Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 197 Journal of Neuropsychology (2008), 2, 197–225 q 2008 The British Psychological Society The British Psychological Society www.bpsjournals.co.uk Structure and function in acquired prosopagnosia: Lessons from a series of 10 patients with brain damage Jason J. S. Barton* Neurology, Ophthalmology and Visual Sciences, Psychology, University of British Columbia, Canada Acquired prosopagnosia varies in both behavioural manifestations and the location and extent of underlying lesions. We studied 10 patients with adult-onset lesions on a battery of face-processing tests. Using signal detection methods, we found that discriminative power for the familiarity of famous faces was most reduced by bilateral occipitotemporal lesions that involved the fusiform gyri, and better preserved with unilateral right-sided lesions. Tests of perception of facial structural configuration showed severe deficits with lesions that included the right fusiform gyrus, whether unilateral or bilateral. This deficit was most consistent for eye configuration, with some patients performing normally for mouth configuration. Patients with anterior temporal lesions had better configuration perception, though at least one patient showed a more subtle failure to integrate configural data from different facial regions. Facial imagery, an index of facial memories, was severely impaired by bilateral lesions that included the right anterior temporal lobe and marginally impaired by fusiform lesions alone; unilateral right fusiform lesions tended to spare imagery for facial features. These findings suggest that (1) prosopagnosia is more severe with bilateral than unilateral lesions, indicating a minor contribution of the left hemisphere to face recognition, (2) perception of facial configuration critically involves the right fusiform gyrus and (3) access to facial memories is most disrupted by bilateral lesions that also include the right anterior temporal lobe. This supports assertions that more apperceptive variants of prosopagnosia are linked to fusiform damage, whereas more associative variants are linked to anterior temporal damage. Next, we found that behavioural indices of covert recognition correlated with measures of overt familiarity, consistent with theories that covert behaviour emerges from the output of damaged neural networks, rather than alternative pathways. Finally, to probe the face specificity of the prosopagnosic defect, we tested recognition of fruits and vegetables: While face specificity was not found in most of our patients, the data of one patient suggested that this may be possible with more focal lesions of the right fusiform gyrus. * Correspondence should be addressed to Professor Jason J. S. Barton, Neuro-ophthalmology Section D, VGH Eye Care Center, 2550 Willow Street, Vancouver, BC V5Z 3N9, Canada (e-mail: jasonbarton@shaw.ca). DOI:10.1348/174866407X214172 Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 198 Jason J. S. Barton Prosopagnosia, the inability to recognize familiar faces, was probably first described as a consequence of cerebral damage by Quaglino and Borelli in 1867 (Della Sala & Young, 2003). Prosopagnosia is both a symptom and a syndrome. As a symptom, it can be one of many deficits in patients with widespread cognitive dysfunction, as in Alzheimer’s disease (Cronin-Coulomb et al., 2000; Mendez, Martin, Smyth, & Whitehouse, 1992; Roudier et al., 1998), Huntington’s disease (Janati, 1985), Parkinson’s disease (Dewick, Hanley, Davies, Playfer, & Turnbull, 1991), autism (Sasson, 2006) and schizophrenia (Feinberg, Rifkin, Schaffer, & Walker, 1986; Onitsuka et al., 2003). As a syndrome, while it has sometimes been attributed to a similar combination of generalized cognitive and visual disturbances (Bay, 1953; Cohn, Neumann, & Wood, 1977), it is now recognized as a selective functional entity generated by discrete neurologic lesions to specific anatomic structures involved in face processing, as had been proposed decades ago (Bodamer, 1947; Hoff & Potzl, 1937). This more specific entity of prosopagnosia is better labelled a syndrome than a disorder. A review of the cases described in the last 50 years shows considerable variety in many facets of this problem. Prosopagnosic patients vary in their lesions, the severity of their prosopagnosia, the type of prosopagnosia, their associated visual