An Age-Dependent Index For Depth Of Anesthesia Under The Propofol

Accurate monitoring of anesthesia depth is one of the important guarantees for successful operation.With the advent of the aging age,the number of operations for the elderly is increasing.However,due to the influence of brain degeneration,the brain function of the elderly is significantly different from that of adults.The situation that existing monitoring instruments cannot judge the depth of anesthesia in the elderly is increasingly prominent,and there are great risks in the elderly anesthesia.Therefore,this paper deeply studied the age-dependent electroencephalogram features in the process of anesthesia from adults to the elderly,and developed new methods to provide a new technical means for monitoring brain states during general anesthesia.Firstly,we recorded electroencephalogram signals from 142 patients between the age range of 30-90 years during propofol anesthesia in the operating room of the inpatient department of the Second Hospital of Tianjin Medical University as our sample.The noises from power frequency interference,baseline drift and electromyogram need to be removed by using notch filter,wavelet transform and bandpass filtering.This study used spectral method and quantitative analysis method to analyze the changes of the two groups in three states which were awake state,anesthesia state and recovery state.The result shows that the electroencephalogram power of the elderly group in other frequency bands and full frequency band was lower than that of adults group,and the changes in the elderly group were more obvious than those of adults group except for gamma.There is the most significant difference in alpha,and relative power values of older people are higher in awake state and lower in anesthesia state and recovery state than those of adults.This indicates thinning of the cerebral cortex and a reduction in the number of synapses in older people.Secondly,this study used coherence and wavelet coherence methods to compare and analyze electroencephalogram oscillation coupling of the two groups.Since previous studies have suggested that there is significant effect in the coherence of alpha frequency band for consciousness judgment,this study focuses on calculating this frequency band.The experimental results showed that there were significant differences in the coupling characteristics between channels,especially during anesthesia and convalescence.Then,this study analyzed the age-dependent electroencephalogram features from the perspective of non-linearity.Based on the permutation entropy,this study proposed a new nonlinear entropy method for age-dependency analysis: enhanced permutation entropy and dispersion entropy.By analyzing the age-dependency of the three kinds of entropy algorithms in three states,it is found that the three kinds of entropy values of the elderly group are higher in the awake state and recovery state as well as lower in anesthesia state than those of the adult group.At the same time,this study analyzed the age-dependency of the recursive method during propofol anesthesia in different ages.Quantitative analysis showed that there is no age-dependency in order recurrence plot on the condition that it could clearly distinguish the anesthetic state,so it could well describe the depth of anesthesia in different age groups.Finally,using the above electroencephalogram features as training samples,this study combined the least square support vector machine method with the stack denoising autoencoder to carry out index fitting of the electroencephalogram features of anesthesia patients trained.Among them,stacked denoised auto-encoder can improve the robustness of the signal and make full use of the effective information,while the least squares support vector machine had better generalization ability and nonlinear processing ability than traditional support vector machine.The experimental results showed that the bispectral index fitting degree of the indicators obtained by this model was better than that by the basic method.