Yasuda Masaharu, Hikosaka Okihide
Laboratory of Sensorimotor Research, National Eye Institute, National Institutes of HealthBethesda, MD, USA.
Front Neuroanat. 2017 Apr 11;11:35. doi: 10.3389/fnana.2017.00035. eCollection 2017.
We reach a goal immediately after detecting the target, or later by withholding the immediate action. Each time, we choose one of these actions by suppressing the other. How does the brain control these antagonistic actions? We hypothesized that the output of basal ganglia (BG), substantia nigra pars reticulata (SNr), suppresses antagonistic oculomotor signals by sending strong inhibitory output to superior colliculus (SC). To test this hypothesis, we trained monkeys to perform two kinds of saccade task: Immediate (visually guided) and delayed (visually-withheld but memory-guided) saccade tasks. In both tasks, we applied one-direction-reward (1DR) procedure to modify the level of goal-reaching motivation. We identified SNr neurons that projected to SC by their antidromic activation from SC. We stimulated SC on both sides because SNr neurons projecting to the ipsilateral SC (ipsiSC) and those projecting to the contralateral SC (contraSC) might have antagonistic functions. First, we found that ipsiSC-projecting neurons were about 10 times more than contraSC-projecting neurons. More importantly, ipsiSC-projecting SNr neurons were roughly divided into two groups which would control immediate and delayed saccades separately. The immediate-type SNr neurons were clearly inhibited by a visual target on the contralateral side in both visual- and memory-1DR tasks. The inhibition would disinhibit SC neurons and facilitate a saccade to the contralateral target. This is goal-directed in visual-1DR task, but is erroneous in memory-1DR task. In contrast, the delayed-type SNr neurons tended to be excited by a visual target (especially on the contralateral side), which would suppress the immediate saccade to the target. Instead, they were inhibited before a delayed (memory-guided) saccade directed to the contralateral side, which would facilitate the saccade. ContraSC-projecting SNr neurons were more variable with no grouped features, although some of them may contribute to the saccade to the ipsilateral target. Finally, we found that some ipsiSC-projecting SNr neurons were inhibited more strongly when reward was expected, which was associated with shortened saccade reaction times. However, many SNr neurons showed no reward-expectation effect. These results suggest that two separate oculomotor circuits exist in BG, both of which contribute to goal-directed behavior, but in different temporal contexts.
我们在检测到目标后立即达成目标，或者通过抑制立即行动稍后达成目标。每次，我们通过抑制另一种行动来选择其中一种行动。大脑是如何控制这些拮抗行动的呢？我们假设基底神经节（BG）黑质网状部（SNr）的输出通过向上丘（SC）发送强烈的抑制性输出，来抑制拮抗的眼球运动信号。为了验证这一假设，我们训练猴子执行两种扫视任务：立即（视觉引导）和延迟（视觉抑制但记忆引导）扫视任务。在这两种任务中，我们应用单向奖励（1DR）程序来改变达成目标的动机水平。我们通过从SC的逆向激活来识别投射到SC的SNr神经元。我们刺激双侧SC，因为投射到同侧SC（ipsiSC）和对侧SC（contraSC）的SNr神经元可能具有拮抗功能。首先，我们发现投射到ipsiSC的神经元比投射到contraSC的神经元多大约10倍。更重要的是，投射到ipsiSC的SNr神经元大致分为两组，它们将分别控制立即和延迟扫视。在视觉和记忆1DR任务中，立即型SNr神经元在对侧出现视觉目标时均明显受到抑制。这种抑制会解除对SC神经元的抑制，并促进向对侧目标的扫视。这在视觉1DR任务中是目标导向的，但在记忆1DR任务中是错误的。相比之下，延迟型SNr神经元倾向于被视觉目标（尤其是对侧的目标）兴奋，这会抑制对目标的立即扫视。相反，在指向对侧的延迟（记忆引导）扫视之前，它们会受到抑制，这会促进扫视。投射到contraSC的SNr神经元变化更大，没有分组特征，尽管其中一些可能有助于向同侧目标的扫视。最后，我们发现一些投射到ipsiSC的SNr神经元在预期奖励时受到更强的抑制，这与缩短的扫视反应时间有关。然而，许多SNr神经元没有表现出奖励预期效应。这些结果表明，BG中存在两个独立的眼球运动回路，它们都有助于目标导向行为，但在不同的时间背景下发挥作用。
