The synthesis of the aortic valve closure sound of the dog by the mean filter of forward and backward predictor

Synthesis of Normal Heart Sounds Using Generative Adversarial Networks and Empirical Wavelet Transform

Currently, there are many works in the literature focused on the analysis of heart sounds, specifically on the development of intelligent systems for the classification of normal and abnormal heart sounds. However, the available heart sound databases are not yet large enough to train generalized machine learning models. Therefore, there is interest in the development of algorithms capable of generating heart sounds that could augment current databases. In this article, we propose a model based on generative adversary networks (GANs) to generate normal synthetic heart sounds. Additionally, a denoising algorithm is implemented using the empirical wavelet transform (EWT), allowing a decrease in the number of epochs and the computational cost that the GAN model requires. A distortion metric (mel–cepstral distortion) was used to objectively assess the quality of synthetic heart sounds. The proposed method was favorably compared with a mathematical model that is based on the morphology of the phonocardiography (PCG) signal published as the state of the art. Additionally, different heart sound classification models proposed as state-of-the-art were also used to test the performance of such models when the GAN-generated synthetic signals were used as test dataset. In this experiment, good accuracy results were obtained with most of the implemented models, suggesting that the GAN-generated sounds correctly capture the characteristics of natural heart sounds.

An Approach for Automatic Identification of Fundamental and Additional Sounds from Cardiac Sounds Recordings

This paper presents an approach for automatic segmentation of cardiac events from non-invasive sounds recordings, without the need of having an auxiliary signal reference. In addition, methods are proposed to subsequently differentiate cardiac events which correspond to normal cardiac cycles, from those which are due to abnormal activity of the heart. The detection of abnormal sounds is based on a model built with parameters which are obtained following feature extraction from those segments that were previously identified as normal fundamental heart sounds. The proposed algorithm achieved a sensitivity of 91.79% and 89.23% for the identification of normal fundamental, S₁ and S₂ sounds, and a true positive (TP) rate of 81.48% for abnormal additional sounds. These results were obtained using the PASCAL Classifying Heart Sounds challenge (CHSC) database.

Multilayer 3D electrodes for neural implants

OBJECTIVE

In many applications, multielectrode arrays employed as neural implants require a high density and a high number of electrodes to precisely record and stimulate the activity of the nervous system while preserving the overall size of the array.

APPROACH

Here we present a multilayer and three-dimensional (3D) electrode array, together with its manufacturing method, enabling a higher electrode density and a more efficient signal transduction with the biological tissue.

MAIN RESULTS

The 3D structure of the electrode array allows for a multilayer placement of the interconnects within a flexible substrate, it narrows the probe size per the same number of electrodes, and it maintains the electrode contacts at the same level within the tissue. In addition, it augments the electrode surface area, leading to a lower electrochemical impedance and a higher charge storage capacity. To characterize the recordings capabilities of the multilayer 3D electrodes, we measured visually evoked cortical potentials in mice and analysed the evolution of the peak prominences and latencies according to different light intensities and recording depths within the brain. The resulting signal-to-noise ratio is improved compared to flat electrodes. Finally, the 3D electrodes have been imaged inside a clarified mouse brain using a light-sheet microscope to visualize their integrity within the tissue.

SIGNIFICANCE

The multilayer 3D electrodes have proved to be a valid technology to ensure tissue proximity and higher recording/stimulating efficiencies while enabling higher electrode density and reducing the probe size.

Detection of the valvular split within the second heart sound using the reassigned smoothed pseudo Wigner-Ville distribution

Background

In this paper, we developed a novel algorithm to detect the valvular split between the aortic and pulmonary components in the second heart sound which is a valuable medical information.

Methods

The algorithm is based on the Reassigned smoothed pseudo Wigner–Ville distribution which is a modified time–frequency distribution of the Wigner–Ville distribution. A preprocessing amplitude recovery procedure is carried out on the analysed heart sound to improve the readability of the time–frequency representation. The simulated S2 heart sounds were generated by an overlapping frequency modulated chirp–based model at different valvular split durations.

Results

Simulated and real heart sounds are processed to highlight the performance of the proposed approach. The algorithm is also validated on real heart sounds of the LGB–IRCM (Laboratoire de Génie biomédical–Institut de recherches cliniques de Montréal) cardiac valve database. The A2–P2 valvular split is accurately detected by processing the obtained RSPWVD representations for both simulated and real data.

Modeling of heart systolic murmurs based on multivariate matching pursuit for diagnosis of valvular disorders

Heart murmurs are pathological sounds produced by turbulent blood flow due to certain cardiac defects such as valves disorders. Detection of murmurs via auscultation is a task that depends on the proficiency of physician. There are many cases in which the accuracy of detection is questionable. The purpose of this study is development of a new mathematical model of systolic murmurs to extract their crucial features for identifying the heart diseases. A high resolution algorithm, multivariate matching pursuit, was used to model the murmurs by decomposing them into a series of parametric time-frequency atoms. Then, a novel model-based feature extraction method which uses the model parameters was performed to identify the cardiac sound signals. The proposed framework was applied to a database of 70 heart sound signals containing 35 normal and 35 abnormal samples. We achieved 92.5% accuracy in distinguishing subjects with valvular diseases using a MLP classifier, as compared to the matching pursuit-based features with an accuracy of 77.5%.

