Proaction Laboratory, Faculty of Psychology and Educational Sciences, University of Coimbra, 3004-531 Coimbra, Portugal
Proaction Laboratory, Faculty of Psychology and Educational Sciences, University of Coimbra, 3004-531 Coimbra, Portugal.
J Neurosci. 2021 May 26;41(21):4678-4685. doi: 10.1523/JNEUROSCI.2857-20.2021. Epub 2021 Apr 13.
Human object recognition is dependent on occipito-temporal cortex (OTC), but a complete understanding of the complex functional architecture of this area must account for how it is connected to the wider brain. Converging functional magnetic resonance imaging evidence shows that univariate responses to different categories of information (e.g., faces, bodies, and nonhuman objects) are strongly related to, and potentially shaped by, functional and structural connectivity to the wider brain. However, to date, there have been no systematic attempts to determine how distal connectivity and complex local high-level responses in occipito-temporal cortex (i.e., multivoxel response patterns) are related. Here, we show that distal functional connectivity is related to, and can reliably index, high-level representations for several visual categories (i.e., tools, faces, and places) within occipito-temporal cortex; that is, voxel sets that are strongly connected to distal brain areas show higher pattern discriminability than less well-connected sets do. We further show that in several cases, pattern discriminability is higher in sets of well-connected voxels than sets defined by local activation (e.g., strong amplitude responses to faces in fusiform face area). Together, these findings demonstrate the important relationship between the complex functional organization of occipito-temporal cortex and wider brain connectivity. Human object recognition relies strongly on OTC, yet responses in this broad area are often considered in relative isolation to the rest of the brain. We employ a novel connectivity-guided voxel selection approach with functional magnetic resonance imaging data to show higher sensitivity to information (i.e., higher multivoxel pattern discriminability) in voxel sets that share strong connectivity to distal brain areas, relative to (1) voxel sets that are less strongly connected, and in several cases, (2) voxel sets that are defined by strong local response amplitude. These findings underscore the importance of distal contributions to local processing in OTC.
人类物体识别依赖于枕颞叶皮层（OTC），但要完全理解该区域复杂的功能结构，必须考虑到它与大脑其他区域的连接方式。趋同的功能磁共振成像证据表明，对不同类别的信息（例如，面部、身体和非生物物体）的单变量反应强烈相关，并且可能受到与大脑其他区域的功能和结构连接的影响。然而，迄今为止，还没有人系统地试图确定 OTC 中远端连接和复杂的高级局部反应（即多体素反应模式）之间的关系。在这里，我们表明，远端功能连接与 OTC 中几个视觉类别（即工具、面部和地点）的高级表示相关，并且可以可靠地对其进行索引；也就是说，与远端大脑区域强烈连接的体素集比连接较弱的体素集具有更高的模式可辨别性。我们进一步表明，在几种情况下，连接良好的体素集的模式可辨别性高于由局部激活定义的体素集（例如，梭状回面孔区中对面孔的强烈振幅反应）。总之，这些发现表明 OTC 复杂的功能组织与大脑广泛连接之间存在重要关系。人类物体识别强烈依赖于 OTC，但该广泛区域的反应通常被认为与大脑的其他部分相对孤立。我们采用一种新的基于连接的体素选择方法，使用功能磁共振成像数据来显示在与远端大脑区域具有强烈连接的体素集中，信息的敏感性更高（即多体素模式可辨别性更高），与（1）连接较弱的体素集相比，在某些情况下，（2）与由强烈局部响应幅度定义的体素集相比。这些发现强调了远端对 OTC 中局部处理的重要贡献。
