|
EEG is by the EEG plethysmometer the brain itself weak bioelectrical enlarge record become a graph, and brain electrophysiological activity in the overall surface of the cerebral cortex or scalp. EEG (EEG) recorded the electrical activity of the brain contains a large number of physiological information extracted EEG features and for in-depth research to help us further explore the brain. In recent years, the automatic interpretation and evaluation of the implementation EEG has been researchers continue to explore areas of research, the EEG feature extraction and automatic classification of EEG data is the basis for automatic interpretation and evaluation of EEG inspection and quantitative analysis has an important significance. Objective: This article by subjects EEG data to objectively record Mahalanobis distance discriminant Fisher discriminant and Bayes discriminant analysis, come up with a feature extraction and classification decisions and accurate method applied to EEG data and applications in the the drinking events EEG data analysis, in order to distinguish the category of intake of alcohol after head various parts of the electrode, so as to explore the the EEG characteristics change in the state of alcohol consumption. Methods: depending on the intensity of alpha waves 21 conductive electrode is divided into four categories, namely, discriminant analysis 6 normal state subjects 21 lead electrode EEG data. 6 subjects Drinking 200ml (7.2 ml), alcohol intake every 20 minutes 6 drinking events draw a relatively accurate analytical method for discriminant analysis. Results: The EEG data of the of 6 normal state subjects Mahalanobis distance discriminant, Fisher Discrimination and Bayes discriminant analysis and forecast the electrode classification accuracy rate of 64.4%, 72.3%, 22.7%. 6 subjects drinking events in EEG data do Fisher discriminant analysis, drawn 21 conductive pole class specific changes with the experimental conditions and the amount of alcohol intake: quiet, eyes closed, no drinking, each electrode is basically correct classification. Drink 200ml central region electrode (C3, CZ, C4) sentenced to the rear head (P3, PZ, P4, and O1, OZ, O2, T5, T6) increase in the number of the former head (FZ, F3, F4, FP1, FPZ, FP2, F7, F8) and the side of the head electrode (T3, T4) were sentenced to the increased number of the central area. After drinking 400ml, Chuo-electrode was sentenced to reduce the number of back of the head, the former head and the side of the head electrode Chuo sentenced to reduce the number. Drink 600ml central region electrode sentenced to reduce the number of rear head, sentenced to the increase in the number of the front head portion, the front head electrode sentenced to increase to the number of the central area, the side head electrode sentenced to the increase in number of former head. Drink 800ml, the former head electrode sentenced to reduce the number of the central area, and the rear head electrode sentenced to the central region of the increase in the number. Drinking 1000ml, front head and rear head electrode was sentenced to increase the number of the central area of ??the electrode. Conclusions: the Fisher discriminant law can be better applied to EEG data feature extraction and classification decision-making. The electrical activity of the brain after drinking a significant reaction, changing with increasing alcohol consumption, the conductive pole category. Category changed significantly after drinking 200ml, the category changes stabilized after drinking 600ml and 800ml. Four types of electrodes in the head, after the head electrode Category variations smallest. Small changes in male subjects electrode category, the female subjects electrodes category changed greatly.
|