Zicheng Yang, Xiaolan Chen, Xiaohong Sun, Huimeng Lei
These findings indicate that reward contingency dynamically reorganizes the temporal alignment of lOFC-PN activity and its coupling with behavior, shifting from reward-aligned under Pavlovian learning to action-aligned under instrumental learning. This contingency-dependent temporal reorganization may support flexible adaptation of neural processing to changing reward contingencies during associative learning.
BACKGROUND: Animals must learn not only which cues predict reward, but also when to act to obtain it. Although the orbitofrontal cortex (OFC) encodes reward-related information, whether reward contingency is sufficient to reorganize the temporal alignment of OFC activity remains unclear.
METHODS: We recorded calcium signals from lateral OFC pyramidal neurons (lOFC-PNs) in wild-type (WT) and Sapap3 knockout (KO) mice using fiber photometry, during sequential learning of Pavlovian and instrumental contingencies in a single-odor reward association task. We further used a cue-free lick test and linear mixed-effects modeling to dissociate the contribution of reward contingency from that of licking behavior.
RESULTS: Licking behavior and lOFC-PN activity were comparable between WT and KO mice, supporting pooled analysis. Under Pavlovian contingencies, licking and lOFC-PN activity were concentrated around reward delivery. Under instrumental contingencies, both shifted to the pre-reward delay period, and the timing of lOFC-PN activity became progressively coupled to anticipatory licking as learning advanced. A cue-free lick test and linear mixed-effects modeling confirmed that these contingency-related differences reflected reward contingency rather than licking per se.
CONCLUSION: These findings indicate that reward contingency dynamically reorganizes the temporal alignment of lOFC-PN activity and its coupling with behavior, shifting from reward-aligned under Pavlovian learning to action-aligned under instrumental learning. This contingency-dependent temporal reorganization may support flexible adaptation of neural processing to changing reward contingencies during associative learning.