Research On The Detection Technology Of The Brain Vigilance Based On EEG And ECG

Vigilance is defined as the ability to maintain attention for prolonged periods of time. Withthe development of automation technology, the roles of more and more workers have been shiftedfrom active controllers to that of system supervisors monitoring machines to make correspondingreaction to the occurrence of certain signals. Vigilance decrement can result in the supervisorsmissing of some signals which in turn may cause irreparable consequences. For example,astronauts in the space, facing the low pressure, oxygen-poor and full of all kinds of harmful raysof the environment, undertaking high load of work tasks as a result of the limitation of size ofspacecraft, and at the same time any small problem of high automated spacecraft may causeserious result so the astronauts should maintain a certain level of vigilance. Therefore, to establishan astronaut vigilance detection system, to explore the rule of vigilance variation with time, andthe future research on technology to improve vigilance or on arrangement of the astronauts`work,have very great practical value.In this context, we designed addition and subtraction experiment of three digits as mentalload to inducing vigilance changes. Both KSS scale and DSSQ scale are used as tools ofsubjective evaluation, and the average response time, error indicators of psychomotor vigilancetask marked signals: vigilance all over the time is divided into three stages. Then we extractedindicators which can reflect vigilance change from EEG and HRV(Heart Rate Variability) in thetime domain, frequency domain and nonlinear domain, lead optimization and dimensionalityreduction of the19leads of EEG were also done. On the basis, we used support vector machinefor pattern classification and compared the difference of classification effect among using EEGalone, using HRV alone and using both EEG and HRV.In this paper, we get the conclusions: Prolonged mental workload leaded to the decline ofbrain vigilance level. Compared with the initial stage of the experiment, also is the period of calmvigilance, reaction time of PVT experiment increased by5%to12%in the period of first declinedvigilance, it increased by15%to25%in the period of second declined vigilance; the changes inthe indicators of heart rate variability showed that heart rate variability is on the decline as a whole,and sympathetic activity gradually increased with the fluctuations; the indicators in EEG reflectedthe degree of excitement of cortical EEG gradually declined and the level of consciousness alsoreduced.19leads of EEG became5leads by using fisher discrimination and PCA on this basis made the characteristic number drop to eight. During the SVM classification results, theclassification accuracy of simply using HRV was78.18%, the accuracy of simply using EEG was81.01%, and the accuracy of HRV combined with EEG was88.28%, proving the multi-parameterclassification better.