Facial identity and facial speech processing: Familiar faces and voices in the McGurk effect

An experiment was conducted to investigate the claims made by Bruce and Young (1986) for the independence of facial identity and facial speech processing. A well-reported phenomenon in audiovisual speech perception—theMcGurk effect (McGurk & MacDonald, 1976), in which synchronous but conflicting auditory and visual phonetic information is presented to subjects—was utilized as a dynamic facial speech processing task. An element of facial identity processing was introduced into this task by manipulating the faces used for the creation of the McGurk-effect stimuli such that (1) they were familiar to some subjects and unfamiliar to others, and (2) the faces and voices used were either congruent (from the same person) or incongruent (from different people). A comparison was made between the different subject groups in their susceptibility to the McGurk illusion, and the results show that when the faces and voices are incongruent, subjects who are familiar with the faces are less susceptible to McGurk effects than those who are unfamiliar with the faces. The results suggest that facial identity and facial speech processing are not entirely independent, and these findings are discussed in relation to Bruce and Young’s (1986) functional model of face recognition.     

听障学生汉语唇读面部加工方式的差异：来自眼动的证据

为探讨高低唇读理解能力听障学生唇读面部加工方式的差异,研究采用视频—图片匹配范式并结合眼动技术,考察高低唇读能力组语前-语中-语后和整体面部加工方式。结果发现,虽然两组都表现出社会协调模式,但高唇读能力组社会协调分数更高,且眼部维持时间更长。表明高唇读能力者整体加工和眼部、口形并行加工能力强,支持凝视假说和社会协调模式;低唇读能力者整体加工效率低,更依赖口形,未能通过补偿策略获得良好的补偿效果。     

陌生面孔的印象加工:基于面孔特征的视角

个体往往会基于面孔特征对他人进行特质推断和社会评价,这种推进和评价会影响个体的行为决策。结合印象形成的双加工理论和面孔加工的双通路模型,本文提出陌生面孔的印象加工主要包括两类方式:一是基于社会分类线索及相似特征的加工,激活个体的刻板印象;二是基于熟悉面孔的相似性特征线索,通过过度泛化激活与熟悉他人相一致的印象。本文进一步阐述了两类面孔线索印象加工的过程和影响陌生面孔印象加工的相关因素。在此基础上,本文提出现有的研究对两类加工方式的交互作用及其加工阶段尚未作深入探讨,同时指出知觉者本人的人格特质和面孔信息对印象加工的影响也需要在未来研究中进一步探索。     

Understanding covert recognition

An implementation of Bruce and Young's (1986) functional model of face recognition is used to examine patterns of covert face recognition previously reported in a prosopagnosic patient, PH. Although PH is unable to recognize overly the faces of people known to him, he shows normal patterns of face processing when tested indirectly. A simple manipulation of one set of connections in the implemented model induces behaviour consistent with patterns of results from PH obtained in semantic priming and interference tasks. We compare this account with previous explanations of covert recognition and demonstrate that the implemented model provides the most natural and parsimonious account available. Two further patients are discussed who show deficits in person perception. The first (MS) is prosopagnosic but shows no covert recognition. The second (ME) is not prosopagnosic, but cannot access semantic information relating to familiar people. The model provides an account of recognition impairments which is sufficiently general also to be useful in describing these patients.     

Incomplete Faces Don't Show the Whole Picture: Repetition Priming from Jumbled Faces

Four experiments examined repetition priming of familiarity decisions to faces of famous people by the prior exposure of intact or jumbled faces. In Experiments 1 and 2 the primes were either the identical picture of the target face or a picture of the face with the internal features jumbled up. (In Experiment 2 the external features were also removed from all faces.) Compared with response times to previously unseen faces, familiarity decisions were made more rapidly if the subject had seen and identified the famous face in the pre-training stage; this was independent of whether they saw an intact or jumbled face. Priming was not shown if the face was not recognized earlier. Experiment 3 demonstrated that, if faces were not recognized spontaneously in the pre-training stage, being prompted as to their identity by the experimenter still did not yield priming at test-a result that replicated a previous study using incomplete faces (Brunas-Wagstaff, Young, & Ellis, 1992). Experiment 4 showed that it was the situation in which the information was given that was critical in determining whether priming occurred. The findings of this study are related to mechanisms for repetition priming of faces and used to discuss the necessity of modifications to the Bruce and Young (1986) model such as those offered by Burton, Bruce, and Johnston (1990).     

