Two routes to face perception: evidence from psychophysics and computational modeling

The aim of this study was to separately analyze the role of featural and configural face representations. Stimuli containing only featural information were created by cutting the faces into their parts and scrambling them. Stimuli only containing configural information were created by blurring the faces. Employing an old-new recognition task, the aim of Experiments 1 and 2 was to investigate whether unfamiliar faces (Exp. 1) or familiar faces (Exp. 2) can be recognized if only featural or configural information is provided. Both scrambled and blurred faces could be recognized above chance level. A further aim of Experiments 1 and 2 was to investigate whether our method of creating configural and featural stimuli is valid. Pre-activation of one form of representation did not facilitate recognition of the other, neither for unfamiliar faces (Exp. 1) nor for familiar faces (Exp. 2). This indicates a high internal validity of our method for creating configural and featural face stimuli. Experiment 3 examined whether features placed in their correct categorical relational position but with distorted metrical distances facilitated recognition of unfamiliar faces. These faces were recognized no better than the scrambled faces in Experiment 1, providing further evidence that facial features are stored independently of configural information. From these results we conclude that both featural and configural information are important to recognize a face and argue for a dual-mode hypothesis of face processing. Using the psychophysical results as motivation, we propose a computational framework that implements featural and configural processing routes using an appearance-based representation based on local features and their spatial relations. In three computational experiments (Experiments 4–6) using the same sets of stimuli, we show how this framework is able to model the psychophysical data.     

Emotion recognition: The role of featural and configural face information

Several studies investigated the role of featural and configural information when processing facial identity. A lot less is known about their contribution to emotion recognition. In this study, we addressed this issue by inducing either a featural or a configural processing strategy (Experiment 1) and by investigating the attentional strategies in response to emotional expressions (Experiment 2). In Experiment 1, participants identified emotional expressions in faces that were presented in three different versions (intact, blurred, and scrambled) and in two orientations (upright and inverted). Blurred faces contain mainly configural information, and scrambled faces contain mainly featural information. Inversion is known to selectively hinder configural processing. Analyses of the discriminability measure (A′) and response times (RTs) revealed that configural processing plays a more prominent role in expression recognition than featural processing, but their relative contribution varies depending on the emotion. In Experiment 2, we qualified these differences between emotions by investigating the relative importance of specific features by means of eye movements. Participants had to match intact expressions with the emotional cues that preceded the stimulus. The analysis of eye movements confirmed that the recognition of different emotions rely on different types of information. While the mouth is important for the detection of happiness and fear, the eyes are more relevant for anger, fear, and sadness.     

Influence of subjective difficulty on the degree of configural and featural processing in face recognition1

This study aimed to assess the effect of subjective difficulty on the degree of featural or configural processing in face recognition. It could be assumed that featural processing is analytic processing ( Peterson & Rhodes, 2003 ), while configural processing is automatic processing ( Dunning & Stern, 1994 ). It has been suggested that task difficulty affected the automaticity of processing. Task difficulty was manipulated using a number of alternatives or time pressure. Subjective difficulty could also affect the automaticity of processing. If so, then subjective difficulty may in turn affect the degree of featural or configural processing. Participants in a difficult condition were given instructions mentioning that the face‐recognition task was difficult, while participants in the control condition were given no such instruction. The amount of retrieved featural information in the former condition was less than that in the latter. These results suggested that the degree of featural processing decreased when the participants found the task difficult.     

Configural information is processed differently in human action

Recent evidence indicates that observers' sensitivity to configural information in dynamic human action is disrupted when action is inverted, whereas sensitivity to featural action information is not. The current research involved two experiments that expand upon this basic finding. Experiment 1 revealed that featural and configural action information are processed similarly in static representations of action as in dynamic action. Experiment 2 indicated that configural processing is uniquely sensitive to orientation only in human action as compared to a similar control stimulus. These findings further support the idea that the perception of action recruits specialized orientation-specific configural processing, and parallel similar findings in face perception and visual expertise.     

