Detection of heart murmurs using wavelet analysis and artificial neural networks

This paper presents the algorithm and technical aspects of an intelligent diagnostic system for the detection of heart murmurs. The purpose of this research is to address the lack of effectively accurate cardiac auscultation present at the primary care physician office by development of an algorithm capable of operating within the hectic environment of the primary care office. The proposed algorithm consists of three main stages. First; denoising of input data (digital recordings of heart sounds), via Wavelet Packet Analysis. Second; input vector preparation through the use of Principal Component Analysis and block processing. Third; classification of the heart sound using an Artificial Neural Network. Initial testing revealed the intelligent diagnostic system can differentiate between normal healthy heart sounds and abnormal heart sounds (e.g., murmurs), with a specificity of 70.5% and a sensitivity of 64.7%.     

A robust heart sounds segmentation module based on S-transform

This paper presents a new module for heart sounds segmentation based on S-transform. The heart sounds segmentation process segments the PhonoCardioGram (PCG) signal into four parts: S1 (first heart sound), systole, S2 (second heart sound) and diastole. It can be considered one of the most important phases in the auto-analysis of PCG signals. The proposed segmentation module can be divided into three main blocks: localization of heart sounds, boundaries detection of the localized heart sounds and classification block to distinguish between S1 and S2. An original localization method of heart sounds are proposed in this study. The method named SSE calculates the Shannon energy of the local spectrum calculated by the S-transform for each sample of the heart sound signal. The second block contains a novel approach for the boundaries detection of S1 and S2. The energy concentrations of the S-transform of localized sounds are optimized by using a window width optimization algorithm. Then the SSE envelope is recalculated and a local adaptive threshold is applied to refine the estimated boundaries. To distinguish between S1 and S2, a feature extraction method based on the singular value decomposition (SVD) of the S-matrix is applied in this study. The proposed segmentation module is evaluated at each block according to a database of 80 sounds, including 40 sounds with cardiac pathologies.     

A generating model of realistic synthetic heart sounds for performance assessment of phonocardiogram processing algorithms

A new model which is capable of generating realistic synthetic phonocardiogram (PCG) signals is introduced based on three coupled ordinary differential equations. The new PCG model takes into account the respiratory frequency, the heart rate variability and the time splitting of first and second heart sounds. This time splitting occurs with each cardiac cycle and varies with inhalation and exhalation. Clinical PCG statistics and the close temporal relationship between events in ECG and PCG are used to deduce values of PCG model parameters.In comparison with published PCG models, the proposed model allows a larger number of known PCG features to be taken into consideration. Moreover it is able to generate both normal and abnormal realistic synthetic heart sounds. Results show that these synthetic PCG signals have the closest features to those of a conventional heart sound in both time and frequency domains. Additionally, a sound quality test carried out by eight cardiologists demonstrates that the proposed model outperforms the existing models.This new PCG model is promising and useful in assessing signal processing techniques which are developed to help clinical diagnosis based on PCG.     

Heart energy signature spectrogram for cardiovascular diagnosis

A new method and application is proposed to characterize intensity and pitch of human heart sounds and murmurs. Using recorded heart sounds from the library of one of the authors, a visual map of heart sound energy was established. Both normal and abnormal heart sound recordings were studied. Representation is based on Wigner-Ville joint time-frequency transformations. The proposed methodology separates acoustic contributions of cardiac events simultaneously in pitch, time and energy. The resolution accuracy is superior to any other existing spectrogram method. The characteristic energy signature of the innocent heart murmur in a child with the S3 sound is presented. It allows clear detection of S1, S2 and S3 sounds, S2 split, systolic murmur, and intensity of these components. The original signal, heart sound power change with time, time-averaged frequency, energy density spectra and instantaneous variations of power and frequency/pitch with time, are presented. These data allow full quantitative characterization of heart sounds and murmurs. High accuracy in both time and pitch resolution is demonstrated. Resulting visual images have self-referencing quality, whereby individual features and their changes become immediately obvious.     

Ultrasonic pregnancy diagnosis in the camel

Application of ultrasound for pregnancy diagnosis has been tested and evaluated in 15 Iranian camels (Camelus dromedarius), all of which ultimately calved. Transabdominal examinations were unsuccessful, while intrapelvic application resulted in the reception of sounds characteristic for foetal life, similar to those found in other domestic animals. Signals of foetal heart, pulse of umbilical vessels and uterine artery as well as foetal movement could be recognized as distinct sounds and have been recorded for further studies. An attempt was made to verify the findings of the ultrasonic diagnosis through rectal palpation. The ultrasonic technique resulted in 12 correct and three incorrect diagnoses.     

