A CNN based system to blind sources separation of MEG signals

In this paper a cellular neural network (CNN) based system to perform a real-time, parallel processing of magetoencephalographic data is proposed. In particular, a nonlinear approach to blind sources separation, instead of the linear procedure performed by independent component analysis, is introduced. Moreover, the characteristic spatial distribution of the cells in the CNN system has been exploited to reproduce the topology of the acquisition channels over the scalp.