The Tuning for Complex Visual Stimuli in V1

描述（由申请人提供）：这项工作的长期目标是了解视觉形式感知的神经机制。本项目旨在研究初级视觉皮层（V1）对平滑轮廓、连接和纹理边界等中等复杂视觉刺激的处理。最近的证据表明，V1不仅仅是一组过滤器，它被动地从所有视觉输入中提取最简单的元素（例如，短定向线段），就像之前认为的那样。更确切地说，复杂刺激的基本成分彼此之间相互作用强烈，从而产生V1反应，这种反应可能与对单个成分的反应有很大不同。目前的项目旨在测试这样一个假设，即这种复杂的视觉处理可以从VI的柱状结构和内在电路中预测出来。特别是，有人提出，皮质柱可以调谐到复杂刺激的基本成分（例如，特定视网膜位置、方向、颜色对比度等的线段），通过内在V1电路以可预测的方式促进和抑制彼此，以产生复合整体的调谐。此外，任何给定V1神经元的复杂调谐都可以从其在皮层上相对于被复杂刺激激活的列的地理位置来预测。这一假设将通过研究不同类型复杂刺激的V1处理来验证，包括平滑轮廓、连接和纹理边界，这些对场景分割在感知上很重要。本研究将使用警觉猴V1的光学成像和电极记录相结合的方法。光学成像可以绘制神经元皮层的位置，这些神经元对每个复杂的刺激及其单独的成分做出反应。电极记录和相互关联，在光学图像的引导下，将用于测量同一组神经元之间的皮质内相互作用。基于这些相互作用计算的复杂调谐可以与测量值进行比较，从而验证我们的假设。提出的光学成像和电极记录的结合将使阐明V1调谐的皮层机制的几何结构成为可能，这是仅通过电极记录无法实现的目标。这项研究的结果将为理解复杂视觉刺激的皮层处理提供一个广泛的框架，这是包括盲人视觉假体设计在内的临床应用的重要一步。

DESCRIPTION (provided by applicant): The long-term goal of this work is to understand the neural mechanisms of visual form perception. The current project aims to study the processing of modestly complex visual stimuli such as smooth contours, junctions and texture boundaries in primary visual cortex (V1). Recent evidence suggests that V1 is much more than just a bank of filters that passively extracts the simplest elements (e.g. short oriented line segments) from all visual input, as was earlier believed. Rather, the elemental components of complex stimuli interact strongly with each other to give V1 responses that could differ dramatically from the responses to the individual components. The current project is intended to test the hypothesis that such complex visual processing can be predicted from the columnar architecture and intrinsic circuitry of VI. In particular, it is proposed that cortical columns tuned to elemental components of a complex stimulus (e.g. line segments at particular retinotopic positions, orientations, color contrast etc.) facilitate and inhibit each other in predictable ways through intrinsic V1 circuitry to generate tuning for the composite whole. Further, that the complex tuning of any given V1 neuron can be predicted from its geographical position on cortex relative to columns activated by the complex stimulus in question. This hypothesis will be tested by studying the V1 processing for different families of complex stimuli including smooth contours, junctions and texture boundaries that are perceptually important for scene segmentation. A combination of optical imaging and electrode recordings in alert monkey V1 will be used for this study. Optical imaging will allow for mapping the cortical positions of neurons responding to each complex stimulus and its individual components. Electrode recordings and cross correlations, guided by optical images, will be used to measure the intracortical interactions between the same groups of neurons. The complex tuning computed on the basis of these interactions can then be compared with the measured values and thus test our hypothesis. The proposed combination of optical imaging and electrode recording will make it possible to elucidate the geometry of cortical mechanisms underlying V1 tuning, a goal that is not reachable through electrode recordings alone. The results of this study will provide a broad framework for understanding the cortical processing of complex visual stimuli, an essential step for clinical applications including the design of visual prostheses for the blind.