Probing prefrontal influences on the emergence of visual category representations

为了理解这个世界，我们的视觉大脑将大量可能的输入模式归类为相对较少的有意义的类别，例如面孔或场景。长期以来，这些类别的皮层表征一直被认为仅仅来自前馈处理。然而，越来越清楚的是，视觉皮层中的类别表征是由来自其他大脑区域的自上而下的连接形成的，包括前额叶皮层（PFC）。PFC对视觉处理的影响提出了一个问题，即PFC自上而下的控制在多大程度上是类别表征出现所必需的。为了回答这个问题，本项目将研究当PFC的自上而下控制受损时的视觉类别表征。具体来说，重复经颅磁刺激（TMS）将用于破坏健康参与者的PFC活动长达30分钟。在此期间，我们将探讨视觉类别表征的多模态方法。通过分析fMRI和EEG记录，我们将精确地确定在正常处理和PFC自上而下控制受损的情况下，不同视觉类别的大脑表征出现的位置和时间。将这些记录与计算前馈模型进行比较，我们将量化前额叶控制覆盖视觉自下而上处理的程度。最后，在行为分类实验中，我们将使用一个“双站点”TMS设计，同时破坏视觉皮层和PFC。这些实验将因果地证明前馈处理如何与PFC的反馈收敛，以锻造指导我们的行为的类别表征。我们的项目将提供一个全面的因果关系如何PFC影响视觉类别处理在神经元，计算和行为水平。该项目的进展有可能为我们提供更好的皮层通信模型，并激发对这种通信障碍后果的新研究。

To make sense of the world, our visual brain groups a vast amount of possible input patterns into a relatively small number of meaningful categories, such as faces or scenes. Cortical representations of these categories have long been assumed to emerge solely from feedforward processing. However, it has become increasingly clear that category representations in visual cortex are shaped by top-down connections from other brain areas, including the prefrontal cortex (PFC). The influence of PFC on visual processing raises the question to which extent top-down control from PFC is necessary for category representations to emerge. To answer this question, this project will study visual category representations when top-down control from PFC is impaired. Specifically, repetitive transcranial magnetic stimulation (TMS) will be used to disrupt PFC activity in healthy participants for up to 30 minutes. During this time period, we will probe visual category representations in a multi-modal approach. Analyzing fMRI and EEG recordings, we will precisely establish where and when brain representations of different visual categories emerge under normal processing and under impaired top-down control from the PFC. Comparing these recordings to computational feedforward models, we will quantify the extent to which prefrontal control overrides visual bottom-up processing. Finally, in behavioral categorization experiments, we will use a “dual-site” TMS design to simultaneously disrupt visual cortex and PFC. These experiments will causally demonstrate how feedforward processing converges with feedback from PFC to forge the category representations that guide our behavior. Our project will provide a comprehensive causal account of how PFC impacts visual category processing at the neuronal, computational, and behavioral level. Advances made in this project have the potential to provide us with better models of cortical communication and inspire new research into the consequences of impairments to this communication.