Object, face and body representations in the human brain

这项研究的目的是了解我们如何看到我们所看到的：大脑如何分析落在眼睛视网膜上的光线，以编码一个充满物体，人和事物的世界？ 在过去的一年里，我们继续调查i）自下而上（感觉驱动）和自上而下（内部驱动）的大脑加工之间的相互作用，重点是视觉心理意象，工作记忆和任务的效果，和ii）复杂视觉刺激的感知，最近集中在视觉场景。 i）自上而下的处理。我们的视觉感知是自下而上的感官信息和自上而下的信号之间相互作用的产物，这些信号引导对输入的解释并反映先验知识和意图。 在没有感觉输入的情况下，心理意象完全依赖于这种自上而下的信号，并提供了一个研究其对感觉皮层区域影响的机会。使用功能磁共振成像，我们进行了详细的比较视觉图像和知觉的个别复杂的对象。我们发现，（1）我们可以解码参与者在多个大脑区域看到或想象的特定物体的身份，（2）图像和感知信息在整个视觉处理流中的分布不同。这些发现表明，虽然意象和感知涉及相同的大脑区域，但在意象和感知下运行的神经动力学是不同的。 在工作记忆过程中，物体信息在没有感觉输入的情况下也会被记住。我们一直在研究哪些皮层区域在工作记忆中是活跃的，以及它们包含哪些信息（视觉信息与概念信息）。 最后，不同的任务需要从视觉刺激中提取不同类型的信息，依赖于自上而下的信号。我们一直在研究复杂视觉刺激的表征如何根据参与者正在执行的任务而变化。在大脑的多个区域，我们发现了关于任务的信息。此外，这个信息是不同的任务，强调视觉刺激的物理特性和任务，强调概念属性。 2.对复杂视觉刺激的感知。现实世界的场景是非常复杂和异构的，但我们能够毫不费力地识别和分类它们。虽然先前的研究已经确定了一个似乎专门用于场景处理的大脑区域，但仍然不清楚这个区域的确切作用是什么。我们向参与者展示了大量复杂的现实世界场景，并使用数据驱动的fMRI方法来识别该区域的表征性质。我们发现，在这个区域中的场景表示主要反映场景的空间属性（例如，无论它们是开放的还是封闭的），而不是语义属性（即场景类别）。 此外，我们研究了复杂视觉场景的不同元素是如何在场景观看过程中参与的不同大脑区域中表现出来的。具体来说，我们创建了由空间背景上的单个物体组成的人工视觉场景，使我们能够将不同大脑区域中代表的空间和物体信息分开。有些区域主要包含关于物体的信息，一些空间属性，而一个区域包含关于空间和物体的信息。这些特性与我们最近开发的描述大脑中视觉空间信息处理的神经解剖学框架一致。 阐明大脑如何使我们能够识别物体，场景，面孔和身体，为我们对周围世界的内部表征的本质提供了重要的见解。理解这些表征对于试图确定许多精神健康和神经系统疾病的潜在缺陷至关重要。

The goal of this research is to understand how we see what we see: how does the brain analyze the light falling on the retina of the eye to encode a world full of objects, people and things? During the past year we have continued to investigate i) the interaction between bottom up (sensory driven) and top-down (internally driven) processing in the brain, focusing on visual mental imagery, working memory and effect of task, and ii) perception of complex visual stimuli, focusing most recently on visual scenes. i) Top-down processing. Our visual perception is the product of an interaction between bottom-up sensory information and top-down signals guiding interpretation of the input and reflecting prior knowledge and intent. Mental imagery, in the absence of sensory input, relies entirely on this top-down signal, and provides an opportunity to investigate its impact on sensory cortical areas. Using fMRI we conducted a detailed comparison of visual imagery and perception for individual complex objects. We found that (1) we can decode the identity of the specific object participants view or imagine in multiple brain regions, and (2) imagery and perceptual information are distributed differently throughout the visual processing stream. These findings suggest that while imagery and perception engage the same brain regions, the neural dynamics operating under imagery and perception are different. During working memory, object information is also held in mind in the absence of sensory input. We have been investigating which cortical regions are active during working memory and what information they contain (visual versus conceptual information). Finally, different tasks require different types of information to be extracted from visual stimuli, relying on top down signals. We have been investigating how the representations of complex visual stimuli vary according to the task a participant is performing. In multiple regions throughout the brain, we found information about task. Further, this information was different for tasks that emphasized physical properties of the visual stimuli and tasks that emphasized conceptual properties. ii) Perception of complex visual stimuli. Real-world scenes are incredibly complex and heterogeneous, yet we are able to identify and categorize them effortlessly. While prior studies have identified a brain region that appears specialized for scene processing, it remains unclear exactly what the precise role of this region is. We presented participants with large numbers of complex real-world scenes and used a data-driven fMRI approach to identify the nature of representations in this region. We found that scene representations in this region primarily reflect the spatial properties of scenes (e.g. whether they are open or closed) and not the semantic properties (i.e. scene category). Further, we examined how different elements of complex visual scenes are represented across the different brain regions engaged during scene viewing. Specifically, we created artificial visual scenes comprising a single object on a spatial background, enabling us to tease apart the spatial and object information represented in different brain regions. Some regions contained information primarily about objects, some spatial properties, while one region contained information about both space and objects. These properties were consistent with a recent neuroanatomical framework we developed describing the processing of visuospatial information in the brain. Elucidating how the brain enables us to recognize objects, scenes, faces and bodies provides important insights into the nature of our internal representations of the world around us. Understanding these representations is vital in trying to determine the underlying deficits in many mental health and neurological disorders.