Neural Mechanisms of Reward Processing

描述(由候选人提供)：当只在延迟一段时间后才期望获得奖励时，其现值通过相对于延迟的长度贴现其价值来合理地确定。旨在确定人类和动物身上这种时间折扣函数性质的实验发现，时间在短期内比在长期内更受重视：这种现象被称为双曲线时间折扣。双曲贴现的确切性质仍然存在激烈的辩论，没有经验数据能够清楚地区分不同的假设。在这项提案中，开发了一系列功能磁共振实验，以区分双曲线贴现如何产生于大脑功能的两种最流行的理论。当受试者在不同延迟的不同程度的奖励结果之间做出选择时，我们检验了几个假设。我们的实验同时使用自然奖励(果汁)和金钱奖励来确定相关的大脑结构是否适用于各种奖励模式。除了这项实验工作外，我们还将使用强化学习方法进行计算研究，以确定人们为什么会随着时间的推移而双曲线贴现是否有规范的原因。

DESCRIPTION (provided by candidate): When rewards are expected only after some delay, their present value is reasonably determined by discounting their worth relative to the length of the delay. Experiments aimed at determining the nature of such time discount functions in people and animals have found that time is valued significantly more in the near-term than in the long-term: a phenomenon known as hyperbolic time discounting. The exact nature of hyperbolic discounting remains hotly debated, with no empirical data able to clearly distinguish between alternative hypotheses. In this proposal, a series of functional MRI experiments are developed to distinguish between two of the most popular theories for how hyperbolic discounting arises from brain function. We test several hypotheses as subjects make choices between rewarding outcomes of differing magnitudes at different delays. Our experiments employ both natural rewards (fruit juice) and monetary rewards to determine whether the associated brain structures generalize across reward modalities. In addition to this experimental work, we will conduct computational research using reinforcement learning methods to determine whether there are normative reasons for why people discount hyperbolically through time.