Overnight airway obstruction severity prediction centered on acoustic properties of smart phone: validation with esophageal pressure.

OBJECTIVE: Obstructive sleep apnea is characterized by a number of airway obstructions. Esophageal pressure manometry (EPM) based estimation of consecutive peak to trough differences (ΔPes) is the gold standard method to quantify the severity of airway obstructions. However, the procedure is rarely available in sleep laboratories due to invasive nature. There is a clinical need for a simplified, scalable technology that can quantify the severity of airway obstructions. In this paper, we address this and propose a pioneering technology, centered on sleep related respiratory sound (SRS) to predict overnight ΔPes signal. APPROACH: We recorded streams of SRS using a bedside iPhone 7 smartphone from subjects undergoing diagnostic polysomnography (PSG) studies and EPM was performed concurrently. Overnight data was divided into epochs of 10 s duration with 50% overlap. Altogether, we extracted 42 181 such epochs from 13 subjects. Acoustic features and features from the two PSG signals serve as an input to train a machine learning algorithm to achieve mapping between non-invasive features and ΔPes values. A testing dataset of 14 171 epochs from four new subjects was used for validation. MAIN RESULTS: The SRS based model predicted the ΔPes with a median of absolute error of 6.75 cmH2O (±0.59, r = 0.83(±0.03)). When information from the PSG were combined with the SRS, the model performance became: 6.37cmH2O (±1.02, r = 0.85(±0.04)). SIGNIFICANCE: The smart phone based SRS alone, or in combination with routinely collected PSG signals can provide a non-invasive method to predict overnight ΔPes. The method has the potential to be automated and scaled to provide a low-cost alternative to EPM.

Corrections to "Automatic Croup Diagnosis Using Cough Sound Recognition".

Presents corrections to shareholder information from this paper, "Automatic croup diagnosis using cough sound recognition," (Sharan, R.V., et al), IEEE Trans. Biomed. Eng., vol. 66, no. 2, pp. 485-495, Feb. 2019.

Corrigendum: Predicting spirometry readings using cough sound features and regression (2018 Physiol. Meas. 39 095001).

The purpose of this submission is to provide missing information to complete the conflict of interest statement associated with the article. The statements provided here augment the already provided information rather than replace it.

An Investigation of Critical Frequency Sub-bands of Snoring Sounds for OSA Diagnosis.

Obstructive Sleep Apnea (OSA) is a result of upper airway narrowing during sleep. The upper airway characteristics are likely to manifest in the acoustic characteristics of snoring sounds as snoring is a result of upper airway structure vibrations. In previous studies, researchers have used different regions of the frequency spectrum to diagnose OSA and determine sites of obstruction as well. However, there is no agreement among researchers about the frequency ranges critical for OSA diagnosis. This paper provides the results of a study of snore sound based OSA diagnosis performance using a multiple acoustic features and multiple classifiers. The results of the study may provide useful insights for researchers to identify frequency sub-bands critical for OSA diagnosis.

Smart Phone based Snoring Sound analysis to Identify Upper Airway Obstructions.

Obstructive sleep apnea (OSA) is characterized by upper airway obstructions known as apnea/hypopnea events. Narrowing of the upper airway during or near the vicinity of apnea/hypopnea causes the spectrum of the snores to shift to higher frequencies. Using an instrumentation quality wideband (WB) microphone (4Hz-100kHz), we previously demonstrated that potentially diagnostically useful frequency shifts could be detected even in regions beyond the human hearing range. WB-microphone based systems are expensive and not available for home use or population screening application. In this paper we explore the feasibility of using smart phones to analyze snoring sounds in the 20Hz-22kHz band to identify events of upper airway obstructions. Modern smart phones have internal microphones with bandwidths up to 22kHz, above the nominal human hearing range, and provide a good platform for sound acquisition and processing. For the work of this paper we used a Samsung Galaxy S3 phone and recorded overnight respiratory sound data from 8 patients undergoing routine Polysomnography (PSG) study in a hospital. Our target was to develop models to classify each standard 30 second epoch of data as non-apnea or apnea. Using 700 epochs we developed logistic regression models with the input as snoring sound features and the outputs as the diagnostic classification of each event (apnea/non-apnea). Models developed within a 20Hz-15kHz band had accuracies of 89-93%, sensitivities 70-78% and kappa index ranging 0.75-0.83 on validation data set. When the same models were developed on the 20Hz-22kHz frequency band the improved performance shows accuracies 94- 97%, sensitivities 93-100%, and kappa ranging 0.86-0.91. The study shows that smart phones based high frequency band (15-22kHz) of snoring sounds carry information about the upper airway obstructions. Our non-contact, smart phone based snoring sound technology has potential to identify upper airway obstructions.

