From associative to controlled learning of non-adjacent dependencies - brain function and structure

婴儿能够在短短几年内自发地学习他们的母语，也就是说，他们似乎具备了非凡的语言学习能力。Friederici，Müller & Oberecker（2011，PloS ONE）使用事件相关电位（ERP）发现，4个月大的德国婴儿在短暂接触后就能区分语法和非语法的非相邻依赖关系（NAD），而成年人在被动倾听中无法学习，但只有当他们必须执行明确的语法判断任务时。这表明从婴儿的不费力的NAD的联想学习到成人的自上而下/控制学习的发展转变。在成年人中，前额叶皮层可能表现出自上而下的学习控制，通过抑制联想自下而上的学习机制，从而限制被动倾听下学习NAD的能力。我们的ERP和fNIRS数据来自CROSSING的第一个资助期，揭示了24个月大的孩子在语言领域（即，所说的非母语句子）但不在非语言域中（即，音调序列），而36个月大的婴儿在这两个领域都没有表现出学习。因此，这些数据表明，从24个月到36个月的NAD学习相关的联想学习机制发生了特定领域的转变。在此基础上，我们提出了以下问题：（1）语言域信号和非语言域信号的本质特征是什么？它们促进了24个月龄儿童被动听力的学习，而非语言域信号促进了被动听力的学习。(2)为了促进3岁以上儿童有控制的成功学习，需要满足哪些学习条件？(3)学习机制中的所谓发展变化是否受到大脑灰质和白色物质成熟的限制？3- 6岁儿童学习NAD的变化是由哪些结构和功能性大脑网络决定的？(4)此外，我们将绘制一个更详细的图片NAD学习与一般认知能力的发展，在2至6岁的儿童。

Infants are able to learn their native language spontaneously within just a few years, that is, they seem to be equipped with remarkable language learning abilities. Using event-related potentials (ERPs), Friederici, Müller & Oberecker (2011, PloS ONE) found that 4-month-old German infants discriminate grammatical from ungrammatical non-adjacent dependencies (NADs) after only brief exposure, while adults were not able to learn during passive listening, but only when they had to perform an explicit grammaticality judgment task. This indicates a developmental shift from infants’ effortless, associative learning of NADs to adults’ top-down/controlled learning. In adults, the prefrontal cortex might exhibit top-down control over the learning, by inhibiting associative bottom-up learning mechanisms and hence limiting the ability to learn NADs under passive listening. Our ERP and fNIRS data from the first funding period of CROSSING reveal NAD learning under passive listening for 24-month-olds in the linguistic domain (i.e., spoken non-native sentences) but not in the non-linguistic domain (i.e., tone sequences), while 36-months-olds did not show learning in either domain. Hence, the data suggest a domain-specific shift of associative learning mechanisms underlying NAD learning from 24 to 36 months of age. Extending this research, we aim to answer the following questions in the presented research proposal: (1) What are the essential features of the signal in the linguistic domain vs. the non-linguistic domain that promote 24-months-olds’ learning under passive listening in the former, but not the latter? (2) Which learning conditions need to be fulfilled to promote controlled and successful learning beyond the age of 3 years? (3) Are the supposed developmental changes in the learning mechanisms constrained by the maturation of the grey and white matter of the brain? Which structural and functional brain networks underlie the changes in learning NADs in 3- to 6-year-olds? (4) In addition, we will draw a more detailed picture of NAD learning in relation to the development of general cognitive competences in 2- to 6-year-old children.