Residual Neural Network precisely quantifies dysarthria severity-level based on short-duration speech segments.

Recently, we have witnessed Deep Learning methodologies gaining significant attention for severity-based classification of dysarthric speech. Detecting dysarthria, quantifying its severity, are of paramount importance in various real-life applications, such as the assessment of patients' progression in treatments, which includes an adequate planning of their therapy and the improvement of speech-based interactive systems in order to handle pathologically-affected voices automatically. Notably, current speech-powered tools often deal with short-duration speech segments and, consequently, are less efficient in dealing with impaired speech, even by using Convolutional Neural Networks (CNNs). Thus, detecting dysarthria severity-level based on short speech segments might help in improving the performance and applicability of those systems. To achieve this goal, we propose a novel Residual Network (ResNet)-based technique which receives short-duration speech segments as input. Statistically meaningful objective analysis of our experiments, reported over standard Universal Access corpus, exhibits average values of 21.35% and 22.48% improvement, compared to the baseline CNN, in terms of classification accuracy and F1-score, respectively. For additional comparisons, tests with Gaussian Mixture Models and Light CNNs were also performed. Overall, the values of 98.90% and 98.00% for classification accuracy and F1-score, respectively, were obtained with the proposed ResNet approach, confirming its efficacy and reassuring its practical applicability.

A Standardized Protocol for Maximum Repetition Rate Assessment in Children.

BACKGROUND/AIMS: Maximum repetition rate (MRR) is often used in the assessment of speech motor performance in older children and adults. The present study aimed to evaluate a standardized protocol for MRR assessment in young children in Dutch. METHODS: The sample included 1,524 children of 2-7 years old with no hearing difficulties and Dutch spoken in their nursery or primary school and was representative for children in the Netherlands. The MRR protocol featured mono-, tri-, and bisyllabic sequences and was computer-implemented to maximize standardization. RESULTS: Less than 50% of the 2-year-olds could produce >1 monosyllabic sequence correctly. Children who could not correctly produce ≥2 monosyllabic sequences could not produce any of the multisyllabic sequences. The effect of instruction ("faster" and "as fast as possible") was small, and multiple attempts yielded a faster MRR in only 20% of the cases. MRRs did not show clinically relevant differences when calculated over different numbers of repeated syllables. CONCLUSIONS: The MRR protocol is suitable for children of 3 years and older. If children cannot produce at least 2 of the monosyllabic sequences, the multisyllabic tasks should be omitted. Furthermore, all fast attempts of each sequence should be analyzed to determine the fastest MRR.

Construct Validity of the Viking Speech Scale.

OBJECTIVE: The Viking Speech Scale (VSS) reliably classifies the speech performance of children with cerebral palsy. This paper aims to establish the construct validity of the VSS by testing the extent to which percentage intelligibility in single word speech and connected speech predicts VSS rating. PATIENTS AND METHODS: This is a secondary analysis of two sets of anonymised data collected for previous research. The full data set comprised 79 children with cerebral palsy from the US (n = 43) and the UK (n = 36): (43 boys, 36 girls); mean age 7.2 years (SD 3.3). Single word intelligibility was measured using the TOCS+ words for US children and Children's Speech Intelligibility Measure for the UK children. Connected speech intelligibility was measured from a subset of repeated sentences in TOCS+ for US children and picture description for the UK children. We used ordinal logistic regression to examine prediction of VSS rating by percentage single word and connected speech intelligibility scores in both samples. RESULTS: Percentage single word intelligibility and connected speech intelligibility predicted VSS rating in univariate and multivariate regression models for both the US and UK samples. CONCLUSION: Intelligibility predicts VSS for both single words and connected speech, establishing the construct validity of VSS.

Data-Driven Classification of Dysarthria Profiles in Children With Cerebral Palsy.

