Abstract:
|
Advances in neuroimaging technology have provided a gateway for studying the functional and structural networks of the brain in vivo. Existing approaches typically examine these modalities in separate analyses, although multimodal methods are emerging that facilitate joint analyses. We present a joint model that assesses the link between functional and structural connectivity in the brain, by integrating two popular types neuroimaging data: functional magnetic resonance imaging (fMRI) and diffusion tensor imaging (DTI). We employ the EM algorithm to estimate the model parameters by maximizing the data likelihood. We illustrate the application of our method to a resting state study of healthy controls and subjects with major depressive disorder, and show that the positive association between SC and FC tends to be strongest for connections within a functional module, as we hypothesized. We also conduct simulation studies, and demonstrate that our method yields accurate estimates of the model parameters.
|