Holistic processing of shape cues in face identification: Evidence from face inversion,composite faces,and acquired prosopagnosia

According to a classical functional architecture of face processing (Bruce & Young, 1986), sex processing on faces is a parallel function to individual face recognition. One consequence of the model is thus that sex categorization on faces is not influenced by face familiarity. However, the behavioural and neuro-psychological evidences supporting this dissociation are yet equivocal. To test the independence between sex processing on faces and familiar face recognition, familiar (learned) faces were morphed with new faces, generating facial continua of visual similarity to familiar faces. First, a pilot experiment shown that subjects familiarized with one extreme of the face continuum roughly perceive one half of the continuum (60 to 100% of visual similarity to familiar faces) as made of familiar faces and the other part as unfamiliar. In the experiment proper, subjects were familiarized with faces and tested in a sex decision task made on faces at the different steps of the continua. Subjects were significantly quicker at telling the sex of faces perceived as familiar (60-100%), and the effect was not observed in a control (untrained) group. These results indicate that familiar face representations are activated before sex categorization is completed, and can facilitate this processing. The nature of the interaction between sex categorization on faces and familiar face recognition is discussed.     

Understanding voice perception

Voices carry large amounts of socially relevant information on persons, much like 'auditory faces'. Following Bruce and Young (1986)'s seminal model of face perception, we propose that the cerebral processing of vocal information is organized in interacting but functionally dissociable pathways for processing the three main types of vocal information: speech, identity, and affect. The predictions of the 'auditory face' model of voice perception are reviewed in the light of recent clinical, psychological, and neuroimaging evidence.     

Dissociation of facial attractiveness and distinctiveness processing in congenital prosopagnosia

According to the influential model of Bruce and Young (1986) socially relevant facial information is processed separately from facial information leading to individual face recognition. In recent years functional imaging has identified a network of distinct occipitotemporal cortex areas for the processing of these two kinds of information. Functionally it is not clear at which processing level the “social” and the “recognition” pathways diverge. The study of subjects with a profound face recognition and learning deficit (congenital prosopagnosia—cPA) promises for a better understanding of this issue. We therefore tested the perception of attractiveness (a cue of prime social importance) and distinctiveness (a facial feature related to recognition) in 14 people with cPA. Although attractiveness ratings were highly consistent with controls, cPA subjects' distinctiveness ratings showed random patterns. This dissociation of normal attractiveness processing and impaired distinctiveness processing in cPA helps to specifies the nature of the impairment in this condition while shedding light on the functional architecture of normal face processing.     

Face aftereffects suggest interdependent processing of expression and sex and of expression and race

Bruce and Young (1986) proposed that functionally different aspects of faces (e.g., sex, identity, and expression) are processed independently. Although interdependent processing of identity and expression and of identity and sex have been demonstrated previously, evidence for interdependent processing of sex and expression is equivocal. Using a visual adaptation paradigm, we show that expression aftereffects can be simultaneously induced in different directions along anger–fear continua for male and female faces (Experiment 1) and for East Asian and Black African faces (Experiment 2). These findings for sex- and race-contingent expression aftereffects suggest that processing of expression is interdependent with processing of sex and race and are therefore problematic for models of face perception that have emphasized independent processing of functionally different aspects of faces. By contrast, our findings are consistent with models of face processing that propose that invariant physical aspects of faces and changeable social cues can be processed interdependently.     

An appreciation of Bruce and Young's (1986) serial stage model of face naming after 25 years

The current status of Bruce and Young's (1986) serial model of face naming is discussed 25 years after its original publication. In the first part of the paper, evidence for and against the serial model is reviewed. It is argued that there is no compelling reason why we should abandon Bruce and Young's claim that recall of a name is contingent upon prior retrieval of semantic information about the person. The current status of the claim that people's names are more difficult to recall than the names of objects is then evaluated. Finally, an account of the anatomical location in the brain of Bruce and Young's three processing stages (face familiarity, retrieval of semantic information, retrieval of names) is suggested. In particular, there is evidence that biographical knowledge about familiar people is stored in the right anterior temporal lobes (ATL) and that the left temporal pole (TP) is heavily involved in retrieval of the names of familiar people. The issue of whether these brain areas play a similar role in object processing is also discussed.     

