One-shot learning is a crucial mechanism for survival and adaptation, enabling rapid acquisition of knowledge from a single experience [1]. Here, we investigated the synaptic plasticity associated with one-shot learning in vivo, focusing on endocannabinoid-mediated long-term potentiation (eCB-LTP) at cortico-striatal synapses [2]. We recorded neural activity in the cortex and striatum of behaving mice subjected to an uncomfortable stimulus in an arena using Neuropixel probes. This task was designed to induce avoidance behavior after a single encounter of the stimulus, thus modeling one-shot learning. After spike-sorting, the recorded neural data was processed to retrieve spiking activities of individual neurons. We then plugged these recorded spiking activities as inputs to a published mechanistic model of the dynamics of molecular networks implied in eCB-LTP, that has previously been validated on a wide range of ex vivo data [2]. Together, this strategy allowed us to infer what individual cortico-striatal neuron pair activities were likely to trigger synaptic plasticity and to correlate the result with the behaviour of the animal. Using these processed data, we first clustered individual cortico-striatal neuron pairs, i.e. potential synapses, based on their predicted synaptic weight variations. We then investigated the correlations between these clusters and the mouse behaviors, including locomotion, grooming, sniffing, and learning. Our analysis revealed that locomotion-related behaviors are the most distinct and unambiguous behaviors associated with eCB-LTP at cortico-striatal synapses. We also identified synapses that undergo strong eCB-LTP when the mice is in contact with the uncomfortable stimulus, a population that is a likely engaged in this one-shot learning task. These findings highlight the link between specific neural activity patterns and particular behavioral responses. Therefore, our approach demonstrated significant potential for the inference of task-specific synapse/neuron pairs populations, which can be important for our understanding of the neural basis of behaviors. Additionally, our method is useful to remove simulation artifacts and irrelevant learning events, thereby enhancing the accuracy and reliability of the data analysis.