Purpose: The objectives of this study were to examine different speech profiles among children with dysarthria secondary to cerebral palsy (CP) and to characterize the effect of different speech profiles on intelligibility. Method: Twenty 5-year-old children with dysarthria secondary to CP and 20 typically developing children were included in this study. Six acoustic and perceptual speech measures were selected to quantify a range of segmental and suprasegmental speech characteristics and were measured from children's sentence productions. Hierarchical cluster analysis was used to identify naturally occurring subgroups of children who had similar profiles of speech features. Results: Results revealed 4 naturally occurring speech clusters among children: 1 cluster of children with typical development and 3 clusters of children with dysarthria secondary to CP. Two of the 3 dysarthria clusters had statistically equivalent intelligibility levels but significantly differed in articulation rate and degree of hypernasality. Conclusion: This study provides initial evidence that different speech profiles exist among 5-year-old children with dysarthria secondary to CP, even among children with similar intelligibility levels, suggesting the potential for developing a pediatric dysarthria classification system that could be used to stratify children with dysarthria into meaningful subgroups for studying speech motor development and efficacy of interventions.

[Myasthenia gravis should be considered in cases of Parkinson's disease and progressive dysphagia]. / Bei Morbus Parkinson und progredienter Dysphagie an Myasthenia gravis denken.

We report on four consecutive patients with Parkinson's disease, in whom anti-acetylcholine receptor (AChR) antibody positive bulbar myasthenia gravis (MG) turned out to be responsible for progressive dysphagia.

Effect of Dysarthria Type, Speaking Condition, and Listener Age on Speech Intelligibility.

PURPOSE: The aim of this study was to examine the effect of loud and slow speech cues on younger and older listeners' comprehension of dysarthric speech, specifically, (a) whether one strategy, as opposed to the other, promoted greater intelligibility gains for different speaker groups; (b) whether older and younger listeners' understandings were differentially affected by these strategies; and (c) which acoustic changes best predicted intelligibility gain in individual speakers. METHOD: Twenty younger and 40 older listeners completed a perceptual task. Six individuals with dysarthria produced phrases across habitual, loud, and slow conditions. The primary dependent variable was proportion of words correct; follow-up acoustic analyses linked perceptual outcomes to changes in acoustic speech features. RESULTS: Regardless of dysarthria type, the loud condition produced significant intelligibility gains. Overall, older listeners' comprehension was reduced relative to younger listeners. Follow-up analysis revealed considerable interspeaker differences in intelligibility outcomes across conditions. Although the most successful speaking mode varied, intelligibility gains were strongly associated with the degree of change participants made to their vowel formants. CONCLUSIONS: Perceptual outcomes vary across speaking modes, even when speakers with dysarthria are grouped according to similar perceptual profiles. Further investigation of interspeaker differences is needed to inform individually tailored intervention approaches.

Progressive dysarthria and augmentative and alternative communication in conversation: establishing the reliability of the Dysarthria-in-Interaction Profile.

BACKGROUND: The Dysarthria-in-Interaction Profile's potential contribution to the clinical assessment of dysarthria-in-conversation has been outlined in the literature, but its consistency of use across different users has yet to be reported. AIMS: To establish the level of consistency across raters on four different interaction categories. That is, how reliable clinicians are when rating a series of videos. A secondary aim was to investigate the relationship between raters' estimates of dysarthric speech intelligibility and their rating of each dyad's overall interaction. METHODS & PROCEDURES: Ten UK speech and language therapists rated independently a series of 40 video samples featuring people with progressive dysarthria in conversation with family members. An equal number of video samples was selected from a collection of recordings featuring four different types of interactional relationship. OUTCOMES & RESULTS: The results show that practising speech and language therapists can rate consistently, and with a high degree of agreement, a series of everyday conversation videos featuring dyads with progressive dysarthria and presenting at different interaction levels. The results also indicate that speech intelligibility does not predict the level of impairment in the interaction in a systematic way suggesting that conversation contains elements that are not directly related to speech intelligibility. CONCLUSIONS & IMPLICATIONS: Further work is required to establish the clinical functionality of this tool, but the results presented here support the development of this conversation profiling system, particularly for people experiencing significant intelligibility problems but remaining highly interactive/communicative.

