Investigating scene processing in the human brain

每天，我们都在理解复杂的视觉环境，以执行简单但必不可少的行为。这些行为统称为导航示能，包括识别我们的周围环境（例如厨房内），选择适当的导航路径（例如通过休息室进入走廊的路线），识别要与之交互的对象（例如电灯开关）和避开障碍物（例如咖啡桌）。尽管我们执行这些行为的重要性和容易性，但我们对大脑，特别是视觉皮层是如何组织来完成这一任务的知之甚少。这种知识差距的一个潜在原因是，关于视觉皮层组织的观点在历史上主要是由两种不同的观点形成的，这两种观点在很大程度上是独立的。第一种观点关注视野中位置的重要性，并基于这样一个事实，即视觉皮层的早期阶段包含了视野的空间地图，称为视野地图。相反，第二种观点认为，大脑包含对某些类别有选择性的特殊区域，如面孔和场景。这些类别选择区域位于视觉皮层的后期阶段，被认为对视野中的位置不敏感，并且独立于这些早期视野图。然而，最近的证据已经确定了整个大脑的视野地图，远远超出了视觉皮层的早期阶段。此外，几个类别选择区域已被证明在空间上重叠多个地图。这种空间重叠表明这两个原则之间的相互作用，并提出了一个基本的和尚未回答的问题：这些视野地图在视觉感知中扮演什么角色？我们的目标是回答这个问题，关于大脑的场景选择区域，称为枕位区（OPA）和视野地图重叠。OPA对场景图像的反应比其他类别更强烈，可能在我们有效导航的能力中发挥作用。有趣的是，我们团队最近的研究发现，至少有五个独立的视野图与OPA重叠。这一发现是令人惊讶的，因为长期以来的想法，类别选择区域是不敏感的视野位置。它还提出了一种可能性，即这些地图中的每一个都执行不同但互补的功能。我们的目标是通过利用不同技术的优势来测试这种可能性。首先，我们将使用功能性磁共振成像（fMRI）来识别OPA和它在个体参与者中重叠的视野图。一旦确定，我们将使用fMRI测量处理不同类型场景期间这些地图中的活动。功能磁共振成像允许人们测量大脑中非常小（~ 2 mm）区域的活动，尽管相对缓慢（例如每2秒）。我们将通过使用一种每毫秒测量一次大脑活动的技术-脑磁图（MEG）来克服功能磁共振成像的相对迟缓。这将提供场景处理的时间方面何时发生的度量，例如当决定穿过房间的路线时。最后，我们将使用一种称为经颅磁刺激（TMS）的技术来测试这些大脑区域处理的因果性质。在TMS期间，快速变化的磁场被传递到头部的一小部分。这种磁场暂时地、无痛地扰乱了受刺激大脑区域的正常活动。与功能磁共振成像和脑磁图（MEG）提供的大脑活动的间接测量不同，经颅磁刺激可以提供个别区域对特定任务的贡献的因果证据。通过一系列新颖而雄心勃勃的项目，我们的目标是建立OPA视野图在复杂的视觉过程中发挥的作用，如导航启示。

Every day, we make sense of complex visual environments to carry out simple, yet essential behaviours. These behaviours, which together are referred to as navigational affordances, include identifying our surroundings (e.g. inside a kitchen), selecting appropriate paths for navigation (e.g. the route through your lounge and into your hallway), recognising objects to interact with (e.g. light switches) and avoiding obstacles (e.g. the coffee table). Despite the importance and ease at which we perform these behaviours, we know relatively little about how the brain, and the visual cortex in particular is organised to accomplish this. One potential reason for this knowledge gap is that ideas about the organisation of visual cortex have been shaped historically by two different perspectives that were largely independent.The first perspective focused on the importance of positions in the visual field and was grounded in the fact that early stages of visual cortex contains spatial maps of the visual field called visual field maps. In contrast, the second perspective proposed that the brain contains special areas selective to certain categories, such as faces and scenes. These category-selective areas, located at later stages of visual cortex, were thought to be insensitive to positions in the visual field and independent from these early visual field maps. However, recent evidence has identified visual field maps throughout the brain, spreading far beyond the early stages of visual cortex. Moreover, several category-selective areas have been shown to spatially overlap multiple maps. This spatial overlap suggests an interaction between these two principles and raises a fundamental and as yet unanswered question: What role do these visual field maps play in visual perception?It is this question that we aim to answer with respect to a scene-selective area of the brain, referred to as the Occipital Place Area (OPA) and the visual field maps it overlaps. OPA responds more strongly to images of scenes than to other categories and may play a role in our ability to navigate effectively. Intriguingly, recent research by our team identified at least five separate visual field maps overlapping OPA. This finding is surprising given the long-held idea that category-selective areas are insensitive to visual field position. It also raises the possibility that each of these maps performs different, yet complementary functions. We aim to test this possibility by capitalising on the strengths of different techniques. First, we will use functional magnetic resonance imaging (fMRI) to identify both OPA and the visual field map it overlaps in individual participants. Once identified, we will measure the activity in these maps during the processing of different types of scenes using fMRI. FMRI allows one to measure the activity of very small (~2mm) areas of the brain albeit relatively slowly (e.g. every 2 s). We will overcome the relative sluggishness of fMRI by using a technique that measures brain activity every millisecond - magnetoencephalography (MEG). This will provide measures of when in time aspects of scene processing occur, such as when deciding on the route to take through a room. Finally, we will test the causal nature of processing in these brain areas using a technique called transcranial magnetic stimulation (TMS). During TMS, a rapidly changing magnetic field is delivered to a small part of the head. This magnetic field temporarily and painlessly disrupts the normal activity of the stimulated brain region. Unlike both fMRI and MEG, which provide indirect measurements of brain activity, TMS can provide causal evidence of the contribution of individual areas to a particular task. Through a series of novel and ambitious projects we aim to establish the role that OPA visual field maps play in complex visual processes such as those involved in navigational affordances.