Local neuronal ensembles that co-reactivate across regions during sleep are preferentially stabilized

Neuronal ensembles, which represent coordinated activity patterns within individual brain regions, play crucial roles in memory. While local ensembles dynamically change in response to experience, factors influencing the stability of neuronal ensemble compositions remain largely unknown. In this study, we analyzed the dynamics of neuronal ensembles in the prelimbic cortex layer 5 (PL5), basolateral nucleus of the amygdala (BLA), and ventral hippocampus CA1 region (vCA1) during conditioning, extinction, retention-of-extinction, and interleaving sleep epochs in fear-conditioned male rats. We found that ensemble compositions in the PL5 were more stable than those in the BLA and vCA1. Ensemble reactivation during non-rapid eye movement (NREM) sleep following extinction sessions did not fully explain whether extinction-related ensembles were preserved until the retention-of-extinction sessions. Although the extinction sessions did not affect the number of ensemble pairs that were inter-regionally coactivated during NREM sleep, patterns of inter-regional ensemble coactivation were reorganized after extinction sessions, suggesting that extinction modifies the fine-scale structure of inter-regional networks. Notably, extinction-related ensembles that coactivated with those in other regions during post-extinction NREM sleep were more likely to be preserved until the retention-of-extinction sessions. In post-extinction NREM sleep, the preserved extinction-related ensembles contributing to inter-regional coactivation were activated more frequently during fast network oscillations. These findings suggest that local ensembles contributing to broader inter-regional activity are preferentially stabilized, supporting a systems-level mechanism of memory consolidation. Significance Statement Neuronal ensembles, relatively small populations of simultaneously activated neurons within a brain region, encode specific information. Since ensemble compositions change with experience, factors that stabilize local ensembles are critical for advancing our understanding of memory mechanisms and developing translational applications to treat memory-related disorders. While previous research has primarily focused on local mechanisms underlying ensemble stabilization, the role of inter-regional interactions has largely remained unexplored. Our study revealed that local ensembles that coactivate with ensembles in other brain regions are more likely to be preserved over time. Moreover, preserved ensembles contributing to inter-regional coactivation are reactivated more frequently during fast network oscillations in sleep, suggesting the important role of inter-regional networks activated during fast network oscillations in stabilizing local ensembles.