or memory deficits and the status of other face-processing skills. Understanding the sources of this variability in terms of both the neuroanatomic substrate and cognitive modelling is a challenge. As with most rare syndromes, the small number of prosopagnosic cases has hampered progress. While detailed single case studies can yield valuable insights, structure–function correlations are best revealed by contrasts between patients. Such contrasts illuminate the relation of differences in behaviour to differences in anatomy, helping us make sense of the variability in a condition as complex as prosopagnosia. This paper reviews the studies we have performed in a series of 10 patients studied over the last 7 years. Variability in prosopagnosia Although the early literature on hemispheric specialization suggested that face processing might be lateralized to the right hemisphere, seminal autopsy studies on two small series of patients concluded that prosopagnosia was caused by bilateral lesions in the medial occipitotemporal cortex (Damasio, Damasio, & van Hoessen, 1982; Meadows, 1974), a conclusion that later received support from neuroimaging studies. However, subsequent studies with pathologic (Landis, Regard, Blieste, & Kleihuis, 1988) or imaging evidence (de Renzi, 1986b; Landis, Cummings, Christen, Bogen, & Imbof, 1986; Michel, Perenin, & Sieroff, 1986; Michel, Poncet, & Signoret, 1989; Schweinberger, Klos, & Sommer, 1995; Sergent & Villemure, 1989; Takahashi, Kawamura, Hirayama, Shiota, & Isono, 1995; Wada & Yamamoto, 2001) reported that prosopagnosia could be caused by a unilateral right-sided lesion (Benton, 1990) – and, on very rare occasions, even a left-sided lesion (Mattson, Levin, & Grafman, 2000). Variation in anatomic location has also been found in the intrahemispheric site of pathology. Most cases have damage in the vicinity of the lingual and fusiform gyri (Damasio et al., 1982; Meadows, 1974), a conclusion also supported by a recent metaanalysis of imaging data (Bouvier & Engel, 2006). However, some cases have damage to more anterior temporal structures (Evans, Heggs, Antoun, & Hodges, 1995; Gainotti, Barbier, & Marra, 2003). This variability in lesion location is consistent with current concepts of the neuroanatomy of face processing: Functional neuroimaging has shown that there is a network of cortical regions activated during face perception, including Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society Structure and function in prosopagnosia 199 the fusiform, superior temporal and occipital face-responsive regions in both hemispheres, but more so on the right (Haxby, Hoffman, & Gobbini, 2000; Rossion et al., 2003). Variability in the types of functional deficit affecting face recognition has also been described in prosopagnosia. Cognitive models of face recognition classically outline a series of processing modules, flowing from early visual analysis to the encoding of facial structure, and the matching of percepts to stored data regarding facial memories, followed by access to supra-modal semantic information-stores (Bruce & Young, 1986; Damasio, Tranel, & Damasio, 1990). Damage to different stages of this model generates different hypothetical patterns of behaviour. Likewise, the clinical data show that patients vary in their behaviour. Patients can achieve normal or abnormal scores on perceptual matching tasks like the Benton Facial Recognition Test (Benton & van Allen, 1972), and their ability to make judgments about other facial properties, such as expression, gaze direction, gender and age, can also be normal (Campbell, Heywood, Cowey, Regard, & Landis, 1990; de Haan & Campbell, 1991; Kracke, 1994; Bartlett & Searcy, 1993) or impaired (Bruyer et al., 1983; Evans et al., 1995; Sergent & Poncet, 1990; Sergent & Villemure, 1989; Tranel, Damasio, & Damasio, 1988). Evidence that a patient has relatively preserved perceptual analysis of faces with these tests has generally been taken as support for the patient having an associative form of prosopagnosia, in which an accurate encoding of facial structure can be achieved but the ability to access facial memories to trigger the act of recognition is disrupted (Damasio et al., 1990; de Renzi, Faglioni, Grossi, & Nichelli, 1991; Tranel & Damasio, 1985). Failures on such tests, on the other hand, suggest an apperceptive form, in which the fault lies in an inability to