A hallmark of the prefrontal cortex (PFC) is flexible representation of task-relevant variables. To investigate roles of different interneuron subtypes in this process, we examined discharge characteristics and inactivation effects of parvalbumin (PV)- and somatostatin (SST)-expressing neurons in the mouse PFC during probabilistic classical conditioning. We found activity patterns and inactivation effects differed between PV and SST neurons: SST neurons conveyed cue-associated quantitative value signals until trial outcome, whereas PV neurons maintained valence signals even after trial outcome. Also, PV, but not SST, neuronal population showed opposite responses to reward and punishment. Moreover, inactivation of PV, but not SST, neurons affected outcome responses and activity reversal of pyramidal neurons. Modeling suggested opposite responses of PV neurons to reward and punishment as an efficient mechanism for facilitating rapid cue-outcome contingency learning. Our results suggest primary roles of mPFC PV neurons in rapid value updating and SST neurons in predicting values of upcoming events.

中文翻译：

---

表达小白蛋白和生长抑素的神经元在前额叶皮层中任务变量的灵活表示中的不同作用。

前额叶皮层（PFC）的标志是与任务相关的变量的灵活表示。若要调查不同的中间神经元亚型在此过程中的作用，我们检查了概率经典条件下小鼠PFC中表达小白蛋白（PV）和生长抑素（SST）的神经元的放电特性和失活作用。我们发现PV和SST神经元之间的活动模式和灭活效果有所不同：SST神经元在试验结果之前传达与线索相关的定量值信号，而PV神经元甚至在试验结果后仍保持价态信号。而且，PV而不是SST的神经元群体对奖赏和惩罚的反应相反。而且，PV的失活而不是SST的失活会影响结果反应和锥体神经元的活动逆转。建模表明PV神经元对奖赏和惩罚的相反反应是促进快速提示结果偶然性学习的有效机制。我们的结果表明，mPFC PV神经元在快速更新值中发挥主要作用，而SST神经元在预测即将发生的事件中具有重要作用。