Local field potential (LFP) oscillations reflect emergent network-level brain signals that mediate behavior. Uncovering the mechanisms whereby these oscillations coordinate in time and space (spatiotemporal dynamics) to represent complex behavior across individuals would provide fundamental insights into how the brain encodes emotions at the mesoscale. Here we developed a multi-kernel machine learning approach to integrate LFP activity acquired concurrently from seven cortical and limbic brain regions into an analytical model that predicts stress-induced behavioral pathology across individual mice subjected to chronic social defeat stress. The prefrontal cortex (PFC) to ventral hippocampus directed spatiotemporal dynamic unmasked by this analytical model is correlated with acute stress behavior and brain-wide neuronal firing. Finally, we show that this precise spatiotemporal dynamic is enhanced in three independent molecular-, pharmacological-, and behavioral-based models of stress-pathology. Thus, our findings show that PFC to VHip-directed spatiotemporal dynamics organized at the mesoscale represent a convergent stress-vulnerability pathway.