Contribution of parallel inputs to speed tuning in higher visual cortex

摘要 为了编码关于我们周围环境的视觉信息，视觉系统使用对特征做出反应的神经元 视觉世界--光线、颜色、对比度、运动、方向、速度、方向等。最早 视觉系统的各个阶段(视网膜、丘脑)，这些特征由相对简单的属性组成，例如 光的空间或时间调制，编码为并行通道。在后期(主要视觉 皮质(V1)，纹状体外区域)，这些简单的属性组合在一起以表示更复杂的视觉 特征，如方向、方向和速度。大脑皮层回路如何结合简单视觉的平行输入 并将它们转换成新的复杂的视觉特征是不完全理解的。 与许多物种高度相关的一种转变是识别视觉运动的速度。这个 根据定义，对象的速度需要处理至少两个输入-其空间位置和时间 位移。这些参数可以通过使用变化的漂移正弦光栅来定量研究 在空间频率(SF)和时间频率(Tf)上系统地进行了分析。SF的每个组合(循环/度) 而Tf(周期/秒)对应于特定的速度(Tf/SF=速度，度/秒)。在视觉上 系统中，有一些神经元在物体以特定速度移动时反应最好；换句话说，它们 都是“速度调整”。一些速度调节的神经元被认为首先出现在V1中。然而，大多数都被发现了 在专用于运动处理的高级视觉区域。一种假设是速度调节是通过一种 从V1到更高视觉区域的偏移SF和TF通道的总和。新的转基因和病毒工具使 细胞类型的特异性以及最近对小鼠视觉系统的改进使这一点成为一种 在小鼠身上测试的易于处理的假设。在小鼠中，前外侧(AL)和后内侧(PM)是较高的视觉区域 选择性地调整到不同的SFS、TFS和速度。AL调到了粗糙的特征和快速的速度，而PM则是 调整到精细的功能和较低的速度。过去的研究表明，AL和PM可以继承SF和TF的调谐 来自V1的首选项，但不知道它们是否也继承了速度调整，或者速度调整是如何出现的。 初步数据还确定了V1‘S主要丘脑输入层(第四层)的两个细胞群，它们是 差异化地调整到SFS和TFS。该提议假设这些V1第4层人群代表 平行输入对高级视觉区域神经元的调谐和产生的影响不同 速度调节。AIMS将首先描述层流中V1、AL和PM神经元的速度调节特性 利用活体唤醒技术对这些第四层人群的反应特性进行评估 细胞外电生理学和双光子钙成像。最后，假设将通过以下方式直接进行验证 在AL和PM中有选择地停用每个第4层群体并确定其对调整的影响一起， 这些研究将对理解大脑皮层的速度调节机制以及如何 复杂的感觉信息跨越层级皮质区域进行转换。

Abstract To encode visual information about our surroundings, the visual system uses neurons that respond to features of the visual world—light, color, contrast, movement, orientation, speed, direction, and more. At the earliest stages (retina, thalamus) of the visual system, these features consist of relatively simple properties, such as spatial or temporal modulation of light, that are encoded as parallel channels. At later stages (primary visual cortex (V1), extrastriate regions), these simple properties are combined to represent more complex visual features, such as orientation, direction, and speed. How cortical circuits combine parallel inputs of simple visual features and transform them into new complex visual features is incompletely understood. One transformation that is highly relevant to many species is the identification of the speed of visual motion. The speed of an object by definition requires the processing of at least two inputs- its spatial location and its temporal displacement. These parameters can be studied quantitatively by using drifting sinusoidal gratings that vary systematically in spatial frequency (SF) and temporal frequency (TF). Each combination of SF (cycles/degree) and TF (cycles/second) corresponds to a particular speed (TF/SF = speed, degrees/second). In the visual system, there are neurons that respond best when objects are moving at a particular speed; in other words, they are “speed tuned.” Some speed tuned neurons are thought to first emerge within V1. However, most are found in higher visual areas dedicated to motion processing. One hypothesis is that speed tuning emerges through a summation of offset SF and TF channels from V1 to higher visual areas. New transgenic and viral tools enabling cell type specificity as well as a recently improved understanding of the mouse visual system now make this a tractable hypothesis to test in mice. In mice, higher visual areas anterolateral (AL) and posteromedial (PM) are selectively tuned to different SFs, TFs, and speeds. AL is tuned to coarse features and fast speeds, while PM is tuned to fine features and slow speeds. Past studies have shown that AL and PM can inherit SF and TF tuning preferences from V1, but it is not known if they also inherit speed tuning or how speed tuning emerges. Preliminary data have also identified two cell populations in V1’s main thalamic input layer (layer 4) that are differentially tuned to SFs and TFs. This proposal hypothesizes that these V1 layer 4 populations represent parallel inputs that contribute differentially to the tuning of neurons in higher visual areas and the generation of speed tuning. The aims will first describe the speed tuning properties of V1, AL, and PM neurons in a laminar specific manner and evaluate the response properties of these layer 4 populations using awake in vivo extracellular electrophysiology and 2-photon calcium imaging. Finally, the hypothesis will be directly tested by selectively inactivating each layer 4 population and determining its effect on tuning in AL and PM. Together, these studies will have significant implications for understanding mechanisms of speed tuning in cortex and how complex sensory information is transformed across hierarchal cortical areas.