Effects of geometric distortions on face-recognition performance

The importance of 'configural' processing for face recognition is now well established, but it remains unclear precisely what it entails. Through four experiments we attempted to clarify the nature of configural processing by investigating the effects of various affine transformations on the recognition of familiar faces. Experiment 1 showed that recognition was markedly impaired by inversion of faces, somewhat impaired by shearing or horizontally stretching them, but unaffected by vertical stretching of faces to twice their normal height. In experiment 2 we investigated vertical and horizontal stretching in more detail, and found no effects of either transformation. Two further experiments were performed to determine whether participants were recognising stretched faces by using configural information. Experiment 3 showed that nonglobal vertical stretching of faces (stretching either the top or the bottom half while leaving the remainder undistorted) impaired recognition, implying that configural information from the stretched part of the face was influencing the process of recognition--ie that configural processing involves global facial properties. In experiment 4 we examined the effects of Gaussian blurring on recognition of undistorted and vertically stretched faces. Faces remained recognisable even when they were both stretched and blurred, implying that participants were basing their judgments on configural information from these stimuli, rather than resorting to some strategy based on local featural details. The tolerance of spatial distortions in human face recognition suggests that the configural information used as a basis for face recognition is unlikely to involve information about the absolute position of facial features relative to each other, at least not in any simple way.     

Configural face encoding and spatial frequency information

Configural relations and a critical band of spatial frequencies (SFs) in the middle range are particularly important for face recognition. We report the results of four experiments in which the relationship between these two types of information was examined. In Experiments 1, 2A, and 2B, the face inversion effect (FIE) was used to probe configural face encoding. Recognition of upright and inverted faces and nonface objects was measured in four conditions: a no-filter condition and three SF conditions (low, medium, and high frequency). We found significant FIEs of comparable magnitudes for all frequency conditions. In Experiment 3, discrimination of faces on the basis of either configural or featural modifications was measured under the same four conditions. Although the ability to discriminate configural modifications was superior in the medium-frequency condition, so was the ability to discriminate featural modifications. We conclude that the band of SF that is critical for face recognition does not contribute preferentially to configural encoding.     

Face viewpoint effects about three axes: the role of configural and featural processing

We directly compared recognition for faces following 0 degrees-75 degrees viewpoint rotation about the yaw, pitch, and roll axes. The aim was to determine the extent to which configural and featural information supported face recognition following rotations about each of these axes. Experiment 1 showed that performance on a sequential-matching task was viewpoint-dependent for all three types of rotation. The best face-recognition accuracy and shortest reaction time was found for roll rotations, then for yaw rotations, and finally the worst accuracy and slowest reaction time was found for pitch rotations. Directional differences in recognition were found for pitch rotations, but not for roll or yaw. Experiment 2 provided evidence that, in all three cases, viewpoint-dependent declines in recognition were primarily driven by the loss of configural information. However, it also appeared that significant featural information was lost following yaw and pitch (but not roll) rotations. Together, these findings show that unfamiliar-face recognition is viewpoint-dependent following rotation about each axis (and in each direction), and that performance is based on the availability of configural and, to a lesser extent, featural information.     

Carry-over effects in face recognition: Processing mode or attentional window?

This study investigates whether or not the dominant processing mode (global or local) in a Navon task is carried over to a face recognition task when the dominant processing mode (configural or featural) in an encoding phase is manipulated. In Experiment 1, Navon figures that were equal in size to a facial photograph were prepared. Participants performed better on the face recognition task after reading large letters in the Navon figures than after reading small letters when configural processing was required in the encoding phase. In Experiment 2, Navon figures that were equal in size to the parts of a face were prepared. The performance after reading large letters did not differ from that after reading small letters. This suggests that both the dominant processing mode in Navon tasks and the size of Navon figures should be considered when discussing the effects of Navon tasks in face recognition.     

