Pathological speech signal analysis and classification using empirical mode decomposition

Tempo-Spectral EEG Biomarkers for Odour Identification

Different odours evoke different activity in the brain. Among the non-invasive methods, electroencephalogram (EEG) is the most widely used mode to measure brain activity. While there has been significant work around EEG signal analysis, studies in the area of EEG with odour as stimuli is nascent. In this paper, we experiment and study different EEG biomarkers with an aim to understand which biomarker shows promise for odour identification. We show, on a widely used and publicly available data-set, through a series of experiments that it is possible to get a Subject Dependent (SD) odour classification accuracy of over 90%, using a set of tempo-spectral EEG biomarkers. We further experiment with Subject Independent (SI) odour classification, which has not been addressed and show that the performance drops to under 50% for SI odour classification.Clinical Relevance - The study shows that the same odour evoke different brain responses from the subject.

Different Performances of Machine Learning Models to Classify Dysphonic and Non-Dysphonic Voices

OBJECTIVE

To analyze the performance of 10 different machine learning (ML) classifiers for discrimination between dysphonic and non-dysphonic voices, using a variance threshold as a method for the selection and reduction of acoustic measurements used in the classifier.

METHOD

We analyzed 435 samples of individuals (337 female and 98 male), with a mean age of 41.07 ± 13.73 years, of which 384 were dysphonic and 51 were non-dysphonic. From the sustained /ε/ vowel sample, 34 acoustic measurements were extracted, including traditional perturbation and noise measurements, cepstral/spectral measurements, and measurements based on nonlinear models. The variance method was used to select the best set of acoustic measurements. We tested the performance of the best-selected set with 10 ML classifiers using precision, sensitivity, specificity, accuracy, and F1-Score measurements. The kappa coefficient was used to verify the reproducibility between the two datasets (training and testing).

RESULTS

The naive Bayes (NB) and stochastic gradient descent classifier (SGDC) models performed best in terms of accuracy, AUC, sensitivity, and specificity for a reduced dataset of 15 acoustic measures compared to the full dataset of 34 acoustic measures. SGDC and NB obtained the best performance results, with an accuracy of 0.91 and 0.76, respectively. These two classifiers presented moderate agreement, with a Kappa of 0.57 (SGDC) and 0.45 (NB).

CONCLUSION

Among the tested models, the NB and SGDC models performed better in discriminating between dysphonic and non-dysphonic voices from a set of 15 acoustic measures.

The acoustic voice quality index, version 2.06 and 3.01, for the Persian-speaking population

INTRODUCTION

Dysphonia assessment includes approaches like acoustic analysis, which is non-invasive and easy to use and provides an understandable numerical output. The Acoustic Voice Quality Index (AVQI) is an acoustic model that can detect dysphonia. The Persian language is spoken by around 70,000,000 native speakers. Since AVQI versions 2.06 and 3.01 have not been validated for the Persian yet, this study investigated their concurrent validity and diagnostic accuracy among the Persian-speaking population.

METHODS

This scale development study was conducted from 2020 to 2021 on 180 normophonic and dysphonic native Persian-speaking residents of Mashhad, Iran. Five raters rated the samples by auditory-perceptual-judgments, including Grade from the Grade-Rough-Breathy-Asthenic-Strained (an ordinal scale) and the overall dysphonia severity from the Persian version Consensus Auditory Perceptual Evaluation of Voice (a continuous scale) to investigate both versions' concurrent validity. The intra- and inter-rater reliability and concurrent validity were evaluated for both scales. Both versions' diagnostic accuracy was assessed by the receiver operating characteristic, and the optimal thresholds were determined.

RESULTS

AVQI-version-2-Persian thresholds of 3.47 and 4.04 provided sensitivity of 88.30% and 85.53% and specificity of 79.07% and 85.58% by the ordinal and continuous scales, respectively. AVQI-version-3-Persian thresholds of 3.07 and 3.03 also rendered sensitivity of 74.47% and 85.53%, and specificity of 97.67% and 91.35% by the ordinal and continuous scales sequentially.

CONCLUSION

The significant values of concurrent validities and diagnostic accuracies of both versions of AVQI-Persian confirmed that it can discriminate between normal and pathological voices among the Persian-speaking population. Hence, it can be used for screening or diagnosis purposes.

Classification of Dysphonic Voices in Parkinson's Disease with Semi-Supervised Competitive Learning Algorithm

This article proposes a novel semi-supervised competitive learning (SSCL) algorithm for vocal pattern classifications in Parkinson’s disease (PD). The acoustic parameters of voice records were grouped into the families of jitter, shimmer, harmonic-to-noise, frequency, and nonlinear measures, respectively. The linear correlations were computed within each acoustic parameter family. According to the correlation matrix results, the jitter, shimmer, and harmonic-to-noise parameters presented as highly correlated in terms of Pearson’s correlation coefficients. Then, the principal component analysis (PCA) technique was implemented to eliminate the redundant dimensions of the acoustic parameters for each family. The Mann−Whitney−Wilcoxon hypothesis test was used to evaluate the significant difference of the PCA-projected features between the healthy subjects and PD patients. Eight dominant PCA-projected features were selected based on the eigenvalue threshold criterion and the statistical significance level (p < 0.05) of the hypothesis test. The SSCL algorithm proposed in this paper included the procedures of the competitive prototype seed selection, K-means optimization, and the nearest neighbor classifications. The pattern classification experimental results showed that the proposed SSCL method can provide the excellent diagnostic performances in terms of accuracy (0.838), recall (0.825), specificity (0.85), precision (0.846), F-score (0.835), Matthews correlation coefficient (0.675), area under the receiver operating characteristic curve (0.939), and Kappa coefficient (0.675), which were consistently better than those results of conventional KNN or SVM classifiers.