Automatic Croup Diagnosis Using Cough Sound Recognition.

OBJECTIVE: Croup, a respiratory tract infection common in children, causes an inflammation of the upper airway restricting normal breathing and producing cough sounds typically described as seallike "barking cough." Physicians use the existence of barking cough as the defining characteristic of croup. This paper aims to develop automated cough sound analysis methods to objectively diagnose croup. METHODS: In automating croup diagnosis, we propose the use of mathematical features inspired by the human auditory system. In particular, we utilize the cochleagram for feature extraction, a time-frequency representation where the frequency components are based on the frequency selectivity property of the human cochlea. Speech and cough share some similarities in the generation process and physiological wetware used. As such, we also propose the use of mel-frequency cepstral coefficients which has been shown to capture the relevant aspects of the short-term power spectrum of speech signals. Feature combination and backward sequential feature selection are also experimented with. Experimentation is performed on cough sound recordings from patients presenting various clinically diagnosed respiratory tract infections divided into croup and non-croup. The dataset is divided into training and test sets of 364 and 115 patients, respectively, with automatically segmented cough sound segments. RESULTS: Croup and non-croup patient classification on the test dataset with the proposed methods achieve a sensitivity and specificity of 92.31% and 85.29%, respectively. CONCLUSION: Experimental results show the significant improvement in automatic croup diagnosis against earlier methods. SIGNIFICANCE: This paper has the potential to automate croup diagnosis based solely on cough sound analysis.

Predicting spirometry readings using cough sound features and regression.

OBJECTIVE: Spirometry is a commonly used method of measuring lung function. It is useful in the definitive diagnosis of diseases such as asthma and chronic obstructive pulmonary disease (COPD). However, spirometry requires cooperative patients, experienced staff, and repeated testing to ensure the consistency of measurements. There is discomfort associated with spirometry and some patients are not able to complete the test. In this paper, we investigate the possibility of using cough sound analysis for the prediction of spirometry measurements. APPROACH: Our approach is based on the premise that the mechanism of cough generation and the forced expiratory maneuver of spirometry share sufficient similarities enabling this prediction. Using an iPhone, we collected mostly voluntary cough sounds from 322 adults presenting to a respiratory function laboratory for pulmonary function testing. Subjects had the following diagnoses: obstructive, restrictive, or mixed pattern diseases, or were found to have no lung disease along with normal spirometry. The cough sounds were automatically segmented using the algorithm described in Sharan et al (2018 IEEE Trans. Biomed. Eng.). We then represented cough sounds with various cough sound descriptors and built linear and nonlinear regression models connecting them to spirometry parameters. Augmentation of cough features with subject demographic data is also experimented with. The dataset was divided into 272 training subjects and 50 test subjects for experimentation. MAIN RESULTS: The performance of the auto-segmentation algorithm was evaluated on 49 randomly selected subjects from the overall dataset with a sensitivity and PPV of 84.95% and 98.51%, respectively. Our regression models achieved a root mean square error (and correlation coefficient) for standard spirometry parameters FEV1, FVC, and FEV1/FVC of 0.593L (0.810), 0.725L (0.749), and 0.164 (0.547), respectively, on the test dataset. In addition, we could achieve sensitivity, specificity, and accuracy of 70% or higher by applying the GOLD standard for COPD diagnosis on the estimated spirometry test results. SIGNIFICANCE: The experimental results show high positive correlation in predicting FEV1 and FVC and moderate positive correlation in predicting FEV1/FVC. The results show possibility of predicting spirometry results using cough sound analysis.

Snore Sound Analysis Can Detect the Presence of Obstructive Sleep Apnea Specific to NREM or REM Sleep.