Lost properties? Retrieval differences between name codes and semantic codes for familiar people

Summary Two experiments compared decision times for matching tasks involving the names and faces of familiar people. In Experiment 1 category and property judgements were compared when subjects were asked to respond to simultaneously presented stimuli: blocks of either faces or surnames. The results showed that the advantage known to exist for category statements extended to these types of materials. In Experiment 2 the same type of property judgement was compared with a first-name-matching task. The finding was that forename judgements required a longer decision time than property judgements when faces were the presented stimuli, bot not when surnames were used. This result lends further support to the Bruce and Young (1986) model of face recognition. It is additional evidence for sequential access to identity-specific codes and name codes from faces and for the functional separation of these two types of information.Robert Johnston is supported by a grant from the Science and Engineering Research Council     

From Monroe to Moreau: an analysis of face naming errors.

Functional models of face recognition and speech production have developed separately. However, naming a familiar face is, of course, an act of speech production. In this paper we propose a revision of Bruce and Young's (1986) model of face processing, which incorporates two features of Levelt's (1989) model of speech production. In particular, the proposed model includes two stages of lexical access for names and monitoring of face naming based on a "perceptual loop". Two predictions were derived from the perceptual loop hypothesis of speech monitoring: (1) naming errors in which a (correct) rare surname is erroneously replaced by a common surname should occur more frequently than the reverse substitution (the error asymmetry effect); (2) naming errors in which a common surname is articulated are more likely to be repaired than errors which result in articulation of a rare surname (the error-repairing effect). Both predictions were supported by an analysis of face naming errors in a laboratory face naming task. In a further experiment we considered the possibility that the effects of surname frequency observed in face naming errors could be explained by the frequency sensitivity of lexical access in speech production. However, no effect of the frequency of the surname of the faces used in the previous experiment was found on face naming latencies. Therefore, it is concluded that the perceptual loop hypothesis provides the more parsimonious account of the entire pattern of the results.     

Event-Related Potentials Related to Normal and Morphed Emotional Faces

S. Bentin and L. Y. Deouell (2000) have suggested that face recognition is achieved through a special-purpose neural mechanism, and its existence can be identified by a specific event-related potential (ERP) correlate, the N170 effect. In the present study, the authors explored the structural significance of N170 by comparing normal vs. morphed stimuli. They used a morphing procedure that allows a fine modification of some perceptual details (first-order relations). The authors also aimed to verify the independence of face identification from other cognitive mechanisms, such as comprehension of emotional facial expressions, by applying an emotion-by-emotion analysis to examine the emotional effect on N170 ERP variation. They analyzed the peak amplitude and latency variables in the temporal window of 120-180 ms. The ERP correlate showed a classic N170 ERP effect, more negative and more posteriorly distributed for morphed faces compared with normal faces. In addition, they found a lateralization effect, with a greater right-side distribution of the N170, but not directly correlated to the morphed or normal conditions. Two cognitive codes, structural and expression, are discussed, and the results are compared with the multilevel model proposed by V. Bruce and A. W. Young (1986, 1998).     

Event-related potentials related to normal and morphed emotional faces

S. Bentin and L. Y. Deouell (2000) have suggested that face recognition is achieved through a special-purpose neural mechanism, and its existence can be identified by a specific event-related potential (ERP) correlate, the N170 effect. In the present study, the authors explored the structural significance of N170 by comparing normal vs. morphed stimuli. They used a morphing procedure that allows a fine modification of some perceptual details (first-order relations). The authors also aimed to verify the independence of face identification from other cognitive mechanisms, such as comprehension of emotional facial expressions, by applying an emotion-by-emotion analysis to examine the emotional effect on N170 ERP variation. They analyzed the peak amplitude and latency variables in the temporal window of 120-180 ms. The ERP correlate showed a classic N170 ERP effect, more negative and more posteriorly distributed for morphed faces compared with normal faces. In addition, they found a lateralization effect, with a greater right-side distribution of the N170, but not directly correlated to the morphed or normal conditions. Two cognitive codes, structural and expression, are discussed, and the results are compared with the multilevel model proposed by V. Bruce and A. W. Young (1986, 1998).     