encode facial structure with a degree of accuracy sufficient to discriminate between different faces. Variability also characterizes the presence of another interesting phenomenon, covert recognition (Bruyer, 1991; Young, 1994). Despite the fact that prosopagnosic patients deny familiarity with most if not all faces shown to them, it is still possible to show some residual familiarity or even recognition with other techniques. These include physiologic methods, such as measuring electrodermal skin responses (Bauer, 1984; Bauer & Verfaellie, 1988; Tranel & Damasio, 1985), or behavioural methods, such as the speed of learning face pairs (Bruyer et al., 1983; McNeil & Warrington, 1991; Schweinberger et al., 1995; Sergent & Signoret, 1992), priming effects (Young, Hellawell, & de Haa, 1988), or interference effects (de Haan, Young, & Newcombe, 1987b). Not all patients show such covert effects (de Haan & Campbell, 1991; Newcombe, Young, & de Haan, 1989; Sergent & Villemure, 1989; Young & Ellis, 1989) and it is of interest to determine why this is the case. Finally, there is also variability in the assertions of how face-specific the recognition deficit is in prosopagnosic patients. Most of these patients can identify objects at some basic level or category, unlike patients with severe generalized visual agnosia. However, they may not be able to identify subtypes (‘subordinate categories’) such as types of cars, food or coins, or specific individuals (‘exemplars’) such as buildings, handwriting or personal clothing (Damasio et al., 1982; de Haan & Campbell, 1991; Lhermitte, Chain, Escourolle, Ducarne, & Pillon, 1972; Whiteley & Warrington, 1977). On the other hand, there are also patients who can identify personal belongings (de Renzi, 1986a), individual animals (Bruyer et al., 1983; McNeil & Warrington, 1993), specific places (Bruyer et al., 1983; Evans et al., 1995), cars (Bruyer et al., 1983; Henke, Schweinberger, Grigo, Klos, & Sommer, 1998), flowers (Evans et al., 1995), vegetables (Henke et al., 1998) and different eyeglasses (Farah et al., 1995a) although others have argued that Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 200 Jason J. S. Barton more detailed measures of reaction time and signal detection parameters may reveal deficits in object processing in these patients too (Gauthier, Behrmann, & Tarr, 1999). Examining the relation of additional deficits in object recognition to lesion anatomy would thus add to the persistent debate on face specificity in prosopagnosia. The patient cohort The Boston cohort consisted of 14 patients with prosopagnosia. Of these patients, 10 had lesions acquired in adult life. Four had childhood-onset prosopagnosia: Some of these lacked visible lesions while others had less localizable or more diffuse damage than is typically seen with adult-onset cases. Therefore, while it has been of interest to determine the degree to which the deficits in childhood-onset cases parallel those of the adult-onset cases, the former have contributed less to structure–function correlations and will not be considered further in this review. As expected, there was a considerable variety in the location of lesions in the 10 patients with adult-onset prosopagnosia. As with all ‘naturally occurring’ human lesions, no two lesions were exactly the same. As a working hypothesis based upon prior evidence that prosopagnosia can be related to either fusiform or anterior temporal damage and with either right or bilateral lesions, we classified our patients according to whether their lesions involved one or more of four distinct regions: The anterior temporal and medial occipitotemporal cortex in the right hemisphere and the anterior temporal and medial occipitotemporal cortex in the left hemisphere. The 10 patients thus formed the following groups (Figures 1 and 2): Bilateral anterior temporal lesions Patient 008 is a 33-year-old woman with bilateral anterior temporal lesions from a closed head injury and surgical resection 10 years prior. She states that she cannot appreciate shadows and shading but denies abnormal colour perception. She has no topographagnosia. Visual acuity was 20/15 in her right eye and 20/400 in her left eye from pre-existing strabismic amblyopia. She read 11 of 14 pseudoisochromatic plates. Goldmann perimetry showed full visual fields. She scored 45/50 on words and 13/50 on faces with the Warrington