Comparative analysis of ANN performance of four feature extraction methods used in the detection of epileptic seizures.

Epilepsy, a prevalent neurological disorder characterized by disrupted brain activity, affects over 70 million individuals worldwide, as reported by the World Health Organization (WHO). The development of computer-aided diagnosis systems has become vital in assessing epilepsy severity promptly and initiating timely treatment. These systems enable the detection of epileptic seizures by analyzing the electrical activity in the EEG recordings of the patients. In addition, it helps doctors to choose suitable treatment by quickly determining the type, duration, and characteristics of seizures and increases the patient's quality of life. The proposed computer-aided diagnosis system in this study comprises three modules: preprocessing, feature extraction, and classification. The initial module employs a low-pass Chebyshev II filter to eliminate noise artifacts from signal recordings. The second module involves deriving feature vectors using Bispectrum Analysis, Empirical Mode Decomposition, Discrete Wavelet Transform, and Wavelet Packet Analysis. The third module employs the Artificial Neural Networks method for epileptic seizure detection. This study not only enables the comparison of feature extraction efficacy among Bispectrum Analysis, Empirical Mode Decomposition, Discrete Wavelet Transform, and Wavelet Packet Analysis techniques, but it also reveals that Bispectrum Analysis and Empirical Mode Decomposition yield the highest accuracy rate. The method achieves 100% accuracy in detecting epileptic seizures. Additionally, sensitivity analysis has been conducted to enhance the success of Discrete Wavelet Transform and Wavelet Packet Analysis methods and to identify significant features.

Escalator Foundation Bolt Loosening Fault Recognition Based on Empirical Wavelet Transform and Multi-Scale Gray-Gradient Co-Occurrence Matrix.

An escalator is an essential large-scale public transport equipment; once it fails, this inevitably affects the operation of the escalator and even leads to safety concerns, or perhaps accidents. As an important structural part of the escalator, the foundation of the main engine can cause the operation of the escalator to become abnormal when its fixing bolts become loose. Aiming to reduce the difficulty of extracting the fault features of the footing bolt when it loosens, a fault feature extraction method is proposed in this paper based on empirical wavelet transform (EWT) and the gray-gradient co-occurrence matrix (GGCM). Firstly, the Teager energy operator and multi-scale peak determination are used to improve the spectral partitioning ability of EWT, and the improved EWT is used to decompose the original foundation vibration signal into a series of empirical mode functions (EMFs). Then, the gray-gradient co-occurrence matrix of each EMF is constructed, and six texture features of the gray-gradient co-occurrence matrix are calculated as the fault feature vectors of this EMF. Finally, the fault features of all EMFs are fused, and the degree of the loosening of the escalator foundation bolt is identified using the fused multi-scale feature vector and BiLSTM. The experimental results show that the proposed method based on EWT and GGCM feature extraction can diagnose the loosening degree of foundation bolts more effectively and has a certain engineering application value.

Deep learning algorithm using bispectrum analysis energy feature maps based on ultrasound radiofrequency signals to detect breast cancer.

Background: Ultrasonography is an important imaging method for clinical breast cancer screening. As the original echo signals of ultrasonography, ultrasound radiofrequency (RF) signals provide abundant tissue macroscopic and microscopic information and have important development and utilization value in breast cancer detection. Methods: In this study, we proposed a deep learning method based on bispectrum analysis feature maps to process RF signals and realize breast cancer detection. The bispectrum analysis energy feature maps with frequency subdivision were first proposed and applied to breast cancer detection in this study. Our deep learning network was based on a weight sharing network framework for the input of multiple feature maps. A feature map attention module was designed for multiple feature maps input of the network to adaptively learn both feature maps and features that were conducive to classification. We also designed a similarity constraint factor, learning the similarity and difference between feature maps by cosine distance. Results: The experiment results showed that the areas under the receiver operating characteristic curves of our proposed method in the validation set and two independent test sets for benign and malignant breast tumor classification were 0.913, 0.900, and 0.885, respectively. The performance of the model combining four ultrasound bispectrum analysis energy feature maps in breast cancer detection was superior to that of the model using an ultrasound grayscale image and the model using a single bispectrum analysis energy feature map in this study. Conclusion: The combination of deep learning technology and our proposed ultrasound bispectrum analysis energy feature maps effectively realized breast cancer detection and was an efficient method of feature extraction and utilization of ultrasound RF signals.

Wave Turbulence and Energy Cascade in the Hippocampus.

Mesoscale cortical activity can be defined as the organization of activity of large neuron populations into collective action, forming time-dependent patterns such as traveling waves. Although collective action may play an important role in the cross-scale integration of brain activity and in the emergence of cognitive behavior, a comprehensive formulation of the laws governing its dynamics is still lacking. Because collective action processes are macroscopic with respect to neuronal activity, these processes cannot be described directly with methods and models developed for the microscale (individual neurons).To identify the characteristic features of mesoscopic dynamics, and to lay the foundations for a theoretical description of mesoscopic activity in the hippocampus, we conduct a comprehensive examination of observational data of hippocampal local field potential (LFP) recordings. We use the strong correlation between rat running-speed and the LFP power to parameterize the energy input into the hippocampus, and show that both the power and non-linearity of collective action (e.g., theta and gamma rhythms) increase with increased speed. Our results show that collective-action dynamics are stochastic (the precise state of a single neuron is irrelevant), weakly non-linear, and weakly dissipative. These are the principles of the theory of weak turbulence. Therefore, we propose weak turbulence a theoretical framework for the description of mesoscopic activity in the hippocampus. The weak turbulence framework provides a complete description of the cross-scale energy exchange (the energy cascade). It uncovers the mechanism governing major features of LFP spectra and bispectra, such as the physical meaning of the exponent α of power-law LFP spectra (e.g., f -α, where f is the frequency), the strengthening of theta-gamma coupling with energy input into the hippocampus, as well as specific phase lags associated with their interaction. Remarkably, the weak turbulence framework is consistent with the theory of self organized criticality, which provides a simple explanation for the existence of the power-law background spectrum. Together with self-organized criticality, weak turbulence could provide a unifying approach to modeling the dynamics of mesoscopic activity.