Study On The Mental Fatigue Detecting Technology And Its Formation Mechanism Based On Brain Functional Network

Nowadays,with the increasing of the pace of life and the social competition pressure,mental fatigue increases significantly and spreads in the social group.Mental fatigue is often caused by prolonged cognitive activity of the brain,and accompanies by the reduction of work performance and vigilance levels,and impaired cognition.In the domain of real-time monitoring,vehicles driving,aircraft piloting,high-risk operations,etc.,many accidents are related to mental fatigue.For example,every year there are about 21%of fatal traffic accidents that have obviously causal relationship with fatigue driving,which have serious threats to people’s lives and property safety.Therefore,it is an important project with great application values to provide an objective and effective indicator for mental fatigue detection and provide a stable and reliable evaluation means for the corresponding countermeasures to mental fatigue.Among the various methods of mental fatigue evaluation,electroencephalogram(EEG)is considered as the most reliable and promising method and "gold standard" for mental fatigue detection,because EEG can directly measure the states of the brain’s neural activities.Traditional mental fatigue evaluation index based on EEG can be classified with linear indicators and nonlinear indicators.Linear indicators mainly base on the power and power ratios of the EEG rhythms,meanwhile nonlinear indicators mainly include various types of entropies,complexity,related dimension,Lyapunov index,etc.However,traditional indicators only contain the amplitude information of the EEG,ignoring the influences of phase information on the results.And these indicators cannot reflect the relations of functional connections between different brain areas,and cannot depict the overall brain function states.Thus,traditional mental fatigue evaluation indexes have great limitations.In this thesis,brain functional network theories were introduced into mental fatigue research,in order to explore its applications in mental fatigue detection and analyze the neural dynamics mechanism of the forming process of mental fatigue.To this end,detailed mental fatigue induction experiment was designed,EEG data at resting state and task state in the process of mental fatigue were precisely collected,and the feasibility of the mental fatigue model is confirmed based on the power and power ratios of the EEG rhythms.Basing on brain functional network theories,functional connections between different brain regions were determined by mutual information,then the values of functional connections were analyzed statistically between five periods of time at resting state and task state respectively,and found that Alphal rhythm(8-10Hz)during task state was the most sensitive to response to mental fatigue,suggesting that brain functional network analysis approach was more suitable in the applications of task state,which can make up for the demerits of traditional linear methods.Then with the two mainstream ways of selecting thresholds,weighted and binary brain networks of the Alphal rhythm at task state were constructed respectively,and the network characteristics were systematically analyzed to study their variation tendencies in the process of mental fatigue.The results indicated that:with the deepening mental fatigue,the maximum eigenvalue,clustering coefficient,and global and local efficiency increased,and the characteristic path length decreased;the evaluating effects of the network features on mental fatigue were better in weighted networks;the overall features for the descriptions of brain functional network characteristics(maximum eigenvalue,characteristic path length and global efficiency)were more sensitive to mental fatigue than the local features(clustering coefficient and local efficiency).The central nodes of the brain functional network in mental fatigue process,F3,F4,C3,C4,P3,P4,Fz,Cz,and Pz,were obtained using weighted degree centrality.Basing on these nine obtained central nodes,the brain functional network scale was shrunk,and the weighted brain functional networks were reconstructed and applied in the mental fatigue detection.The results demonstrated that the variation tendencies of the weighted network features in mental fatigue process were completely consistent with the results obtained before the simplification of the brain functional networks,indicating that the simplified brain functional network can also be used in mental fatigue evaluation.What’s more,the weighted maximum eigenvalue was the best evaluation index for mental fatigue.For the defect that the process of building brain functional network in the practical applications is too specialized,a new method for automatically modelling brain functional network was proposed based on the weights and physical distances of the functional connections and network motif theories.Compared with traditional method of building brain functional network,it was proved that the model can extract the most weights of strong connections in the adjacency matrix,and it was found that this model has self-optimizing function and can distinguish brain functional networks that with different weight sizes,which were verified to be inevitable based on weight distribution.Then this model was applied in mental fatigue detection,and consistent results were obtained compared with traditional method of building brain functional network,suggesting that this model has great application values.Basing on brain functional network theories,the formation mechanism of mental fatigue was analyzed from the perspectives of brain functional connection,network features,small-world property,and fractal feature.The results showed that in order to complete the given task,the brain should activate more brain function regions,enhance the strengths of the functional connections between different brain regions,carry out fast functional integration and separation,and adjust the structure of brain functional networks to be higher clustering coefficient,global and local efficiency,shorter characteristic path length,and to improve the speed of information transmission and information processing,which resulted in the slow destruction of optimized brain functional network structures,the increases of self-similarity and complexity of the brain functional network structures,making the network structure tend to be the randomization,and finally exacerbate mental fatigue process.