to compare findings. To systematically identify existing case history questionnaires available for general clinical practice and provide synthesis and analysis of the questions within. A scoping review was conducted across published and unpublished literature using the PRISMA Extension for Scoping Reviews framework. The broad search included research papers within six electronic databases, textbooks, online publishing sites, voice clinic websites and peak body websites. Search criteria were uniformly applied. Descriptive analysis and content analysis was conducted. At each stage, transparency and replicability was achieved through an independent review process. Identified voice case history questionnaires were few (n=23) with 80% from unpublished sources. A total of 581 unique questions were identified. No single question was common across all 23 questionnaires. The most frequently asked questions, excluding demographics, included medicines taken (n=21), smoking (n=20) and alcohol (n=19). These questions were not reflected in the highest frequency categories: Health Status/Medical Conditions/Reports (n=200), Vocal Symptoms (n=88), Voice Use (n=51) despite these categories representing 58% of all questions asked. Within the highest frequency category, the subcategory of Systemic Diseases was the highest, representing 19% of all questions within the category. This study illuminates the similarities yet many differences that exist across identified voice case history questionnaires in terms of number of questions, number of categories, preference for question-type and structure. It demonstrates the need for standardization of a voice case history questionnaire which would potentially enable more accurate diagnosis and data comparison between voice clinics to aid future research.

Validation of the Acoustic Voice Quality Index in the Russian language

Of this study was to validate the Acoustic Voice Quality Index (AVQI) in Russian language and determine threshold values to differentiate normal and dysphonic voices in russophone population. 41 recordings of patients with laryngeal hemiplegia and hemiparesis appeared after thyroid carcinomas' surgical treatment and laryngeal cancer from the Tomsk Cancer Research Institute were used, as well as 10 new recordings of normophonic people, composing a control group. Voice samples included the first quatrain of the poem «Sail» written by M.Yu. Lermontov, and sustained vowel «a». Four specialists who work with voice disorders (three speech-language pathologists and one phoniatrist) rated degree of dysphonia for all voice sample. Moreover, the AVQI was calculated for each sample. There was a statistically significant correlation between auditory-perceptual assessment and the AVQI (r=0.74), this confirms that AVQI is effective in evaluating voice disorders. The group of people with normal voice had relatively low values of AVQI, whereas the group of patients with laryngeal diseases had relatively high values. Analysis of ROC-curve showed that AVQI has excellent diagnostic accuracy to differentiate normal and dysphonic voices (AUC=0.966). The AVQI is the excellent tool for objective assessment of voice disorders in the Russian population/.

Comparison of teachers' voice disorders before and during COVID-19 pandemic.

&lt;b&gt; Aim:&lt;/b&gt; The aim of the study was to analyze the impact of the COVID-19 pandemic and the related change of the teaching mode from stationary to distance learning on the severity of voice-related ailments among teachers. &lt;/br&gt;&lt;/br&gt; &lt;b&gt; Materials and methods:&lt;/b&gt; A questionnaire survey of teachers was conducted to assess voice disorders during stationary and remote work using the Vocal Tract Discomfort (VTDs) scale and Numeric Rating Scale (NRS), and the respondents' subjective feelings were assessed. Demographic and environmental factors associated with voice work were examined. Data on sickness absence, which were obtained from the "Health Needs Maps 2020" Ministry of Health's, were also analyzed. Responses were subjected to statistical analysis. A P-value level below 0.05 was considered statistically significant. &lt;/br&gt;&lt;/br&gt; &lt;b&gt;Results:&lt;/b&gt; 128 teachers participated in the survey. The overall assessment of voice disorders using VTDs and NRS scales showed no statistically significant differences for complaints between stationary and remote work. Detailed analysis revealed more se-vere voice disorders in teachers working more than 6 months remotely (P = 0.049) and having more than 20 lessons per week (P = 0.012). Subjective assessment confirmed a significantly lower percentage of teachers reporting voice disorders when wor-king remotely compared to stationary (P = 0.043). This resulted in less sickness absence and a 40% decrease in sick leave related to voice disorders in 2020 compared to 2019. &lt;/br&gt;&lt;/br&gt; &lt;b&gt;Conclusions:&lt;/b&gt; During the remote learning period of the COVID-19 pandemic, teachers reported lower severity of voice disor-ders and this influenced the reduction of sickness absences. There were no statistically significant differences in voice complaints as assessed by VTDs and NRS scales for both teaching modes. Several factors affecting the severity of vocal tract disorders were identified - the number of class hours per week (&gt;20) for stationary teaching and a long period of remote teaching (&gt;6 months).

