Discretization of flat Electroencephalography (EEG) using a probability distribution for fuzzy topographic topological mapping (FTTM) application

Abstract

The abnormality of the electrical activity in the brain during an epileptic seizure can be measured by using Electroencephalography (EEG) and Magnetoencephalography (MEG). Flat EEG is EEG data accurately translated to a two-dimensional plane. The probability of the Flat EEG signal centroid relocating to a different location of the brain after a time interval can be represented in matrix form. Then, the aim of this study is to discretize the Flat EEG plane for FTTM model application into the centered cells represented by matrices. This study also attempts to construct an algorithm for populating a matrix with probabilities that a Flat EEG centroid moves to another location. These objectives lead to the construction of a program for generating the matrix of probabilities, given a Flat EEG centroid and a vector representing the change of its location. Additionally, this study contributed to the programming techniques for generating Flat EEG matrix data. The Flat EEG plane containing signal centroids was discretized into centered cells and mapped to the entry of a matrix. The matrix of probabilities that a signal centroid moves to another location was constructed. Apart from that, an algorithm for populating the matrix of probabilities was also constructed. A program to generate the matrix of probabilities was constructed based on the algorithm for populating the matrix of probabilities. After that, the program for generating matrix of probabilities was implemented by computer programming using the Microsoft Visual Studio 2010 software. Thus, matrices of probabilities of size 5 X 5 and 15 X 15 were generated for time t = 0 to t = 1 with varying values of parameter for probability distribution, r. The matrix of larger size was also more precise which means the details were finer than the smaller sized matrix. Therefore, the constructed program for generating matrix of probabilities was able to describe the movement of the signal centroid to another location after the present time. The potential for added slices of Flat EEG data for another time was recommended for the future work of this study.