In this work, we propose a novel approach to perform Dependent Component Analysis (DCA). DCA can be thought as the separation of latent, dependent sources from their observed mixtures which is a more realistic model than Independent Component Analysis (ICA) where the sources are assumed to be independent. In general, the sources can be spatio-temporally dependent and the mixing system may be non-stationary. Here, we propose a DCA algorithm, that combines concepts of particle filters and Markov Chain Monte Carlo (MCMC) methods in order to separate non-stationary mixtures of spatially dependent Gaussian sources.