Copyright © 1999 Massachusetts Institute of Technology
Analysis and retention of facial images is a crucial skill for primates. The survival value of this skill is reflected in our extraordinary MEMORY for faces, in the visual preferences for face stimuli shown by infants, and in our remarkable sensitivity to subtle differences among faces. Striking parallels have emerged between the results of perceptual, developmental, neuropsychological, neurophysiological, and functional neuroimaging studies of face recognition. These indicate that face recognition in primates is a specialized capacity consisting of a discrete set of component processes with neural substrates in ventral portions of occipito-temporal and frontal cortices and in the medial temporal lobes.
Appreciation of the specialness of the face for social organisms can be traced back at least to Darwin's The Expression of Emotion in Man and Animals (1872). Moreover, face agnosia -- a selective deficit in recognizing faces -- has been inferred from the clinical literature since the turn of the century. Intensive research on face perception of normal individuals has a more recent history, linked to a growing interest in INFANT COGNITION and perception. One early milestone was Yin's (1969) demonstration of the inversion effect, the tendency for recognition of faces to be differentially impaired (relative to that of other "mono-oriented" stimuli such as houses) by turning the stimulus upside down. This finding has been widely interpreted to mean that face recognition depends on specialized mechanisms for configural processing (i.e., analysis of small differences in details and spatial relations of features within a prototypical organization). Face "specialness" is also supported by data showing that infants preferentially look at or track facelike arrangements of features relative to jumbles of features or control stimuli. Nonetheless, the protracted development of adult levels of performance indicates that face recognition additionally involves either a long period of NEURAL DEVELOPMENT and/or cognitive processing capacity, specific experience with faces, or both. Exactly what improves with maturation or experience remains unclear (Chung and Thomson 1995).
The fascinating syndrome of face agnosia (prosopagnosia) has spurred considerable controversy regarding the "specialness" of faces. The degree of DOMAIN SPECIFICITY present in prosopagnosia is relevant to whether face recognition is best viewed as a unique capacity, or merely as an example of general mechanisms of OBJECT RECOGNITION (Farah 1996). Although prosopagnosics are aware that faces are faces -- that is, they know the basic level category -- they fail to identify reliably or achieve a sense of familiarity from faces of family members, famous persons, and other individuals they previously knew well. Typically, they also have trouble forming memories of new faces, even if other new objects are learned. However, prosopagnosics may identify individuals by salient details such as clothing and hairstyle, or by nonvisual features such as voice.
Varying patterns of deficits in processing faces occur in brain-damaged individuals, and these differing patterns provide evidence for dissociable component operations in face recognition. Some patients show sparing of ability to judge the age, gender, and even emotional expression of faces whose identity they cannot grasp. Others have difficulty with all aspects of face processing; such patients are unable even to analyze facial features normally, a necessary precondition to identification. Finally, some brain-damaged patients can perceive structural attributes of faces adequately for judgments about emotion and gender, and even judge if a face is familiar, but show a specific inability to recall the associated name.
In both humans and monkeys, faces are analyzed in subregions of the visual-cortical object recognition pathway. In particular, the temporal neocortex in nonhuman primates (notably inferior temporal cortex or "area TE") contains neurons that fire selectively to face stimuli (Gross and Sergent 1992). The question arises as to whether such cells truly respond to visual information unique to faces, or whether their selectivity is more parsimoniously explained as responsiveness to features shared by faces and other object classes; the bulk of the evidence supports the former description. For example, although face-selective neurons vary in the degree of their preference for face stimuli, many respond to both real faces and pictures of faces, but give nearly no response to any other stimuli tested, including other complex objects and pictures in which features making up the face are rearranged or "scrambled." Moreover, for many such neurons, specificity of response is maintained over transformations in size, stimulus position, angle of lighting of the face, blurring, and so forth. Thus, at least some face-selective neurons are sensitive to the global aspects of a face, such as prototypical configuration of stimulus features. Finally, face-selective neurons are present very early in life (Rodman, Gross, and Ó Scalaidhe 1993), consistent with the idea that they represent inborn "prototypes" for faces.
Subsets of face-selective neurons appear to participate in specific aspects of face coding. For example, some respond selectively to distinctive features, such as eyes per se, distance between the eyes, or extent of the forehead. Others do not respond to isolated features but instead are selective for orientation of a face (e.g., profile or frontal view). Still others have responses specific for particular expressions; a final subset are particularly sensitive to eye gaze direction (looking back or looking to the side), an important social signal in both monkeys and humans. Cells selectively responsive to faces have a localized distribution in several senses. First, although face cells make up only a tiny fraction of neurons (1-5%) within TE and adjacent areas as a whole, their concentration is much higher in irregular localized clumps. Second, different types of face-selective cells are found in different regions, such that cells sensitive to facial expression and gaze direction tend to be found within the superior temporal sulcus, whereas cells more generally selective for faces and, purportedly, for individuals tend to be located in TE on the inferior temporal gyrus.