Time-frequency analysis of phonocardiogram signals using wavelet transform: a comparative study

Analysis of phonocardiogram (PCG) signals provides a non-invasive means to determine the abnormalities caused by cardiovascular system pathology. In general, time-frequency representation (TFR) methods are used to study the PCG signal because it is one of the non-stationary bio-signals. The continuous wavelet transform (CWT) is especially suitable for the analysis of non-stationary signals and to obtain the TFR, due to its high resolution, both in time and in frequency and has recently become a favourite tool. It decomposes a signal in terms of elementary contributions called wavelets, which are shifted and dilated copies of a fixed mother wavelet function, and yields a joint TFR. Although the basic characteristics of the wavelets are similar, each type of the wavelets produces a different TFR. In this study, eight real types of the most known wavelets are examined on typical PCG signals indicating heart abnormalities in order to determine the best wavelet to obtain a reliable TFR. For this purpose, the wavelet energy and frequency spectrum estimations based on the CWT and the spectra of the chosen wavelets were compared with the energy distribution and the autoregressive frequency spectra in order to determine the most suitable wavelet. The results show that Morlet wavelet is the most reliable wavelet for the time-frequency analysis of PCG signals.

Nonlinear transient chirp signal modeling of the aortic and pulmonary components of the second heart sound

This paper describes a new approach based on the time-frequency representation of transient nonlinear chirp signals for modeling the aortic (A2) and the pulmonary (P2) components of the second heart sound (S2). It is demonstrated that each component is a narrow-band signal with decreasing instantaneous frequency defined by its instantaneous amplitude and its instantaneous phase. Each component is also a polynomial phase signal, the instantaneous phase of which can be accurately represented by a polynomial having an order of thirty. A dechirping approach is used to obtain the instantaneous amplitude of each component while reducing the effect of the background noise. The analysis-synthesis procedure is applied to 32 isolated A2 and 32 isolated P2 components recorded in four pigs with pulmonary hypertension. The mean +/- standard deviation of the normalized root-mean-squared error (NRMSE) and the correlation coefficient (rho) between the original and the synthesized signal components were: NRMSE = 2.1 +/- 0.3% and rho = 0.97 +/- 0.02 for A2 and NRMSE = 2.52 +/- 0.5% and rho = 0.96 +/- 0.02 for P2. These results confirm that each component can be modeled as mono-component nonlinear chirp signals of short duration with energy distributions concentrated along its decreasing instantaneous frequency.

Analysis-synthesis of the phonocardiogram based on the matching pursuit method

The matching pursuit method of Mallat and Zhang is applied to the analysis and synthesis of phonocardiograms (PCG's). The method is based on a classical Gabor wavelet or time-frequency atom which is the product of a sinusoid and a Gaussian window function. It decomposes a signal into a series of time-frequency atoms by an iterative process based on selecting the largest inner product of the signal (and the subsequent residues) with atoms from a redundant dictionary. The Gaussian window controls the envelope duration and time position of each atom; and the sinusoid represents the frequency. The method was applied to two sets of PCG's: one with very low-noise level and the other with 10% noise energy. Each data base includes 11 PCG's representing the normal and the pathological conditions of the heart. The normalized root-mean-square error (NRMSE) was computed between the original and the reconstructed signals. The results show that the matching pursuit method is very suitable to the transient and complex properties of the PCG's, as it yielded excellent NRMSE's around 2.2% for the two sets of 11 PCG's tested.

Time-frequency analysis of the first heart sound. Part 1: Simulation and analysis

The authors propose a simulated first heart sound (S1) signal that can be used as a reference signal to evaluate the accuracy of time-frequency representation techniques for studying multicomponent signals. The composition of this simulated S1 is based on the hypothesis that an S1 recorded on the thorax over the apical area of the heart is composed of constant frequency vibrations from the mitral valve and a frequency modulated vibration from the myocardium. Essentially, the simulated S1 consists of a valvular component and a myocardial component. The valvular component is modelled as two exponentially decaying sinusoids of 50 Hz and 150 Hz and the myocardial component is modelled by a frequency modulated wave between 20 Hz and 100 Hz. The study shows that the simulated S1 has temporal and spectral characteristics similar to S1 recorded in humans and dogs. It also shows that the spectrogram cannot resolve the three components of the simulated S1. It is concluded that it is necessary to search for a better time-frequency representation technique for studying the time-frequency distribution of multicomponent signals such as the simulated S1.

Spectral composition of heart sounds before and after mechanical heart valve implantation using a modified forward-backward Prony's method

This paper investigates the impact of the lung-thorax and heart-valve system on the overall spectral composition of the externally recorded heart sounds. The study concentrates in the case of the first and the second heart sounds for normal patients and patients before and after implantation of a mechanical valve in the mitral or aortic position. The analysis is performed using a modified forward-backward overdetermined Prony's method (MFBPM) which uses a forward-backward mean filter and a modified procedure for estimating the position of the signal poles. In terms of the normalized cross-correlation coefficient, this method has an average modeling accuracy of 99.62% for representing the first and second heart sounds and an average least square time-domain error of 0.43%. Results obtained from 40 subjects show that the condition of the native mitral or aortic valve affects mostly the distribution of the amplitudes of the spectral components, whereas the number of the spectral components or their respective relative energy remains more or less unchanged. It has been found that the amplitudes of frequency components in the range 120-250 Hz are more affected by abnormalities of native mitral valves. Furthermore, in the case of the second heart sound the region 250-400 Hz has been found to be more affected by abnormalities in the aortic valve. It has also been found that the mechanical prosthetic heart valve affects mostly the spectrum beyond 400 Hz. A clear difference has been observed in the frequency spectrum above 400 Hz between both normally and abnormally functioning native valves and normally functioning mechanical valves. Preliminary results in some malfunctioning cases of mechanical prosthesis suggest that spectral components beyond 400 Hz can be used to monitor the condition of these prostheses.