Segmentation of heart sounds based on dynamic clustering

The heart sound signal is first separated into cycles, where the cycle detection is based on an instantaneous cycle frequency. The heart sound data of one cardiac cycle can be decomposed into a number of atoms characterized by timing delay, frequency, amplitude, time width and phase. To segment heart sounds, we made a hypothesis that the atoms of a heart sound congregate as a cluster in time–frequency domains. We propose an atom density function to indicate clusters. To suppress clusters of murmurs and noise, weighted density function by atom energy is further proposed to improve the segmentation of heart sounds. Therefore, heart sounds are indicated by the hybrid analysis of clustering and medical knowledge. The segmentation scheme is automatic and no reference signal is needed. Twenty-six subjects, including 3 normal and 23 abnormal subjects, were tested for heart sound signals in various clinical cases. Our statistics show that the segmentation was successful for signals collected from normal subjects and patients with moderate murmurs.     

XIVth SYMPOSIUM OF THE INTERNATIONAL BIOACOUSTICS COUNCIL

ABSTRACT

This is the first reported study of corixid water bugs examining whether all species of a genus in one locality can be distinguished by their sounds. More extensive analysis than has been reported for any corixids revealed that, although some species are difficult to distinguish morphologically, inter-species sound differences are very clear.

The sounds of all nine species of Micronecta in the study area near Melbourne, Australia were recorded. Male sounds were recorded in the laboratory, over a minimum water temperature range of 15 to 25°C. Females do not produce sounds. Signals consisted of groups of pulse-trains, except for one species with signals of usually one pulse-train. Signals were species-specific; pulse-train rate alone was sufficient to distinguish between species. There were also species differences in other signal parameters. Males also produced clicks (single pulse-trains) and low-amplitude sounds; there were some species differences in the latter. Similar signals occurred between only one pair of species, which were from different habitats (ponds and rivers). Pulse periods and pulse-train periods were negatively correlated with temperature, with curves of best fit being quadratic. Five species were also recorded in ponds; the sounds and effect of temperature were compared with laboratory recordings.     

Probing real sensory worlds of receivers with unsupervised clustering

The task of an organism to extract information about the external environment from sensory signals is based entirely on the analysis of ongoing afferent spike activity provided by the sense organs. We investigate the processing of auditory stimuli by an acoustic interneuron of insects. In contrast to most previous work we do this by using stimuli and neurophysiological recordings directly in the nocturnal tropical rainforest, where the insect communicates. Different from typical recordings in sound proof laboratories, strong environmental noise from multiple sound sources interferes with the perception of acoustic signals in these realistic scenarios. We apply a recently developed unsupervised machine learning algorithm based on probabilistic inference to find frequently occurring firing patterns in the response of the acoustic interneuron. We can thus ask how much information the central nervous system of the receiver can extract from bursts without ever being told which type and which variants of bursts are characteristic for particular stimuli. Our results show that the reliability of burst coding in the time domain is so high that identical stimuli lead to extremely similar spike pattern responses, even for different preparations on different dates, and even if one of the preparations is recorded outdoors and the other one in the sound proof lab. Simultaneous recordings in two preparations exposed to the same acoustic environment reveal that characteristics of burst patterns are largely preserved among individuals of the same species. Our study shows that burst coding can provide a reliable mechanism for acoustic insects to classify and discriminate signals under very noisy real-world conditions. This gives new insights into the neural mechanisms potentially used by bushcrickets to discriminate conspecific songs from sounds of predators in similar carrier frequency bands.     

Comparison of first and second heart sounds after mechanical heart valve replacement

In this article, the spectral features of first heart sounds (S1) and second heart sounds (S2), which comprise the mechanical heart valve sounds obtained after aortic valve replacement (AVR) and mitral valve replacement (MVR), are compared to find out the effect of mechanical heart valve replacement and recording area on S1 and S2. For this aim, the Welch method and the autoregressive (AR) method are applied on the S1 and S2 taken from 66 recordings of 8 patients with AVR and 98 recordings from 11 patients with MVR, thereby yielding power spectrum of the heart sounds. Three features relating to frequency of heart sounds and three features relating to energy of heart sounds are obtained. Results show that in comparison to natural heart valves, mechanical heart valves contain higher frequency components and energy, and energy and frequency components do not show common behaviour for either AVR or MVR depending on the recording areas. Aside from the frequency content and energy of the sound generated by mechanical heart valves being affected by the structure of the lungs–thorax and the recording areas, the pressure across the valve incurred during AVR or MVR is a significant factor in determining the frequency and energy levels of the valve sound produced. Though studies on native heart sounds as a non-invasive diagnostic method has been done for many years, it is observed that studies on mechanical heart valves sounds are limited. The results of this paper will contribute to other studies on using a non-invasive method for assessing the mechanical heart valve sounds.     

