Extraction and assessment of diagnosis-relevant features for heart murmur classification.

This paper presents a heart murmur detection and multi-class classification approach via machine learning. We extracted heart sound and murmur features that are of diagnostic importance and developed additional 16 features that are not perceivable by human ears but are valuable to improve murmur classification accuracy. We examined and compared the classification performance of supervised machine learning with k-nearest neighbor (KNN) and support vector machine (SVM) algorithms. We put together a test repertoire having more than 450 heart sound and murmur episodes to evaluate the performance of murmur classification using cross-validation of 80-20 and 90-10 splits. As clearly demonstrated in our evaluation, the specific set of features chosen in our study resulted in accurate classification consistently exceeding 90% for both classifiers.

Chaotic Analysis of Hippocampal and Cortical Sleep EEG during Various Vigilance States.

In this paper, we investigate the hippocampal and cortical sleep EEG of adult rats at different sleep stages by employing Lyapunov exponent and third-order cumulant measures to quantify and compare the chaotic and nonlinear behavior of EEG obtained during vigilance states of quiet- waking, slow-wave sleep, and rapid eye movement (REM) sleep. Lyapunov exponent was computed to characterize the EEG for chaos and third-order cumulant was used to measure the deviations from Gaussianity of the signal. Our results show positive Lyapunov exponents for all EEG states indicating a Iow- dimensional chaos for both REM and non-REM system. Furthermore, REM sleep EEG exhibits the largest Lyapunov exponent in both hippocampal and cortical EEG amongst other vigilance states. We also identified non-zero third-order cumulant for all the vigilance states which suggests their non- Gaussian behavior.

Automatic heart sounds detection and systolic murmur characterization using wavelet transform and AR modeling.

This paper describes a signal processing procedure that identifies the first and the second heart sounds (S1 and S2), extracts the systole from the diastole, detects and characterizes the systolic murmur found within. The identification of heart sounds was facilitated by discrete wavelet transform (DWT) approximation using the Coiflet wavelet and followed by using indicators that quantify signal activity and strength. The systole was isolated and divided into smaller short segments where the signal activity measure and absolute amplitude were computed. S1 and S2, and the onset and duration of a systolic murmur were marked. Using the indices derived from AR modeling, a systolic murmur can be characterized by its timing, duration, pitch, and shape either as crescendo, decrescendo, crescendo-decrescendo, or plateau. The performance of the proposed procedure was evaluated and proved with clinically recorded systolic murmur episodes.

A comparison of linear and chaotic measures for rat hippocampal EEG during different vigilance states.

The correlation dimension was used in this paper as a quantifier to describe the chaotic behavior of sleep EEG recorded from the hippocampus of adult rats during vigilance states of quiet-waking, slow-wave sleep, and REM sleep. A modified Grassberger-Procaccia method was implemented to compute the correlation integral using a Euclidean distance normalized by the embedding dimension. The performance of the correlation dimension as a measure to characterize the sleep EEG was compared to the quantitative measures derived from linear autoregressive models. Even though linear and chaotic measures are based on completely different theories and concepts, our experimental results have indicated them both effective in capturing the characteristic differences of sleep EEG during various states. The preliminary results have also shown the correlation dimension being particularly effective in emphasizing the differences in regard to the chaotic behavior between the EEG activity in SWS and QW and REM sleep.

Quantitative analysis of heart sounds and systolic heart murmurs using wavelet transform and AR modeling.

A quantitative approach integrating AR modeling and wavelet transform is presented in this paper to analyze the digitized phonocardiogram. The recognition of the first and the second heart sounds (S(1) and S(2)) were facilitated with wavelet transform without referring to the QRS waveform. We found that the Daubechies wavelet is most effective in identifying S(1) and S(2). In addition, the boundaries of S(1), S(2), and the onset and duration of the systolic murmur thus identified within the systole could be marked using the wavelet-filtered signal's strength. Furthermore, quantitative measures derived from a 2(nd) order AR model were used to delineate the configuration and pitch of the systolic murmur found within through piecewise segmentation. The proposed approach was tested and proved effective in delineating a set of clinically diagnosed systolic murmurs. The suggested AR and wavelet transform combined approach can be generalized with minor adjustments to delineate diastolic murmurs as well.

Coherence estimation of the hippocampal EEG through fast multichannel orthogonal projection.

Hippocampal EEGs at subfields CA1 and the dentate gyrus (DG) are modeled as stationary, multi-channel autoregressive (MAR) process. This work discusses the development of a new MAR modeling algorithm that can efficiently compute MAR coefficient matrices through progressive multichannel orthogonal projection. The resultant MAR coefficients are least square (LS) optimal and utilized to compute the power spectra and the coherence function of hippocampal EEG at CA1 and DG during REM sleep for animals of 15 and 90 days of age. The results show that the new method is easy to implement and provides consistent and smooth spectral and coherence estimates for hippocampal EEG epochs with varying data length.