Medial prefrontal cortex (mPFC) dysfunction has emerged as a consistent locus of pathology in depression and other stress-related psychiatric conditions, as well as an important therapeutic target for deep brain stimulation. Convergent evidence from chronic stress models in rodents and from spectroscopy and postmortem studies in patients suggest that hyperexcitability, elevated metabolic activity, and synapse loss in the mPFC may contribute to anhedonia and other depression-related behaviors, but the underlying mechanisms remain poorly understood. Whether and how changes in the excitability of mPFC pyramidal cells affect synaptic remodeling and mPFC microcircuit function is unknown. To address these questions, we conducted two sets of studies. First, we used stable step function opsins and a chronic stress model to manipulate the excitability of mPFC pyramidal cells and test for effects on the remodeling of postsynaptic dendritic spines, assessed by repeated two-photon imaging through chronically implanted microprisms in Thy1/YFP transgenic mice. Second, we used two-photon calcium imaging to test for associated changes in mPFC network dynamics. We found that postsynaptic spine remodeling was accelerated in hyperexcitable and chronic stress states, leading to an accumulating loss of synapses over time. Gradually accumulating synapse loss was associated with widespread changes in functional connectivity and in graph metrics of network efficiency and clustering that were especially pronounced in highly connected, “hub-like” neurons and were correlated with increased anhedonia- and anxiety-like behavior. We are currently testing whether optogenetic and pharmacological interventions to reverse stress-induced changes in mPFC excitability are sufficient to rescue dysfunctional network dynamics and behavior.