Research On The Brain Function State Intervened By Neural Feedback Training

With the accelerated rhythm of social life,the brain fatigue has widely covered most of the population.As a sub-health state,brain fatigue is seriously affects life and work,quite apart from that,it can alert diseases are brewing.Therefore,how to effectively intervene and improve the brain fatigued state has become a research focus in the world.The neural feedback(NFB)training method is a process of physiological and psychological interaction,making participants themselves produce specific electroencephalogram(EEG)reactions.Accordingly,the auxiliary intervention and improvement of the brain functional state are realized.In this paper,NFB game had an auxiliary role in improving brain functional state,different features were extracted: multi-fractal detrending fluctuation(MFDF),fuzzy entropy(FE)and phase locking value(PLV).Based on various analysis methods reflect the difference of effect,and the genetic algorithm to optimize the relevance vector machine classifier to achieve the classification of brain functional status.In order to obtain multifractal properties,the EEG features that quality index,Hurst index,singular spectrum width were extracted,classification rate all over 80%,indicating the MFDF algorithm can effectively distinguish the brain functional state.In order to excavate the influence of NFB game on the complexity of brain,the transient variation of sequence which was processed by FE method will hardly catch,and cause the information loss,according to this issue,the boundary width of exponential function was defined to capture details in sequence.The improved FE for the classification rate was 89.99%,shows that the improved algorithm can more effectively highlight the complex information of EEG signals.Entropy is dramatically increased for the first game after learning(P<0.05)that the complexity of brain is significantly increased.At the same time,in order to analyze the level of brain synchronization after NFB game,based on the traditional PLV,the Euclidean distance was introduced as a criterion to determine the selection of components,and improve the similarity between the inherent modal function(IMF)and the original signal.The classification accuracy of improved PLV was 82.35%,the synchronization increased significantly(P<0.05)after the game.In order to show classification results of different brain area more intuitively,this paper based on Visual Studio,SQL Server and MATLAB,designed an emotional state assessment system.The system includes the login module,the neural feedback training module,the signal acquisition module,the filter and the feature extraction module,which can effectively assess the brain functional state.