Automatic Detection of Speech Disorder in Dysarthria using Extended Speech Feature Extraction and Neural Networks Classification

Ijitona, Tolulope; Soraghan, John J.; Lowit, Anja; Di-Caterina, G; Yue, Hongyang

doi:10.1049/cp.2017.0360

Cited by 13 publications

(6 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, advances in neural architectures as well as transfer learning techniques applied to deep learning have largely enabled this opening up. Among these works, we can cite [7,8,9,10]. The study presented in this paper is one of the main objectives of a long-term research project, which is the search for linguistic units playing a significant role in speech intelligibility, and therefore in its loss in the event of speech disorders.…”

Section: Introductionmentioning

confidence: 99%

Towards Interpreting Deep Learning Models to Understand Loss of Speech Intelligibility in Speech Disorders — Step 1: CNN Model-Based Phone Classification

Abderrazek¹,

Fredouille²,

Ghio³

et al. 2020

Interspeech 2020

View full text Add to dashboard Cite

Perceptual measurement is still the most common method for assessing disordered speech in clinical practice. The subjectivity of such a measure, strongly due to human nature, but also to its lack of interpretation with regard to local alterations in speech units, strongly motivates a sophisticated tool for objective evaluation. Of interest is the increasing performance of Deep Neural Networks in speech applications, but more importantly the fact that they are no longer considered as black boxes. The work carried out here is the first step in a long-term research project, which aims to determine the linguistic units that contribute most to the maintenance or loss of the intelligibility in speech disorders. In this context, we study a CNN trained on normal speech for a classification task of phones and tested on pathological speech. The aim of this first study is to analyze the response of the CNN model to disordered speech in order to study later its effectiveness in providing relevant knowledge in terms of speech severity or loss of intelligibility. Compared to perceptual severity and intelligibility measures, the results revealed a very strong correlation between these metrics and our classifier performance scores, which is very promising for future work.

show abstract

Section: Introductionmentioning

confidence: 99%

Towards Interpreting Deep Learning Models to Understand Loss of Speech Intelligibility in Speech Disorders — Step 1: CNN Model-Based Phone Classification

Abderrazek¹,

Fredouille²,

Ghio³

et al. 2020

Interspeech 2020

View full text Add to dashboard Cite

show abstract

“…These networks were applied to centroid formants, which represent extended speech characteristics, aiding in the discrimination between dysarthric and non-dysarthric speech. Subsequently, the study employed an experimental database consisting of 200 speech samples from ten individuals with dysarthria and an equal number of speeches from ten age-matched healthy individuals (Ijitona et al, 2017 ). Mani et al created a software program capable of making determinations about specific features through the application of fractal analysis.…”

Section: Literature Surveymentioning

confidence: 99%

AFM signal model for dysarthric speech classification using speech biomarkers

Shabber,

Sumesh

2024

Front. Hum. Neurosci.

View full text Add to dashboard Cite

Neurological disorders include various conditions affecting the brain, spinal cord, and nervous system which results in reduced performance in different organs and muscles throughout the human body. Dysarthia is a neurological disorder that significantly impairs an individual's ability to effectively communicate through speech. Individuals with dysarthria are characterized by muscle weakness that results in slow, slurred, and less intelligible speech production. An efficient identification of speech disorders at the beginning stages helps doctors suggest proper medications. The classification of dysarthric speech assumes a pivotal role as a diagnostic tool, enabling accurate differentiation between healthy speech patterns and those affected by dysarthria. Achieving a clear distinction between dysarthric speech and the speech of healthy individuals is made possible through the application of advanced machine learning techniques. In this work, we conducted feature extraction by utilizing the Amplitude and frequency modulated (AFM) signal model, resulting in the generation of a comprehensive array of unique features. A method involving Fourier-Bessel series expansion is employed to separate various components within a complex speech signal into distinct elements. Subsequently, the Discrete Energy Separation Algorithm is utilized to extract essential parameters, namely the Amplitude envelope and Instantaneous frequency, from each component within the speech signal. To ensure the robustness and applicability of our findings, we harnessed data from various sources, including TORGO, UA Speech, and Parkinson datasets. Furthermore, the classifier's performance was evaluated based on multiple measures such as the area under the curve, F1-Score, sensitivity, and accuracy, encompassing KNN, SVM, LDA, NB, and Boosted Tree. Our analyses resulted in classification accuracies ranging from 85 to 97.8% and the F1-score ranging between 0.90 and 0.97.

show abstract

“…A good example of dysarthria detection has been published previously [32]. The rate of success of a neural network is modest, that is, about 70% when competing with standard diagnostic performance.…”

Section: Voice Analysismentioning

confidence: 99%

A Digital Telehealth System to Compute Myasthenia Gravis Core Examination Metrics: Exploratory Cohort Study

Garbey¹,

Joerger²,

Lesport³

et al. 2023

JMIR Neurotech

View full text Add to dashboard Cite

Background Telemedicine practice for neurological diseases has grown significantly during the COVID-19 pandemic. Telemedicine offers an opportunity to assess digitalization of examinations and enhances access to modern computer vision and artificial intelligence processing to annotate and quantify examinations in a consistent and reproducible manner. The Myasthenia Gravis Core Examination (MG-CE) has been recommended for the telemedicine evaluation of patients with myasthenia gravis. Objective We aimed to assess the ability to take accurate and robust measurements during the examination, which would allow improvement in workflow efficiency by making the data acquisition and analytics fully automatic and thereby limit the potential for observation bias. Methods We used Zoom (Zoom Video Communications) videos of patients with myasthenia gravis undergoing the MG-CE. The core examination tests required 2 broad categories of processing. First, computer vision algorithms were used to analyze videos with a focus on eye or body motions. Second, for the assessment of examinations involving vocalization, a different category of signal processing methods was required. In this way, we provide an algorithm toolbox to assist clinicians with the MG-CE. We used a data set of 6 patients recorded during 2 sessions. Results Digitalization and control of quality of the core examination are advantageous and let the medical examiner concentrate on the patient instead of managing the logistics of the test. This approach showed the possibility of standardized data acquisition during telehealth sessions and provided real-time feedback on the quality of the metrics the medical doctor is assessing. Overall, our new telehealth platform showed submillimeter accuracy for ptosis and eye motion. In addition, the method showed good results in monitoring muscle weakness, demonstrating that continuous analysis is likely superior to pre-exercise and postexercise subjective assessment. Conclusions We demonstrated the ability to objectively quantitate the MG-CE. Our results indicate that the MG-CE should be revisited to consider some of the new metrics that our algorithm identified. We provide a proof of concept involving the MG-CE, but the method and tools developed can be applied to many neurological disorders and have great potential to improve clinical care.

show abstract

Automatic Detection of Speech Disorder in Dysarthria using Extended Speech Feature Extraction and Neural Networks Classification

Cited by 13 publications

References 15 publications

Towards Interpreting Deep Learning Models to Understand Loss of Speech Intelligibility in Speech Disorders — Step 1: CNN Model-Based Phone Classification

Towards Interpreting Deep Learning Models to Understand Loss of Speech Intelligibility in Speech Disorders — Step 1: CNN Model-Based Phone Classification

AFM signal model for dysarthric speech classification using speech biomarkers

A Digital Telehealth System to Compute Myasthenia Gravis Core Examination Metrics: Exploratory Cohort Study

Contact Info

Product

Resources

About