Analysis of the MEG source localization method combining signal space separation and magnetic field tomography

摘要

Magnetic Field Tomography (MFT) is a MEG localization method which requires least inappropriate assumptions but most computational load. To reduce the massive computational load of MFT, the combination method of MFT with Signal Space Separation (SSS) was proposed in our previous work. In this paper, the localization ability of the combination of MFT and SSS (MFT+SSS) is evaluated more precisely. Simple Numerical simulations revealed the improvement of the localization of MFT+SSS. Besides that, weighted Tikhonov regularization for the ill-posed SSS inversion is proposed. This regularization method establishes a remarkable robustness against the measurement noise.