Social Processing and Neural Plasticity

社会感知、记忆、抽象和行为的核心运作来自于专门的大脑回路，这些回路是通过童年及以后的丰富经验形成的。经验的这种作用意味着个体的大脑在能力上不同，在某种程度上，在组织和生理上也不同。在脑损伤和疾病的情况下，大脑的差异直接转化为能力的变化。例如，颞叶非常特定的部位受损可能导致无法识别单个面孔，而不会影响对其他物体的识别。其他区域的损伤可能会使受试者难以识别面部表情，语音语调，甚至促使他们相信他们的配偶是骗子。类似地，精神疾病中回路的异常发展，例如那些刺激社会互动的回路，可能导致在理解社会信息方面的削弱性知觉缺陷。 虽然研究人员已经了解了大约50年，视觉大脑专门用于社会信息，但这种专门化的方式，随着时间的推移而不断完善，并支持我们的流体社会行为仍然知之甚少。 大脑天生具有适应性，特别是在早期生活中。 在我们的实验室中，我们采用的方法使我们能够随着时间的推移跟踪神经活动模式，包括几天，几周，甚至几个月。这套工具包括功能性磁共振成像（fMRI）、单细胞记录和光学记录，使我们能够研究社会刺激的视觉感知学习如何在大脑中表达。我们的大部分工作集中在面部斑块上，这是颞叶和前额皮质的小而有限的区域，与其他类别的刺激相比，它们对面部表现出更大的fMRI反应。我们研究的纵向性质使我们能够在多个时间尺度上研究可塑性。这项研究得到了进一步的帮助，我们开发了一个虚拟面孔刺激，其动画，面部表情和环境背景是在完全的实验控制。这个刺激工具箱对于系统地研究决定神经放电的因素非常有用，例如在自然视觉几何学的背景下。 在过去的一年里，我们发表了与人脸感知的重要方面相关的论文。例如，我们证明了听觉和视觉信息在发声行为期间整合在面部选择神经元中（Khandhadia等人，2021）。 虽然这种整合在一个与面部表情有关的面部选择区域中很明显，但在另一个与面部身份有关的区域中几乎完全不存在。 我们还提供了证据表明，面部身份的识别利用了学习的，平均面部的内部表征（Koyano等人，2021）。这种规范性的刺激塑造了神经元在面部斑块上的反应，突出了独特的，因此可识别的面部特征。大脑使用面部识别的规范最初是基于心理物理学研究假设的。我们在面部斑块系统中对平均面部的清晰神经表征的展示支持了这一观点，并提出了一种更普遍的模式，大脑通过其自然经验收集和存储信息。 在另一项研究中，我们采用了一种完全不同的方法来理解面部神经反应的决定因素。在这项研究中，我们研究了驱动不同面部斑块中神经元的关键面部元素。面部元素在不同的图像上随机混合，例如，一个人的眼睛与许多其他人的嘴，头和身体随机呈现。这种组合方法的一个显著结果是，给定面部的一个非常小的区域可以在多大程度上决定神经元的反应，以及对这一重要特征的反应对场景的重大变化的鲁棒性。例如，选择性地对一组眼睛做出反应的神经元通常会继续对这些眼睛做出反应，即使它们被放置在不同的面部中，该面部附着在不同的身体上并放置在不同的场景中。这些特征的局部优势是令人惊讶的，特别是因为它们是从与上述基于范数的编码发现相同的神经群体中测量的。 在另一项研究中，我们正在研究自然几何在面孔编码中的作用。 物理几何学的作用，如面部的公制尺寸和观察者的厘米距离，尚未研究。我们使用我们的头像脸来确定面部选择性斑块中的神经元对面部大小等因素高度敏感。 此外，在使用不同的尺寸/距离组合时，我们能够确定关键变量成形尺寸响应是以厘米为单位的物理尺寸，而不是以度为单位的视网膜角度。 这项研究表明，腹侧视觉通路中的编码对象的信息，包括自然的几何形状的对象，以及它在场景中的物理位置。 在最后一项研究中，我们一直在跟踪细胞Ca++荧光反应后，介绍了面孔和其他图像刺激。 与微电极记录类似，这些记录也允许在数周和数月内进行纵向跟踪。本研究旨在了解细胞选择性的本质，以及相邻神经元之间偏好的相似性。 在未来，我们的Ca++荧光工作也将被用来调查是否在面对选择性地区的神经元修改他们的响应配置文件作为受试者学习新面孔。

The core operations of social perception, memory, abstraction, and behavior derive from specialized brain circuits that are shaped through abundant experience during childhood and beyond. This role of experience means that the brains of individuals differ in their capacities and, to some extent, in their organization and physiology. In the case of brain damage and diseases, the differences in the brain translate directly to changes in abilities. For example, damage to very specific parts of the temporal lobe can lead to the inability to recognize individual faces while not affecting the recognition of other objects. Damage in other areas can leave subjects with a difficulty in recognizing facial expression, voice intonation, or even prompt them to believe that their spouse is an impostor. Similarly, abnormal development of the circuits in psychiatric diseases, for example those spurring social interaction, can lead to debilitating perceptual deficits in understanding social information. While researchers have understood for approximately 50 years that the visual brain is specialized for social information, the manner in which this specialization comes about, is refined over time, and supports our fluid social behavior remains poorly understood. The brain is inherently adaptable, particularly in early life. In our laboratory, we employ methods that allow us to track neural activity patterns over time, including days, weeks, and even months. This set of tools, which includes functional magnetic resonance imaging (fMRI), single-cell recordings, and optical recordings, allows us to study how visual perceptual learning of social stimuli is expressed in the brain. We have focused much of this work on face patches, which are small, circumscribed regions of the temporal and prefrontal cortex showing greater fMRI responses to faces than to other categories of stimuli. The longitudinal nature of our studies allows us to investigate plasticity over multiple time scales. This research has been further aided by our development of an avatar face stimulus, whose animation, facial