Blind Source Separation refers to the separation of signals given only linear combinations of those signals, such that: x(t)=A s(t) where x(t) is a matrix of a set of observed signals (mixed signals), x1(t) . . . xn(t), A is an unknown mixing matrix, and s(t) is a set of source signals s1(t) . . . sm(t), assumed to be statistically independent, i.e. p ( s ) = p ( s l , ??? , s m ) = ??? i = 1 m p