Suppr超能文献

奖励适应与学习机制:对比大鼠的奖励值变化和驱动学习。

Reward Adaptation and the Mechanisms of Learning: Contrast Changes Reward Value in Rats and Drives Learning.

机构信息

1 School of Psychology, Cardiff University.

2 School of Psychology, UNSW.

出版信息

Psychol Sci. 2018 Feb;29(2):219-227. doi: 10.1177/0956797617729825. Epub 2017 Nov 2.

Abstract

Formal theories of learning suggest that associations between events are determined by the internal representations of those events. Thus, learning should depend on perceived reward value-even when perceptions differ from objective values. We examined this prediction in flavor-preference learning in rats. In two experiments, simultaneous contrast either increased perceived reward value, which was paired with a distinctive flavor cue (the positive conditioned stimulus, CS+), or decreased the perceived value of the same reward, which was then paired with a second flavor (the negative conditioned stimulus, CS-). Even though the CS+ and CS- were paired with the same objective reward, there was a preference for the CS+ in subsequent tests. Moreover, the size of contrast-produced changes in reward value during training predicted the preference for the CS+ at test. This contrast-produced learning effect illustrates the mechanisms by which associations, which normally track veridical relationships between events in the world, are formed.

摘要

学习的形式理论表明,事件之间的关联取决于这些事件的内部表现。因此,即使感知与客观价值不同,学习也应该取决于感知的奖励价值。我们在大鼠的味觉偏好学习中检验了这一预测。在两项实验中,同时对比要么增加了与独特味道线索(正条件刺激,CS+)配对的感知奖励价值,要么降低了与第二种味道(负条件刺激,CS-)配对的相同奖励的感知价值。即使 CS+和 CS-与相同的客观奖励配对,在随后的测试中也会偏好 CS+。此外,在训练过程中对比产生的奖励价值变化的大小预测了在测试中对 CS+的偏好。这种对比产生的学习效应说明了关联形成的机制,而关联通常会追踪世界事件之间的真实关系。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验