The Normative Study of Acoustic and Aerodynamic Characteristics of Voice among Healthy Adult Turkish Speaker Population.

Phonatory Aerodynamic System (PAS Model 6600) is an evaluation instrument that assesses the effectiveness of surgical interventions, treatments, and therapy for voice disorders. It can be used for the assessment of voice disorders by supporting other perceptual and instrumental methods. It is important to establish normative data, because the use of appropriate norms is necessary for diagnostic and descriptive accuracy. Therefore, this study is aimed primarily at establishing adult normative databases for phonatory aerodynamic measures obtained with the KayPENTAX PAS Model 6600 among healthy adult Turkish speakers and then examining the effect of age, gender, and age-gender interaction variables on these measures. The contribution of the study is considered so important since it will generate normative data for all measurements—except the mean pitch—by the five protocols of PAS for the first time. Two hundred and six healthy Turkish speakers with normal voice (106 women and 100 men) were included in the study and stratified into three age groups. Forty-five phonatory aerodynamic measures across five PAS protocols (vital capacity, maximum sustained phonation, comfortable sustained phonation, variation in sound pressure level, and voicing efficiency) were collected. Age, gender, and age-gender interaction variables were analyzed for 45 PAS parameters. Significant gender and age effect was found for 30 and 19 variables, respectively. Gender-age interaction together was observed for only 6 parameters. Significant differences were not found for the remaining 10 parameters. Significant age and gender effects were observed for 35 phonatory and aerodynamic measures which are essential part of the objective clinical assessment of voice. Consequently, normative data used as reference in voice assessment should be generated according to age and gender differences.

The Effect of Visual Sort and Rate Versus Visual Analog Scales on the Reliability of Judgments of Dysphonia.

Purpose The reliability of auditory-perceptual judgments between listeners is a long-standing problem in the assessment of voice disorders. The purpose of this study was to determine whether a relatively novel experimental scaling method, called visual sort and rate (VSR), yielded stronger reliability than the more frequently used method of visual analog scales (VAS) for ratings of overall severity (OS) and breathiness (BR) in speakers with voicedisorders. Method Fifty speech samples were selected from a database of speakers with voice disorders. Twenty-two inexperienced listeners provided ratings of OS or BR in four rating blocks: VSR-OS, VSR-BR, VAS-OS, and VSR-BR. For the VAS task, listeners rated each speaker for BR or OS using a vertically oriented 100-mm VAS. For the VSR task, stimuli were distributed into sets of samples with a range of speaker severities in each set. Listeners sorted and ranked samples for OS or BR within each set, and final ratings were captured on a vertically oriented 100-mm VAS. Interrater variability, defined as the mean of the squared differences between a listener's ratings and group mean ratings, and intrarater reliability (Pearson r) were compared across rating tasks for OS and BR using paired t tests. Results Results showed that listeners had significantly less interrater variability (better reliability) when using VSR methods compared to VAS for judgments of both OS and BR. Intrarater reliability was high across rating tasks and dimensions; however, ratings of BR were significantly more consistent within individual listeners when using VAS than when using VSR. Conclusions VSR is an experimental method that decreases variability of auditory-perceptual judgments between inexperienced listeners when rating speakers with a range of dysphonic severities and disorders. Future research should determine whether a clinically viable tool may be developed based on VSR principles and whether such benefits extend to experienced listeners.

Effect of Training With Anchors on Auditory-Perceptual Evaluation of Dysphonia in Speech-Language Pathology Students.

