B He1, J Lian, G Li. 1. Department of Electrical Engineering and Computer Science, University of Illinois at Chicago, MC 154, SEO 1120, 851 Morgan Street, Chicago IL 60607, USA. bhe@uic.edu
Abstract
BACKGROUND: A new realistic geometry (RG) spline Laplacian estimation technique has been developed for high-resolution EEG imaging. METHODS: Estimation of the parameters associated with the spline Laplacian is formulated by seeking the general inverse of a transfer matrix. The number of spline parameters, which need to be determined through regularization, is reduced to one in the present approach, thus enabling easy implementation of the RG spline Laplacian estimator. RESULTS: Computer simulation studies have been conducted to test the feasibility of the new approach in a 3-concentric-sphere head model. The new technique has also been applied to human visual evoked potential data with a RG head model. CONCLUSIONS: The present numerical and experimental results demonstrate the feasibility of the new approach and indicate that the RG spline Laplacian can be estimated easily from the surface potentials and the scalp geometry.
BACKGROUND: A new realistic geometry (RG) spline Laplacian estimation technique has been developed for high-resolution EEG imaging. METHODS: Estimation of the parameters associated with the spline Laplacian is formulated by seeking the general inverse of a transfer matrix. The number of spline parameters, which need to be determined through regularization, is reduced to one in the present approach, thus enabling easy implementation of the RG spline Laplacian estimator. RESULTS: Computer simulation studies have been conducted to test the feasibility of the new approach in a 3-concentric-sphere head model. The new technique has also been applied to human visual evoked potential data with a RG head model. CONCLUSIONS: The present numerical and experimental results demonstrate the feasibility of the new approach and indicate that the RG spline Laplacian can be estimated easily from the surface potentials and the scalp geometry.