Examinations of identity invariance in facial expression adaptation

Faces provide a wealth of information essential to social interaction, including both static features, such as identity, and dynamic features, such as emotional state. Classic models of face perception propose separate neural-processing routes for identity and facial expression (Bruce & Young, 1986), but more recent models suggest that these routes are not independent of each other (Calder & Young, 2005). Using a perceptual adaptation paradigm in the present study, we attempted to further examine the nature of the relation between the neural representations of identity and emotional expression. In Experiment 1, adaptation to the basic emotions of anger, surprise, disgust, and fear resulted in significantly biased perception away from the adapting expression. A significantly decreased aftereffect was observed when the adapting and the test faces differed in identity. With a statistical model that separated surface texture and reflectance from underlying expression geometry, Experiment 2 showed a similar decrease in adaptation when the face stimuli had identical underlying prototypical geometry but differed in the static surface features supporting identity. These results provide evidence that expression adaptation depends on perceptual features important for identity processing and thus suggest at least partly overlapping neural processing of identity and facial expression.     

The effects of different types of encoding task on memory for famous faces and names

In this study, incidental memory for familiar faces following different types of encoding task was investigated. Subjects who had been asked to name faces of celebrities at presentation subsequently remembered them significantly better than subjects who had been asked to provide contextual information about the faces, and than subjects who had been asked to distinguish them from unfamiliar faces. This effect persisted regardless of whether the tests required memory for names, faces, or biographical information. It is argued that these results can be explained in terms of the face-processing framework of Bruce and Young (1986) and the theory of episodic memory for faces put forward by Bruce (1982, 1988). However the findings are not consistent with levels of processing (Craik & Lockhart, 1972), nor transfer appropriate processing (Morris, Bransford, & Franks, 1977).     

Repeated Repetition Priming in Face Recognition

Prior exposure to an item can facilitate subsequent recognition of that item. This effect, known as repetition priming, has been found for the recognition of many stimuli including faces (Bruce & Young, 1986). Three experiments are reported, which investigated whether repetition priming is limited to the first repetition or whether subsequent repetitions continually act to increase the speed of face processing. Experiment 1 demonstrated that repetition can reduce categorization time for faces after the first exposure, and this effect is independent of practice effects. Experiment 2 demonstrated that the relationship between reaction time and number of repetitions fits a negative power function. Experiment 3 investigated how delay affects this power function. Delay was found to decrease the negative gradient of the power curve. The effects of priming and delay are discussed in terms of the predictions made by Burton's (1994) interactive activation and competition with learning (IACL) model of face recognition and accounts of automaticity.     

Are Names Difficult to Recall because They Are Unique? A Case Study of a Patient with Anomia

According to Burton and Bruce (1992), names are more difficult to recall than biographical information about people, such as their occupation, because names are unique or highly distinctive. It follows from this that anomic patients who have great difficulty in recalling names should also find it difficult to recall other information that is unique to a particular individual. This paper attempts to evaluate this claim by examining the case of NP, a patient who has severe anomic word-finding difficulties following the rupture and repair of a posterior cerebral artery aneurysm. NP's ability to recall biographical information about people that she cannot name was investigated in a series of experiments. These revealed that she can answer specific questions about the occupations and appearance of well-known people and can recall distinctive meaningful information about them such as the identity of their spouse, even though she is unable to recall their name. It is argued that these results support the view that names are represented in a store separate from that for semantic information about people that we know. The findings are therefore consistent with the sequential stage model of face identification put forward by Bruce and Young (1986) and are explained in terms of the theory of speech production put forward by Levelt (1989).     

面部表情识别与面孔身份识别的独立加工与交互作用机制

面孔识别功能模型认为,面部表情识别与面孔身份识别是两条独立的并行路径。以往诸多研究者都认可并遵循二者分离的原则。但近期研究表明,面部表情识别与面孔身份识别存在交互作用。首先总结和分析已有的面部表情识别的研究成果,评述神经心理学与认知神经科学研究中的论争,然后介绍人脸知觉的分布式神经机制以及相关的实验证据,最后提出面部表情识别与面孔身份识别的多级整合模型,并展望研究前景。     

The face and person perception: insights from social cognition

Social-cognitive investigations of face perception have tended to be motivated by different goals than cognitive and neuropsychological studies-namely, to understand the dynamics of social categorization rather than identity recognition-and the result has been a lack of cross-pollination of insights and ideas between the disciplines. We review the evidence from social cognition, with an eye to discussing how this work aligns with the Bruce and Young (1986) model of face recognition. Acknowledging the invaluable impact the model has exerted on our understanding of face recognition, we suggest that considering the bottom-up constraints of visual processing and the top-down influences of semantic knowledge will contribute to a more comprehensive understanding of face perception.