STUDY OBJECTIVES: Severities of obstructive sleep apnea (OSA) estimated both for the overall sleep duration and for the time spent in rapid eye movement (REM) and non-rapid eye movement (NREM) sleep are important in managing the disease. The objective of this study is to investigate a method by which snore sounds can be analyzed to detect the presence of OSA in NREM and REM sleep. METHODS: Using bedside microphones, snoring and breathing-related sounds were acquired from 91 patients with OSA (35 females and 56 males) undergoing routine diagnostic polysomnography studies. A previously developed automated mathematical algorithm was applied to label each snore sound as belonging to either NREM or REM sleep. The snore sounds were then used to compute a set of mathematical features characteristic to OSA and to train a logistic regression model (LRM) to classify patients into an OSA or non-OSA category in each sleep state. The performance of the LRM was estimated using a leave-one-patient-out cross-validation technique within the entire dataset. We used the polysomnography-based diagnosis as our reference method. RESULTS: The models achieved 80% to 86% accuracy for detecting OSA in NREM sleep and 82% to 85% in REM sleep. When separate models were developed for females and males, the accuracy for detecting OSA in NREM sleep was 91% in females and 88% to 89% in males. Accuracy for detecting OSA in REM sleep was 88% to 91% in females and 89% to 91% in males. CONCLUSIONS: Snore sounds carry sufficient information to detect the presence of OSA during NREM and REM sleep. Because the methods used include technology that is fully automated and sensors that do not have a physical connection to the patient, it has potential for OSA screening in the home environment. The accuracy of the method can be improved by developing sex-specific models.

Automatic picking of snore events from overnight breath sound recordings.

Snoring is one of the earliest symptoms of Obstructive Sleep Apnea (OSA). However, the unavailability of an objective snore definition is a major obstacle in developing automated snore analysis system for OSA screening. The objectives of this paper is to develop a method to identify and extract snore sounds from a continuous sound recording following an objective definition of snore that is independent of snore loudness. Nocturnal sounds from 34 subjects were recorded using a non-contact microphone and computerized data-acquisition system. Sound data were divided into non-overlapping training (n = 21) and testing (n = 13) datasets. Using training dataset an Artificial Neural Network (ANN) classifier were trained for snore and non-snore classification. Snore sounds were defined based on the key observation that sounds perceived as `snores' by human are characterized by repetitive packets of energy that are responsible for creating the vibratory sound peculiar to snorers. On the testing dataset, the accuracy of ANN classifier ranged between 86 - 89%. Our results indicate that it is possible to define snoring using loudness independent, objective criteria, and develop automated snore identification and extraction algorithms.

Cough sound analysis for diagnosing croup in pediatric patients using biologically inspired features.

This paper aims to diagnose croup in children using cough sound signal classification. It proposes the use of a time-frequency image-based feature, referred as the cochleagram image feature (CIF). Unlike the conventional spectrogram image, the cochleagram utilizes a gammatone filter which models the frequency selectivity property of the human cochlea. This helps reveal more spectral information in the time-frequency image making it more useful for feature extraction. The cochleagram image is then divided into blocks and central moments are extracted as features. Classification is performed using logistic regression model (LRM) and support vector machine (SVM) on a comprehensive real-world cough sound signal database containing 364 patients with various clinically diagnosed respiratory tract infections divided into croup and non-croup. The best results, sensitivity of 88.37% and specificity of 91.59%, are achieved using SVM classification on a combined feature set of CIF and the conventional mel-frequency cepstral coefficients (MFCCs).

Hidden Markov modelling of intra-snore episode behavior of acoustic characteristics of obstructive sleep apnea patients.

