额叶-尾状核环路：数字归纳推理新机制研究与ACT-R认知计算建模

归纳推理是人类适应生存最重要的高级思维功能之一，是人脑智能的核心组成。但是，由于其自身复杂性，其脑机制尚不清楚。我们前期研究表明，额叶-尾状核环路在数字归纳推理中起着重要的作用，特别是，背外侧前额叶与规则的假设形成有关，尾状核与规则的应用有关。然而，尾状核被作为整体结构，其亚区并行环路在归纳推理中的信息加工机制尚不清楚，尚且缺乏统一的认知计算模型。因此，本课题拟对健康和尾状核亚区损伤受试者，以数字归纳推理任务为刺激材料，采用tDCS技术和fMRI多体素模式分析方法，研究额叶-尾状核并行环路在归纳推理中的功能分离及神经可塑性机制；并结合ACT-R认知模拟的方法，构建其统一的认知计算模型。本项目旨在阐明人类归纳推理的脑内信息加工机制，有望为受脑启发的类脑智能系统研发及高级认知功能康复提供认知理论基础及新思路。

Inductive reasoning is one of the most important higher-level cognitive functions for human survival and is the critical component of human intelligence. However, the neural mechanisms are still unclear due to its complexity. Our previous studies have demonstrated that fronto-caudate loops play an important role in numerical inductive reasoning. Specifically, the dorsolateral prefrontal cortex (DLPFC) implicates in hypothesis generation and the caudate nucleus in the rule application. However, previous studies have considered the caudate nucleus as a uniform structure. Thus, it remains unsolved that the mechanisms of information processes of inductive reasoning in the caudate subdivision loops. Besides, it still lacks the unified cognitive computational model. Therefore, the present proposal aims to investigate the dissociation and plasticity mechanisms of inductive reasoning and its cognitive computational model within the fronto-caudate subdivision parallel loops. Both healthy volunteers and patients with caudate subdivisions lesions will be recruited and administrated to the transcranial direct current stimulation (tDCS) and the numerical inductive reasoning task functional magnetic resonance imaging (fMRI) examinations. Finally, a unified computational model will be constructed by using the cognitive architecture of adaptive control of thought-rational (ACT-R). This project will clarify the information processes of inductive reasoning and is expected to provide the cognitive theoretical basis and new ideas for the development of brain-inspired human-level intelligent systems and the rehabilitation of higher cognitive function.