Perceptual ratings of subgroups of ataxic dysarthria.

BACKGROUND: The speech characteristics of ataxic dysarthria are known to be quite diverse. The varied presentation of this dysarthria challenges researchers and clinicians alike, and brings into question whether it is a single entity. While the possibility of subtypes of ataxic dysarthria has been suggested, the nature of these putative groups remains unclear. AIMS: The purpose of this pilot study was to determine if perceptual speech characteristics would align with a pattern of unusual variability or unusual consistency across speech subsystems and speaking tasks. METHODS & PROCEDURES: A framework of speech characteristics was created from the existing literature that clustered speech attributes according to notions of instability (unusual variability) or inflexibility (unusual consistency). These speech features were used to develop a perceptual rating form. Ten experienced speech-language pathologists listened to pre-recorded, exemplary samples of 10 speakers with ataxic dysarthria and rated the perceptual speech features. OUTCOMES & RESULTS: Results suggested that five speakers fit the pattern of instability, one speaker aligned with inflexibility and four speakers had a mixed presentation. Intra-rater reliability was satisfactory. CONCLUSIONS & IMPLICATIONS: This study adds to the sparse, yet growing, literature to support the existence of subgroups in ataxic dysarthria. The more frequent occurrence of the instability profile is consistent with primary disruption to the timing function of the cerebellar circuit. Identification of subgroups has important clinical and research implications and further research is warranted.

Free-classification of perceptually similar speakers with dysarthria.

PURPOSE: In this investigation, the construct of perceptual similarity was explored in the dysarthrias. Specifically, we employed an auditory free-classification task to determine whether listeners could cluster speakers by perceptual similarity, whether the clusters mapped to acoustic metrics, and whether the clusters were constrained by dysarthria subtype diagnosis. METHOD: Twenty-three listeners blinded to speakers' medical and dysarthria subtype diagnoses participated. The task was to group together (drag and drop) the icons corresponding to 33 speakers with dysarthria on the basis of how similar they sounded. Cluster analysis and multidimensional scaling (MDS) modeled the perceptual dimensions underlying similarity. Acoustic metrics and perceptual judgments were used in correlation analyses to facilitate interpretation of the derived dimensions. RESULTS: Six clusters of similar-sounding speakers and 3 perceptual dimensions underlying similarity were revealed. The clusters of similar-sounding speakers were not constrained by dysarthria subtype diagnosis. The 3 perceptual dimensions revealed by MDS were correlated with metrics for articulation rate, intelligibility, and vocal quality, respectively. CONCLUSIONS: This study shows (a) feasibility of a free-classification approach for studying perceptual similarity in dysarthria, (b) correspondence between acoustic and perceptual metrics to clusters of similar-sounding speakers, and (c) similarity judgments transcended dysarthria subtype diagnosis.

Effects of listeners' working memory and noise on speech intelligibility in dysarthria.

The purpose of this study was to investigate the effects of listeners' working memory (WM), types of noise and signal to noise ratios (SNRs) on speech intelligibility in dysarthria. Speech intelligibility was judged by using a word transcription task. A three-way mixed design (2 × 3 × 2) was used with the WM group (high/low group) as a between-subject factor and the types of noise (multi-talker babble/environmental noise) and SNRs (0, +10 and +20 dB) as within-subject factors. The dependent measure was the percentage of correctly transcribed words. The results revealed that the high WM group performed significantly better than the low WM group and listeners performed significantly better at higher levels of SNRs on the speech intelligibility test. The findings of this study suggested that listeners' cognitive abilities and SNRs should be considered as important factors when evaluating speech intelligibility in dysarthria.

Vowel acoustics in dysarthria: speech disorder diagnosis and classification.

