| General anesthesia（GA）is a drug-induced reversible state while maintaining stable vital signs.If general anesthesia is not achieved or maintained during surgery,unexpected intraoperative awareness can occur,potentially contributing to post traumatic stress disorder syndrome.Therefore,the search for reliable biomarkers is of great significance for the identification and monitoring of the state of brain consciousness during general anesthesia.In this study,non-invasive,high time-resolution electroencephalogram（EEG）signals were used to analyze steady-state and dynamic brain networks to reveal the neural mechanism of general anesthesia induced by propofol,and to develop accurate and reliable depth monitoring methods to assess changes in consciousness during anesthesia.The specific contents of this study are as follows:1.Research on the mechanism of steady-state brain network under GA.This chapter firstly constructs coherence networks for the resting state（Rest）,anesthesia-induced loss of consciousness（LOC）,and recovery of consciousness（ROC）of general anesthesia patients based on EEG,to capture Rest,LOC,ROC status network topology,and network property differences.The results show that,compared with Rest and ROC states,the fronto-occipital connectivity of the LOC state is disrupted,and the network properties are characterized by increased characteristic path length,decreased clustering coefficient,global efficiency,and local efficiency,and blocked information transfer between brain regions.Finally,the Rest and LOC states are classified based on network properties and spatial pattern of the network,and the classification accuracy of feature fusion is as high as 95%.Steady-state brain network results provide quantitative indicators for clinical anesthesia status recognition2.Research on the mechanism of dynamic brain network under GA.This chapter firstly calculates the time-varying fuzzy entropy of single-channel EEG signals to verify the superiority of multi-channel cross-fuzzy entropy,and finds that fuzzy entropy can only identify possible dysfunctions,but cannot reveal temporal fluctuations.Then,the time-varying network was constructed by the cross-fuzzy entropy and coherence methods respectively.It was found that the time-varying network constructed by the cross-fuzzy entropy showed the connectivity fluctuation from the fronto-occipital connection to the interior of the frontal lobe at about 30 seconds in the LOC stage.The LOC phase also showed significant changes after 30 s.However,both the time-varying network and the network properties of the coherence method show insignificant fluctuations.Finally,through the inflection point detection,it was found that the inflection points of the timevarying network attributes based on the cross-fuzzy entropy were about 31 s,which was basically the same as the clinical detection time point of the patient’s loss of consciousness.Based on this,an algorithm was developed to detect the time point of the patient’s loss of consciousness.The results of dynamic brain network analysis in general anesthesia show the superiority of cross-fuzzy entropy in capturing the fluctuation of consciousness during anesthesia induction,which provides a new idea for developing objective and reliable monitoring index of anesthesia depth.In conclusion,this paper analyzed the steady-state and dynamic brain network and found that brain network characteristics can effectively identify different brain states of consciousness,aiming to provide quantitative indicators for better clinical anesthesia management;we have also developed a method for monitoring the depth of anesthesia based on cross-fuzzy entropy,which was aimed to more accurately reflect the continuous fluctuation of the level of consciousness. |
PDF Full Text Request