Faces elicit different scanning patterns depending on task demands

Eye movements were recorded while participants discriminated upright and inverted faces that differed with respect to either configural or featural information. Two hypotheses were examined: (1) whether featural and configural information processing elicit different scanning patterns; (2) whether fixations on a specific region of the face dominate scanning patterns. Results from two experiments were compared to examine whether participants’ prior knowledge of the kind of information that would be relevant for the task (i.e., configural vs featural) influences eye movements. In Experiment 1, featural and configural discrimination trials were presented in random order such that participants were unaware of the information that would be relevant on any given trial. In Experiment 2, featural and configural discrimination trials were blocked and participants were informed of the nature of the discriminations. The results of both experiments suggest that faces elicit different scanning patterns depending on task demands. When participants were unaware of the nature of the information relevant for the task at hand, face processing was dominated by attention to the eyes. When participants were aware that relational information was relevant, scanning was dominated by fixations to the center of the face. We conclude that faces elicit scanning strategies that are driven by task demands.     

Featural and configural face processing in adults and infants: a behavioral and electrophysiological investigation

We sought to elucidate the behavioral and electrophysiological correlates of face processing, in adults and infants, by manipulating either the featural or configural information within the face. Two different experiments are reported. In these experiments, event-related potentials (ERPs) were recorded from the scalp while adult, 8-month-old, and 4-month-old participants completed configural-change and featural-change face tasks. The infants also completed a behavioral visual paired-comparison task with featural and configural face changes. ERP results reveal hemispheric differences in processing featural but not configural changes for the N170 in adults. Furthermore, featural and configural changes are processed differently within the right and left hemispheres. The right hemisphere N170 is significantly greater for configural compared to featural changes. The left hemisphere N170, however, exhibits the opposite effect. Infant data suggest that similar to adults, 8-month-old, but not 4-month-old participants, exhibit similar hemispheric differences between featural and configural changes for the P400 component. Behavioral results suggest increased sensitivity to both featural and configural face changes in 8-month-olds compared to 4-month-olds.     

Losing face: impaired discrimination of featural and configural information in the mouth region of an inverted face

Given that all faces share the same set of features—two eyes, a nose, and a mouth—that are arranged in similar configuration, recognition of a specific face must depend on our ability to discern subtle differences in its featural and configural properties. An enduring question in the face-processing literature is whether featural or configural information plays a larger role in the recognition process. To address this question, the face dimensions task was designed, in which the featural and configural properties in the upper (eye) and lower (mouth) regions of a face were parametrically and independently manipulated. In a same–different task, two faces were sequentially presented and tested in their upright or in their inverted orientation. Inversion disrupted the perception of featural size (Exp. 1), featural shape (Exp. 2), and configural changes in the mouth region, but it had relatively little effect on the discrimination of featural size and shape and configural differences in the eye region. Inversion had little effect on the perception of information in the top and bottom halves of houses (Exp. 3), suggesting that the lower-half impairment was specific to faces. Spatial cueing to the mouth region eliminated the inversion effect (Exp. 4), suggesting that participants have a bias to attend to the eye region of an inverted face. The collective findings from these experiments suggest that inversion does not differentially impair featural or configural face perceptions, but rather impairs the perception of information in the mouth region of the face.     

人脸特征信息反转效应的实验研究

结构信息和特征信息的研究是近年来人脸认知的热点,特征信息是否存在反转效应不同的理论对此有不同的看法。通过对平均人脸特征信息阈限的研究,发现:平均人脸特征信息的阈限具有明显的反转效应;特征信息的识别难度是决定特征信息的识别是否出现反转效应的重要因素。研究支持整体假设,认为成人是将人脸作为一个整体进行认知的。     

