Research On EEG Inverse Problem Based On Genetic Algorithm

Research on brain function is among the most challenges in natural science. Due to having no hurt to human, research on electrical activity in brain is the focus. In the present several decades, the research on inverse problem of EEG has been greatly interested in and been payed attention to in the world. Inverse problem of EEG means we use EEG data to get the information of equivalent dipole sources that can reflect the activity of EEG. Optimization methods are available to this estimation.The genetic algorithm is a kind of searching method, which simulates the natural evolution. It is simple and easy to implement, especially it do not need the special field knowledge, so it has been used in very broad fields. Now the genetic algorithm has got a lot of fruits and more scholars begin to pay attention to it.The main work of this paper is using the genetic algorithm to solve the inverse problem of EEG. The research work is focused on the following several aspects:(1) Some problems in BEM numerical solution'process and related techniques are analyzed. The isolated problem with an adjustable parameter is introduced to solve the ill-conditioning problem caused by the low conductivity value of the skull.(2) Some work has been done in the researching of theory and application of the genetic algorithm. Based on the study of the basic structure of the genetic algorithm, some improvement on the design of the fitness function and genetic operators is given.(3) The adaptive genetic algorithm has been used in solving the inverse problem of EEG. Some conclusion has been drawn that its running speed and the ability to avoid local optimization have been greatly developed.(4) The adaptive genetic algorithm program has been developed in solving Multi-peak optimization. Through simulated calculation, some conclusions have been drawn that the genetic algorithm with adaptive genetic operators can run with higher speed in at first and more availably to avoid local optimal solution in the middle and end than the simple genetic algorithm.