Obstructive sleep apnea (OSA) is a breathing disorder that can cause serious medical consequences. It is caused by full (apnea) or partial (hypopnea) obstructions of the upper airway during sleep. The gold standard for diagnosis of OSA is the polysomnography (PSG). The main measure for OSA diagnosis is the apnea-hypopnea index (AHI). However, the AHI is a time averaged summary measure of vast amounts of information gathered in an overnight PSG study. It cannot capture the dynamic characteristics associated with apnea/hypopnea events and their overnight distribution. The dynamic characteristics of apnea/hypopnea events are affected by the structural and functional characteristics of the upper airway. The upper airway characteristics also affect the upper airway collapsibility. These effects are manifested in snoring sounds generated from the vibrations of upper airway structures which are then modified by the upper airway geometric and physical characteristics. Hence, it is highly likely that the acoustical behavior of snoring is affected by the upper airway structural and functional characteristics. In the current work, we propose a novel method to model the intra-snore episode behavior of the acoustic characteristics of snoring sounds which can indirectly describe the instantaneous and temporal dynamics of the upper airway. We model the intra-snore episode acoustical behavior by using hidden Markov models (HMMs) with Mel frequency cepstral coefficients. Assuming significant differences in the anatomical and physiological upper airway configurations between low-AHI and high-AHI subjects, we defined different snorer groups with respect to AHI thresholds 15 and 30 and also developed HMM-based classifiers to classify snore episodes into those groups. We also define a measure called instantaneous apneaness score (IAS) in terms of the log-likelihoods produced by respective HMMs. IAS indicates the degree of class membership of each episode to one of the predefined groups as well as the instantaneous OSA severity. We then assigned each patient to an overall AHI band based on the majority vote of each episode of snoring. The proposed method has a diagnostic sensitivity and specificity between 87-91%.

Wavelet augmented cough analysis for rapid childhood pneumonia diagnosis.

Pneumonia is the cause of death for over a million children each year around the world, largely in resource poor regions such as sub-Saharan Africa and remote Asia. One of the biggest challenges faced by pneumonia endemic countries is the absence of a field deployable diagnostic tool that is rapid, low-cost and accurate. In this paper, we address this issue and propose a method to screen pneumonia based on the mathematical analysis of cough sounds. In particular, we propose a novel cough feature inspired by wavelet-based crackle detection work in lung sound analysis. These features are then combined with other mathematical features to develop an automated machine classifier, which can separate pneumonia from a range of other respiratory diseases. Both cough and crackles are symptoms of pneumonia, but their existence alone is not a specific enough marker of the disease. In this paper, we hypothesize that the mathematical analysis of cough sounds allows us to diagnose pneumonia with sufficient sensitivity and specificity. Using a bedside microphone, we collected 815 cough sounds from 91 patients with respiratory illnesses such as pneumonia, asthma, and bronchitis. We extracted wavelet features from cough sounds and combined them with other features such as Mel Cepstral coefficients and non-Gaussianity index. We then trained a logistic regression classifier to separate pneumonia from other diseases. As the reference standard, we used the diagnosis by physicians aided with laboratory and radiological results as deemed necessary for a clinical decision. The methods proposed in this paper achieved a sensitivity and specificity of 94% and 63%, respectively, in separating pneumonia patients from non-pneumonia patients based on wavelet features alone. Combining the wavelets with features from our previous work improves the performance further to 94% and 88% sensitivity and specificity. The performance far surpasses that of the WHO criteria currently in common use in resource-limited settings.

Signal shape feature for automatic snore and breathing sounds classification.

Snore analysis techniques have recently been developed for sleep studies. Most snore analysis techniques require reliable methods for the automatic classification of snore and breathing sounds in the sound recording. In this study we focus on this problem and propose an automated method to classify snore and breathing sounds based on the novel feature, 'positive/negative amplitude ratio (PNAR)', to measure the shape of the sound signal. The performance of the proposed method was evaluated using snore and breathing recordings (snore: 22,643 episodes and breathing: 4664 episodes) from 40 subjects. Receiver operating characteristic (ROC) analysis showed that the proposed method achieved 0.923 sensitivity with 0.918 specificity for snore and breathing sound classification on test data. PNAR has substantial potential as a feature in the front end of a non-contact snore/breathing-based technology for sleep studies.

HMM-based snorer group recognition for Sleep Apnea diagnosis.

This paper presents an Hidden Markov Models (HMM)-based snorer group recognition approach for Obstructive Sleep Apenea diagnosis. It models the spatio-temporal characteristics of different snorer groups belonging to different genders and AHI severity levels. The current experiment includes selecting snore data from subjects, identifying snorer groups based on gender and AHI values (AHI < 15 and AHI > 15), detecting snore episodes, MFCC computation, training and testing HMMs. A set of multi-level classification rules is employed for incremental diagnosis of OSA. The proposed method, with a relatively small data set, produces results nearly comparable to any existing methods with single feature class. It classifies snore episodes with 62.0% (male), 67.0% (female) and recognizes snorer group with 78.5% accuracy. The approach makes its diagnosis decision at 85.7% (sensitivity), 71.4% (specificity) for males and 85.7% (sensitivity and specificity) for females.

