It is important to show brain activities and their relations visually for the elucidation of the processing mechanism and for the diagnosis of diseases in the human brain. We developed a three-dimensional displaying tool by estimating dipoles to show the activities and by analysing information flow between them to show the relations. At first, we estimate dipoles (via 3-layered concentric spherical model, 2-dipole estimation) from evoked potentials. Secondary, using derived 2 dipole locations and moments as loci and quantities of brain activities, we applied stationary analysis for the information flow between the two time-series of the 1st and the 2nd dipole moments. Therefore, we obtain bi-directional information flows between the neuronal activities localized in 3D space of the brain with respect to somatosensory evoked potentials measured with 21 electrodes arranged according to the international 10–20 standard. Furthermore, we tried non-stationary analysis for the information flow with simulation data.