Estimating neural currents from neuromagnetic measurements

Abstract

This thesis concerns three new methods for estimating the electrical activity of the human brain from the magnetic fields measured outside the head. The first method models the electrical activity as a combination of focal current dipoles and applies global optimization to estimate their locations without need for an initial solution. The second method is the minimum current estimate that is capable of modeling both focal and distributed electrical activity. The third method measures the movements of the head during the measurements and uses this information for more accurate estimates of the neural currents. The usability of the methods is demonstrated using simulations and measurements. The minimum current estimate is also applied to four experiments of cognitive processes in the human brain: one studying visuomotor interaction, second studying visual attention, third studying integration of auditory and visual representations of letters, and fourth studying observation of sign language in signing and non-signing subjects.