The Motor Imagery EEG Processing Technology Research

Motor imagery EEG signal processing technology is an important part of thebrain-computer interface technology. If the imagery EEG signal processing is achieved, thedeep study of the brain will be accelerated. Even if it will promote the medical rehabilitation,assisted living and military security. Brain-computer interface is a control system which usingthe external devices like computer replace the neuromuscular pathway. Through thebrain-computer interface, human brain can transfer the information with the outside world andcontrol the external devices directly. At present, the offline analysis of the motor image EEGcan identify the motion of the left hand, right hand, foot and tongue. But the effect of onlineanalysis which carried on the four categories of motor imagery is dissatisfied. On the basis ofsummarizing the research situation, this paper does some research on technical and analyticaltools of the motion imagery EEG processing technology, mainly in the preprocessing, featureextraction methods and classification of the motor imagery EEG. Combination of the virtualreality technology, two interaction experiments of motor imagery brain-computer aredesigned. The experiment results are satisfied.The feature extraction and classification of the imagery EEG signal is the main part ofthe imagery EEG signal processing technology. On the basis of common spatial patternalgorithm of the feature extraction methods in traditional motor imagery EEG, kernel methodis added in this paper to prevent ignoring the useful signal. The experiment results aresatisfied. Alpha waves are added in the extracting and processing of the motor imagery EGGbased on the feature extraction of the left and right hand motor imager EEG. The amplitude ofthe alpha wave occurs changes significantly, when eyes status are changed (open or close).The feature of the alpha wave is used in this paper. Alpha wave is used as a control factor inthe motor imagery experiment which can reduce the fatigue of subjects and improve theaccuracy of the system. By analysis the characteristics of the EEG collected by64-leadelectrodes, physiological characteristics generated in imagery EEG and characteristicspectrum extracted in practical application of algorithms are combined. The electrode whichhas a greater impact on left and right hand motor imagine is selected. By doing this, the effectof irrelevant electrode errors are reduced, the amount of computation is decreased. Lineardiscriminant analysis classifier, BP neural network classifier, support vector machine classifier applied in motor imagery the EEG processing technology are introduced in thispaper. The standard data provided by the brain-computer interface contest are analyzed andcompared in offline situation. Linear discriminant analysis classifier are selected in theinteractive experimental system. The virtual reality experiment system designed in this paperhad been used by three subjects. The result is good and the requirements of the design isreached.Imagery EEG signal processing technology and virtual reality technology combined toincrease the flexibility of environmental feedback, more enhanced visual effects imagery EEGstudies, which will help the system feedback regulation. Opened and closed the alpha wave asa control switch, imagine the difficulty of movement is reduced through simple actions toreduce the training burden, facilitate the user can concentrate on.