Application Research Of Anesthesia Monitoring And Sleep Stage Classification Based On EEG

Electroencephalogram(EEG)indicates the electrical activity of the brain.EEG signals are highly random and may contain valuable information about the different physiological states of the brain.However,because EEG signals are non-linear and unstable,it is difficult to obtain useful information directly only by observing time-domain information.Therefore,advanced signal processing techniques can be used to extract important features for scientific research.In this paper,we analyzed EEG signals from two research directions:depth of anesthesia monitoring and sleep stage classification.Monitoring the depth of anesthesia(DoA)with EEG is an ongoing challenge in anesthesia studies.In this paper,we propose a novel method based on Long Short-Term Memory(LSTM)and sparse denoising autoencoder(SDAE)to combine the hybrid features of EEG to monitor the DoA.The EEG signals were preprocessed using filtering etc..And then more than ten features including sample entropy,permutation entropy,spectral and alpha-ratio are extracted from EEG signal and we integrated the optional features such as permutation entropy and alpha-ratio to extract the essential structure and learn the efficient temporal model for monitoring the DoA.The experimental results evaluated on the datasets demonstrated that our proposed method provided better performance than the methods using permutation entropy,alpha-ratio,LSTM and other traditional indices.In recent years,EEG signals have been widely used in the monitoring of sleep quality.In addition to using EEG for sleep monitoring,in this paper,we also used electromyographic signals(EMG)and electrooculographic signals(EOG).No need to manually extract features,after inputting the signal to our proposed 2CNN-CNN structure,it can directly output the current sleep state.Experimental results show that the proposed method can achieve higher classification accuracy than the several methods compared.The main innovations and contributions of this article are as follows:(1)In this paper,we proposed a method for monitoring the depth of anesthesia based on mixed features,which combines dozens of features such as permutation entropy,sample entropy,wavelet entropy,average spectrum,alpha-ratio,etc.,and inputs them into the neural network to monitor the current anesthesia status.Compared with other combinations of experiments,this kind of feature combination achieved better results.(2)In this paper,we proposed an anesthesia depth monitoring neural network based on LSTM and SDAE.The network consists of three layers of SDAE and two layers of LSTM.This method can not only reduce the noise in the EEG signal and improve the robustness of the system,but also use the temporal information in the EEG signal and combine the previous information to predict the current depth of anesthesia.The experimental results show that this network model can monitor the depth of anesthesia well and achieve a higher monitoring accuracy than traditional methods.(3)In this paper,we combined EEG,EEG,and EMG signals,and designed a 2CNN-CNN sleep stage classification method.First,through two convolutional neural networks of different sizes,extract the time and frequency domain information from the PSG signal,then concatenate the two convolutional neural networks together,and then stitch a layer of CNN to train and classify the classification results.The experimental results show that the classification results obtained by our proposed method are better than the classification results of a single CNN or stitching LSTM.