A robust model for spatiotemporal dependencies

Fabian J. Theis, Peter Gruber, Ingo R. Keck, Elmar W. Lang

Research output: Contribution to journalArticlepeer-review

10 Scopus citations

Abstract

Real-world data sets such as recordings from functional magnetic resonance imaging (fMRI) often possess both spatial and temporal structures. Here, we propose an algorithm including such spatiotemporal information into the analysis, and reduce the problem to the joint approximate diagonalization of a set of autocorrelation matrices. We demonstrate the feasibility of the algorithm by applying it to fMRI analysis, where previous approaches are outperformed considerably.

Original languageEnglish
Pages (from-to)2209-2216
Number of pages8
JournalNeurocomputing
Volume71
Issue number10-12
DOIs
StatePublished - Jun 2008
Externally publishedYes

Keywords

  • Autodecorrelation
  • Blind source separation
  • Functional magnetic resonance imaging
  • Independent component analysis

Fingerprint

Dive into the research topics of 'A robust model for spatiotemporal dependencies'. Together they form a unique fingerprint.

Cite this