Recognition Test and 25/54 on the Benton Face Recognition Test. Her lesions spare the middle and posterior portions of the fusiform gyri bilaterally, but also involve right lateral prefrontal cortex (Figure 3). Bilateral medial occipitotemporal lesions Subject 004 is a 38-year-old man shot in the occiput at age 20, and treated with craniotomy and evacuation of an intracranial haematoma, leaving bilateral posterior occipital lesions, which are slightly larger and more posterior on the left. He has topographagnosia and complains of altered colour perception. Visual acuity was 20/20 in both eyes. He read 10 of 14 pseudoisochromatic plates and had left hemiachromatopsia. Goldmann perimetry showed complete loss of superior visual fields bilaterally, with slight inferior extension on the right. He scored 46/50 on the word portion and 33/50 on the face portion of the Warrington Recognition Test and 32/54 on the Benton Face Recognition Test. Subject 010 is a 41-year-old man with bilateral posterior occipitotemporal lesions from a car accident 20 years prior that caused a subdural haematoma. He had been Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society Structure and function in prosopagnosia 201 Figure 1. Bilateral lesions shown on coronal template drawings. Top shows the lesion of subject 008, with bilateral anterior temporal lesions. The next group, patients 004 and 010, have bilateral occipitotemporal lesions, involving the fusiform gyrus. The bottom group, patients 007 and 011, have predominantly right-sided lesions, involving both anterior temporal and occipitotemporal cortex and smaller left occipitotemporal lesions. cortically blind for a few weeks. He now has prosopagnosia, a right hemianopia, some mild object agnosia and complaints of partial dyschromatopsia. Visual acuity was 20/20 in both eyes and he read 14 of 14 pseudoisochromatic plates. On the Warrington Recognition Test, he scored 48/50 on the word portion and 24/50 on the face portion. He was impaired (37/54) on the Benton Face Recognition Test. Figures 1 and 3 show that he has extensive damage to both fusiform gyri, and an additional right superior frontal lesion. Mixed right occipitotemporal, right anterior temporal and left occipitotemporal lesions Subject 007 is a 37-year-old woman who had viral encephalitis at age 17 and has an extensive right-sided lesion, extending from anterior temporal and orbitofrontal cortex to more posterior occipitotemporal cortex, with minor left parahippocampal damage (O’Connor, Butters, Miliotis, Eslinger, & Cermak, 1992). She has topographagnosia but Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 202 Jason J. S. Barton Figure 2. Unilateral lesions shown on coronal template drawings. Top shows the lesion of patient 013, with anterior temporal lobectomy. Bottom shows the lesions of the remaining four patients, with lesions of the occipitotemporal cortex. denies problems with colour perception. Visual acuity was 20/20 in both eyes. She read 12 of 14 pseudoisochromatic plates but had hemi-achromatopsia limited to the left upper quadrant. Goldmann perimetry was normal. There were no signs of left hemineglect or simultanagnosia. She scored 43/50 on the word portion and 29/50 on the face portion of the Warrington Recognition Test, and 39/54 on the Benton Face Recognition Test. Subject 011 is a 54-year-old man who had a car accident at age 18 and was admitted with coma, skull fractures and a right subdural haematoma, treated with right craniectomy and partial excision of the right anterior temporal lobe. Subsequently, he had focal left-sided seizures and required shunting for hydrocephalus. He remains on phenytoin, phenobarbital and leviracetam. He had some reading difficulties initially, which improved, and has good memory function. He has no symptoms of topographagnosia. Visual acuity was 20/20 in both eyes and he read 8 of 14 pseudoisochromatic plates. Goldmann perimetry showed a homonymous left superior quadrantanopia and smaller right inferior paracentral hemifield scotoma. On the Warrington Recognition Test, he performed normally with words, recognizing 45/50, Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society Structure and function in prosopagnosia 203 Figure 3. Status of fusiform gyrus in representative patients. Images on the left are from functional MR scanning of a control patient, showing face activation of the fusiform face