Albright T D, Stoner G R
Salk Institute for Biological Studies, La Jolla, CA 92037, USA.
Proc Natl Acad Sci U S A. 1995 Mar 28;92(7):2433-40. doi: 10.1073/pnas.92.7.2433.
The primate visual motion system performs numerous functions essential for survival in a dynamic visual world. Prominent among these functions is the ability to recover and represent the trajectories of objects in a form that facilitates behavioral responses to those movements. The first step toward this goal, which consists of detecting the displacement of retinal image features, has been studied for many years in both psychophysical and neurobiological experiments. Evidence indicates that achievement of this step is computationally straightforward and occurs at the earliest cortical stage. The second step involves the selective integration of retinal motion signals according to the object of origin. Realization of this step is computationally demanding, as the solution is formally underconstrained. It must rely--by definition--upon utilization of retinal cues that are indicative of the spatial relationships within and between objects in the visual scene. Psychophysical experiments have documented this dependence and suggested mechanisms by which it may be achieved. Neurophysiological experiments have provided evidence for a neural substrate that may underlie this selective motion signal integration. Together they paint a coherent portrait of the means by which retinal image motion gives rise to our perceptual experience of moving objects.
灵长类动物的视觉运动系统执行着许多对于在动态视觉世界中生存至关重要的功能。这些功能中突出的一项是能够以一种便于对这些运动做出行为反应的形式恢复并呈现物体的轨迹。朝着这个目标迈出的第一步，即检测视网膜图像特征的位移，已经在心理物理学和神经生物学实验中研究了多年。有证据表明，这一步骤在计算上很简单，并且发生在最早的皮层阶段。第二步涉及根据起源物体对视网膜运动信号进行选择性整合。这一步骤在计算上要求很高，因为该解决方案在形式上是欠约束的。根据定义，它必须依赖于利用指示视觉场景中物体内部和物体之间空间关系的视网膜线索。心理物理学实验已经记录了这种依赖性，并提出了可能实现它的机制。神经生理学实验为可能是这种选择性运动信号整合基础的神经基质提供了证据。它们共同描绘了一幅连贯的画面，展示了视网膜图像运动如何产生我们对移动物体的感知体验。
