视知觉恒常性的研究:神经机制与计算建模

视知觉恒常性是人类视觉系统维持稳定的物体认知的一项重要能力。尽管以往研究表明知觉恒常性是由大脑通过对变化的观察条件进行补偿调节得到，但其内在神经机制仍然未知。本课题通过两个方面来研究该问题：1)利用功能性磁共振成像、事件相关电位技术结合心理物理方法，探讨物体大小、深度知觉的神经机制，尤其是在视距改变的条件下高级视皮层对初级视皮层的反馈调节机制；2)利用数学建模，对物体大小、深度知觉恒常性建立符合大脑神经生理结构、模拟其神经机制的神经网络模型。本课题将加深对知觉恒常性的内在神经机制的理解，为进一步探索物体认知的视觉加工机制奠定基础，同时也为物体识别等相关应用研究提供理论指导。

Visual perceptual constancy is a remarkable ability of human visual system to maintain perceptual stability of an object. Although previous studies have shown that the brain can compensate for change in observation condition to achieve perceptual constancy, its underlying neural mechanisms remain unknown. In this project, we investigate the constancy problem through two aspects. First, studying the neural mechanisms of size and depth perception of an object by combing the methods of functional magnetic resonance imaging (fMRI), electroencephalograph (EEG) technique and psychophysics. Particularly, we are interested in the mechanisms of feedback regulation from higher visual areas to primary visual cortex, under the condition of changing viewing distance. Second, constructing a computational neural model of size and depth constancies based on the neurophysiological structure of the brain. The model will be tested by simulating the results from other neurophysiological studies and from the former study we proposed. The project will develop a further and deeper understanding of the underlying neural mechanisms of perceptual constancy, pave the way for exploring the visual processing mechanisms of object perception, and provide guidance of related application research such as object recognition.