Neural organization of learning and thought in the cortico-basal ganglia system

我们于2002至2004年进行了以下两项获资助的核心项目，为期三年：源自CM/Pf复合体的丘脑纹状体投射的功能作用：我们证明CM/Pf丘脑中的神经元向纹状体传递有关具有生物学意义的、吸引注意力的外部事件的信息（Matsumoto等人，2001；Minamimoto等人，2002）。基于我们最近对基底节区和丘脑的研究发现，我们提出了一个网络模型，即皮质-基底节区环路作为基于奖励的决策和行动选择，而丘脑-纹状体投射对基于奖励的决策和行动偏见起补充作用（Kimura et al. 2004）。所提出的模型得到了实验的支持（Minamimoto et al., 2005）。奖励依赖的动作选择和学习：我们证明了纹状突起神经元编码执行顺序运动任务的过程信息（Kimura et al., 2003; Ueda et al., 2003）。通过对人类大脑的功能性MRI成像研究，我们发现尾状核在基于奖励的决策策略学习中发挥着关键作用，以选择更高价值的行动（Haruno et al. 2004）。通过记录猕猴执行基于奖励的决策任务时纹状神经元的活动，我们最近发现纹状投射神经元编码特定于动作的奖励值（Samejima et al.，提交）。

We conducted the following two core projects supported by the grant for the three-year term from 2002 to 2004.Functional roles of the thalamo-striate projections originating from CM/Pf complex: We demonstrated that the neurons in the CM/Pf thalamus convey information about biologically significant, attention-drawing external events to the striatum (Matsumoto et al., 2001; Minamimoto et al., 2002). Based on our recent findings on the basal ganglia and thalamus, we proposed a network model that the cortico-basal ganglia loops work as reward-based decision and action selection, while the thalamo-striate projection plays complementary roles to the reward-based decision and action bias (Kimura et al. 2004). The proposed model was supported experimentally (Minamimoto et al., 2005).Reward-dependent action selection and learning: We demonstrated that the striate projection neurons encode information about procedure of performing sequential movement task (Kimura et al., 2003; Ueda et al., 2003). Using functional MRI imaging study of human brain, we showed that the caudate nucleus play a critical roles in reward-based learning of decision strategy to select higher-valued action (Haruno et al. 2004). By recording striate neuron activity of macaque monkeys performing reward-based decision task, we recently found that the striate projection neurons encode action-specific reward value (Samejima et al., submitted).