Moving mesh method for reconstructing some spread sources in the brain

The purpose of this paper is to propose a new algorithm for the analysis of biomagnetic field data obtained from magnetoencephalography (MEG) measurements. This new method overcomes two major problems faced by the current method of data analysis. The first problem is the need to determine the number of sites of brain activity before calculations can be performed. The second problem is inability of the analysis to provide any information regarding the volume of the brain activity. The new data analysis method, called the Moving Mesh Method (MMM), is capable of analyzing MEG data without the need to determine the number of sources beforehand. In addition, the MMM determines the location of brain activity as a three dimensional volume, instead of as a point source of activity. The MMM uses an iterative method of calculating the position of the sources to achieve greater accuracy, and a regularized g-inverse matrix to stabilize its solution. The feasibility of the MMM was examined by two methods. First, a computer simulation was used to confirm the MMM's capability to analyzing MEG data. In the second experiment, the MMM was applied to analyze somatosensory evoked fields obtained using a new imaging system (Shimadzu Biomagnetic Imaging System, Model-100). From the interpretation of the results, we have concluded that the MMM is a feasible method of biomagnetic data analysis.