Detection of Cognitive Decline Due to Mental Fatigue Using Electroencephalogram

Mental fatigue is associated with human errors. It is thus important to search for indicators that can easily evaluate it. The aim of this study is to construct of a system that can evaluate mental fatigue in a simple method. Therefore, we investigated the influence of the accuracy of linear discriminant analysis in two classes before and after the mental load. Also, a subjective questionnaire and the autonomic nerve activity based on heart rate variability analysis were examined to obtain evidence that sufficient mental fatigue occur in the experiment. In the experiment, we employed the experiment of about 3 hours a mental arithmetic task, and the measurement items ware an electrocardiogram, electroencephalogram, and a questionnaire survey. The result showed that the fatigue value of the subjective questionnaire survey increased when the number of trials increased, and the HF component calculated by heart rate variability analysis decreased. Therefore, the mental fatigue had occurred sufficiently in this experiment. Subsequently, linear discriminant analysis was performed in two classes before and after the onset of the experiment. As a result, the determination accuracy of the linear discriminant analysis became the most highest by combining of the power spectrum of the α-wave band and P300.