area. Top row shows comparable coronal slices from two patients with involvement of the fusiform gyrus. Patient 009 has a small hypointensity in the fusiform gyrus (yellow arrow), atrophy of the surrounding white matter and compensatory enlargement of the right lateral ventricle. Patient 010 has large right and left lesions that include the fusiform gyri. Next two rows show coronal and axial slices from two patients with sparing of the middle and posterior fusiform gyrus. Patient 013 had a large right temporal lobectomy; the posterior aspect of his lesion extends over lateral occipitotemporal cortex but does not involve the fusiform gyrus at the level of the FFA. The same is true of patient 008, who has bilateral anterior temporal lesions. but poorly with faces, scoring 33/50. On the Benton Face Recognition Test, he is equally poor, scoring 36/54. His MRI shows extensive right-sided lesions of prefrontal, anterior temporal and occipitoparietal cortex, and left-sided damage to lateral occipitotemporal cortex (Figure 1). Subject 011 is a well-known subject of numerous previous studies (Farah, McMullen, & Meyer, 1991; Farah et al., 1995a; Levine & Calvanio, 1989; Levine, Warach, & Farah, 1985). Unilateral right anterior temporal lesions Subject 013 is a 41-year-old man seen 11 years after a right temporal lobectomy for complex partial seizures since age 17, with continuing use of carbamazepine and lamotrigine for seizure control. He complains of difficulty in recognizing faces and remembering names of people, streets and movies. Visual acuity was 20/20 and he had a partial left upper quadrantanopia. On the Warrington Recognition Test, he performed normally for words (44/50), but poorly with faces (32/50). He was impaired on the Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 204 Jason J. S. Barton Benton Face Recognition Test, scoring 33/54. MRI showed an extensive temporal resection. Unilateral (right) medial occipitotemporal lesions Subject 005 is a 59-year-old man seen 10 months after a right medial occipitotemporal stroke. He complains of decreased brightness but not loss of colour perception. He has topographagnosia. Visual acuity was 20/25 in the right eye and 20/20 in the left eye. He had a complete left homonymous hemianopia on Goldmann perimetry. He read 12 of 14 Ishihara pseudoisochromatic plates. He scored 42/50 on words and 33/50 on faces of the Warrington Recognition Test and 35/54 on the Benton Face Recognition Test. Subject 006 is a 52-year-old man seen 7 months after a right medial occipitotemporal haemorrhage and resection of an oligodendroglioma. He was treated with irradiation and chemotherapy. He has topographagnosia but denies problems with colour vision. Visual acuity was 20/25 in both eyes, and he had a complete left homonymous hemianopia on Goldmann perimetry. He read 12 of 14 pseudoisochromatic plates. He scored 49/50 on the word portion and 30/50 on the face portion of the Warrington Recognition Test and 32/54 on the Benton Face Recognition Test. Subject 009 is a 49-year-old man with a right medial occipitotemporal stroke 3 months prior to testing. His face recognition defects were not recognized until he was tested. Under some circumstances, he displayed ‘provoked overt recognition’ (Morrison, Bruce, & Burton, 2001): When told the category of occupation of famous individuals, he could name several faces that he had previously claimed were unfamiliar. Visual acuity was 20/20 in both eyes. He read 12/14 pseudoisochromatic plates and Goldmann perimetry confirmed a complete left hemianopia. He scored 33/50 on the word portion and 33/50 on the face portion of the Warrington Recognition Test, and 43/54 on the Benton Face Recognition Test. Figure 3 shows that he has lost white and grey matter in the fusiform gyrus, and a study with functional MRI has confirmed that he no longer has activation of the fusiform face area by faces (de Gelder, Frissen, Barton, & Hadjikhani, 2003). Subject 012 is a 55-year-old man seen 6 months after a right medial occipitotemporal and thalamic infarct (Figure 4). He did not complain of face-recognition difficulties, but when asked how he recognized his daughter, he stated ‘by her personality’ rather than her face. Visual acuity was 20/20 in the right eye and 20/25 in the left eye and he read 10/14 pseudoisochromatic plates. He had a complete left homonymous hemianopia on Goldmann perimetry. The