Towards Advanced Data Analysis by Combining Soft Computing and Statistics

Volume 285 of the series Studies in Fuzziness and Soft Computing pp 305-317

Mining Local Connectivity Patterns in fMRI Data

  • Kristian LoeweAffiliated withDepartment of Knowledge and Language Processing, University of Magdeburg Email author 
  • , Marcus GrueschowAffiliated withLaboratory for Social and Neural Systems Research, Department of Economics, University of Zürich
  • , Christian BorgeltAffiliated withEuropean Centre for Soft Computing

* Final gross prices may vary according to local VAT.

Get Access


A core task in the analysis of functional magnetic resonance imaging (fMRI) data is to detect groups of voxels that exhibit synchronous activity while the subject is performing a certain task. Synchronous activity is typically interpreted as functional connectivity between brain regions. We compare classical approaches like statistical parametric mapping (SPM) and some new approaches that are loosely based on frequent pattern mining principles, but restricted to the local neighborhood of a voxel. In particular, we examine how a soft notion of activity (rather than a binary one) can be modeled and exploited in the analysis process. In addition, we explore a fault-tolerant notion of synchronous activity of groups of voxels in both the binary and the soft/fuzzy activity setting. We apply the methods to fMRI data from a visual stimulus experiment to demonstrate their usefulness.