Suppr超能文献

MST 神经元对镶嵌刺激的反应。

Responses of MST neurons to plaid stimuli.

机构信息

Montreal Neurological Institute, McGill University, Montreal, Quebec, Canada.

出版信息

J Neurophysiol. 2013 Jul;110(1):63-74. doi: 10.1152/jn.00338.2012. Epub 2013 Apr 17.

Abstract

The estimation of motion information from retinal input is a fundamental function of the primate dorsal visual pathway. Previous work has shown that this function involves multiple cortical areas, with each area integrating information from its predecessors. Compared with neurons in the primary visual cortex (V1), neurons in the middle temporal (MT) area more faithfully represent the velocity of plaid stimuli, and the observation of this pattern selectivity has led to two-stage models in which MT neurons integrate the outputs of component-selective V1 neurons. Motion integration in these models is generally complemented by motion opponency, which refines velocity selectivity. Area MT projects to a third stage of motion processing, the medial superior temporal (MST) area, but surprisingly little is known about MST responses to plaid stimuli. Here we show that increased pattern selectivity in MST is associated with greater prevalence of the mechanisms implemented by two-stage MT models: Compared with MT neurons, MST neurons integrate motion components to a greater degree and exhibit evidence of stronger motion opponency. Moreover, when tested with more challenging unikinetic plaid stimuli, an appreciable percentage of MST neurons are pattern selective, while such selectivity is rare in MT. Surprisingly, increased motion integration is found in MST even for transparent plaid stimuli, which are not typically integrated perceptually. Thus the relationship between MST and MT is qualitatively similar to that between MT and V1, as repeated application of basic motion mechanisms leads to novel selectivities at each stage along the pathway.

摘要

从视网膜输入中估计运动信息是灵长类动物背侧视觉通路的基本功能。先前的工作表明,该功能涉及多个皮层区域,每个区域都从前一个区域整合信息。与初级视觉皮层(V1)中的神经元相比,中颞(MT)区域的神经元更能准确地表示棋盘格刺激的速度,对这种模式选择性的观察导致了两阶段模型,其中 MT 神经元整合了成分选择性 V1 神经元的输出。这些模型中的运动整合通常通过运动拮抗作用来补充,从而提高速度选择性。MT 区域投射到运动处理的第三阶段,即内侧上颞(MST)区域,但令人惊讶的是,人们对 MST 对棋盘格刺激的反应知之甚少。在这里,我们表明,MST 中增加的模式选择性与两阶段 MT 模型所实施的机制的更高普遍性相关:与 MT 神经元相比,MST 神经元更能整合运动成分,并表现出更强的运动拮抗作用的证据。此外,当用更具挑战性的非动力学棋盘格刺激进行测试时,相当一部分 MST 神经元具有模式选择性,而 MT 中则很少有这种选择性。令人惊讶的是,即使对于透明棋盘格刺激,MST 中也发现了运动整合,而这些刺激在感知上通常不会被整合。因此,MST 和 MT 之间的关系在质量上与 MT 和 V1 之间的关系相似,因为基本运动机制的重复应用会导致沿通路的每个阶段产生新的选择性。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验