ARMA-based spectral bandwidth for evaluation of bowel motility by the analysis of bowel sounds.

Approximately 10%-20% of adults and adolescents suffer from irritable bowel syndrome (IBS) worldwide. IBS is characterized by chronic gastrointestinal dysfunction which may reflect in altered motility. Currently, the diagnosis of IBS is made through expensive invasive radiographic and endoscopic examinations. However these are inconvenient and unsuited for community screening. Bowel sounds (BSs) can be easily recorded with non-invasive and low-cost equipment. Recently, several researchers have pointed out changes in features obtained from BS according to the pathological condition of bowel motility. However a widely accepted, simple automatic BS detection algorithm still has to be found, and the appropriate recording period needs to be investigated for further evaluation of bowel motility. In this study we propose a novel simple automatic method to detect the BSs based on the 3 dB bandwidth of the frequency peaks in the autoregressive moving average spectrum. We use the measure, sound-to-sound interval (SSI) obtained by the proposed method, to capture bowel motility. In this paper, we show that the proposed method for automatic detection could achieve a sensitivity of 87.8±5.88%, specificity of 91.7±4.33% and area under the curve of 0.923 when working on 16 healthy volunteers during mosapride administrations. Furthermore, we show that the measured SSI averaged over a period of 30 min can clearly capture bowel motility. Our findings should have the potential to contribute toward developing automated BS-based diagnosis of IBS.

Cough sound analysis can rapidly diagnose childhood pneumonia.

Pneumonia annually kills over 1,800,000 children throughout the world. The vast majority of these deaths occur in resource poor regions such as the sub-Saharan Africa and remote Asia. Prompt diagnosis and proper treatment are essential to prevent these unnecessary deaths. The reliable diagnosis of childhood pneumonia in remote regions is fraught with difficulties arising from the lack of field-deployable imaging and laboratory facilities as well as the scarcity of trained community healthcare workers. In this paper, we present a pioneering class of technology addressing both of these problems. Our approach is centred on the automated analysis of cough and respiratory sounds, collected via microphones that do not require physical contact with subjects. Cough is a cardinal symptom of pneumonia but the current clinical routines used in remote settings do not make use of coughs beyond noting its existence as a screening-in criterion. We hypothesized that cough carries vital information to diagnose pneumonia, and developed mathematical features and a pattern classifier system suited for the task. We collected cough sounds from 91 patients suspected of acute respiratory illness such as pneumonia, bronchiolitis and asthma. Non-contact microphones kept by the patient's bedside were used for data acquisition. We extracted features such as non-Gaussianity and Mel Cepstra from cough sounds and used them to train a Logistic Regression classifier. We used the clinical diagnosis provided by the paediatric respiratory clinician as the gold standard to train and validate our classifier. The methods proposed in this paper could separate pneumonia from other diseases at a sensitivity and specificity of 94 and 75% respectively, based on parameters extracted from cough sounds alone. The inclusion of other simple measurements such as the presence of fever further increased the performance. These results show that cough sounds indeed carry critical information on the lower respiratory tract, and can be used to diagnose pneumonia. The performance of our method is far superior to those of existing WHO clinical algorithms for resource-poor regions. To the best of our knowledge, this is the first attempt in the world to diagnose pneumonia in humans using cough sound analysis. Our method has the potential to revolutionize the management of childhood pneumonia in remote regions of the world.

Automatic identification of wet and dry cough in pediatric patients with respiratory diseases.

