A network inversion technique for estimating equivalent dipole descriptionof visual evoked potential

For the activation study of the brain, dipole localization from the scalp p otential is one of the most promising techniques to realize a reasonable te mporal resolution which cannot be realized in functional MR and PET. The go al of our study is to estimate inversely the electrical brain activity in t he form of several dipoles from the scalp potential, using a network invers ion technique. As a basic approach, we have inversely estimated several dip oles from the potential distribution on a spherical surface, in the homogen eous sphere model. In the training phase, by expanding the neural network input dimensions bei ng redundant, the network can easily learn the forward mapping. In the inve rsion phase, the space of the expanded-network-in put-vector can be narrowe d by introducing a penalty term. Additionally, a consensus term was used to force several dipoles to have a similar orientation. We estimate that this is applicable to the localization of several dipoles that reflect the actu al brain activity, especially in the visual evoked potentials.