Mechanisms of Avian Visual Perception, Cognition, and Action

描述(申请人提供)：视觉是现代认知科学中经久不衰的难题之一，而阐明视觉对象感知的机制是一个特别的挑战。到达视网膜的反射光中没有任何东西包含有关物体的位置、形状、大小或数量的直接信息。了解大脑如何处理视觉刺激以获取这些信息将是视觉科学的一大进步。由于鸟类是活跃和移动的，鸟类视觉系统必须解决推动人类和机器视觉研究理论进步的相同问题。在如此小的、非哺乳动物的系统中研究这些问题，大大有助于发展统一的一般视觉理论。利用我们基于辨别学习的成熟的心理物理技术，我们建议通过研究八哥和鸽子在受控分析环境中如何处理视觉刺激来评估这个问题。这将是 首次对两种鸟类的视觉处理进行平行比较。本应用程序的总体目标是调查和比较这些物种如何解决与照明和阴影处理、边缘处理和图形分组以及行为和动作的识别和分类相关的视觉认知和对象感知的中心问题。我们的具体目标包括：1)等同于我们对这些物种视觉加工的比较研究程序；2)考察表面阴影的加工及其对物体和深度知觉的贡献；3)考察形状知觉中的边缘和图形分组过程；以及4)考察视觉调节的动作和运动识别的机制。目前的项目是我们长期目标的一部分，目的是了解不同模式识别系统感知、识别、分类、整合、理解和响应复杂视觉信息的最接近的心理、计算和神经机制。以鸟类为代表的视觉力量、小体积和不同神经组织的独特组合为更好地理解这一极其重要的感觉通道以及它如何在不同类别的动物中发挥作用和实施提供了一个特殊的科学机会。这种比较将解决是否只有几种生物学上可信的机制来快速处理物体的视觉，或者在高度移动的生物体中是否有多种途径获得功能视觉。由于存在着强大的进化压力，需要将视觉信息高效快速地计算在鸟类的小而高度视觉的大脑上，我们的结果将直接有助于为患有各种视觉障碍的人类治疗、矫正解决方案或假肢的实践开发，并有助于设计用于自我引导的生物模拟机器人设备的视觉传感器。

DESCRIPTION (provided by applicant): Vision is one of the enduring puzzles of modern cognitive science and elucidating the mechanisms of visual object perception presents a particular challenge. Nothing in the reflected light that arrives at the retina contains direct information about the location, shape, size, or number of objects in view. Understanding how the brain processes visual stimuli to derive this information would be a major advance in visual science. Because birds are active and mobile, avian visual systems must resolve the same problems that have driven theoretical advances in human and machine vision research. Examining these issues in such small, non-mammalian systems adds substantially to the development of a unified general theory of vision. Using our well-developed psychophysical techniques based on discrimination learning, we propose to evaluate this problem by examining in how starlings and pigeons process visual stimuli in controlled analytical contexts. This will be the first parallel comparison of visual processing in two avian species. The overall objective of this application is to investigate and compare how these species solve central problems of visual cognition and object perception as related to the processing of illumination and shading, edge processing and figural grouping, and the recognition and classification of behavior and action. Our specific aims include: 1) equating our procedures for the comparative examination of visual processing in these species; 2) examining the processing of surface shading and its contributions to object and depth perception; 3) examining edge and figural grouping processes in shape perception; and 4) examining the mechanisms of visually-mediated action and movement recognition. The current project is a part of our long-term objective to understand the proximate psychological, computational and neural mechanisms by which different pattern recognition systems perceive, recognize, categorize, integrate, understand, and respond to complex visual information. The unique combination of visual power, small size, and different neural organization exemplified by birds offers a special scientific opportunity for better understanding this extremely important sensory modality and how it functions and is implemented across different classes of animals. This comparison will address whether there are only a few biologically plausible mechanisms for rapid visual processing of objects or whether there are multiple routes to functional vision in highly mobile organisms. Because there is strong evolutionary pressure for highly efficient and rapid computation of visual information placed on the small, highly visual brain of birds, our results will contribute directly to the pracical development of treatments, corrective solutions or prostheses for humans with a variety of visual disorders and to the design of visual sensors for self-guided bio-mimetic robotic devices.