Cough is the most common symptom of several respiratory diseases. It is a defense mechanism of the body to clear the respiratory tract from foreign materials inhaled accidentally or produced internally by infections. The identification of wet and dry cough is an important clinical finding, aiding in the differential diagnosis especially in children. Wet coughs are more likely to be associated with lower respiratory track bacterial infections. At present during a typical consultation session, the wet/dry decision is based on the subjective judgment of a physician. It is not available for the non-trained person, long term monitoring or in the assessment of treatment efficacy. In this paper we address these issues and develop an automated technology to classify cough into 'wet' and 'dry' categories. We propose novel features and a Logistic regression model (LRM) for the classification of coughs into wet/dry classes. The performance of the method was evaluated on a clinical database of pediatric coughs (C = 536) recorded using a bed-side non-contact microphone from N = 78 patients. Results of the automatic classification were compared against two expert human scorers. The sensitivity and specificity of the LRM in picking wet coughs were between 87 and 88% with 95% confidence interval on training/validation dataset (310 cough events from 60 patients) and 84 and 76% respectively on prospective dataset (117 cough events from 18 patients). The kappa agreement with two expert human scorers on prospective dataset was 0.51. These results indicate the potential of the method as a useful clinical tool for cough monitoring, especially at home settings.

Artificial neural networks for breathing and snoring episode detection in sleep sounds.

Obstructive sleep apnea (OSA) is a serious disorder characterized by intermittent events of upper airway collapse during sleep. Snoring is the most common nocturnal symptom of OSA. Almost all OSA patients snore, but not all snorers have the disease. Recently, researchers have attempted to develop automated snore analysis technology for the purpose of OSA diagnosis. These technologies commonly require, as the first step, the automated identification of snore/breathing episodes (SBE) in sleep sound recordings. Snore intensity may occupy a wide dynamic range (> 95 dB) spanning from the barely audible to loud sounds. Low-intensity SBE sounds are sometimes seen buried within the background noise floor, even in high-fidelity sound recordings made within a sleep laboratory. The complexity of SBE sounds makes it a challenging task to develop automated snore segmentation algorithms, especially in the presence of background noise. In this paper, we propose a fundamentally novel approach based on artificial neural network (ANN) technology to detect SBEs. Working on clinical data, we show that the proposed method can detect SBE at a sensitivity and specificity exceeding 0.892 and 0.874 respectively, even when the signal is completely buried in background noise (SNR < 0 dB). We compare the performance of the proposed technology with those of the existing methods (short-term energy, zero-crossing rates) and illustrate that the proposed method vastly outperforms conventional techniques.

Multi-feature snore sound analysis in obstructive sleep apnea-hypopnea syndrome.

Snoring is the most common symptom of obstructive sleep apnea hypopnea syndrome (OSAHS), which is a serious disease with high community prevalence. The standard method of OSAHS diagnosis, known as polysomnography (PSG), is expensive and time consuming. There is evidence suggesting that snore-related sounds (SRS) carry sufficient information to diagnose OSAHS. In this paper we present a technique for diagnosing OSAHS based solely on snore sound analysis. The method comprises a logistic regression model fed with snore parameters derived from its features such as the pitch and total airway response (TAR) estimated using a higher order statistics (HOS)-based algorithm. Pitch represents a time domain characteristic of the airway vibrations and the TAR represents the acoustical changes brought about by the collapsing upper airways. The performance of the proposed method was evaluated using the technique of K-fold cross validation, on a clinical database consisting of overnight snoring sounds of 41 subjects. The method achieved 89.3% sensitivity with 92.3% specificity (the area under the ROC curve was 0.96). These results establish the feasibility of developing a snore-based OSAHS community-screening device, which does not require any contact measurements.

Interhemispheric asynchrony correlates with severity of respiratory disturbance index in patients with sleep apnea.

Obstructive sleep apnea (OSA) hypopnea syndrome is a disorder characterized by airway obstructions during sleep; full obstructions are known as apnea and partial obstructions are called hypopnea. Sleep in OSA patients is significantly disturbed with frequent apnea/hypopnea and arousal events. We illustrate that these events lead to functional asymmetry of the brain as manifested by the interhemispheric asynchrony (IHA) computed using EEG recorded on the scalp. In this paper, based on the higher order spectra of IHA time series, we propose a new index [interhemispheric synchrony index (IHSI)], for characterizing brain asynchrony in OSA. The IHSI computation does not depend on subjective criteria and can be completely automated. The proposed method was evaluated on overnight EEG data from a clinical database of 36 subjects referred to a hospital sleep laboratory. Our results indicated that the IHSI could classify the patients into OSA/non-OSA classes with an accuracy of 91% (ρ = 0.0001), at the respiratory disturbance index threshold of 10, suggesting that the brain asynchrony carries vital information on OSA.