Detection of the human brain activity with fundamental solution method

Abstract

Human brain activity mapping is a fundamental task for the neurophysiological research and for the diagnostic purposes too. Non-invasive techniques such as Electroencephalography (EEG) and magnetoencephalography (MEG) allow the reconstruction of the cerebral electrical currents providing useful information on the neuronal activity in the human brain. Based on a typical inverse problem the M/EEG imaging techniques require to solve more times a forward one. In this paper we discuss on a numerical tool based on the Method of the Fundamental Solutions(MFS) to efficiently solve the M/EEG forward problem going over the BEM state-of-the-art procedure. Inspired by the Leave-One-Out Cross Validation (LOOCV) strategy we improve the accuracy of the numerical solver gaining promising results.