Replicating the communication (synaptic action) in human brains could be a critical step toward building computers that emulate brain-like performance, or neuromorphic computing. We study the physics of correlated oxide nanometer-scale devices which behave like biological synaptic systems. Using high-resolution imaging and spectroscopy techniques, we probe and manipulate electronic and thermodynamic properties of an artificial synapse made of nanostructured correlated oxides. As the size of a synapse decreases, the impact of its surroundings increases and fluctuations become fundamentally more pronounced, often far from equilibrium. The research aims to explore energetic resource constraints and the response of correlated oxide synapses to weak electric pulses in the presence of thermodynamic fluctuations, and how signal transmission and energy budget can be optimized in these conditions. The goal is to create more reliable and energy-efficient artificial synaptic networks for neuromorphic computing applications.