C.G.Puntonet *, C.Bauer * , **, E.W.Lang **, M.R.Alvarez *, B.Prieto *

This paper presents a new adaptive algorithm for the on­ line linear and non­linear separation of signals with non­ uniform, symmetrical probability distributions. The procedure is based on the interpretation and properties of the vectorial spaces of sources and mixtures, using a multiple linearization in the mixture space. The main characteristics of the procedure are its simplicity, its immunity to symmetrically­distributed additive noise, and the rapid convergence experimentally validated when the method is applied to the separation of multiple EEG signals.