PURPOSE: The purpose of this study was to determine the extent to which vowel metrics are capable of distinguishing healthy from dysarthric speech and among different forms of dysarthria. METHOD: A variety of vowel metrics were derived from spectral and temporal measurements of vowel tokens embedded in phrases produced by 45 speakers with dysarthria and 12 speakers with no history of neurological disease. Via means testing and discriminant function analysis (DFA), the acoustic metrics were used to (a) detect the presence of dysarthria and (b) classify the dysarthria subtype. RESULTS: Significant differences between dysarthric and healthy control speakers were revealed for all vowel metrics. However, the results of the DFA demonstrated some metrics (particularly metrics that capture vowel distinctiveness) to be more sensitive and specific predictors of dysarthria. Only the vowel metrics that captured slope of the second formant (F2) demonstrated between-group differences across the dysarthrias. However, when subjected to DFA, these metrics proved unreliable classifiers of dysarthria subtype. CONCLUSION: The results of these analyses suggest that some vowel metrics may be useful clinically for the detection of dysarthria but may not be reliable indicators of dysarthria subtype using the current dysarthria classification scheme.

Development of The Viking Speech Scale to classify the speech of children with cerebral palsy.

Surveillance registers monitor the prevalence of cerebral palsy and the severity of resulting impairments across time and place. The motor disorders of cerebral palsy can affect children's speech production and limit their intelligibility. We describe the development of a scale to classify children's speech performance for use in cerebral palsy surveillance registers, and its reliability across raters and across time. Speech and language therapists, other healthcare professionals and parents classified the speech of 139 children with cerebral palsy (85 boys, 54 girls; mean age 6.03 years, SD 1.09) from observation and previous knowledge of the children. Another group of health professionals rated children's speech from information in their medical notes. With the exception of parents, raters reclassified children's speech at least four weeks after their initial classification. Raters were asked to rate how easy the scale was to use and how well the scale described the child's speech production using Likert scales. Inter-rater reliability was moderate to substantial (k>.58 for all comparisons). Test-retest reliability was substantial to almost perfect for all groups (k>.68). Over 74% of raters found the scale easy or very easy to use; 66% of parents and over 70% of health care professionals judged the scale to describe children's speech well or very well. We conclude that the Viking Speech Scale is a reliable tool to describe the speech performance of children with cerebral palsy, which can be applied through direct observation of children or through case note review.

Disorders of communication: dysarthria.

Dysarthria is a motor speech disorder which can be classified according to the underlying neuropathology and is associated with disturbances of respiration, laryngeal function, airflow direction, and articulation resulting in difficulties of speech quality and intelligibility. There are six major types of dysarthria: flaccid dysarthria associated with lower motor neuron impairment, spastic dysarthria associated with damaged upper motor neurons linked to the motor areas of the cerebral cortex, ataxic dysarthria primarily caused by cerebellar dysfunction, and hyperkinetic dysarthria and hypokinetic dysarthria, which are related to a disorder of the extrapyramidal system. The sixth is generally termed a mixed dysarthria and is associated with damage in more than one area, resulting in speech characteristics of at least two groups. The features of the speech disturbance of these six major types of dysarthria are distinctive and can assist with diagnosis. Dysarthria is a frequent symptom of many neurological conditions and is commonly associated with progressive neurological disease. It has a profound effect upon the patient and their families as communication is integrally related with expressing personality and social relationships. Speech and language therapy can be used to encourage the person to use the speech that is already available to them more effectively, can increase the range and consistency of sound production, can teach strategies for improving intelligibility and communicative effectiveness, can guide the individual to use methods that are less tiring and more successful, and can introduce the appropriate Augmentative and Alternative Communication approaches as and when required.

Las disartrias / The dysarthrias

The dysarthrias are a group of motor alterations of the speech as consequence of a neurological impairment that represent the 54% of the total of disorders that impair oral communication. Several types of dysarthrias could be determined according to the location of the causing lesion, the neurological mechanisms implicated in its generation, and functional characteristics of the speech. Clinical evaluation of the dysarthrias should include the clinical history, physical examination and the motor assessment of the speech. Prognosis is determined by the etiology and severity of the disorder. The objective of the treatment is to satisfy daily communicational needs of the patient in consideration to the underlying neurological condition (AU)

Classifications of vocalic segments from articulatory kinematics: healthy controls and speakers with dysarthria.