Electrophysiological correlates of face recognition have also been obtained from humans (Allison et al. 1994). A large evoked potential (called the N200) is generated by faces (but not other stimulus types) at small sites in the ventral occipitotemporal cortex. These sites or "modules" may be comparable to the clumps of face neurons found in monkeys. Longer-latency face-specific potentials were also recorded from the anterior portions of ventral temporal cortex activated by face recognition in POSITRON EMISSION TOMOGRAPHY (PET) studies. Moreover, N200s to inverted faces recorded from the right hemisphere were smaller and longer than for normally oriented faces; the left hemisphere generated comparable N200s under both conditions. These studies thus provide correlates of both the "inversion effect" and of HEMISPHERIC SPECIALIZATION for some aspects of face processing noted in the clinical literature and in tachistoscopic studies of face recognition in normal humans.
Brain imaging studies in normal humans provide converging evidence for the involvement of the ventral occipito-temporal cortices in face recognition and for the existence of dissociable component operations. Sergent et al. (1992), for example, used PET to compare brain activation while subjects either made discriminations of the gender of faces (perceptual task) or judged their identity. In the perceptual task, selective activation was found in the right ventral occipito-temporal cortex, to a lesser extent in the same area on the left, and in a more lateral left focus as well. These areas overlap, but are generally anterior to, domains activated by other categories of objects. Judgments of face identity, requiring reactivation of stored information about individuals, also activated the right parahippocampal gyrus, anterior temporal cortex, and temporal pole on both sides. These studies and those of Haxby and colleagues have additionally implicated lateralized portions of frontal cortex in face encoding, perceptual judgments of faces, and subsequent recognition of faces. In particular, a right frontal focus appears to be involved in recognizing facial emotion. Finally, the HIPPOCAMPUS appears to participate, along with parahippocampal cortices, primarily at the time of encoding new faces.
Neuropathological data are generally consistent with results of imaging and evoked potential studies regarding anatomical substrates for face recognition. Initially, prosopagnosia was associated clinically with right posterior cortical damage. In the 1980s, a number of cases came to autopsy. The damage in these lay very ventrally and medially at the occipitotemporal junction, in roughly the same region activated in recent PET studies. However, in all such cases this area (or underlying white matter) was damaged bilaterally, and consequently bilateral damage became thought of as a necessary precondition to prosopagnosia. Recently, cases with proposagnosia and right cortical damage alone, along with the results of imaging and evoked potential studies reviewed above, have reaffirmed the critical role of the right hemisphere in face recognition in humans.
Many explanations have been given for the apparent "specialness" of faces. Face perception and recognition may indeed be unique behavioral capacities and reflect dedicated neural circuits that can be selectively damaged. Such uniqueness of faces might result both from their behavioral significance and from the fact that they differ structurally from most other object classes, necessitating different perceptual strategies (such as encoding on the basis of prototypical configuration), strategies selectively lost in prosopagnosia. An alternative explanation is that faces are processed and stored in a manner similar to that for other objects, but faces are simply harder to tell apart than other kinds of objects; this view is consistent with observations that prosopagnosia is often accompanied by some degree of general object agnosia. A related account holds that face processing requires subtle discriminations between highly similar exemplars within a category, and that it is this capacity, not processing of the facial configuration, that is disrupted in prosopagnosia. Interestingly, recent studies show that face processing is still disproportionately impaired when discriminations of face and nonface stimuli are equated for difficulty, so this view has lost some force. A final suggestion is that face processing represents acquisition of EXPERTISE associated with very protracted experience with a category of complex visual stimuli (Carey 1992). Prosopagnosics with deficits in other object recognition domains in which they had previously acquired expertise over long periods (e.g., a show dog expert who lost the ability to differentiate breeds) support this idea.
Growing acceptance of faces as a distinct stimulus type has led, along with growing evidence for component processes, to emergence of theoretical accounts of face recognition tied to central ideas in COMPUTATIONAL NEUROSCIENCE. For example, drawing on the notion of a computational model advanced by David MARR, Bruce and Young (1986) analyzed face processing as a set of seven types of information code (or representation). In their scheme, which fits well with neuropsychological dissociations, everyday face recognition involves the use of "structural" codes to access identity-specific semantic information, and then finally the attachment of a name to the percept. Other recent theoretical accounts have modeled face recognition using artificial neural network architectures derived from parallel-distributed processing accounts of complex systems. Future advances in understanding face recognition will likely require further incorporation of data on self-organizing (developmental and environmental) and modulatory (emotional and motivational) aspects of face processing into existing models.
Allison, T., C. McCarthy, A. Nobre, A. Puce, and A. Belger. (1994). Human extrastriate visual cortex and the perception of faces, words, numbers and colors. Cereb. Cortex 5:544-554.
Bruce, V., and A. Young. (1986). Understanding face recognition. Br. J. Psych. 77:305-327.
