EEG Timing Feature Selection And Its Brain Network Construction And Analysis

Working memeory(WM)is a form of memory in a short time frame.In higher cognitive activities such as learning,memory,thinking and problem solving,the brain needs to process and store information,which requires work.Memorize this mechanism for processing.In the working memory,there is a central execution system.The three components of the visual space memory and the voice loop jointly determine the encoding,storage form and extraction of information in the working memory.Studies have shown that most people with mental illness have cognitive impairment,and working memory disorder is the main cognitive disorder in patients with mental illness.EEG is often used for disease diagnosis and early intervention because of its low cost,non-invasiveness and portability.Traditional EEG analysis methods sometimes include domain analysis,frequency domain analysis,time-frequency analysis,etc.,but these methods are only suitable for linear analysis.In recent years,nonlinear dynamics and complex network theory have been applied to the analysis of EEG signals,in an attempt to better analyze the characteristics of EEG signals and the changes of brain network topology under different states.The traditional EEG brain network construction method is to use the EEG signal electrode as the network node,and the relationship between the nodes as the network side,but this network construction method does not fully utilize the EEG high time resolution characteristics.Zhang et al first proposed a method for constructing pseudo-periodic time series networks.On the basis of this,the majority of scholars further studied and proposed different construction methods,in order to make full use of the high time resolution of EEG and better understand the time of EEG.characteristic.This paper presents a new method for constructing a complex network of time series.The time series complex network constructed by the fine data is analyzed.Firstly,the complex network method is used to analyze the difference of the time series network attributes constructed by normal people and patients.Secondly,the space between the channels is analyzed.The similarity of the complex network topology,further finding the difference between normal people and patients.EEG data is analyzed from both time and space perspectives,and the difference between the disease of schizophrenia patients and the time series network constructed by normal subjects is explored.The specific content mainly includes:(1)Using the microstate to analyze the EEG signal,verify that the microstate method can be used on the acquired data.This study analyzes the microstate number,average duration and state transition probability of microstates,and proves that the microstate is suitable for the EEG data used,which provides theoretical support for the construction of time series complex networks.(2)Segment the EEG data based on the micro-state method to construct a complex network of time series.When constructing the time series network,the time segment of the micro-state division is selected as the node of the network,and the features are extracted from the segmented time segments.The effective feature vector is selected as the feature of each time segment,and the correlation coefficient between the time segment feature vectors is used as the network.The edge of the channel can thus obtain a complex network of time series under multiple sparsity of each channel.(3)Analyze the time series complex network differences constructed.Firstly,the global and local attributes of time-series complex networks are analyzed,and the network genus is analyzed and compared.The characteristics of time-series complex networks constructed by normal people and schizophrenia patients are understood,and the complex networks constructed by the analysis are further analyzed.The channel analyzes the network properties with large differences at the lesion and can reflect the difference in time series.Secondly,it analyzes the similarity of the constructed time series complex network network topology.By comparing the similarity between normal people and schizophrenia patients,the similarity is shown,and normal people and schizophrenia can be analyzed.The relationship between similarities between patient channel networks.