Study On Discriminant Analysis Of Electroencephalogram Data

EEG is a sort of curves recorded the weak bioelectricity of brain themselves after amplifying by the EEG instrument, and it reflects collectively the electrophysiological activity of brain cells. Brain electrical activities recorded by EEG include large numbers of physiological and pathological information, so extracting and researching EEG feature is helpful to further explore the brain. In recent years, the automatic interpretation and evaluation of EEG has been continuously explored by researchers. However, achieving extraction of EEG characteristics and automatic classification of electroencephalogram dates is the foundation of automatic interpretation and evaluation, which have important significance for EEG examination and quantitative analysis.Research objective: In the text, we have done Mahalanobis,Fisher discriminant and Bayes discriminant analysis to EEG data of experiment objects which are recorded impersonally, come up with a relatively accurate method used in feature extraction and classification decisions. We have also done discriminant analysis to EEG data of six objects who intake of alcohol 7.2 ml every 20 minutes with this method.Research methods: In accordance with the strength ofαwave, the head electrodes are divided into four species. In use of part of 21 electrodes EEG data of 6 people, we have done 3 kinds of discriminant analysis, also have done discriminant analysis to EEG data of six objects every 20 minutes intake of alcohol 7.2 ml of six drinking incident. Research results: In use of part of EEG data of 6 people, we have done 3 kinds of discriminant analysis, the electrode classification accuracy rates are 64.4%, 72.3%, 22.7%. We also have done Fisher discriminant analysis to the EEG data, and come up with the changes of electrode categories with experimental conditions and the amount of alcohol: The electrodes classification is right in no drink. After drinking 200 ml, the number of electrodes which the Central (C3,CZ,C4) was sentenced to the back (P3,PZ,P4,O1,OZ,O2,T5,T6) increased, the former (FZ,F3,F4,FP1,FPZ,FP2,F7,F8) and sides(T3,T4) were sentenced to the central increased. After drinking 400 ml, the number of electrodes which the central was sentenced to the back reduced, the former and sides were sentenced to the central reduced. After drinking 600 ml, the number of electrodes which the central was sentenced to the back reduced, the former was sentenced to the central increased, and the sides were sentenced to the former increased. After drinking 800 ml, the number of electrodes which the former was sentenced to the central reduced, the back was sentenced to the central increased.After drinking 1000 ml, the number of electrodes which the former and back were sentenced to the central increased. Conclusions: Fisher discriminant would be better applied to the feature extraction and classification decisions of EEG data. EEG activity shows a significant response after alcohol intaked, electrode categories is constantly changed.After drinking 200 ml categories changed obviously, and drinking 600 ml and 800 ml categories changes become calm. The back is changed obviously in four.The changes are obvious in men and not obvious in women.