PURPOSE: In this study, the authors classified vocalic segments produced by control speakers (C) and speakers with dysarthria due to amyotrophic lateral sclerosis (ALS) or Parkinson's disease (PD); classification was based on movement measures. The researchers asked the following questions: (a) Can vowels be classified on the basis of selected measures of articulatory motions? and (b) Can classification models that are constructed from control productions classify vowels produced by speakers with dysarthria that is related to ALS and PD? METHOD: Nineteen C, 7 PD, and 8 ALS speakers participated in this study. The severity of dysarthria varied across individuals and between the 2 disorder groups. The stimuli were 6 vowels produced in 10 words embedded into sentences read at a comfortable reading rate. Movement data were collected using the x-ray microbeam. Movement measures included distances traveled, durations, and average speeds of vowel-related movement strokes. Vowels and words were classified by linear discriminant analysis with measures of articulatory motion as input variables. RESULTS: The study showed that vocalic segments could be classified using articulatory movement characteristics with up to 80% accuracy. The classification accuracy of the movement-based models depended largely on the number of articulators involved and, to a lesser extent, on the movement measure (e.g., distance, duration, speed). Classification of PD vowels was similar to that of the C group, suggesting a simple scaling of gestures as an explanation of the movement deficit in this disease. Classification performance for ALS vowels appeared to be different from that of C and PD productions. CONCLUSION: Classification of vowels was possible on the basis of their articulatory motions. ALS vowels appeared categorically different from those of C and PD speakers.

Applying the WHO ICF framework to communication assessment and goal setting in Huntington's disease: a case discussion.

PURPOSE: Huntington's Disease (HD) is a fatal, hereditary neurodegenerative disorder that is characterized by a triad of motor, cognitive and psychiatric symptoms that impact on both communicative effectiveness and the treatment techniques used to maximize communicative participation. The purpose of this article is to describe the application of the World Health Organization's International Classification of Functioning, Disability and Health (ICF) to communication assessment and goal setting for a 37 year old man with HD. METHOD: Communication assessment consisted of qualitative interviews based on the ICF framework with the client and his mother, supplemented with quantitative communication assessments. Analysis and conceptualization of assessment information was completed with a modified version of the Rehabilitation Problem Solving Form (RPS-Form) (Steiner et al., 2002). RESULTS: While impairments in body structures and functions were evident, analysis revealed that environmental factors such as family support were both key barriers and facilitators to communicative participation. CONCLUSIONS: This case illustrates the potential of using the ICF to conceptualize communicative functioning and disability in HD and particularly highlights the importance of consideration of personal and environmental factors in revealing contributions to activity limitations and participation restrictions. Further investigation of applications of ICF for individuals with HD is suggested. LEARNING OUTCOMES: As a result of this activity, readers will recognize the application of the WHO ICF to assessment and goal setting in a complex hereditary condition, Huntington Disease. As a result of this activity, readers will describe the use of the Rehabilitation Problem Solving Form (RPS-Form) for communication disorders. As a result of this activity, readers will identify relevant issues in comprehensive communication assessment of a fatal, degenerative neurological condition such as the advantages and challenges of clinical use of the ICF and its coding. As a result of this activity, readers will introduce life storybooks as a potential legacy item in degenerative diseases.

An acoustic study of the relationships among neurologic disease, dysarthria type, and severity of dysarthria.

PURPOSE: This study examined acoustic predictors of speech intelligibility in speakers with several types of dysarthria secondary to different diseases and conducted classification analysis solely by acoustic measures according to 3 variables (disease, speech severity, and dysarthria type). METHOD: Speech recordings from 107 speakers with dysarthria due to Parkinson's disease, stroke, traumatic brain injury, and multiple system atrophy were used for acoustic analysis and for perceptual judgment of speech intelligibility. Acoustic analysis included 8 segmental/suprasegmental features: 2nd formant frequency slope, articulation rate, voiceless interval duration, 1st moment analysis for fricatives, vowel space, F0, intensity range, and Pairwise Variability Index. RESULTS: The results showed that (a) acoustic predictors of speech intelligibility differed slightly across diseases and (b) classification accuracy by dysarthria type was typically worse than by disease type or severity. CONCLUSIONS: These findings were discussed with respect to (a) the relationship between acoustic characteristics and speech intelligibility and (b) dysarthria classification.

