Abstract
In this work, we present a nonlinear principal component analysis (PCA) that identifies underlying sources causing the expression of spatial modes or patterns of activity in neuroimaging time series where these sources can interact to produce second‐order modes. This nonlinear PCA uses a neural network architecture that embodies a specific form for the mixing of sources that is based on a second‐order approximation to any general nonlinear mixing. The modes obtained have a unique rotation and scaling that does not depend on the biologically implausible constraints adopted by conventional PCA. Interactions among sources render the expression of any mode or brain system sensitive to the expression of others. The example considers interactions among functionally specialized brain systems (using a fMRI study of colour and motion processing). Hum. Brain Mapping 8:92–97, 1999. © 1999 Wiley‐Liss, Inc.
Keywords: functional neuroimaging, fMRI, eigenimages, PCA spatial modes, nonlinear unmixing
Full Text
The Full Text of this article is available as a PDF (265.6 KB).
REFERENCES
- Friston KJ, Frith C, Liddle P, Frackowiak RSJ. 1993. Functional connectivity: The principal component analysis of large data sets. J Cereb Blood Flow Metab 13:5–14. http://www.ncbi.nlm.nih.gov:80/htbin-post/Entrez/query?uid=93107235&form=6&db=m&Dopt=r [DOI] [PubMed] [Google Scholar]
- Friston KJ, Poline J‐B, Holmes AP, Frith CD, Frackowiak RSJ. 1996a. A multivariate analysis of PET activation studies. Human Brain Mapp 4:140–151. [DOI] [PubMed] [Google Scholar]
- Friston KJ, Frith CD, Fletcher P, Liddle PF, Frackowiak RSJ. 1996b. Functional topography: multidimensional scaling and functional connectivity in the brain. Cerebral Cortex 6:156–164. [DOI] [PubMed] [Google Scholar]
- Friston KJ, Ashburner J, Frith CD, Poline J‐B, Heather JD, Frackowiak RSJ. 1996c. Spatial registration and normalisation of images. Hum Brain Mapp 3:165–189. [Google Scholar]
- Karhunen J, Joutsensalo J. 1994. Representation and separation of signals using nonlinear pca type learning. Neural Networks 7:113–127. [Google Scholar]
- Kramer MA. 1991. Nonlinear principal component analysis using auto‐associative neural networks. AIChE J 37:233–243. [Google Scholar]
- McIntosh AR, Bookstien FL, Haxby JV, Grady CL. 1996. Spatial pattern analysis of functional brain images using partial least squares. NeuroImage 3:143–157. http://www.ncbi.nlm.nih.gov:80/htbin-post/Entrez/query?uid=98005345&form=6&db=m&Dopt=r [DOI] [PubMed] [Google Scholar]
- Softky W, Kammen D. 1991. Correlations in high dimensional or asymmetric data sets: Hebbian neuronal processing. Neural Networks 4:337–347. [Google Scholar]
- Wold S. 1992. Nonlinear partial least squares modelling. Chemometrics Int Lab Syst 14:71–84. [Google Scholar]
- Worsley KJ, Friston KJ. 1995. Analysis of fMRI time‐series revisited—again. NeuroImage 2:173–181. http://www.ncbi.nlm.nih.gov:80/htbin-post/Entrez/query?uid=98003485&form=6&db=m&Dopt=r [DOI] [PubMed] [Google Scholar]