题目：Thalamic integration of large-scale cortical networks facilities behavioral flexibility

主讲人：王斌（华南师范大学心理学院）

时间：2024年5月16日  下午15:00

地点：心理学院大楼213

主讲人简介：王斌博士，心理学院副教授，硕士生导师。2018年于德国科隆大学获得博士学位。之后在德国波鸿大学进行博士后研究工作。王斌博士长期聚焦于与认知灵活性相关的运动执行，感觉决策等认知过程的神经计算机制，相关的研究成果以第一作者或通讯作者发表在神经科学领域国际权威SCI期刊，包括Nature Communications, Progress in Neurobiology, Journal of Neuroscience, NeuroImage, Plos Computational Biology, Cerebral Cortex等。

讲座内容：

In uncertain environments, agents can employ a mixture of two distinct reinforcement learning (RL) strategies: a flexible, forward-looking strategy referred to as model-based RL, and a computationally-efficient, stimulus-response strategy known as model-free RL. Despite their significance in adaptive behavior, the neural mechanisms governing the arbitration between these strategies remain elusive. Combining multimodal imaging data from humans with a probabilistic reversal task. Here, we show that the mediodorsal thalamus (MD) in humans arbitrates prefrontal (PFC) model-based and striatal model-free RL processes. MD participates in both, yet model-based MD is more laterally localized than model-free, each showing distinct response patterns and functional/structural connections with dorsal and ventral PFC, respectively. These insights highlight the thalamus's crucial role in regulating large-scale learning systems.