Empirical mode decomposition based ECG features in classifying and tracking ventricular arrhythmias

Ventricular arrhythmias (VA) are life-threatening pathophysiological conditions that seriously impact the normal functioning of the heart. Ventricular tachycardia (VT) and ventricular fibrillation (VF) are the two well known types of VA. VF is the lethal of the VAs and could be characterized by its organizational progression over time. The success of cardiac resuscitation strongly depends on the type of VA, its evolution over time and response to therapy. Due to the time critical nature of VF, computationally efficient quantification of VAs and swift feedback are essential. This work attempted to arrive at computationally efficient and data-driven techniques based on Empirical Mode Decomposition for classifying and tracking VAs over time. The approaches are divided into two aims: (1) 'in-hospital' scenarios for characterizing the dynamics of VA episodes to assist clinicians in planning long-term therapy options, and (2) 'out-of-hospital' scenarios for providing near real-time feedback to detect/track the progression of VAs over time to assist medical personnel select/modify therapy options. Using an ECG database of 61 60-s VA segments obtained for classifying VT vs. VF and sub-classifying VF into organized VF (OVF) and disorganized VF (DVF), maximum classification accuracies of 96.7% (AUC = 0.993) and 87.2% (AUC = 0.968) were obtained for classifying VT vs. VF and OVF vs. DVF during 'in-hospital' analysis. Additionally, two near real-time approaches were presented for 'out-of-hospital' analysis where average accuracies of 71% and 73% were achieved for VT/VF and OVF/DVF classification, as well as demonstrating strong potential for monitoring VA progressions over time.

A Survey on Machine Learning Approaches for Automatic Detection of Voice Disorders

The human voice production system is an intricate biological device capable of modulating pitch and loudness. Inherent internal and/or external factors often damage the vocal folds and result in some change of voice. The consequences are reflected in body functioning and emotional standing. Hence, it is paramount to identify voice changes at an early stage and provide the patient with an opportunity to overcome any ramification and enhance their quality of life. In this line of work, automatic detection of voice disorders using machine learning techniques plays a key role, as it is proven to help ease the process of understanding the voice disorder. In recent years, many researchers have investigated techniques for an automated system that helps clinicians with early diagnosis of voice disorders. In this paper, we present a survey of research work conducted on automatic detection of voice disorders and explore how it is able to identify the different types of voice disorders. We also analyze different databases, feature extraction techniques, and machine learning approaches used in these research works.

Developing a New Computer-Aided Clinical Decision Support System for Prediction of Successful Postcardioversion Patients with Persistent Atrial Fibrillation

We propose a new algorithm to predict the outcome of direct-current electric (DCE) cardioversion for atrial fibrillation (AF) patients. AF is the most common cardiac arrhythmia and DCE cardioversion is a noninvasive treatment to end AF and return the patient to sinus rhythm (SR). Unfortunately, there is a high risk of AF recurrence in persistent AF patients; hence clinically it is important to predict the DCE outcome in order to avoid the procedure's side effects. This study develops a feature extraction and classification framework to predict AF recurrence patients from the underlying structure of atrial activity (AA). A multiresolution signal decomposition technique, based on matching pursuit (MP), was used to project the AA over a dictionary of wavelets. Seven novel features were derived from the decompositions and were employed in a quadratic discrimination analysis classification to predict the success of post-DCE cardioversion in 40 patients with persistent AF. The proposed algorithm achieved 100% sensitivity and 95% specificity, indicating that the proposed computational approach captures detailed structural information about the underlying AA and could provide reliable information for effective management of AF.

HEART SOUND CLASSIFICATION AND RECOGNITION BASED ON EEMD AND CORRELATION DIMENSION

Analysis of heart sound is of great importance to the diagnosis of heart diseases. Most of the feature extraction methods about heart sound have focused on linear time-variant or time-invariant models. While heart sound is a kind of highly nonstationary and nonlinear vibration signal, traditional methods cannot fully reveal its essential properties. In this paper, a novel feature extraction approach is proposed for heart sound classification and recognition. The ensemble empirical mode decomposition (EEMD) method is used to decompose the heart sound into a finite number of intrinsic mode functions (IMFs), and the correlation dimensions of the main IMF components (IMF1~IMF4) are calculated as feature set. Then the classical Binary Tree Support Vector Machine (BT-SVM) classifier is employed to classify the heart sounds which include the normal heart sounds (NHSs) and three kinds of abnormal signals namely mitral stenosis (MT), ventricular septal defect (VSD) and aortic stenosis (AS). Finally, the performance of the new feature set is compared with the correlation dimensions of original signals and the main IMF components obtained by the EMD method. The results showed that, for NHSs, the feature set proposed in this paper performed the best with recognition rate of 98.67%. For the abnormal signals, the best recognition rate of 91.67% was obtained. Therefore, the proposed feature set is more superior to two comparative feature sets, which has potential application in the diagnosis of cardiovascular diseases.