Breath Sound Signals Research Articles

Application study of apnea-hypopnea duration for assessing adult obstructive sleep apnea.

Obstructive sleep apnea (OSA) is a common sleep disordered breathing disorder, which can cause serious damage to multiple human systems. Although polysomnography (PSG) is the current gold standard for diagnosis, it is complex and expensive. Therefore, it is of great significance to find a simple, economical and rapid primary screening and diagnosis method to replace PSG for the diagnosis of OSA. The purpose of this study is to propose a new method for the diagnosis and classification of OSA, which is used to automatically detect the duration of sleep apnea hypopnea events (AHE), so as to estimate the ratio(S) of the total duration of all-night AHE to the total sleep time only based on the sound signal of sleep respiration, and to identify OSA. We performed PSG tests on participants and extracted relevant sleep breathing sound signal data. This study is carried out in two stages. In the first stage, the relevant PSG report data of eligible subjects were recorded, the total duration of AHE in each subject's data was extracted, and the S value was calculated to evaluate the severity of OSA. In the second stage, only the sleep breath sound signal data of the same batch of subjects were used for automatic detection, and the S value in the sleep breath sound signal was extracted, and the S value was compared with the PSG diagnosis results to calculate the accuracy of the experimental method. Among 225 subjects. Using PSG as the reference standard, the S value extracted from the PSG diagnostic data report can accurately diagnose OSA(accuracy rate 99.56%) and distinguish its severity (accuracy rate 95.11%). The accuracy of the S value detected in the sleep breathing sound signal in the diagnosis of severe OSA reached 100%. The results show that the experimental parameter S value is feasible in OSA diagnosis and classification. OSA can be identified and evaluated only by sleep breathing sounds. This method helps to simplify the diagnostic grading of traditional OSA and lays a foundation for the subsequent development of simple diagnostic grading equipment.

Hybrid method for noise rejection from breath sound using transient artifact reduction algorithm and spectral subtraction.

Computerized breath sound based diagnostic methods are one of the emerging technologies gaining popularity in terms of detecting respiratory disorders. However, the breath sound signal used in such automated systems used to be too noisy, which affects the quality of the diagnostic interpretations. To address this problem, the proposed work presents the new hybrid approach to reject the noises from breath sound. In this method, 80 chronic obstructive pulmonary disease (COPD), 75 asthmatics and 80 normal breath sounds were recorded from the participants of a hospital. Each of these breath sound data were decontaminated using hybrid method of Butterworth band-pass filter, transient artifact reduction algorithm and spectral subtraction algorithm. The study examined the algorithms noise rejection potential over each category of breath sound by estimating the noise rejection performance metrics, i.e.,mean absolute error (MAE), mean square error (MSE), peak signal to noise ratio (PSNR), and signal to noise ratio (SNR). Using this algorithm, the study obtained a high value of SNR of 70 dB and that of PSNR of 72 dB. The study could definitely a suitable one to suppress noises and to produce noise free breath sound signal.

Nonlinear signal processing, spectral, and fractal based stridor auscultation: A machine learning approach

The work reported in the paper analyses the adventitious stridor breath sound (ST) and the normal bronchial breath sound (BR) using spectral, fractal, and nonlinear signal processing methods. The sixty breath sound signals are subjected to power spectral density (PSD) and wavelet analyses to understand the temporal evolution of the frequency components. The energy envelope of the PSD plot of ST shows three peaks labelled as A (256 Hz), B (369 Hz), and C (540 Hz), of which A alone is present in BR at 265 Hz. The appearance of B and C in the PSD plot of ST is due to the obstructions in the trachea and upper airways caused by lesions. The phase portrait analysis of the time series data of ST and BR gives information about the randomness and the sample entropy of the dynamical system. The study reveals that the fractal dimension and sample entropy values are higher for BR, which may be due to the musical ordered behaviour of ST. The machine learning techniques based on the features extracted from the PSD data and phase portrait parameters offer good predictability, besides the classification of BR and ST, and thereby revealing its potential in pulmonary auscultation.

Untangling the graph based features for lung sound auscultation

Auscultation is a historically significant and currently relevant method employed in medical diagnosis. The technological advancement joining hands with mathematical principles led to the development of sophisticated biomedical instruments. The paper proposes a novel graph-based auscultation method, incorporating the principles of machine learning techniques (MLTs), for catering the requirements of remote auscultation, which is highly relevant during the outbreak of the pandemic COVID-19. The proposed technique, delineated by analyzing the complex network constructed from the sixty bronchial and stridor breath sound signals (BS and SS), unveils the trachea and upper respiratory tract condition through the dynamics of airflow during the production of the signals. The time and frequency domain analyses of the time series signals unwraps the discontinuous and musical nature of BS and SS signal, respectively. The graph features extracted from the complex network, constructed using correlation mapping approach, throws light into the conditions of the trachea. The higher value of the number of edges, graph density and transitivity in BS, with low-frequency components and high correlation between the nodes, accounts for the airflow dynamics through the trachea of uniform and larger diameter. The lowering of these features and enhancement of degree centrality, closeness centrality and entropy in SS, with higher frequency components and low correlation between the nodes, reflects the pathological condition of constrictions in the trachea. The MLTs, principal component and linear discriminant analyses, could effectively classify and accurately predict the signals using the features of the graph, suggesting its potential in auscultation.

Phase portrait for high fidelity feature extraction and classification: A surrogate approach.

