Norm-based face space model revisited: The role of familiarity, distance-to-norm and vector direction

对一张脸的年龄、可爱度、表达的情感或吸引力的评估是在与之前遇到的其他脸的关系中进行的。因此，这种判断在很大程度上依赖于面孔的心理表征之间的关系。根据有影响力的基于规范的多维人脸空间模型（nMDFS, Valentine, 1991），这些表征位于中心规范或原型周围，与规范的距离定义了面部典型性。人脸的位置由从范数到其空间点的向量唯一定义，因此由向量长度（距离到范数，DTN）和向量方向（VD）的组合定义。大部分关于面部感知的文献都是基于该模型的假设（例如，适应性研究和身份后效通常在nMDFS的背景下进行解释）。有趣的是，在nMDFS的背景下，熟悉度对心理面孔表征的作用被忽视了。由于识别研究表明，与不熟悉的面孔相比，熟悉面孔的表征在质量上有所不同（更稳健，与图像无关）（Burton, Jenkins, & Schweinberger, 2011），我们预测熟悉的面孔会被认为比不熟悉的面孔更不典型，并在一项试点研究中证实了这一点。在研究1中，我们将研究熟悉度和DTN对nMDFS中人脸表征的联合作用，并评估反映熟悉度（N250）和DTN （P200）的人脸敏感脑电图成分。在研究2中，我们将通过nMDFS中的相似性评级来评估熟悉面孔与不熟悉面孔的DTN的作用。虽然DTN可以预测不熟悉面孔的典型化，但其他属性，如面孔的吸引力可能由DTN和VD共同定义：有证据表明，典型化可以在一定程度上预测吸引力，但与此同时，VD可以增加吸引力，例如，增强性二态性（Little, Jones, & DeBruine, 2011）。为了进一步发展和规范nMDFS，我们将以吸引力为例（研究3）研究DTN和VD的神经生理相关性，这有可能澄清先前关于面部吸引力的研究中有争议的结果。利用先进的图像处理方法和大脑活动的电生理记录，本项目旨在系统地研究nMDFS背景下面部感知的行为和神经认知机制。我们期望这个项目将(1)提高我们对面孔是如何被心理表征的理解，(2)进一步揭示熟悉和不熟悉的面孔表征的神经生理基础。

The evaluation of a face regarding its age, likability, expressed emotion or attractiveness occurs in relation to other faces, which have been encountered previously. Thus, such judgments heavily depend on the relation between mental representations of faces. According to the influential norm-based multidimensional face space model (nMDFS, Valentine, 1991) these representations are located around a central norm or prototype, with the distance to the norm defining face typicality. A face's location is uniquely defined by a vector from the norm to its point in space, and thus by a combination of vector length (distance-to-norm, DTN) and vector direction (VD). A vast part of the face perception literature is based on assumptions of this model (for instance, adaptation studies and identity aftereffects in particular are typically interpreted in the context of nMDFS).Interestingly, the role of familiarity for mental face representations has been neglected in the context of nMDFS. Because recognition studies suggest that representations for familiar faces are qualitatively different (more robust and image-independent) than those for unfamiliar ones (Burton, Jenkins, & Schweinberger, 2011), we predicted that familiar faces would be perceived as less typical than unfamiliar ones, and confirmed this in a pilot-study. In Study 1, we will investigate the combined roles of familiarity and DTN on face representations in nMDFS, and assess face-sensitive EEG components which reflect familiarity (N250) and DTN (P200). In Study 2, we will assess the role of DTN of familiar faces compared to unfamiliar faces via similarity ratings in nMDFS.While DTN may predict the typicality of unfamiliar faces, other attributes such as the attractiveness of a face may be defined by both DTN and VD: Evidence suggests that typicality can predict attractiveness to some degree, but at the same time attractiveness can be increased by VD, for example, enhanced sexual dimorphism (Little, Jones, & DeBruine, 2011). Aiming at a further development and specification of the nMDFS, we will investigate neuro-physiological correlates of DTN and VD on the example of attractiveness (Study 3), which has the potential of clarifying controversial results of previous research on facial attractiveness.Using both advanced methods of image processing and electrophysiological recordings of brain activity, the present project aims at a systematic investigation of the behavioural and neurocognitive mechanisms underlying face perception in the context of nMDFS. We expect that this project will (1) improve our understanding of how faces are mentally represented and (2) further uncover the neurophysiological basis of familiar and unfamiliar face representations.