One of the major issues confronting theoretical neuroscientists is: how can one systematically coarse-grain a neuronal network while retaining the salient dynamic features of that network? In this talk I'll describe a method which allows for the analysis of the dynamics of heterogeneous networks via a diagrammatic expansion in terms of subnetworks of the original network. I will also apply this analysis to the fly olfactory system, and describe one of the roles played by synaptic depression in this system.