Consciousness-oriented Cognitive EEG Feature Recognition Research

Consciousness is the human brain’s perception and cognition of the internal and external environment,and anesthetics can cause a person to move from a conscious state to an unconscious state.The neural mechanism of loss of consciousness induced by propofol,a widely used anesthetic,is still not well understood,and it is still difficult to assess the dynamic changes in the level of consciousness during anesthesia.In this thesis,we investigate the brain network connectivity and level of consciousness during anesthesia,starting from the electrocorticography(ECo G)scale and using various methods of consciousness level assessment,including spectral analysis,functional connectivity analysis,and brain network analysis.Firstly,ECo G data of five clinical epilepsy patients who receives propofol general anesthesia are used as the research object in this paper to introduce the source of the data and the method of pre-processing,so as to provide high-quality data support for further analysis and research.Compared with electroencephalography(EEG)data,the signal quality of ECo G data is more stable and distinctive,and therefore has higher theoretical research value.Secondly,this paper analyzes several perspectives,including spectral energy,functional connectivity and brain networks,to explore the changes in the level of consciousness during anesthesia in a more comprehensive manner,which helps to better understand the network connectivity mechanism of propofol-induced loss of consciousness.It is found that during propofol-induced loss of consciousness,the patients’ EEG spectral energy changes significantly,and in all frequency ranges,the spectral energy in the anesthetized state is higher than that in the awake state;the functional connections between brain regions become tighter,and the brain network topology changes significantly,showing a decrease in global efficiency and an increase in local efficiency.In addition,the nonlinear functional connectivity analysis method has better performance compared with the linear functional connectivity analysis method.Specifically,the aggregation degree of the brain network significantly increases but the information transmission and interaction efficiency between brain regions decreases after the linear analysis methods such as Pearson correlation coefficient and phase lag index,while the network aggregation degree does not significantly change but the information transmission and interaction efficiency between brain regions significantly increased after the nonlinear analysis methods of mutual information.Finally,based on the research results of the above-mentioned theoretical part,we design and develop a brain network-based consciousness level assessment software system with embedded power spectrum,Pearson correlation coefficient,and mutual information,which can reflect the intraoperative consciousness level changes dynamically in real time.In addition,the system also has a trend review function,which can reproduce the intraoperative monitoring environment and realize the function of postoperative review and offline analysis,which can clinically assist doctors in postoperative analysis.