A critical analysis of linear inverse solutions to the neuroelectromagnetic inverse problem

摘要: This paper explores the possibilities of using linear inverse solutions to reconstruct arbitrary current distributions within human brain. The authors formally prove that due underdetermined character problem, only class measurable can be totally retrieved are those minimal norm. reconstruction smooth or averaged versions currents is also explored. A solution explicitly attempts averages proposed and compared with minimum norm Laplacian solution. In contrast majority previous analysis carried out in field, comparisons, avoid use measures designed for case dipolar sources. To allow evaluation distributed concept resolution kernels. Two summarizing measures, source identifiability visibility, applied comparison. From this study concluded: (1) unable produce adequate estimates at many brain sites (2) better than estimating position single point However, they systematically underestimate their amplitude strength especially deeper areas. Based on these result, it appears unlikely a three-dimensional (3-D) tomography electromagnetic activity based methods without significant amount priori information.