Research On The Classification And Improvement Of Fatigue State Based On EEG

With the rapid development of science and technology,people enjoy great convenience in science and technology.At the same time,fast pace of life and excessive work pressure can easily lead to mental fatigue and affect people’s normal work and life.Therefore,it is an important research subject in the current scientific field to accurately detect mental fatigue status and adopt effective fatigue improvement measures.However,the cost of multi-channel EEG equipment is high and the wear process is complicated.Single-channel EEG acquisition equipment has not yet proposed an effective method for removal of eyelid artifacts,which severely restricts the development of brain fatigue testing.At the same time,the method for improving the fatigue condition needs further study.In response to the above issues,the specific work of this article is as follows:First of all,a single-channel EEG acquisition PC is written based on C# language to realize the real-time acquisition of EEG signals.Aiming at the characteristics of electro-optic artifacts in single-channel EEG signals,an effective eye artifact removal algorithm is studied.Based on empirical mode decomposition and independent component analysis,A combination of variational modal decomposition and independent component analysis is proposed for the removal of eye artifacts.Compared with the eigenmode decomposition and independent component analysis,the algorithm for removing eye-shoe artifacts can effectively avoid the occurrence of band aliasing,and the effect of artifact removal is better.Secondly,according to the features of EEG signals,nine features that can effectively distinguish the state of brain fatigue are extracted,and the features are selected by the random forest algorithm.The recognition rate and stability of the classifier classification before and after the comparison features were selected,and finally the five EEG features after feature selection were used for the detection and classification of fatigue state.Classification algorithm selects AdaBoost algorithm,Compare this algorithm with support vector machines,neural networks,and random forest classification algorithms.Byobserving the ROC curve and comparing the area under the curve and the classifier recognition rate,it is verified that the AdaBoost classification algorithm performs better than the other three algorithms.Finally,according to the current research status of fatigue state improvement methods,two methods of listening to music and music point electrical stimulation are designed to improve the fatigue state.Collected two subjects before and after the two methods of improvement of brain electrical signals,The improvement of the characteristics of the EEG signal before and after improvement and the percentage of the arousal group predicted by the AdaBoost algorithm will be used as evaluation indicators for improvement.