This paper proposes a novel surrogate method of classification of breath sound signals for auscultation through the principal component analysis (PCA), extracting the features of a phase portrait. The nonlinear parameters of the phase portrait like the Lyapunov exponent, the sample entropy, the fractal dimension, and the Hurst exponent help in understanding the degree of complexity arising due to the turbulence of air molecules in the airways of the lungs. Thirty-nine breath sound signals of bronchial breath (BB) and pleural rub (PR) are studied through spectral, fractal, and phase portrait analyses. The fast Fourier transform and wavelet analyses show a lesser number of high-intense, low-frequency components in PR, unlike BB. The fractal dimension and sample entropy values for PR are, respectively, 1.772 and 1.041, while those for BB are 1.801 and 1.331, respectively. This study reveals that the BB signal is more complex and random, as evidenced by the fractal dimension and sample entropy values. The signals are classified by PCA based on the features extracted from the power spectral density (PSD) data and the features of the phase portrait. The PCA based on the features of the phase portrait considers the temporal correlation of the signal amplitudes and that based on the PSD data considers only the signal amplitudes, suggesting that the former method is better than the latter as it reflects the multidimensional aspects of the signal. This appears in the PCA-based classification as 89.6% for BB, a higher variance than the 80.5% for the PR signal, suggesting the higher fidelity of the phase portrait-based classification.

Nonlinear time series and principal component analyses: Potential diagnostic tools for COVID-19 auscultation.

The development of novel digital auscultation techniques has become highly significant in the context of the outburst of the pandemic COVID 19. The present work reports the spectral, nonlinear time series, fractal, and complexity analysis of vesicular (VB) and bronchial (BB) breath signals. The analysis is carried out with 37 breath sound signals. The spectral analysis brings out the signatures of VB and BB through the power spectral density plot and wavelet scalogram. The dynamics of airflow through the respiratory tract during VB and BB are investigated using the nonlinear time series and complexity analyses in terms of the phase portrait, fractal dimension, Hurst exponent, and sample entropy. The higher degree of chaoticity in BB relative to VB is unwrapped through the maximal Lyapunov exponent. The principal component analysis helps in classifying VB and BB sound signals through the feature extraction from the power spectral density data. The method proposed in the present work is simple, cost-effective, and sensitive, with a far-reaching potential of addressing and diagnosing the current issue of COVID 19 through lung auscultation.

Stethoscope-Sensed Speech and Breath-Sounds for Person Identification With Sparse Training Data

A novel person identification (PID) technique is developed in this study, which exploits a new biometric called bronchial breath sound and speech signal acquired by a stethoscope. In addition to investigating the acoustic characteristics of breath sounds for PID, we evaluate three identification methods, including support vector machines (SVM), artificial neural networks (ANN), and i-vector approach. Recognizing the requirement that the amount of sound data collected from each person should be as small as possible, this work studies data augmentation (DA) techniques that avoid the system training process from the overfitting problem when the training sound data is insufficient. In addition, we apply feature engineering techniques to find the informative subset of breath sound features which is beneficial for PID. Our experiments were conducted using a dataset composed of 16 subjects, including an equal number of male and female participants. In the test phase, both Support Vector Machine combined with feature selection and Artificial Neural Networks approaches yielded the promising accuracies of 98%.

A simple method for monitoring sleeping conditions by all-night breath sound measurement

In this work, a simple method for monitoring sleeping conditions by all-night breath sound measurement is proposed. Our research target is to develop and validate a high performance method to classify sleeping conditions into several stages based on breath sound. Subjects were recorded in home/group house using bluetooth breath sound sensor. Breath sound-based features extracted from time and spectral domains can accurately discriminate between snore or non-snore and respiratory variance of breath sound, and so on. This breath sound signal analysis method enables detection and analysis of apnea/hypopnea. This work is the first study that successfully used simple method to monitor sleeping conditions. The experiment validate the effectiveness of the proposed method.

Detection of swallows with silent aspiration using swallowing and breath sound analysis

In this study, the feasibility of acoustical analysis for detection of swallowing silent aspiration is investigated. As a pilot study, we analyzed the breath sounds of 21 dysphagic individuals, 11 of which demonstrated aspiration during the fiberoptic endoscopic evaluation of swallowing (FEES) or videofluoroscopic swallowing study (VFSS). We found that the low frequency components of the power spectrum of the breath sounds after a swallow show higher magnitude when there is aspiration. Thus, we divided the frequency range below 300 Hz into three sub-bands and calculated the average power of the breath sound signal in each sub-band as the characteristic features for the stage 1 classification into two groups of aspirated and non-aspirated patients. Then, for the aspirated group, the unsupervised fuzzy k-means clustering algorithm was deployed to label the breath sounds immediately after a swallow as aspiration or non-aspiration. The results were compared with the FEES/VFSS assessments provided by the speech language pathologists. The results are encouraging: more than 86 % accuracy in detection of silent aspiration. While the proposed method should be verified on a larger dataset, the results are promising for the use of acoustical analysis as a clinical tool to detect silent aspiration.

Adaptive reduction of heart sounds from lung sounds using fourth-order statistics.

When recording lung sounds, an incessant noise source occurs due to heart sounds. This noise source severely contaminates the breath sound signal and interferes in the analysis of lung sounds. In this paper, an adaptive heart-noise reduction method, based on fourth-order statistics (FOS) of the recorded signal, without requiring recorded "noise-only" reference signal, is presented. This algorithm uses adaptive filtering to preserve the entire spectrum. Furthermore, the proposed filter is independent of Gaussian uncorrelated noise and insensitive to the step-size parameter. It converges fast with small excess errors and, due to the narrow-band nature of heart noise (HN), it requires a very small number of taps. Results from experiments with healthy subjects indicate a local HN reduction equal to or greater than 90%.