Categorization of environmental sounds

Sounds in the natural environment are non-stationary, in that their spectral dynamics is time-dependent. We develop measures to analyze the spectral dynamics of environmental sound signals and find that they fall into two categories—simple sounds with slowly varying spectral dynamics and complex sounds with rapidly varying spectral dynamics. Based on our results and those from auditory processing we suggest rate of spectral dynamics as a possible scheme to categorize sound signals in the environment.     

Geophone detection of subterranean termite and ant activity

A geophone system was used to monitor activity of subterranean termites and ants in a desert environment with low vibration noise. Examples of geophone signals were recorded from a colony of Rhytidoponera taurus (Forel), a colony of Camponotus denticulatus Kirby, and a termite colony (undetermined Drepanotermes sp.) under attack by ants from a nearby C. denticulatus colony. The geophone recordings were compared with signals recorded from accelerometers in a citrus grove containing Solenopsis invicta Buren workers. Because of their small size, all of these insects produce relatively weak sounds. Several different types of insect-generated sounds were identified in the geophone recordings, including high-frequency ticks produced by R. taurus and C. denticulatus, and patterned bursts of head bangs produced by Drepanotermes. The S. invicta produced bursts of ticks with three different stridulation frequencies, possibly produced by three different-sized workers. Overall, both systems performed well in enabling identification of high-frequency or patterned pulses. The geophone was more sensitive than the accelerometer to low-frequency signals, but low-frequency insect sound pulses are more difficult to distinguish from background noises than high-frequency pulses. The low cost of multiple-geophone systems may facilitate development of future applications for wide-area subterranean insect monitoring in quiet environments.     

Activation of Auditory Cortex by Anticipating and Hearing Emotional Sounds: An MEG Study

To study how auditory cortical processing is affected by anticipating and hearing of long emotional sounds, we recorded auditory evoked magnetic fields with a whole-scalp MEG device from 15 healthy adults who were listening to emotional or neutral sounds. Pleasant, unpleasant, or neutral sounds, each lasting for 6 s, were played in a random order, preceded by 100-ms cue tones (0.5, 1, or 2 kHz) 2 s before the onset of the sound. The cue tones, indicating the valence of the upcoming emotional sounds, evoked typical transient N100m responses in the auditory cortex. During the rest of the anticipation period (until the beginning of the emotional sound), auditory cortices of both hemispheres generated slow shifts of the same polarity as N100m. During anticipation, the relative strengths of the auditory-cortex signals depended on the upcoming sound: towards the end of the anticipation period the activity became stronger when the subject was anticipating emotional rather than neutral sounds. During the actual emotional and neutral sounds, sustained fields were predominant in the left hemisphere for all sounds. The measured DC MEG signals during both anticipation and hearing of emotional sounds implied that following the cue that indicates the valence of the upcoming sound, the auditory-cortex activity is modulated by the upcoming sound category during the anticipation period.     

基于盒维数的心音信号分形特征研究

在传统盒维数的基础上,从尺度变化的角度,提出一种计算心音信号时域波形分形维数的新的二进盒维数算法,并给出了算法思想和估算方法;然后用该方法对正常心音和几种典型的病态心音的分形维数进行计算,并对其分形特征进行了研究．研究结果表明：心音信号具有明显的分形特征,分形维数能够反映心音信号的复杂程度,并且能够明显地区分正常心音和病态心音．     

Familiarity with speech affects cortical processing of auditory distance cues and increases acuity

Several acoustic cues contribute to auditory distance estimation. Nonacoustic cues, including familiarity, may also play a role. We tested participants' ability to distinguish the distances of acoustically similar sounds that differed in familiarity. Participants were better able to judge the distances of familiar sounds. Electroencephalographic (EEG) recordings collected while participants performed this auditory distance judgment task revealed that several cortical regions responded in different ways depending on sound familiarity. Surprisingly, these differences were observed in auditory cortical regions as well as other cortical regions distributed throughout both hemispheres. These data suggest that learning about subtle, distance-dependent variations in complex speech sounds involves processing in a broad cortical network that contributes both to speech recognition and to how spatial information is extracted from speech.     

Analytical and numerical solutions of the potential and electric field generated by different electrode arrays in a tumor tissue under electrotherapy

Background

During the cardiac cycle, the heart normally produces repeatable physiological sounds. However, under pathologic conditions, such as with heart valve stenosis or a ventricular septal defect, blood flow turbulence leads to the production of additional sounds, called murmurs. Murmurs are random in nature, while the underlying heart sounds are not (being deterministic).

