Abstract #3813

Nonlinear kernel canonical correlation analysis (kCCA) in fMRI

Zhengshi Yang¹, Xiaowei Zhuang¹, Tim Curran², and Dietmar Cordes^1,3

¹Cleveland Clinic Lou Ruvo Center for Brain Health, Las vegas, NV, United States, ²Department of Psychology and Neuroscience, University of Colorado Boulder, Boulder, CO, United States, ³Department of Radiology, University of Colorado-Denver, Denver, CO, United States

Kernel representation is an efficient method to extract nonlinear features without significantly increased computational complexity. Linear kernel CCA has been applied to fMRI data but the performance of nonlinear kernel CCA is still not clear. Here we investigate the accuracy of five types of kernel on simulated fMRI data and then apply Gaussian kernel CCA on real fMRI data. It provides a more sensitive and specific way to detect activation pattern.

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

Join Here