Processing object transformations in human visual cortex

描述(由申请人提供)：人类视觉系统面临着识别对象的计算困难的问题，尽管图像位置、比例、视点和照明可能有很大的变化。保持物体恒定的能力对我们的生存至关重要，它使我们能够感知一个稳定的、可预测的和熟悉的环境并与之互动。尽管它很重要，但视觉系统如何实现对象的恒定在很大程度上是未知的。我们提出了一系列行为和功能磁共振成像(FMRI)实验，旨在表征在观看条件变化时物体识别的神经机制。在第一系列实验中，我们的目标是表征神经对图像变换的反应，包括旋转、比例和位置。第二组实验将以第一组实验的结果为基础，评估注意力和期望对行为和神经反应的作用。最后一组实验将试图明确学习和经验在物体感知和变换不变神经反应发展中的作用。在所有的实验中，我们将使用一种新的功能磁共振自适应技术来表征特定视觉区域的调谐功能。

DESCRIPTION (provided by applicant): The human visual system is faced with the computationally difficult problem of recognizing objects despite potentially large changes in image position, scale, viewpoint, and illumination. The ability to maintain 'object constancy' is crucial to our survival, allowing us to perceive and interact with a stable, predictable, and familiar environment. Despite its importance, how the visual system achieves object constancy is largely unknown. We propose a series of behavioral and functional MRI (fMRI) experiments aimed at characterizing the neural mechanisms underlying object recognition during changes in viewing conditions. In the first series of experiments, we aim to characterize neural responses to image transformations including rotation, scale, and position. The second set of experiments will build on the results of the first by assessing the role of attention and expectation on behavioral and neural responses. The final set of experiments will attempt to specify the role of learning and experience in the development of object perceptions and transformation-invariant neural responses. In all of the experiments, we will use a novel fMRI adaptation technique to characterize tuning functions in specified visual areas.