Smooth WMN-FOCUSS method for EEG dipoles localisation

The new WMN-FOCUSS method is a combination between the weighted minimum norm ldquoWMNrdquo method and the iterative FOCal underdetermined system solver ldquoFOCUSSrdquo one. The initialisation step is very important since it is an iterative method. In this paper, we propose an amelioration of the WMN-FOCUSS by smoothing the current density distribution given by WMN before the initialisation of the iterative process. To reduce the divergence risk, smoothing the first iteration of the WMN-FOCUSS method is also proposed. To evaluate the standard and the smooth WMN-FOCUSS methods, we present a study of the localisation error and the robustness against the noise. The found results show that our method is able to give a good reconstruction of the simulated dipole with small localisation error even with an addition of a white gaussian noise.