Neural mechanisms that mediate the ability to make value-guided decisions have received substantial attention in humans and animals^1,2,3,4,5,6. Experiments in animals typically involve long training periods. By contrast, choices in the real world often need to be made between new options spontaneously. It is therefore possible that the neural mechanisms targeted in animal studies differ from those required for new decisions, which are typical of human imaging studies. Here we show that the primate medial frontal cortex (MFC)⁷ is involved in making new inferential choices when the options have not been previously experienced. Macaques spontaneously inferred the values of new options via similarities with the component parts of previously encountered options. Functional magnetic resonance imaging (fMRI) suggested that this ability was mediated by the MFC, which is rarely investigated in monkeys³; MFC activity reflected different processes of comparison for unfamiliar and familiar options. Multidimensional representations of options in the MFC used a coding scheme resembling that of grid cells, which is well known in spatial navigation^8,9, to integrate dimensions in this non-physical space¹⁰ during novel decision-making. By contrast, the orbitofrontal cortex held specific object-based value representations^1,11. In addition, minimally invasive ultrasonic disruption¹² of MFC, but not adjacent tissue, altered the estimation of novel choice values.

调节做出价值导向决策的能力的神经机制在人类和动物^1,2,3,4,5,6中受到了广泛关注。动物实验通常涉及长时间的训练。相比之下，现实世界中的选择往往需要在新的选择之间自发地做出。因此，动物研究中针对的神经机制可能不同于人类成像研究中典型的新决策所需的神经机制。在这里，我们展示了灵长类动物内侧额叶皮层 (MFC) ⁷is involved in making new inferential choices when the options have not been previously experienced. 猕猴通过与先前遇到的期权组成部分的相似性自发地推断出新期权的价值。功能性磁共振成像 (fMRI) 表明这种能力是由 MFC 介导的，这在猴子³中很少被研究；MFC 活动反映了对不熟悉和熟悉选项的不同比较过程。^{MFC 中选项的多维表示使用类似于网格单元的编码方案，这在空间导航8,9}中是众所周知的，以整合此非物理空间中的维度¹⁰在新的决策过程中。相比之下，眶额皮质持有特定的基于对象的值表示^1,11。此外，MFC 的微创超声破坏¹²，而不是邻近组织，改变了对新选择值的估计。