Visual Adaptation and Neuronal Selectivity

灵长类动物大脑视觉处理的一个显著特征是，存在对高度特定刺激的图片做出反应的神经元，例如动物或人类熟悉的人脸或复杂物体。虽然这些神经元的存在显然与学习和记忆有关，但正如大量病变实验所表明的那样，它们也位于复杂的一般处理流的末端，其生理原理尚未完全阐明。也许出于这个原因，它们在视觉感知、图片记忆和更一般的物体识别中对特征或物体反应神经元的确切作用一直很难确定。 大脑如何代表大量类似的复杂物体，如人脸的概念框架表明，视觉系统能够根据刺激类别的特定统计数据进行自我调整。一种在人类心理物理学文献中根深蒂固，但相对较少受到关注的框架有时被称为基于规范的编码。再来看看面孔的例子，许多证据表明，大脑通过将面孔与内部存储的平均原型进行比较来处理面孔。这一平均面孔将作为解释其他面孔结构的标准。一张眉毛比平均大，鼻子比平均细的脸会用这些相对术语编码，而不是按照它的绝对标准编码。我们之前的神经生理学和心理物理学研究支持这一观点，尽管这一领域的研究才刚刚开始。 在实验室里，我们目前正在启动一些项目，试图阐明颞叶下皮质的编码原理。我们特别感兴趣的是选择性反应是如何随着经验而产生的，以及改变一个人所接触到的面孔的性质(例如种族、性别)是否可以系统地塑造大脑这一部分的神经元参与面孔分析的方式。我们正在使用计算机程序来生成由人脸和非人脸图案组成的变形刺激集。我们的目标是观察神经调节的变化是否发生在反映灵长类面部知觉特征的学习模式的短和/或长时间尺度上。最近的技术进步使得一次监测一个神经元成为可能，其中大脑中有超过1000亿个神经元。在此期间，我们希望确定单个神经元在复杂图像的分析中是否具有坚定不移的作用，或者特定的模式是否导致其选择性的持续调整。 在一个相关的子项目中，我们正在调查对一个人脸进行几秒钟的适应是否会影响后续人脸的处理方式。这项研究的动机是观察到，面孔误解可能是这种刺激序列的心理结果。这些感知通常被称为适应性后效，可以让一张脸看起来像另一张脸，这些变化是以每一张脸偏离平均原型的方式系统地相关的。这些后遗症的形式已经产生了关于面部编码神经元在类似适应条件下可能如何反应的假设。我们的初步证据表明，猴子大脑下颞叶皮质中的神经元可能直接参与了这一过程。我们正在试图了解这种适应范式是否可以提供对灵长类大脑颞叶神经表征的本质的洞察，特别是这种神经元在物体感知和识别的各个方面的作用是否可以被识别。

A striking feature of visual processing in the primate brain is the existence of neurons that are reactive to pictures of highly specific stimuli, such as faces or complex objects with which an animal or human is familiar. While the existence of these neurons is clearly related to learning and memory, as a large body of lesion experiments have shown, they also sit at the end of a complex general processing stream whose physiological principles have yet to be completely elaborated. Perhaps for this reason, their precise role of such feature- or object-responsive neurons in visual perception, picture memory, and more general object recognition has been difficult to ascertain. Conceptual frameworks for how the brain represents a large number of similar complex objects, such as faces, suggest that the visual system is able to tailor itself to the specific statistics of a stimulus category. One framework that has its roots deep in the human psychophysics literature, but which has received relatively little attention, is sometimes called norm-based coding. Considering again the example of faces, much evidence suggests that the brain processes faces by comparing them to an internally stored average prototype. This average face would serve as a norm for interpreting the structure of other faces. A face with larger than average eyebrows and a thinner than average nose would be encoded in those relative terms, rather than according to its absolute metrics. Our previous neurophysiological and psychophysical work supports this view, though research in this area is only beginning to take hold. In the laboratory, we are presently starting projects that attempt to elucidate coding principles in the inferotemporal cortex. We are particularly interested in how selective responses come about with experience, and whether changing the nature of the faces to which one is exposed (e.g. race, gender) can systematically shape the way in which neurons in this part of the brain participate in face analysis. We are using computer programs to generate morphed stimulus sets consisting of both face and non-face patterns. The goal is to observe whether changes in neural tuning occur over short and/or long time scales that reflect the pattern of learning that characterizes primate face perception. Recent technological advances have made it possible to monitor a single neuron, of which there are over 100 billion in the brain, for several weeks at a time. During this period, we hope to establish whether a single neuron has an unwavering role in the analysis of complex images, or whether the specific patterns lead to continual adjustments in its selectivity. In a related subproject, we are investigating whether the adaptation to one face for several seconds affects the way that a subsequent face is processed. The motivation for this study is the observation that face misperceptions can be the psychological result of this stimulus sequence. These perceptions, commonly known as adaptational aftereffects, can make one face appear as another, with the changes being systematically related way each of the faces deviates from the average prototype. The form of these aftereffects has generated hypotheses about how face-encoding neurons might respond under similar adaptation conditions. Our preliminary evidence shows that neurons in the inferior temporal cortex of the monkey brain are likely to be directly involved in this process. We are trying to understand whether such adaptation paradigms can provide insight into the nature of the neural representations in the temporal lobe of the primate brain, and specifically whether the role of such neurons in various aspects of object perception and recognition can be identified.