Time to investigate human visual shape perception and recognition

这项研究旨在确定视觉信息的性质和时间进程，以确定人类视觉对象识别(VOR)的早期加工路线，以及根据物品的独特身份对其进行视觉识别的类别专门机制的运作。我们还旨在研究随着加工过程在人类大脑皮层视觉区域的层级上移，不断增大的接受野大小对功能的影响。这些研究将有助于解决与我们理解人类VOR系统的一些最基本特征有关的悬而未决的问题。为了确定加工时间进程，我们将使用我们开发的一种新的心理物理技术，即随机时间抽样[参见参考文献]。2]。这种方法可以揭示加工效率是如何随着时间和加工的时间特征而演变的，这些加工特征在个体之间具有极高的可重复性，也可以高度区分刺激类别和/或任务参与者执行的任务。对于支持VOR的视觉信息，将使用两种方法。一种是使用离散小波分解(DWT)分析来自不同类别(人脸、地点、物体和单词)的大量图像的空间特征。离散小波变换的兴趣在于图像由一个局部算子(Daubechies小波)处理，该算子提供关于图像对比度的尺度(即空间范围)和方向的局部信息。这些功能特征与人类初级视觉皮质(V1)的功能特征相匹配。DWT图像描述将用于描述刺激类别之间的视觉对比以及在这些类别中识别刺激所需的辨别类型。另一种方法是理想的观察者模型，它要么具有无限的处理能力，要么能够访问类似于人类低级视觉编码的视觉信息(即对比度敏感函数、倾斜效应和随着偏心率的增加而视觉敏感度降低)。人类参与者和理想观察者的结果之间的差异揭示了人类视觉系统固有的处理限制和/或偏见。提出的研究将为人类VOR的基本特性提供丰富、可靠的数据集。预计这一结果将有力地刺激未来的研究，并有助于形成比目前更好的解释能力的VOR理论。COVID19应急计划：由于COVID19，我们实验室的三个测试室已经重新组织，以便我们在任何时候都可以与参与者保持至少2米的距离。目前可用的设备和我们同时使用的NSERC研究工具和仪器申请中要求的设备都与该组织完全兼容，因此我们估计未来由于大流行而导致的实验室操作中断的可能性极小。

The research proposed aims to determine the nature and time course of the visual information subtending the early routing of processing toward category-specialized mechanisms for visual object recognition (VOR) in humans and the operation of category-specialized mechanisms for the visual recognition of items in terms of their unique identity. We also aim to investigate the functional impact of growing receptive field sizes as processing moves up the hierarchy of visual areas in the human cortex. These studies will help resolve outstanding issues pertaining to our understanding of some of the most fundamental features of the human VOR system. To determine processing time course, we will be using a new psychophysical technique we have developed, which is random temporal sampling [see ref. 2]. This method can reveal how processing effectiveness evolves through time as well as temporal features of processing which are extremely replicable across individuals and also highly discriminant of the stimulus class and/or task participants perform. Regarding the visual information subtending VOR, two approaches will be used. One is the analysis of the spatial features of a large collection of images from different categories (faces, places, objects, and words) using discrete wavelet decomposition (DWT). The interest of DWT is that images are processed by a local operator (Daubechies wavelet) which provides localized information about the scale (i.e. spatial extent) and orientation of image contrasts. These functional features match those of the human primary visual cortex (V1). The DWT image descriptions will serve to characterize the visual contrasts between stimulus categories as well as the types of discrimination required for stimulus identification in these categories. The other method is the ideal observer model with either unlimited processing capacity or with an access to visual information restricted similarly to human low-level visual encoding (i.e. contrast-sensitivity function, oblique effect and reduction of visual acuity with increasing eccentricity). Divergences between the results of human participants and the ideal observers reveal processing limitations and/or biases intrinsic to the human visual system. The studies proposed will offer a rich, robust set of data on fundamental properties of VOR in humans. This outcome is expected to strongly stimulate future research and to contribute to the formulation of VOR theories with better explanatory power than those available today. COVID19 contingency plan: Due to COVID19, the three testing rooms in our lab have been reorganized such that we can maintain a minimum 2 meter distance from participants at all times. Both the equipment presently available and that requested in our concurrent NSERC Research tools and instruments application are fully compatible with this organization and we thus estimate future interruptions in the lab operations due to the pandemic as highly unlikely.