Carey, S. (1992). Becoming a face expert. Phil. Trans. R. Soc. Lond. B 335:95-103.
Chung, M-S., and D. M. Thomson. (1995). Development of face recognition. Br. J. Psych. 86:55-87.
Farah, M. J. (1996). Is face recognition "special"? Evidence from neuropsychology. Beh. Brain Res. 76:181-189.
Gross, C. G., and J. Sergent. (1992). Face recognition. Curr. Opin. Neurobiol. 2:156-161.
Haxby, J. V., L. G. Ungerleider, B. Horwitz, J. M. Maisog, S. I. Rapoport, and C. L. Grady. (1996). Face encoding and recognition in the human brain. Proc. Nat. Acad. Sci. 93:922-927.
Rodman, H. R., C. G. Gross, and S. P. Ó Scalaidhe. (1993). Development of brain substrates for pattern recognition in primates: physiological and connectional studies of inferior temporal cortex in infant monkeys. In B. de Boysson-Bardies, S. de Schonen, P. Jusczyk, P. MacNeilage, and J. Morton, Eds., Developmental Neurocognition: Speech and Face Processing in The First Year of Life. Dordrecht: Kluwer Academic, pp. 63-75.
Sergent, J., S. Ohta, and B. MacDonald. (1992). Functional neuroanatomy of face and object processing: a positron emission tomography study. Brain 115:15-36.
Yin, R. K. (1969). Looking at upside-down faces. J. Exp. Psychol. 81:141-145.
Benton, A. L. (1980). The neuropsychology of facial recognition. American Psychologist 35:176-186.
Damasio, A. R., D. Tranel, and H. Damasio. (1990). Face agnosia and the neural substrates of memory. Ann. Rev. Neurosci. 13:89-109.
Desimone, R., T. D. Albright, C. G. Gross, and C. Bruce. (1984). Stimulus selective properties of inferior temporal neurons in the macaque. J. Neurosci. 4:2051-2062.
Dror, I., F. L. Florer, D. Rios, and M. Zagaeski. (1996). Using artificial bat sonar neural networks for complex pattern recognition: recognizing faces and the speed of a moving target. Biol. Cybern. 74:331-338.
Farah, M. J. (1991). Patterns of co-occurrence among the associative agnosias: implications for visual object representations. Cog. Neuropsychol. 8:1-19.
Farah, M. J., K. L. Levinson, and K. L. Klein. (1994). Face perception and within-category discrimination in prosopagnosia. Neuropsychologia 33:661-674.
Field, T. M., R. Woodson, R. Greenberg, and D. Cohen. (1982). Discrimination and imitation of facial expressions by neonates. Science 218:179-181.
Flude, B. M., A. W. Ellis, and J. Kay. (1989). Face processing and name retrieval in an anomic aphasic: names are stored separately from semantic information about familiar people. Brain and Cog. 11:60-72.
George, M. S., T. A. Ketter, D. S. Gill, J. V. Haxby, L. G. Ungerleider, P. Herscovitch, and R. I. Post. (1993). Brain regions involved in recognizing emotion or identity: an oxygen-15 PET study. J. Neuropsychiat. Clin. Neurosci. 5:384-394.
Gross, C. G., H. R. Rodman, P. M. Gochin, and M. W. Colombo. (1993). Inferior temporal cortex as a pattern recognition device. In E. Baum, Ed., Computational Learning and Cognition. Philadelphia: SIAM Press.
Haxby, J. V., C. L. Grady, B. Horwitz, L. G. Ungerleider, M. Mishkin, R. E. Carson, P. Herscovitch, M. B. Schapiro, and S. I. Rapoport. (1991). Dissociation of object and spatial vision processing pathways in human extrastriate cortex. Proc. Natl. Acad. Sci. 88:1621-1625.
Heywood, C. A., and A. Cowey. (1992). The role of the "face cell" area in the discrimination and recognition of faces by monkeys. Phil. Trans. Roy. Soc. Lond. B 335:31-38.
Perrett, D. I., P. A. J. Smith, D. D. Potter, A. J. Mistlin, A. S. Head, A. D. Milner, and M. A. Jeeves. (1985). Visual cells in the temporal cortex sensitive to face view and gaze direction. Proc. Roy. Soc. Lond B 223:293-317.
Rodman, H. R. (1994). Development of inferior temporal cortex in the monkey. Cereb. Cortex 5:484-498.
Rolls, E. T., and G. C. Baylis. (1986). Size and contrast have only small effects on the responses to faces of neurons in the cortex of the superior temporal sulcus of the monkey. Exp. Brain Res. 65:38-48.
Yamane, S., S. Kaji, and K. Kawano. (1988). What facial features activate face neurons in inferotemporal cortex of the monkey. Exp. Brain Res. 73:209-214.
Copyright © 1999 Massachusetts Institute of Technology