Path length entropy analysis of diastolic heart sounds

Ensembled Transfer Learning and Multiple Kernel Learning for Phonocardiogram based Atherosclerotic Coronary Artery Disease Detection

Conventional machine learning has paved the way for a simple, affordable, non-invasive approach for Coronary artery disease (CAD) detection using phonocardiogram (PCG). It leaves a scope to explore improvement of performance metrics by fusion of learned representations from deep learning. In this study, we propose a novel, multiple kernel learning (MKL) for their fusion using deep embeddings transferred from pre-trained convolutional neural network (CNN). The proposed MKL, finds optimal kernel combination by maximizing the similarity with ideal kernel and minimizing the redundancy with other basis kernels. Experiments are performed on 960 PCG epochs collected from 40 CAD and 40 normal subjects. The transferred embeddings attain maximum subject-level accuracy of 89.25% with kappa of 0.7850. Later, their fusion with handcrafted features using the proposed MKL gives an accuracy of 91.19% and kappa 0.8238. The study shows the potential of development of high accuracy CAD detection system by using easy to acquire, non-invasive PCG signal.

Complexity Change in Cardiovascular Disease

With the fast development of wearable medical device in recent years, it becomes critical to conduct research on continuously measured physiological signals. Entropy is a key metric for quantifying the irregularity and/or complexity contained in human physiological signals. In this review, we focus on exploring how entropy changes in various physiological signals in cardiovascular diseases. Our review concludes that the direction of entropy change relies on the physiological signals under investigation. For heart rate variability and pulse index, the entropy of a healthy person is higher than that of a patient with cardiovascular diseases. For diastolic period variability and diastolic heart sound, the direction of entropy change is reversed. Our conclusion should not only give valuable guidance for further research on the application of entropy in cardiovascular diseases but also provide a foundation for using entropy to analyze the irregularity and/or complexity of physiological signals measured by wearable medical device.

Analysis of the influence of memory content of auditory stimuli on the memory content of EEG signal

One of the major challenges in brain research is to relate the structural features of the auditory stimulus to structural features of Electroencephalogram (EEG) signal. Memory content is an important feature of EEG signal and accordingly the brain. On the other hand, the memory content can also be considered in case of stimulus. Beside all works done on analysis of the effect of stimuli on human EEG and brain memory, no work discussed about the stimulus memory and also the relationship that may exist between the memory content of stimulus and the memory content of EEG signal. For this purpose we consider the Hurst exponent as the measure of memory. This study reveals the plasticity of human EEG signals in relation to the auditory stimuli. For the first time we demonstrated that the memory content of an EEG signal shifts towards the memory content of the auditory stimulus used. The results of this analysis showed that an auditory stimulus with higher memory content causes a larger increment in the memory content of an EEG signal. For the verification of this result, we benefit from approximate entropy as indicator of time series randomness. The capability, observed in this research, can be further investigated in relation to human memory.

Improved Heart Sound Detection and Signal-to-Noise Estimation Using a Low-Mass Sensor

GOAL

The purpose of this paper is to improve the detection of high-frequency sounds from the heart for better identification of turbulent blood flow in partially occluded coronary arteries. This paper also describes a method for the quantitative assessment of data quality.

METHODS

A very light-weight dual accelerometer has been developed that places a small mechanical load on the chest. When used in conjunction with a novel correlation-based analysis, this dual-signal transducer provides an estimate to the signal-to-noise ratio (SNR) of the acoustic signal.

RESULTS

The new transducer has significantly better SNR properties than the traditional cardiac microphones. This improvement is due to increased sensitivity to high-frequency signals not a reduction in noise and is likely the result of reduced mechanical loading on the chest.

CONCLUSION

Substantial improvement in the detection of high-frequency heart sounds is possible as is quantitative assessment of data quality.

SIGNIFICANCE

The new transducer and analysis will lead to substantial improvements in the acoustic detection of partially occluded arteries associated with coronary artery disease. It is finally possible to obtain a measurement of the quality of heart sound signals as they are being recorded.