Meeting Banner
Abstract #4143

A New Model for Canonical Correlation Analysis with Spatial Constraints

Martin Miguel Merener 1 , Richard Byrd 2 , Rajesh R. Nandy 3 , and Dietmar Cordes 1,4

1 Physics, Ryerson University, Toronto, Ontario, Canada, 2 Computer Science, University of Colorado Boulder, Boulder, CO, United States, 3 School of Public Health, University of North Texas, Fort Worth, TX, United States, 4 Department of Psychology and Neuroscience, University of Colorado Boulder, CO, United States

This study provides important improvements in fMRI data analysis techniques for the detection of active brain areas. We propose and study a family of constraints for CCA, which naturally generalizes two interesting previously studied models. The solutions for these models can be found numerically and efficiently. For several choices of these constraints, the performance of the method in determining active voxels is excellent as measured via ROC simulations, and provide a significant improvement compared to previously published models in constrained CCA.

This abstract and the presentation materials are available to members only; a login is required.

Join Here