Independent component analysis in PET

Independent component analysis (ICA) is a statistical and computational technique that represents a multidimensional random vector as a linear combination of non-gaussian random variables (independent components) that are as independent as possible (Hyvärinen & Oja, 2000). ICA is a non-gaussian version of factor analysis, somewhat similar to principal component analysis (PCA), allowing blind separation of signal components. It has become a standard analysis technique in machine learning and signal processing. In analysis of PET data, ICA has been used for automatic image segmentation (Margadán-Méndez et al., 2004, 2010; Juslin et al., 2005, 2007), especially in order to extract input function from dynamic image (Lee et al., 2001, 2002; Naganawa et al., 2005a, 2005b, 2007, 2008; Chen et al., 2007; Mabrouk et al., 2013, 2014).

The FastICA algorithm is a computationally efficient method for performing the estimation of ICA (Hyvärinen, 1999).

See also:


Hyvärinen A. Fast and robust fixed-point algorithms for independent component analysis. IEEE Trans Neural Netw. 1999; 10(3):626-634. doi: 10.1109/72.761722.

Hyvärinen A, Oja E. Independent component analysis: algorithms and applications. Neural Netw. 2000; 13(4-5): 411-430. doi: 10.1016/s0893-6080(00)00026-5.

Hyvärinen A, Karhunen J, Oja E: Independent Component Analysis. Wiley, 2001. ISBN 0-471-22131-7.

Stone JV: Independent Component Analysis: A Tutorial Introduction. MIT Press, 2004. ISBN 0-262-69315-1.


Updated at: 2019-11-12
Created at: 2019-11-12
Written by: Vesa Oikonen