Multiple Source Estimation Method Combined with Genetic Algorithm and Simulated Annealing
Yumie Ono, Atsushi Ishiyama, Naoko Kasai · IEEJ Transactions on Fundamentals and Materials · 2002
Recently, the magnetoencephalograph (MEG) measurement is expected as a means to find a higher-order function in the brain. It is physiologically considered that multiple areas in the brain are activated at the latency. Therefore the source localization method to which two or more signal sources can be estimated in short time by high accuracy is requested. Then we developed a method to estimate multiple-dipole sources by combining the genetic algorithm and the simulated annealing. In this combined method, the genetic algorithm is managed for a global search, and then the simulated annealing is applied for a detailed estimation. To assess this method, the simulation was carried out with the data of multiple-dipole sources model including white noise. The results of the simulation suggest that the presented method can be applied to the real MEG data and is useful for multiple source localization in shorter time and with high accuracy.