Abstract:
|
Motivated by multi-subject and multi-trial experiments in neuroimaging studies, we develop a modeling framework for community detection in groups of related networks. The proposed model, which we call the random effects stochastic block model, is flexible and facilitates the study of group differences and subject specific variations in the community structure. We propose two methods to estimate the parameters of the model, a variational-EM algorithm and two nonparametric two-step methods based on spectral and matrix factorization. The methodology is applied to publicly available fMRI datasets from multi-subject experiments involving Schizophrenia, ADHD and Autism patients along with healthy controls. Our methods reveal an overall putative community structure representative of the groups as well as subject-specific variations within each group.
|