Associative models of discrimination and generalisation: an evaluation in Pavlovian conditoning and causal judgement paradigms

学习是人类从基本生存到日常生活的根本。这种普遍性在自然界中是相匹配的，在进化规模的各个部分的动物都表现出学习能力。成功的科学学习研究包括识别目标行为和制定合适的学习程序。我们的项目关注的是人们如何学习预测事件，并使用收敛操作来衡量绩效。基于巴甫洛夫使用唾液反射来衡量学习的开创性贡献，联想理论现在已经扩展到动物反射的研究之外，表明人类行为的非常不同的方面可以通过联想模型来解决。联想理论的一个关键问题是说明多重刺激是如何被感知和处理的。例如，巴甫洛夫钟发生在包含大量背景声音和视觉刺激的环境中。此外，对任何刺激的反应都取决于环境。在不同的化合物（或环境）中对相同的刺激做出不同的反应是解决许多问题的基础。本研究旨在验证巴甫洛夫眨眼条件反射和因果判断范式中竞争性联想模型对刺激化合物表征的预测。这些表面上非常不同的程序，一个基于简单的反射，另一个基于“高阶”认知过程，有许多共同的属性。通过使用这些不同的范例，我们将检验结论的普遍性。

Learning is fundamental to humans from basic survival to everyday live. This ubiquity is matched in nature where animals in all parts of the evolutionary scale demonstrate learning. Successful scientific study of learning involves identification of the target behaviour and development of suitable study procedures. Our project is concerned with how people learn to predict events and uses convergent operations to measure performance. Based upon Pavlov's seminal contribution to use salivary reflexes to measure learning, associative theory has now been extended way beyond the study of reflexes in animals, showing that very different aspects of human behaviour can be addressed by associative models. One of the key problems in associative theory is to specify how multiple stimuli are perceived and processed. For example, Pavlov's bell occurred in a context containing numerous background sounds and visual stimuli. Furthermore, responses to any stimulus depend on the context. Responding differently to the same stimulus in different compounds (or contexts) is fundamental to the solution of many problems. Our proposal aims to test the predictions of competing associative models for the representation of stimulus compounds in Pavlovian eyeblink conditioning and causal judgement paradigms. These superficially very different procedures, one based on simple reflexes the other on "higher-order" cognitive processes, share many attributes. By using these different paradigms we will test the generality of our conclusions.