特征捆绑的知觉学习：行为和脑机制研究

本课题结合行为研究和脑功能成像方法（ERP和fMRI），采用创新的研究范式，在不同的时间尺度和特征联合类型上，对特征捆绑的知觉学习机制进行系统的研究。研究从特征捆绑学习与注意关系入手，重点关注学习能否使得特征整合过程在知觉早期自动化进行，以及学习过程中选择性注意机制是否改变的问题。在行为上主要通过测量特征捆绑学习后对各单特征迁移性的方法，在脑电上主要利用N2pc成分，来检测各种特征联结能否在知觉早期被学习，以及学习发生的时间过程；并结合fMRI的精细空间定位和功能联结方法全面考察学习前后脑活动变化情况，系统地研究各种不同的特征捆绑学习的内在脑机制；初步建立特征捆绑知觉学习的认知加工模型。本课题将知觉学习与特征捆绑这两个脑与认知领域中的热点课题紧密地结合起来，不仅有助于我们加深了解知觉学习的基本原理，还有助于我们理解捆绑问题的发生机制，对理解和掌握其他类型的学习机制也具有重要理论和实践意义。

Using both behavioral measures and neuroimaging methods (ERP and fMRI), this project aims to investigate the mechanisms of perceptual learning in "feature binding". Different kinds of conjunctions are examined in different time scales with new paradigms. Focusing the relationship between "feature binding" and attention, we mainly concern about the following questions: Can feature integration be processed automatically in early perception after perceptual training? And will the mechanism of selective attention alter as a result of perceptual learning? To examine whether feature conjunction can be learnt as a functional unit, we test how the learning effect transfer to single feature in behavioral studies, and test whether the conjunction item can elicit N2pc component when it is task irrelevant. Short-term and long-term training are adopted to reveal the time course of learning. Functional localization and functional connectivity methods are used in fMRI studies to investigate the brain mechanisms of learning of different conjunction. A model will be set up basing on the coming results. Combining two hot topics in cognitive neuroscience, this project will broaden our understanding of both perceptual learning and feature binding, and be theoretically and practically useful for comprehending the mechanism of other kinds of learning.