眼睛区域构型信息与特征信息的跨维共变增益效应及其加工特异性

面孔知觉可能在区域尺度上发生多维信息整合, 但迄今无特异性实验证据。本研究在两个实验中操纵面孔眼睛区域或嘴巴区域的单维构型或特征信息, 测量人们觉察单维变化或跨维共变的敏感度, 以此检测面孔区域尺度上的多维信息整合有何现象与规律, 进而揭示面孔知觉的多维信息整合机制。实验获得3个发现：(1)正立面孔眼睛区域的信息变化觉察呈现出“跨维共变增益效应”, 跨维信息共变觉察的敏感度显著高于任意一种单维信息变化觉察的敏感度; (2)“跨维共变增益效应”只在正立面孔的眼睛区域出现, 在倒置面孔的眼睛区域、正立面孔的嘴巴区域或倒置面孔的嘴巴区域都没有出现, 因此具有面孔区域特异性和面孔朝向特异性; (3)就单维信息变化觉察而言, 眼睛区域的敏感度不会受到面孔倒置的损伤, 但是嘴巴区域的敏感度会受到面孔倒置的显著损伤。综合可知, 面孔知觉确实会发生区域尺度上的信息整合, 它不是普遍性的信息量效应, 而是特异性的眼睛效应(只发生在正立面孔的眼睛区域)。这是面孔整体加工(face holistic processing)在单维信息分辨和多维信息整合之间建立联系的必经环节。这提示我们对全脸多维信息知觉整合的理解需要从传统的面孔整体加工假设升级到以眼睛为中心的层级化多维信息整合算法(a hierarchical algorithm for multi-dimensional information integration)。     

What is in a view? The role of featural information in the recognition of unfamiliar faces across viewpoint transformation

In recognising a face the visual system shows a remarkable ability in overcoming changes in viewpoint. However, the mechanisms involved in solving this complex computational problem, particularly in terms of information processing, have not been clearly defined. Considerable evidence indicates that face recognition involves both featural and configural processing. In this study we examined the contribution of featural information across viewpoint change. Participants were familiarised with unknown faces and were later tested for recognition in complete or part-face format, across changes in view. A striking effect of viewpoint resulting in a reduction in profile recognition compared with the three-quarter and frontal views was found. However, a complete-face over part-face advantage independent of transformation was demonstrated across all views. A hierarchy of feature salience was also demonstrated. Findings are discussed in terms of the problem of object constancy as it applies to faces.     

Understanding the role of configural processing in face emotion recognition in Parkinson’s disease

This investigation examined whether impairment in configural processing could explain deficits in face emotion recognition in people with Parkinson’s disease (PD). Stimuli from the Radboud Faces Database were used to compare recognition of four negative emotion expressions by older adults with PD (n = 16) and matched controls (n = 17). Participants were tasked with categorizing emotional expressions from upright and inverted whole faces and facial composites; it is difficult to derive configural information from these two types of stimuli so featural processing should play a larger than usual role in accurate recognition of emotional expressions. We found that the PD group were impaired relative to controls in recognizing anger, disgust and fearful expressions in upright faces. Then, consistent with a configural processing deficit, participants with PD showed no composite effect when attempting to identify facial expressions of anger, disgust and fear. A face inversion effect, however, was observed in the performance of all participants in both the whole faces and facial composites tasks. These findings can be explained in terms of a configural processing deficit if it is assumed that the disruption caused by facial composites was specific to configural processing, whereas inversion reduced performance by making it difficult to derive both featural and configural information from faces.     

The face inversion effect—Parts and wholes: Individual features and their configuration

The face inversion effect (FIE) is a reduction in recognition performance for inverted faces (compared to upright faces) that is greater than that typically observed with other stimulus types (e.g., houses). The work of Diamond and Carey, suggests that a special type of configural information, “second-order relational information” is critical in generating this inversion effect. However, Tanaka and Farah concluded that greater reliance on second-order relational information did not directly result in greater sensitivity to inversion, and they suggested that the FIE is not entirely due to a reliance on this type of configural information. A more recent review by McKone and Yovel provides a meta-analysis that makes a similar point. In this paper, we investigated the contributions made by configural and featural information to the FIE. Experiments 1a and1b investigated the link between configural information and the FIE. Remarkably, Experiment 1b showed that disruption of all configural information of the type considered in Diamond and Carey's analysis (both first and second order) was effective in reducing recognition performance, but did not significantly impact on the FIE. Experiments 2 and 3 revealed that face processing is affected by the orientation of individual features and that this plays a major role in producing the FIE. The FIE was only completely eliminated when we disrupted the single feature orientation information in addition to the configural information, by using a new type of transformation similar to Thatcherizing our sets of scrambled faces. We conclude by noting that our results for scrambled faces are consistent with an account that has recognition performance entirely determined by the proportion of upright facial features within a stimulus, and that any ability to make use of the spatial configuration of these features seems to benefit upright and inverted normal faces alike.     

