Title: A dimension reduction method for group analysis of functional neuroimaging data
Authors: Mihye Ahn - University of Nevada Reno (United States) [presenting]
Abstract: Recently, much attention has been paid to the analysis of functional imaging data to delineate the intrinsic functional connectivity pattern among different brain regions within each subject. However, only few approaches for integrating functional connectivity pattern from multiple subjects have been proposed. The goal is to develop a reduced-rank model framework for analyzing the whole-brain voxel-wise functional images across multiple subjects in the frequency domain. Considering the neighboring voxels with different weights, the frequency and spatial factors can be extracted. Imposing sparsity on the frequency factors enables us to identify the dominant frequencies. In addition, the spatial maps can be used for detecting group difference, when the comparison between different groups is of specific interest. A simulation study shows that the proposed method achieves less spatial variability and better estimates of frequency and spatial factors than to some existing methods. Finally, we apply the proposed method to ADNI data.