Extensions to the Speech Disorders Classification System (SDCS).

This report describes three extensions to a classification system for paediatric speech sound disorders termed the Speech Disorders Classification System (SDCS). Part I describes a classification extension to the SDCS to differentiate motor speech disorders from speech delay and to differentiate among three sub-types of motor speech disorders. Part II describes the Madison Speech Assessment Protocol (MSAP), an ∼ 2-hour battery of 25 measures that includes 15 speech tests and tasks. Part III describes the Competence, Precision, and Stability Analytics (CPSA) framework, a current set of ∼ 90 perceptual- and acoustic-based indices of speech, prosody, and voice used to quantify and classify sub-types of Speech Sound Disorders (SSD). A companion paper provides reliability estimates for the perceptual and acoustic data reduction methods used in the SDCS. The agreement estimates in the companion paper support the reliability of SDCS methods and illustrate the complementary roles of perceptual and acoustic methods in diagnostic analyses of SSD of unknown origin. Examples of research using the extensions to the SDCS described in the present report include diagnostic findings for a sample of youth with motor speech disorders associated with galactosemia, and a test of the hypothesis of apraxia of speech in a group of children with autism spectrum disorders. All SDCS methods and reference databases running in the PEPPER (Programs to Examine Phonetic and Phonologic Evaluation Records) environment will be disseminated without cost when complete.

Perceptual and acoustic reliability estimates for the Speech Disorders Classification System (SDCS).

A companion paper describes three extensions to a classification system for paediatric speech sound disorders termed the Speech Disorders Classification System (SDCS). The SDCS uses perceptual and acoustic data reduction methods to obtain information on a speaker's speech, prosody, and voice. The present paper provides reliability estimates for the two perceptual methods (narrow phonetic transcription; prosody-voice coding) and the acoustic analysis methods the SDCS uses to describe and classify a speaker's speech competence, precision, and stability. Speech samples from 10 speakers, five with significant motor speech disorder and five with typical speech, were re-measured to estimate intra-judge and inter-judge agreement for the perceptual and acoustic methods. Each of the speakers completed five speech tasks (total = 50 datasets), ranging in articulatory difficulty for the speakers, with consequences for the difficulty level of data reduction. Point-to-point percentage of agreement findings for the two perceptual methods were as high or higher than reported in literature reviews and from previous studies conducted within the laboratory. Percentage of agreement findings for the acoustics tasks of segmenting phonemes, editing fundamental frequency tracks, and estimating formants ranged from values in the mid 70% to 100%, with most estimates in the mid 80% to mid 90% range. Findings are interpreted as support for the perceptual and acoustic methods used in the SDCS to describe and classify speakers with speech sound disorders.

Classification of speech and language profiles in 4-year-old children with cerebral palsy: a prospective preliminary study.

PURPOSE: In this study, the authors proposed and tested a preliminary speech and language classification system for children with cerebral palsy. METHOD: Speech and language assessment data were collected in a laboratory setting from 34 children with cerebral palsy (CP; 18 male, 16 female) with a mean age of 54 months (SD = 1.8). Measures of interest were vowel area, speech rate, language comprehension scores, and speech intelligibility ratings. RESULTS: Canonical discriminant function analysis showed that 3 functions accounted for 100% of the variance among profile groups, with speech variables accounting for 93% of the variance. Classification agreement varied from 74% to 97% based on 4 different classification paradigms. CONCLUSIONS: The results of this study provide preliminary support for the classification of speech and language abilities of children with CP into 4 initial profile groups. Further research is necessary to validate the full classification system.