Analysis And Research Of Mental Fatigue Based On EEG

The control of mental fatigue is the key to reduce the occurrence of high-risk accidents.Successful detection of mental fatigue is of great significance in practical engineering applications.EEG signals are characterized by high resolution and low cost,which contain a large amount of physiological information.Research on mental fatigue based on EEG signals has become an important research method.Based on EEG signals,this thesis studied the characteristics of mental fatigue,proposed a new feature suitable for classification of mental fatigue,designed and developed a data collection and detection system for mental fatigue.The detailed research content consists of the following three aspects.The first part is about the research on the characteristics of mental fatigue based on EEG.Changes of EEG caused by fatigue is a current hot research issue,but the experimental design in different studies will lead to differences in the characteristics that appear in the fatigue state,for which there is currently no consensus.This thesis designed the N-back experiment and collected data.The mental fatigue was studied under the N-back experimental paradigm.Through the analysis of the collected EEG data,subjective indicators such as reaction time and response accuracy used as auxiliary indicators,we found that the α and β rhythms could be used to validly represent the mental fatigue state.The prefrontal lobe was the most discriminative brain area.The second part is about research on the classification of mental fatigue based on EEG.The selection of EEG features is a core issue in the classification of mental fatigue state and awake state.Entropy can respond to the degree of confusion of information.Many scholars use different types of entropy for the study of mental fatigue,such as differential entropy and spectrum entropy.Therefore,this thesis proposed fusion features,which combined rhythm entropy with power spectrum features on the basis of previous studies.Through mental fatigue detection on the data of 12 subjects,we found that the classification accuracy based on the proposed fusion features can reach 85.2%,which was7.9% higher than the differential entropy and 9.9% higher than the power spectral density.The third part is about the design of mental fatigue data collection and detection system.A brain-computer interface system was designed and developed based on the Windows platform for collecting EEG data and detecting fatigue status.Among them,the data acquisition module and the rhythm recognition module were integrated through application software,and the stimulation module interacted with the rhythm recognition module through UDP.The system included the design of the N-back experimental paradigm and the spatial N-back experimental paradigm.The EEG data of the subjects was collected through the EEG acquisition device,and the mental fatigue state of the subject was detected through the real-time recognition of the rhythm,and music was used for warning.Through feedback,the fatigue state of subjects can be effectively detected and relieved,indicating that the system can effectively achieve the purpose of detection.