EEG Analysis Based On Artificial Potential Field

EEG is the record of the brain's electrical activity. It carries redundant physiological and pathological information inside the brain. If the brain is normal, the neural cells originate the electrical signal regularly. If the brain is abnormal, such as seizure, encephaloma, and cerebral hemorrhage, the neural cells originate the electrical signal irregularly. The different way of the neural cells originate the electrical signal can be reflected in the pattern of EEG. Therefore, EEG is a very important tool in the fields of the diagnosis of neural diseases and the functional measure of the brain. And the EEG signal is a classical example of nonlinear signal, the EEG analysis also plays a significant roll in the field of processing graphic signal of computer. So EEG analysis is the important study direction in clinical and computer fields. But even though the same environment, the same age, the same health, etc, the EEG of the two persons are still different. The EEG analysis is an extremely difficult study. EEG analysis needs amount of expert knowledge in clinical and computer fields.The feature, the advantage and disadvantage of each EEG analysis method which exited are discussed. A new idea that uses the artificial potential field method for EEG analysis is mentioned, and then the analysis of the feasibility for the research is also given in this thesis. Artificial potential is the theory basis and methods to solve the problem in the thesis. To satisfy the need of solving problems in society world, the artificial potential uses the conception"extract potential"and"electric potential"in physical for reference to create some rational artificial potential that likes"extract potential"in the physical. At first, the series of the data created by digitizing the image and graphic form a data field. Artificial potential field is the field produced by the data which is in the data field, just as electric field is produced by electric charges. The data in the data space are influenced by each other. We recommend an artificial potential field function and discuss the meaning of the categories in the function. The contour view of the artificial potential field produced by all data in a set by the means of potential function shows clustering apparently.The paper applies artificial potential field method to the EEG analysis and gets four kinds of fields associated with the amplitude, frequency, phase and variance respectively. Then features can be gotten from these four fields and these features use not only the time-frequency, but also the phase and variance information. Preliminary experiments were done by the EEG signal and very promising results were attained.