Perception of novel faces: the parts have it!

It has been suggested that, as a result of expertise, configural information plays a predominant role in face processing. We investigated this idea using novel and learned faces. In experiment 1, sixteen participants matched two subsequently presented blurred or scrambled faces, which could be either upright or inverted, in a sequential same -different matching task. By means of blurring, featural information is hampered, whilst scrambling disrupts configural information. Each face was unfamiliar to the participants and was presented for 1000 ms. An ANOVA on the d' values revealed a significant advantage for scrambled faces. In experiment 2, fourteen participants were tested with the same design, except that the second face was always intact. Again, the ANOVA on the d' values revealed a significant advantage for scrambled faces. In experiment 3 half of the faces were learned in a familiarisation block prior to the experiment. The ANOVA of these d' values revealed a significant interaction of familiarity and condition, showing that blurred stimuli were better recognised when the faces were familiar. These results suggest that recognition of novel faces, compared to learned faces, relies relatively more on the processing of featural information. In the course of familiarisation the importance of configural information increases.     

Picture-plane inversion leads to qualitative changes of face perception

Presenting a face stimulus upside-down generally causes a larger deficit in perceiving metric distances between facial features ("configuration") than local properties of these features. This effect supports a qualitative account of face inversion: the same transformation affects the processing of different kinds of information differently. However, this view has been recently challenged by studies reporting equal inversion costs of performance for discriminating featural and configural manipulations on faces. In this paper I argue that these studies did not replicate previous results due to methodological factors rather than largely irrelevant parameters such as having equal performance for configural and featural conditions at upright orientation, or randomizing trials across conditions. I also argue that identifying similar diagnostic features (eyes and eyebrows) for discriminating individual faces at upright and inverted orientations by means of response classification methods does not dismiss at all the qualitative view of face inversion. Considering these elements as well as both behavioral and neuropsychological evidence, I propose that the generally larger effect of inversion for processing configural than featural cues is a mere consequence of the disruption of holistic face perception. That is, configural relations necessarily involve two or more distant features on the face, such that their perception is most dependent on the ability to perceive simultaneously multiple features of a face as a whole.     

Face imagery is based on featural representations

The effect of imagery on featural and configural face processing was investigated using blurred and scrambled faces. By means of blurring, featural information is reduced; by scrambling a face into its constituent parts configural information is lost. Twenty-four participants learned ten faces together with the sound of a name. In following matching-to-sample tasks participants had to decide whether an auditory presented name belonged to a visually presented scrambled or blurred face in two experimental conditions. In the imagery condition, the name was presented prior to the visual stimulus and participants were required to imagine the corresponding face as clearly and vividly as possible. In the perception condition name and test face were presented simultaneously, thus no facilitation via mental imagery was possible. Analyses of the hit values showed that in the imagery condition scrambled faces were recognized significantly better than blurred faces whereas there was no such effect for the perception condition. The results suggest that mental imagery activates featural representations more than configural representations.     

Face recognition in 4- to 7-year-olds: processing of configural, featural, and paraphernalia information.

We tested 4- to 7-year-old children's face recognition by manipulating the faces' configural and featural information and the presence of superfluous paraphernalia. Results indicated that even with only a single 5-s exposure to the target face, most children could use configural and feature cues to make identity judgments. Repeated exposure and experimenter feedback enabled other children to do so as well. Even after attaining proficiency at identifying the target face, however, children's recognition was impaired when a superfluous hat was added to the face. Thus, although young children can process featural and configural face information, their memories are highly susceptible to disruption from superfluous paraphernalia.