Application of Polynomial Spline Independent Component Analysis to fMRI Data
Atsushi Kawaguchi, Young K., Xuemei Huang · InTech eBooks · 2012
In independent component analysis (ICA), it is assumed that the components of the observed k-dimensional random vector x = (x1, . . . , xk) are linear combinations of the components of a latent k-vector s = (s1, . . . , sk) such that s1, . . . , sk are mutually independent. This is denoted by x = As, (1) where A is a k× k full-rank non-random mixing matrix. The main objective then is to extract the mixingmatrix through a set of observations x1, x2, . . . , xn. For a detailed description of this method, including its motivation, existence and relationship with other well known statistical methods such as principal component analysis, factor analysis, see Hyvarinen et al. (2001).