expressions, and environmental context is under complete experimental control. This stimulus toolbox has been of great use for systematically studying the factors that determine neural firing, for example in the context of the geometry of natural vision. In the past year, we have published papers related to important aspects of face perception. For example, we demonstrated that auditory and visual information are integrated among face-selective neurons during vocal behavior (Khandhadia et al, 2021). While this integration was pronounced in one face selective region concerned with facial expression, it was nearly entirely absent in a different region concerned with face identity. We also provided evidence that the recognition of facial identity utilizes learned, internal representation of the average face (Koyano et al, 2021). This normative stimulus shapes the responses of neurons across face patches, highlighting the distinctive, and hence recognizable, facial features. The brains use of norms for face recognition was originally hypothesized based on psychophysical studies. Our demonstration of a clear neural representation of the average face within the face patch system bolsters this view and suggests a more general mode by which the brain gathers and stores information through its natural experience. In a different study, which is now accepted for publication (Waidmann et al, 2022), we took an entirely different approach to understanding the determinants of neural responses to faces. In that study, we investigated critical facial elements for driving neurons in different face patches. Facial elements were randomly intermixed on different images, for example with the eyes from one individual presented randomly with the mouths, heads, and bodies from many other individuals. One striking result from this combinatorial approach was the extent to which a very small region of a given face can determine the neuron's response, and the extent to which responses to this important feature are robust to major changes to the scene. For example, neurons responding selectively to one set of eyes often continue to respond to those eyes even when they are placed within a different face, which is attached to a different body and placed in a different scene. The local domination of such features was surprising, particularly as they were measured from the same neural populations as the norm-based coding findings mentioned above. In another study , we are investigating the role of natural geometry in the encoding of faces. The role of physical geometry, such as the metric size of a face and distance to the observer in centimeters, has not been studied. We used our avatar face to establish that neurons in face selective patches are highly sensitive to factors such as face size. Moreover, in using different size/distance combinations, we were able to determine that the critical variable shaping size responses was the physical size in centimeters, rather than the retinal angle in degrees. This study suggests that the encoding of objects in the ventral visual pathway incorporates information about the natural geometry of objects, as well as its physical location within a scene. In a final study, we have been tracing the cellular Ca++ fluorescence responses following the presentation of faces and other image stimuli. Similar to the microelectrode recordings, these recordings also permit longitudinal tracking over weeks and months. The present study aims to understand the nature of cellular selectivity, as well as the similarity of preference among neighboring neurons. In the future, our Ca++ fluorescence work will also be used to investigate whether neurons in face selective areas modify their response profiles as the subjects learn new faces.