Visual Adaptation, Selective Attention, and Shape Coding

经典的视觉后效效应为表征参与局部图像特征编码的神经单位的感受野特性提供了一个强有力的工具，在V1细胞生理指定的感受野与知觉之间提供了联系。然而，后效的应用主要局限于研究局部特征的低水平编码，因为后效具有早期加工的特征(例如，视网膜复制、方位和空间频率的窄调谐)。最近，Suzuki和他的同事以及其他人报告了在特定的时间条件下(在短暂的适应测试间隔之后用简短的测试模式探索短暂的适应)，可以诱导出全局形状属性的后效，如长宽比、倾斜、锥度、曲率和凸度。这些简短的形状后效带有高级编码的签名(例如，位置、比例和/或定义表面特征的容差)。因此，简单的形状后效作为描述处理阶段的工具很有希望，在该处理阶段中，定向轮廓的全局构型被系统地编码。对这种中级视觉编码的详细表征将扩展对视觉图案处理如何从局部图像特征的编码发展到各种中间几何特征的编码，并最终发展到有意义对象的表示的编码的理解。这项提议有三个方面的好处。在目标1中，将仔细指定产生全局形状后效(相对于局部图像后效)的时间要求，以细化简短的形状后效，作为探索全局形状编码的工具。在目标2中，将使用简单的形状后效来从心理物理的角度来描述接受场的性质和性质 在对全局形状属性进行编码的阶段操作的种群编码。在对象经常重叠的真实世界环境中，如果不参考注意选择机制，则不能考虑全局形状属性的位置容忍编码。与这一想法一致的是，短暂的形状后遗症已被证明对注意力敏感。因此，在目标3中，实验旨在调查观察者的注意力和图像参数 交互以选择用于全局形状编码的各个形状。因此，提出的三个方面的研究将提供一个全面的心理物理特征的接受场和编码属性的中高级形状表征，其激活取决于物理属性的刺激和注意。结果将开始在较高皮质视觉区域中的生理指定的感受野之间提供链接(例如， 在腹侧的“图形处理流”，如V4和IT)和对全局形状属性的感知。

Classic visual aftereffects have provided a powerful tool for psychophysically characterizing the receptive field properties of neural units involved in coding local image features, providing a link between physiologically specified receptive fields of cells in V 1 and perception. However, the application of aftereffects has been limited primarily to studying low-level coding of local features because the aftereffects bore signatures of early processing (e.g., retinotopy, narrow tunings for orientation and spatial frequency). Recently, Suzuki and colleagues and others have reported that under specific temporal conditions (brief adaptation probed with a brief test pattern following a brief adapt-to-test interval), aftereffects can be induced for global shape attributes such as aspect ratio, skew, taper, curvature, and convexity. These brief shape aftereffects bear signatures of high-level coding (e.g., tolerance for position, scale, and/or defining surface features). Brief shape aftereffects are thus promising as a tool for characterizing a processing stage in which global configurations of oriented contours are systematically coded. Detailed characterization of this "mid-level" visual coding will expand understanding of how visual pattern processing evolves from the coding of local image features to the coding of various intermediate geometric features, and eventually to the coding of representations of meaningful objects. The oal of this proposal is three-fold. In Aim 1, the temporal requirements for producing global shape aftereffects (as opposed to local image aftereffects) will be carefully specified to refine brief shape aftereffects as a tool for probing global shape coding. In Aim 2, brief shape aftereffects will be used to psychophysically characterize the receptive field properties and the nature of population coding operating at the stage of coding global shape attributes. In a real-world environment where objects routinely overlap, position-tolerant coding of global shape attributes cannot be considered without reference to an attentional selection mechanism. Consistent with this idea, brief shape aftereffects have been shown to be sensitive to attention. Thus, in Aim 3, the experiments are designed to investigate how observers' attention and image parameters interact to select individual shapes for global shape coding. Thus, the proposed three-fold investigation will provide a thorough psychophysical characterization of the receptive field and coding properties of the mid- to high-level shape representation whose activation depends on both the physical properties of the stimuli and attention. The results will begin to provide a link between the physiologically specified receptive fields in the higher cortical visual areas (e.g., in the ventral "pattern processing stream" such as V4 and IT) and perception of global shape attributes.