2008 Summer Project Week:fMRIconnectivity

From NAMIC Wiki
Revision as of 20:06, 10 June 2008 by Ybkim (talk | contribs)
Jump to: navigation, search
Home < 2008 Summer Project Week:fMRIconnectivity



Key Investigators

  • MIT: Bryce Kim, Polina Golland
  • BWH: Jungsu Oh, Marek Kubicki


Objective

Our objective is to study functional connectivity of schizophrenia patients versus normal with unsupervised data-driven analysis methods, such as ICA and clustering.


Approach, Plan

Our approach for investigating functional connectivity of schizophrenia patients is to apply probabilistic independent component analysis (PICA) and clustering based on Gaussian mixture model (GMM).

Our plan for the project week is to first run such data-driven analysis methods on the data and then perform group analysis. We will also compare the results between PICA and GMM and investigate the factors that contribute to the differences.

Progress

We have applied standard fMRI preprocessing steps on the data and regressed out the effects of white matter and ventricles. We have also customized PICA and GMM tools for our analysis purpose and obtained some preliminary results.



References

  • C. Beckmann and S. Smith. "Probabilistic independent component analysis for functional magnetic resonance imaging." IEEE Transactions on Medical Imaging, 23(2):137–152, 2004.