Purpose Auditory-perceptual evaluation is essential for the clinical assessment of voice disorders. Unstable perceptual voice evaluation has been shown for inexperienced listeners as compared to expert listeners. We examined the effects of perceptual training with external auditory anchors with and without immediate feedback on the evaluation of roughness and breathiness of natural, nonsynthesized speech stimuli (reading of a standard passage) in speech-language pathology students. Method Perceptual voice evaluation and training with anchors using a visual analog scale was implemented with a computer software. Forty-eight speech-language pathology students were randomly assigned into three groups, feedback group (Group F), no feedback group (Group NF), and control group (Group C), attending one training session and four assessment sessions (before training, immediately after training, and 1 and 7 weeks after training). Group F received training with anchors with immediate feedback, Group NF received training without immediate feedback, and Group C received sham training (exposure session). Results Training with anchors significantly increased the rating accuracy (agreement with expert ratings) on both roughness and breathiness for Group F, with the effects lasting for 7 weeks. No significant changes in rating accuracy with training were observed for Group NF and Group C. No improvements in intra- and interrater reliability as well as intrarater agreement were observed in all three groups, whereas interrater agreement on breathiness (but not roughness) significantly increased for all groups, with the effect lasting for 7 weeks only for Group F. Conclusions These findings suggested that perceptual training with external auditory anchors and the use of immediate feedback could be effective for facilitating the development of perceptual voice evaluation skills in speech-language pathology students. Further studies involving more extensive training with stimuli covering a full range of dysphonia severity categories and improvements in design of the training protocol are recommended to verify these results.

Non-Linear Image Distortions in Flexible Fiberoptic Endoscopes and their Effects on Calibrated Horizontal Measurements Using High-Speed Videoendoscopy

Laryngeal images obtained via high-speed videoendoscopy are an invaluable source of information for the advancement of voice science because they can capture the true cycle-to-cycle vibratory characteristics of the vocal folds in addition to the transient behaviors of the phonatory mechanism, such as onset, offset, and breaks. This information is obtained through relating the spatial and temporal features from acquired images using objective measurements or subjective assessments. While these images are calibrated temporally, a great challenge is the lack of spatial calibration. Recently, a laser-projection system allowing for spatial calibration was developed. However, various sources of optical distortions deviate the images from reflecting the reality. The main purpose of this study was to evaluate the effect of the fiberoptic flexible endoscope distortions on the calibration of images acquired by the laser-projection system. Specifically, it is shown that two sources of nonlinear distortions could deviate captured images from reality. The first distortion stems from the wide-angle lens used in flexible endoscopes. It is shown that endoscopic images have a significantly higher spatial resolution in the center of the field of view than in its periphery. The difference between the two could lead to as high as 26.4% error in calibrated horizontal measurements. The second distortion stems from variation in the imaging angle. It is shown that the disparity between spatial resolution in the center and periphery of endoscopic images increases as the imaging angle deviates from the perpendicular position. Furthermore, it is shown that when the imaging angle varies, the symmetry of the distortion is also affected significantly. The combined distortions could lead to calibrated horizontal measurement errors as high as 65.7%. The implications of the findings on objective measurements and subjective visual assessments are discussed. These findings can contribute to the refinement of the methods for clinical assessment of voice disorders. Considering that the studied phenomena are due to optical principles, the findings of this study, especially those related to the effects of the imaging angle, can provide further insights regarding other endoscopic instruments (eg, distal-chip and rigid endoscopes) and procedures (eg, gastroendoscopy and colonoscopy).

Microstimulation in Different Parts of the Periaqueductal Gray Generates Different Types of Vocalizations in the Cat

In the cat four different types of vocalization, mews, howls, cries, and hisses were generated by microstimulation in different parts of the periaqueductal gray (PAG). While mews imply positive vocal expressions, howls, hisses, and cries represent negative vocal expressions. In the intermediate PAG, mews were generated in the lateral column, howls, and hisses in the ventrolateral column. Cries were generated in two other regions, the lateral column of the rostral PAG and the ventrolateral column of the caudal PAG. In order to define the specific motor patterns of the mews, howls, and cries, the following muscles were recorded during these vocalizations; larynx (cricothyroid, thyroarytenoid, and posterior cricoarytenoid), tongue (genioglossus), jaw (digastric), and respiration muscles (diaphragm, internal intercostal, external, and internal abdominal oblique). During these mews, howls, and cries we analyzed the frequency, intensity, activation cascades power density, turns, and amplitude analysis of the electromyograms (EMGs). It appeared that each type of vocalization consists of a specific circumscribed motor coordination. The nucleus retroambiguus (NRA) in the caudal medulla is known to serve as the final premotor interneuronal output system for vocalization. Although neurochemical microstimulation in the NRA itself also generated vocalizations, they only consisted of guttural sounds, the EMGs of which involved only small parts of the EMGs of the mews, howls, and cries generated by neurochemical stimulation in the PAG. These results demonstrate that positive and negative vocalizations are generated in different parts of the PAG. These parts have access to different groups of premotoneurons in the NRA, that, in turn, have access to different groups of motoneurons in the brainstem and spinal cord, resulting in different vocalizations. The findings would serve a valuable model for diagnostic assessment of voice disorders in humans.

