Symmetry as a cue to object and scene representations in human visual cortex

人类视觉系统必须将到达视网膜的2D图像转换为视觉场景的3D表示，我们可以使用它来感知世界并与世界互动。视觉对称性是这个过程的一个重要线索，并且已经发现在视觉反应脑区引起强烈的反应。在视觉感知的背景下，大多数对称性研究都使用了有限范围的刺激参数：（1）反射而不是其他类型的对称性，（2）单个图像位置的对称轴，（3）正面平行图像平面的对称性。拟议的研究计划建立在最近的进展，以解决这些限制，并提高我们的理解对称性作为一个线索，对象和场景感知。项目1将探讨对称性的敏感性是如何通过对称类型、平铺和偏斜来调制的。我之前的工作表明，视觉大脑区域对反射以外的对称性很敏感，并且对称性在平面上平铺的纹理会产生强大的视觉反应，但目前还不清楚平铺和对称类型如何调节大脑区域的活动。倾斜操作的动机是，虽然对称性通常发生在自然世界中，对称的图像平面投影到视网膜上很少发生在自然观看的事实。因此，我们自然会问，当世界上的对称性在视网膜上产生“倾斜”投影时，视觉系统能在多大程度上克服失真。为了解决这一假设，即这些操纵差异驱动不同的视觉皮层的敏感子网络，我将使用脑成像方法来测量的幅度，时间和皮质源的对称性与高的时间和空间精度的反应。这个项目将提供一个更完整的表征对称性调整的视觉系统，并有可能连接不同的分支文献。项目2将讨论这样一个假设，即对对称性的敏感性是婴幼儿在视觉环境中接触对称性时学习的结果，因此人类视觉皮层对对称性的反应应该在早期发育中发生变化。我将通过比较婴儿和成人的对称反应来验证这一点。据我们所知，这将是第一个研究婴儿或幼儿对称感知的脑成像研究。这项研究将揭示在早期视觉发育中可能存在的对对称性的任何敏感性，并提供关于对称性编码如何随着暴露于自然世界而变化的线索。这项研究计划解决了对称性和3D物体和场景感知文献之间的一个重要差距，使用良好的控制刺激结合最先进的脑成像技术。这些发现有可能加深我们对对称性在自然视觉中的作用的理解，并为视觉神经科学和计算机视觉的相关研究提供信息。

The human visual system must convert the 2D image arriving at the retina into a 3D representation of the visual scene that we can use to perceive and interact with the world. Visual symmetry is an important cue to this process and has been found to elicit strong responses in visually responsive brain areas. Most studies of symmetry in the context of visual perception have used a limited range of stimulus parameters: (1) Reflection rather than other types of symmetry, (2) symmetry axes at a single image location, and (3) symmetries in the fronto-parallel image plane. The proposed research program builds on recent advances in order to address these limitations and enhance our understanding of symmetry as a cue to object and scene perception. Project 1 will explore how sensitivity to symmetry is modulated by symmetry type, tiling and skew. My prior work shows that visual brain areas are sensitive to symmetries other than reflection, and that textures in which symmetries are tiled across the plane produce robust visual responses, but it is unclear how tiling and symmetry type modulate activity across brain areas. The skew manipulation is motivated by the fact that although symmetries commonly occur in the natural world, symmetrical image-plane projections onto the retina rarely occur under natural viewing. It is thus natural to ask to what extent the visual system can overcome distortions introduced when symmetries in the world produce "skewed" projections onto the retina. To address the hypothesis that these manipulations differentially drive distinct symmetry-sensitive sub-networks in human visual cortex, I will use brain imaging methods to measure the amplitude, timing and cortical sources of responses to symmetry with high temporal and spatial precision. This project will provide a more complete characterization of symmetry tuning in the visual system and has the potential to connect distinct branches of the literature. Project 2 will address the hypothesis that sensitivity to symmetry is the result of learning as infants and young children are exposed to symmetries in their visual environment, and that responses to symmetry in human visual cortex should therefore change over early development. I will test this by comparing symmetry responses in infants and adults. To our knowledge, this will be the first brain imaging study to investigate symmetry perception in infants or young children. This study will reveal any sensitivity to symmetry that may exist in early visual development, and provide cues as to how encoding of symmetry changes with exposure to the natural world. This research program addresses an important gap between the symmetry and 3D object and scene perception literatures using well-controlled stimuli combined with state-of-the-art brain imaging techniques. The findings have the potential to deepen our understanding of the role of symmetry in natural vision, and inform related research in visual neuroscience and computer vision.