2014/09/09 by Max Hinne, Hinne, Max, Alex Lenkoski +5 · 1 citation
Computer Science · Neuroscience · #Blind Source Separation Techniques #FOS: Biological sciences #FOS: Computer and information sciences #Functional Brain Connectivity Studies #Memory and Neural Mechanisms #Methodology (stat.ME) #Neural dynamics and brain function #Neurons and Cognition (q-bio.NC)
paper · pdf · doi:10.48550/arxiv.1409.2676
openalex publication_date 2014/09/09 · openalex created_date 2022/10/03 · openalex updated_date 2026/07/28
Bayesian estimation of Gaussian graphical models has proven to be challenging\nbecause the conjugate prior distribution on the Gaussian precision matrix, the\nG-Wishart distribution, has a doubly intractable partition function. Recent\ndevelopments provide a direct way to sample from the G-Wishart distribution,\nwhich allows for more efficient algorithms for model selection than previously\npossible. Still, estimating Gaussian graphical models with more than a handful\nof variables remains a nearly infeasible task. Here, we propose two novel\nalgorithms that use the direct sampler to more efficiently approximate the\nposterior distribution of the Gaussian graphical model. The first algorithm\nuses conditional Bayes factors to compare models in a Metropolis-Hastings\nframework. The second algorithm is based on a continuous time Markov process.\nWe show that both algorithms are substantially faster than state-of-the-art\nalternatives. Finally, we show how the algorithms may be used to simultaneously\nestimate both structural and functional connectivity between subcortical brain\nregions using resting-state fMRI.\n