Vocal Acoustic Analysis

Vocal acoustic analysis is becoming a useful tool for the classification and recognition of laryngological pathologies. This technique enables a non-invasive and low-cost assessment of voice disorders, allowing a more efficient, fast, and objective diagnosis. In this work, ANN and SVM were experimented on to classify between dysphonic/control and vocal cord paralysis/control. A vector was made up of 4 jitter parameters, 4 shimmer parameters, and a harmonic to noise ratio (HNR), determined from 3 different vowels at 3 different tones, with a total of 81 features. Variable selection and dimension reduction techniques such as hierarchical clustering, multilinear regression analysis and principal component analysis (PCA) was applied. The classification between dysphonic and control was made with an accuracy of 100% for female and male groups with ANN and SVM. For the classification between vocal cords paralysis and control an accuracy of 78,9% was achieved for female group with SVM, and 81,8% for the male group with ANN.

Predictive Value of Auditory Perceptual Assessment of Voice in Diagnosis of Functional Voice Disorders and Minimal Associated Pathological Lesions

Background: Auditory perceptual assessment is considered the preliminary assessment tool of disordered voice quality. The validity of perceptual measurement depends on characteristics of the raters and the stimuli. It is associated with minimal cost and no specialized instrumentation is required. Because auditory-perceptual evaluation of dysphonia has been heavily criticized on the basis of its reliability and validity, we conducted this study to assess the reliability of auditory perceptual assessment as one of the objective measures for assessment of voice disorders and to determine the predictive relationships between instrumental measures as endoscopic examination and perceptual ratings of impaired voice quality. Subjects and Methods: The study included 50 patients complaining of dysphonia with age ranges from 5 -61 years. Auditory perceptual assessment (APA) of patients’ voices was done and voice recordings were assessed by three experienced phoniatricians in a double blind, randomized fashion. The examiners (Ex1.2.3) assessed the recorded voice quality by using the modified GRBAS scale. Laryngoscopic examination was performed using indirect laryngoscope, flexible fibrotic laryngoscopy as well as rigid laryngoscope. Results: Results showed considerable similarities among raters in assessing the disordered voice quality especially on assessing the overall grade of dysphonia and the “Strain” voice quality with strong inter-rater reliability (Cronbach's Alpha) was 0.780 and 0.710 respectively. However, data showed some differences with poor inter-rater reliability especially for ratings of glottal attack and associated laryngeal function as cough (Cronbach's Alpha) was 0.383 and 0.531 respectively. The study also showed significant relations on comparing between instrumental (video endoscopic) diagnosis and perceptual predictive diagnosis of patients’ disordered voice quality according to two of the raters with P <0.03&0.04. Conclusion: This study showed considerable variability of the inter-rater reliability across parameters of disordered voice quality with high inter-rater reliability for some parameters and low for others. Yet, there was a strong association between instrumental diagnostic measures and the perceptual predictive diagnosis denoting a clear relationship between listener’s judgments of disordered voice quality and the underlying vocal pathology; a thing that is in favor for the validity of APA as a popular and fundamental tool for the assessment of vocal impairment.

Mean fundamental frequency in connected speech and sustained vowel with and without a sentence-frame

