Cognitive regulation alters social and dietary choice by changing attribute representations in domain-general and domain-specific brain circuits

Cognitive regulation of decision making represents a crucial tool for altering behavior to fit momentary goals (e.g. eat healthy, be kinder). Are some people generally more successful using cognitive regulation or does it depend on the choice domain? Why? Do regulatory goals act on centralized, integrative value computations regardless of domain, or do they operate by changing value representations at the level of targeted choice-relevant attributes (e.g. a food’s taste and health in dietary choices)? To address these questions, we used an innovative combination of behavioral computational modeling and multivariate decoding of fMRI responses to identify the neural loci of regulation-related shifts in value representations across distinct regulatory goals and choice domains (dietary choice, altruistic choice). Surprisingly, we found little evidence that regulatory goals altered integrative value representations in the ventromedial prefrontal cortex (VMPFC), which represented all choice-relevant attributes across domains, regardless of regulatory goals. Instead, regulatory effects converged in a region of the dorsolateral prefrontal cortex (DLPFC). This area flexibly encoded goal-consistent values and predicted regulatory success for the majority of choice-relevant attributes in both dietary and social contexts, using unique neural codes for each attribute. However, we also identified an important domain-specific exception: goal-dependent encoding and prediction of regulatory success for prosocial attributes localized to the precuneus and temporo-parietal junction, rather than the DLPFC. Taken together, our results suggest that cognitive regulation of decision making operated by changing specific attribute representations rather than integrated values. Evidence of both domain-general and domain-specific neural loci for flexible attribute representation also reveals important divisions of labor in adaptive cognitive control of decision making, explaining when and why regulatory success generalizes (or doesn’t) across contexts and domains.

认知调节决策制定是调整行为以适应即时目标（例如，健康饮食、更加友善）的关键工具。某些人是否在运用认知调节方面普遍更为成功，抑或这取决于选择的领域？原因何在？调节目标是否在不受领域限制的情况下作用于中心化的综合价值计算，还是通过改变目标选择相关属性（例如，饮食选择中的食物口感与健康）的价值表征来实现其作用？为解答这些问题，我们采用了行为计算建模与多变量解码功能性磁共振成像（fMRI）响应的创新结合方法，以识别不同调节目标和选择领域（饮食选择、利他选择）中与调节相关的价值表征变化的神经定位。令人惊讶的是，我们发现调节目标在腹内侧前额叶皮层（VMPFC）中并未显著改变综合价值表征，VMPFC代表了所有领域的选择相关属性，无论调节目标为何。相反，调节效应集中在背外侧前额叶皮层（DLPFC）的一个区域。该区域灵活地编码与目标一致的价值，并预测了大部分饮食和社会情境中与选择相关属性调节成功的神经编码。然而，我们也发现了一个重要的领域特异性例外：针对亲社会属性的目标依赖性编码和预测调节成功定位于顶枕叶和颞顶交界区，而非DLPFC。综合来看，我们的结果表明，认知调节决策制定是通过改变特定属性表征而非综合价值来实现的。同时，关于领域通用和领域特定神经定位的灵活属性表征的证据，也揭示了在适应认知控制决策制定中的劳动分工的重要性，解释了何时以及为什么调节成功能够在不同情境和领域中泛化（或不能）。