Functional MRI Analysis by a Novel Spatiotemporal ICA Algorithm
Data sets acquired from functional magnetic resonance imaging (fMRI) contain both spatial and temporal structures. In order to blindly extract underlying activities, the common approach however only uses either spatial or temporal independence. More convincing results can be achieved by requiring the transformed data to be as independent as possible in both domains. First introduced by Stone, spatiotemporal independent component analysis (ICA) is a promising algorithm for fMRI decomposition. We propose an algebraic spatiotemporal ICA algorithm with increased performance and robustness. The feasibility of the algorithm is demonstrated in an application to the analysis of an fMRI data sets of a human brain performing an auditory task.
KeywordsIndependent Component Analysis Independent Component Analysis Blind Source Separation Stimulus Component Independent Component Analysis Algorithm
Unable to display preview. Download preview PDF.
- 1.Cardoso, J.-F., Souloumiac, A.: Blind beamforming for non gaussian signals. IEE Proceedings - F 140(6), 362–370 (1993)Google Scholar
- 6.Theis, F.J., Gruber, P., Keck, I.R., Meyer-Bäse, A., Lang, E.W.: Spatiotemporal blind source separation using double-sided approximate joint diagonalization. In: Proc. EUSIPCO 2005 accepted, Antalya, Turkey (2005)Google Scholar