Introduction: Sustained vowel phonation is a common patient task in the assessment of voice disorders. Free from articulatory influence, it allows for acoustic voice analyses unfeasible for connected speech. Previous research indicates that the fundamental frequency (fo) of a self-chosen vowel is not necessarily representative to mean fo of the same subject’s connected speech.Objective: The purpose was to examine mean fo in text reading and to compare it with sustained vowel phonation with and without a sentence-frame. It was hypothesized, that a short phrase of context embedding the prolonged vowel would affect mean fo of the vowel to be closer to that of text reading.Method: Thirteen vocally healthy women, age 22–27 was audio recorded during three tasks: reading of a standard text, producing an isolated, sustained /a/ and repeating three times a frame sentence with an imbedded /a/ and the third time to sustain the /a/. Mean fo of the three conditions was analyzed by means of the Praat computer program.Results: The isolated vowel condition differed significantly from the mean fo of the sentence-framed one, the latter being closest to mean fo of the reading task. The isolated vowel condition also differed significantly from text reading. The difference in mean fo between sentence-framed vowel and text reading was not significant.Conclusion: Data support that the use of a sentence-frame is helpful to get a client’s fo elicited from sustained vowel production in voice assessments more representative to that of text reading as compared to an isolated vowel.

Acoustical Assessment of Voice Disorder With Continuous Speech Using ASR Posterior Features

Traditionally acoustical assessment of voice disorder relies on simple and homogeneous speech samples like sustained vowels. Continuous speech is believed to be more representative of the daily function of voice and more preferable in clinical practice. This paper describes an attempt on automating voice assessment with continuous speech utterances. The proposed system makes use of a novel type of features that are derived from phone posterior probabilities outputted by a deep neural network based automatic speech recognition ASR system. These ASR-based voice features are designed to effectively quantify the mismatch between disordered voice and normal voice. Prediction of voice disorder severity is carried out first at utterance-level and subsequently the prediction scores for individual utterances from a subject are combined to give an overallassessment on the subject. With a low-dimension ASR-based feature vector, the utterance-level prediction accuracy is comparable to that with conventional features with a much higher dimension. By jointly using the ASR features and conventional voice features, a subject-level prediction accuracy of over $\text{80}{\%}$ on three severity classes can be achieved. Subjects with mild disorder and those with severe disorder could be perfectly distinguished by the proposed method.

Voice assessment practices of speech and language therapists in Ireland

ABSTRACT Evidence supports the need for a multiparametric voice assessment incorporating objective and subjective assessment types. European guidelines and the American Speech-Language-Hearing Association recommend a comprehensive voice assessment protocol; however, currently in Ireland, no national standards exist. This study investigates voice assessment practices of speech and language therapists (SLTs) in Ireland, with a particular interest in the use of objective instrumentation. It further elucidates what may act as barriers to the use of instrumental techniques, which has not been addressed in earlier studies. An online questionnaire was distributed to SLTs who work with voice disordered clients, via social media outlets of the Irish Association of Speech and Language Therapy (IASLT) and the Voice Special Interest Group to obtain both qualitative and quantitative data. Forty-five questionnaires were returned; the results of 33 completed questionnaires are presented here. The results suggest that subjective measures (auditory assessment protocols) are more commonly used by SLTs in Ireland than instrumental techniques. Limited access to equipment, the cost of equipment and low prioritization of voice clients in a large caseload were most frequently named as barriers to the use of instrumental assessment. The SLTs acknowledged the need for and expressed interest in more training on the use of instrumental techniques in the assessment of voice disorders. The results provoke discussion surrounding evidence-based practice in voice assessment and have implications for how instrumental techniques are incorporated in the curriculum of SLT training courses and in the continual professional development.

Emulating the perceptual capabilities of a human evaluator to map the GRB scale for the assessment of voice disorders

This paper presents the design of an automatic voice quality analysis system for the assessment of voice pathologies, which emulates the perceptual capabilities of a human evaluator according the GRB scale. For this purpose, a novel methodology based on multiple sets of characteristics, ordinal classification and Gaussian regression is proposed. In particular, a reduced subset of characteristics is identified, and the regressor is used to convert the discrete perceptual scale to a continuum, more in agreement to the nature of the problem under study. The robustness of the system is evaluated in several cross-dataset experiments. Similarly, a clinical evaluation of the predictions provided by the system is carried out. Results indicate that the proposed methodology is proficient in modelling the perceptual capabilities of the human evaluator. They also show that it is possible to extend the GRB scale to a continuum through regression techniques while maintaining the consistency of the results. On average, the deviation between the labels assessed by the expert and the ones provided by the system is of about 0.5 units (in a scale from 0 to 3) for G and B, and of 0.7 units for R. Similarly, the deviation of the labels predicted by the system in the clinical assessment trials is about 0.3 units for G, 0.4 units for B, and 0.5 units for R.