Innovation

We show that a new analytical technique, which we call Digital Subtraction Phonocardiography (DSP), can be used to separate the random murmur component of the phonocardiogram from the underlying deterministic heart sounds.

Methods

We digitally recorded the phonocardiogram from the anterior chest wall in 60 infants and adults using a high-speed USB interface and the program Gold Wave http://www.goldwave.com. The recordings included individuals with cardiac structural disease as well as recordings from normal individuals and from individuals with innocent heart murmurs. Digital Subtraction Analysis of the signal was performed using a custom computer program called Murmurgram. In essence, this program subtracts the recorded sound from two adjacent cardiac cycles to produce a difference signal, herein called a "murmurgram". Other software used included Spectrogram (Version 16), GoldWave (Version 5.55) as well as custom MATLAB code.

Results

Our preliminary data is presented as a series of eight cases. These cases show how advanced signal processing techniques can be used to separate heart sounds from murmurs. Note that these results are preliminary in that normal ranges for obtained test results have not yet been established.

Conclusions

Cardiac murmurs can be separated from underlying deterministic heart sounds using DSP. DSP has the potential to become a reliable and economical new diagnostic approach to screening for structural heart disease. However, DSP must be further evaluated in a large series of patients with well-characterized pathology to determine its clinical potential.     

The first heart sound during the isovolumetric contraction

We make the first attempt to construct a qualitative theory covering the whole process of the major part of the first heart sound from an electrical activation to the phonocardiographic observations at the thorax. We calculate the amplitudes and frequencies of the radiated pressures during the isovolumetric contraction period generated by the muscular wall of the left ventricle and by the valves considered as a spherical shell and two-dimensional membranes, respectively. The analysis shows that both the hemodynamic and the valvular theory are able to explain most of the characteristic features of the first heart sound (linear relation between the amplitudes of the radiated pressure and the slope of the left ventricular pressure-time curve; directional polarity of the amplitudes; equidistant frequency peaks with a decline in amplitudes). However, existing magnitudes of the set of physiological parameters involved seems to favour the hemodynamic theory of the first heart sound. The aortic valve can be neglected as a source of sound. The initial conditions (like valve closure velocity), according to our theory, cannot be important. The predicted time-plot and frequency spectrum of the radiated pressure show a general resemblance with the recorded ones. It is essential to have considerably more quantitative acoustic data both for normal and diseased hearts for subsequent theoretical development.     

A method for computerized modification of certain natural animal sounds for communication study purposes

A new method for computerized modification of sound signals is presented. With digital signal processing in the time domain it is possible to alter the amplitude, the frequency and the time scale of natural sounds independently. The method can be applied to natural sounds with reasonably pure tonal quality.     

Cluster analysis and classification of heart sounds

Acoustic heart signals, generated by the mechanical processes of the cardiac cycle, carry significant information about the underlying functioning of the cardiovascular system. We describe a computational analysis framework for identifying distinct morphologies of heart sounds and classifying them into physiological states. The analysis framework is based on hierarchical clustering, compact data representation in the feature space of cluster distances and a classification algorithm. We applied the proposed framework on two heart sound datasets, acquired during controlled alternations of the physiological conditions, and analyzed the morphological changes induced to the first heart sound (S1), and the ability to predict physiological variables from the morphology of S1. On the first dataset of 12 subjects, acquired while modulating the respiratory pressure, the algorithm achieved an average accuracy of 82 ± 7% in classifying the level of breathing resistance, and was able to estimate the instantaneous breathing pressure with an average error of 19 ± 6%. A strong correlation of 0.92 was obtained between the estimated and the actual breathing efforts. On the second dataset of 11 subjects, acquired during pharmacological stress tests, the average accuracy in classifying the stress stage was 86 ± 7%. The effects of the chosen raw signal representation, distance metrics and classification algorithm on the performance were studied on both real and simulated data. The results suggest that quantitative heart sound analysis may provide a new non-invasive technique for continuous cardiac monitoring and improved detection of mechanical dysfunctions caused by cardiovascular and cardiopulmonary diseases.     

A Quantitative Analysis of the Sounds of Hector's Dolphin

We developed an automatic, computer-based system in which digital signal processing techniques were used to measure 31 variables from digitized Hector's dolphin (Cephalorhynchus hectori) sounds. Principal component analyses of these data were used to investigate the relationships between sounds. Hector's dolphins make only a very few types of pulsed “clicks”, most of which are centred around 125 kHz. None of these had an average frequency of less than 82 kHz, and the only audible sounds were made up of high-frequency clicks repeated at such high rates that the repetition rate was audible to us as a tonal “cry” or “squeal”. In comparison to signal levels recorded from other cetaceans, all the Hector's dolphin signals were low-level; the maximum received sound pressure level was 163 dB (re 1μPa).