The Analysis And Evaluation Of Unconscious Patience’s EEG Based On Nonlinear Dynamics

The analysis and evaluation of unconscious patient’s EEG is one of the hot topicsin the field of rehabilitation medical engineering research in the day and age, which isof great significance for pathological diagnosis and rehabilitation of unconsciouspatient.EEG is a kind of bioelectric signal that can reflect human’s thinking activity andcan be recorded by electrode put into brain or surface of the brain. A large number ofexperiments have shown that EEG is also a kind of chaotic signal which hassignificant features of nonlinear dynamics. Therefore, we can make use of a variety ofnonlinear dynamical parameters to reveal many kinds of human’s thinking activities.In this paper, by analysing research status and processing method of EEG, theEEG experiments are designed, which can reflect human’s thinking activity on thetheoretical basis of nonlinear dynamics; Multi-scale Lempel-Ziv complexity,permutation partition Lempel-Ziv complexity, C0 complexity and basic scale entropyare ued as the nonlinear dynamical parameters which can characterize feature of EEG,and they are proved that they can reflect many kinds of human’s thinking activities byexperiments; We also seek the numerical differences of these nonlinear dynamicalparameters between the normal group and patient group and do some comparativeanalysis in various stimulations. The main work is as below:(1) From the generation mechanism of EEG, the paper illustrates theclassification and processing method of EEG and disignes two types of experiments:validation experiment and comparative experiment. The aim of validation experimentis to prove that EEG has different features in different thouht models and releventnonlinear dynamical parameters can reflect these features. The validation experimenthas three models, including quietude with eyes closed, mental arithmetic with eyesclosed and memory. The aim of comparative experiment is to analyse the differencesof nonlinear dynamical parameters between normal group’s EEG and patient group’sEEG in the same stimulation and the variation in different simulation for each other.The comparative experiment has three models, including resting state, stimulation ofcalling name, command of raising hand.(2) In the preprocessing of EEG stage, in order to eliminate the noise mixing in EEG generated by interference source, the soft thresholding denoising method basedon SURE is selected, because this method has better de-noising effect than otherthreshold methods in the algorithm analysis. To eliminate the disturbance mixing inEEG generated by independent source, which can’t be eliminated by wavelettransform’s method, the algorithm to filter artifacts based on blind source separationof maximum signal noise ratio is first used, because both the separation effect andrunning efficiency of the algorithm are better than FastICA and Infomax that wealways use. To conclude: the noise is well eliminated by the soft thresholdingdenoising method based on SURE and the disturbance mixing in EEG is successfullyseparated by the algorithm to filter artifacts based on blind source separation ofmaximum signal noise ratio.(3) On the basis of analysing nonlinear dynamical theory, Multi-scaleLempel-Ziv complexity, permutation partition Lempel-Ziv complexity, C0 complexityand basic scale entropy are first used to analyse human’s thinking activity. 4 kinds ofEEG nonlinear dynamical parameters of FP1、FP2、P3、P4、F7、F8 are calculated indifferent thought models and then campared with each other by One-Way ANVOAwith SPSS 19.0 for windows. The result shows that the differences in differentthought models of normal group can be well reflected by these 4 nonlinear dynamicalparameters.(4) In the clinical trial data processing, Multi-scale Lempel-Ziv complexity,permutation partition Lempel-Ziv complexity, C0 complexity and basic scale entropyare used on the analysis and evaluation of unconscious patient’s EEG again. 4 kindsof EEG nonlinear dynamical parameters of C3、C4、T3、T4 are calculated in differentstimulations. And we analyse the differences of nonlinear dynamical parametersbetween normal group’s EEG and patient group’s EEG in the same stimulation andthe variation in different simulation for each other by One-Way ANVOA andindependent-samples t test with SPSS 19.0 for windows. The comparative resultshows that EEG can be used in unconscious patient’s stimulative response and state ofconsciousness analysis, and these 4 nonlinear dynamical parameters can be one of thestandards in analysis and evaluation.