Complexity Analysis of EEG Signals for Fatigue Driving Based on Sample Entropy

Sample entropy describes the complexity of time series by information growth rate, which has been widely used in EEG signal analysis. In order to explore the brain activity during fatigue driving, we built a simulated automobile driving experimental platform based on Unity3D software, and designed an experiment that simulating fatigue driving process to collects EEG signals of the brain from 17 healthy subjects. The changes of the complexity of the EEG signals in different rhythms are studied by comparing the sample entropy of different regions during the sober and fatigue states, respectively. The results show that the sample entropy of the EEG signals of the brain in the delta, theta, alpha, beta and gamma rhythms decrease during fatigue in which the beta rhythm and gamma rhythm decrease significantly. The sample entropy of frontal region of the brain in beta rhythm decrease significantly during fatigue state, and alpha, beta and gamma rhythm of central region of brain also decrease significantly during fatigue state, while there is no significant change in other brain regions. This experiment shows that the randomness of nerve cell activity is small and the complexity of brain decreases during fatigue state, which mainly manifest in that the beta rhythm of frontal and central regions is significantly decreased, which can provide a theoretical support for fatigue driving detection.