prosopagnosic defect: Tests of overt familiarity Our first consideration was, does the severity of prosopagnosia vary with the type of lesion? Confirmation of the patient’s complaint that they cannot recognize faces usually relies on tests like the Famous Faces test (Albert, Butters, & Levin, 1979), which asks them to recognize images of celebrities. Usually, these data have been used to indicate the presence or absence of prosopagnosia rather than to gauge severity. Partly this reflects some of the inherent difficulties in quantifying failures of recognition. Scores need to take into account pre-morbid familiarity: With celebrities, this will depend on the patient’s pre-morbid interest and engagement in popular culture and politics. A test using images of family and friends may circumvent this problem but the ability to Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society Structure and function in prosopagnosia 205 Figure 4. Additional lesions in patient 012. Axial MRI scans show damage to the right thalamus, splenium and periventricular white matter damage around the posterior aspect of the left lateral ventricle. compare one patient with another is limited by the individual nature of these images and the controls used. On a technical note, the common practice of showing famous faces alone does not permit one to distinguish discriminative power from criterion bias. While both are likely present in prosopagnosia, it is the loss of discriminative power that indicates a perceptual deficit. In our test of famous face recognition (Barton, Cherkasova, & O’Connor, 2001a), we used well-known or iconic celebrities (e.g. Groucho Marx, John F. Kennedy, Gandhi), randomly presented with an equal number of anonymous faces and asked the patients to indicate which faces were familiar. By plotting true-positive rate against false-positive rate, we could use signal detection methods to calculate d0 , an index of their discriminative power. Following the test, we presented patients with the names of the famous people portrayed in the battery, and asked them to indicate which celebrities were either unknown to them entirely or known by name but not by face: These items were removed from the calculation of their accuracy. Our normal cohort of 14 patients (mean age ¼ 29.5 years, SD ¼ 8:9, range ¼ 21– 52 years) had a mean d0 of 2.78 (SD ¼ 0:42, range ¼ 2.19–3.88; Figure 5). As expected, none of the prosopagnosic cohort scored in the normal range, thus corroborating their experience with personally familiar faces. A contrast of the five patients with unilateral lesions to the five with bilateral lesions suggested that bilateral lesions might be associated with more severe prosopagnosia. The patients with unilateral right-sided lesions scored a d0 of 0.78 on average, whereas the mean for those with bilateral lesions was 0.23. Prosopagnosia was particularly severe in the two patients with bilateral lesions of the medial occipitotemporal cortex, the ‘classic lesion’ described in the older literature (Damasio et al., 1982; Meadows, 1974), who showed no discriminative power at all. However, the difference between all bilateral and unilateral lesions was not significant (p , 0:14), mainly due to two anomalous patients, 012 and 007. Closer inspection of their imaging reveals possible reasons for their anomalous performance, though. Subject 007, although classified as having bilateral damage, has the least amount of left hemispheric damage of all the bilateral patients (Figure 1). Subject 012, although Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 206 Jason J. S. Barton Figure 5. Famous face familiarity. True hits (proportion of famous faces that patients claimed were familiar) are plotted against false alarms (proportion of anonymous faces that patients claimed were familiar), after normalization to z scores. Scores that lie along the solid diagonal line, where the true hit rate equals the false-alarm rate, indicate no ability to discriminate famous from anonymous faces (d 0 ¼ 0). The subtraction of the false-alarm rate from the true hit rate gives d0 , which thus is proportional to the distance from the diagonal lines: The two dashed lines indicate a d0 of 1 and a d0 of 2. Points in the upper left hand corner indicate high discriminative power, as is true for the control patients (‘c’). Colours indicate lesion type: Bilateral anterior temporal (black), bilateral occipitotemporal (blue), unilateral anterior temporal (pink), unilateral occipitotemporal (red), mixed right anterior temporal, right occipitotemporal and left occipitotemporal (green). Patients with unilateral lesions (red and purple numbers) tend to have better d0 than patients with bilateral occipitotemporal lesions (green and blue numbers). classified as having unilateral damage, has more widespread damage than the other patients with unilateral lesions, with not only involvement of the right thalamus and splenium by his infarct, but also probably pre-existing periventricular white matter changes in his left hemisphere (Figure 4). Our tentative conclusions are that unilateral lesions are more likely than bilateral lesions to leave a patient with some residual weak overt face familiarity. This pattern suggests that, though our data are consistent with reports of unilateral right-sided lesions being sufficient to cause prosopagnosia (de Renzi, 1986b; Landis et al., 1986), left hemispheric structures (possibly homologous) likely have a minor contribution to face recognition. Such a conclusion is consistent with functional neuroimaging data which show that, while faces activate the right fusiform gyrus most consistently and strongly, they also activate the left fusiform gyrus in many patients (Kanwisher, McDermott, & Chun, 1997; Puce, Allison, Asgari, Gore, & McCarthy, 1996; Sergent, Ohta, & MacDonald, 1992). Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society Structure and function in prosopagnosia 207 Perceptual processing of faces: Discriminating changes in configuration While tests of famous face recognition may not only confirm the diagnosis but also gauge the severity of the deficit, they are not informative about the potential reasons for the deficit, or where it might fit in a cognitive model of processing modules. Tests of anonymous face matching like the Benton Face Recognition Test are useful probes of perceptual ability, but failure on this test is not informative about what the patient is failing to perceive. Furthermore, there are criticisms that prosopagnosic patients who achieve normal scores on this test may not be processing faces normally, since they generally take longer to do so (Farah, 1990). In addition, face-matching tests can be failed by patients who are not prosopagnosic (Carlesimo & Caltagirone, 1995; de Renzi, Faglioni, & Spinnler, 1968; Parry, Young, Saul, & Moss, 1991), raising concerns that such matching assesses more than just the mechanisms required for face recognition. Computerized photoprocessing tools provide us with the means to manipulate facial images in discrete ways, allowing us to determine what specific types of changes patients can detect in a face. Such tests have been used to probe the basis of the faceinversion effect. Since most humans learn to recognize faces in the upright position, it is hypothesized that the experience-dependent face-expert mechanism develops an orientation dependence (Farah, Tanaka, & Drain, 1995b). Studies suggest that the configuration of facial features is particularly difficult to process in inverted faces (Barton, Keenan, & Bass, 2001b; Freire, Lee, & Symons, 2000; Leder & Bruce, 2000; Malcolm, Leung, & Barton, 2004; Searcy & Bartlett, 1996). Furthermore, recent data have shown a correlation between the sensitivity of normal patients to such configuration changes and their face-recognition performance (Gaspar, 2007). These findings motivated us to examine whether prosopagnosic patients were impaired in processing facial configuration. Our manipulations (Figure 6) involved alterations to inter-ocular distance and to the distance between the nose and the mouth (Barton, Press, Keenan, & O’Connor, 2002). We used an oddity paradigm, in which each trial presented three faces simultaneously, two identical and one containing the alteration. The subject was asked to state which of the three faces was the different one. We tested patients with both limited viewing duration (2 seconds) and unlimited time. Our 14 control patients (mean age ¼ 23 years, range ¼ 16–43 years) perceived the configurational changes efficiently and rapidly. With unlimited viewing, their mean Figure 6. Examples of stimuli in the test of configuration perception. The middle face is the base face, the left face has reduced inter-ocular distance and the right face has the mouth shifted up. Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 208 Jason J. S. Barton reaction time was less than 2 seconds for either eye or mouth position change (Table 1), and mean accuracy did not differ between trials with 2-second viewing or unlimited viewing (Figure 7). Table 1. Reaction times (ms) for perceiving facial configuration Group Patient Eye Mouth Controls Mean SD Range 005 006 009 012 004 010 007 011 013 008 1,803 (759) (1,131–3,409) 15,471 3,705 27,199 17,162 18,236 9,927 13,412 12,655 3,847 2,518 1,729 (809) (1,024–4,023) 12,832 4,200 8,645 16,950 13,045 8,595 11,728 12,872 5,904 3,941 Right occipital Bilateral occipital Mixed occipitotemporal Right anterior temporal Bilateral anterior temporal Among the patients, the most dramatic contrasts were for eye configuration. Those whose lesions included right occipitotemporal damage were severely impaired in this ability at both 2 seconds and unlimited viewing time, performing essentially at chance (33% correct). In contrast, the two patients (008 and 013) whose lesions spared the middle and posterior fusiform gyri showed only marginal difficulty with 2-second viewing and normal performance at unlimited viewing time. The difference between the two patient groups was highly significant at both 2-second (tð9Þ ¼ 7:16, p , :0001) and unlimited viewing (tð9Þ ¼ 6:84, p , :0001). The data for mouth changes were more variable. With unlimited viewing, the two patients without fusiform damage performed normally, but so did three of the eight patients with lesions of the right fusiform gyrus (patients 005, 009 and 004), and there was no significant difference between these two patient groups. The data for patients 005, 009 and 004 thus showed an interesting asymmetry between mouth and eye performance. If we consider the unweighted difference (eye score-mouth score) as a conservative index of symmetry, this was 2 0.01 (SD ¼ 0:09) for 2-second and 0.03 (SD ¼ 0:12) for unlimited viewing in controls. Patients 009 and 004 were significantly asymmetric in favour of mouth discrimination at both 2-second and unlimited viewing, while subject 005 was asymmetric at unlimited viewing only. (In contrast, only subject 008 was significant asymmetric in favour of eye discrimination at 2-second viewing.) The finding of disproportional impairment of configuration processing in the eye region vs. the mouth region in these three patients with fusiform gyral damage is reminiscent of a recent report of a prosopagnosic patient that used lower face but not periocular information to identify faces (Caldara et al., 2005). Unlike the case with familiarity for famous faces, there was little difference between the four patients with unilateral and the four with bilateral damage to the fusiform gyri. At unlimited viewing duration, there was only a slight difference (tð7Þ ¼ 2:75, p , :03) in favour of marginally better performance by those with unilateral lesions. Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society Structure and function in prosopagnosia 209 Figure 7. Configuration perception. These graphs plot for each patient their accuracy in indicating which of three faces contained a change in either eye position (blue bars) or mouth position (red bars). The top graph shows data for trials with 2-second viewing duration, the bottom graph shows data for trials with unlimited viewing. In each graph, the means of normal control data are given on the left, with horizontal dotted lines indicating the 95% prediction limits derived from this data, while data for a nonprosopagnosic patient with left hemianopia (N.01) is shown on the right. The solid horizontal line indicates 33% correct, which is chance performance. The reaction time data for trials with unlimited duration paralleled the accuracy data (Table 1). The two patients with lesions limited to the anterior temporal lobes (013 and 008) had not only good accuracy but short reaction times, whereas, with the exception of subject 006, all those whose lesions included the right fusiform gyrus had not only reduced accuracy but also very prolonged mean reaction times, many 5–10 times the normal mean. Could some of these deficits be related to low-level factors such as hemianopia? Conceivably, failure to process the spatial relations among features might be impaired by hemifield defects from associated damage to the optic radiations or striate cortex. Indeed, there was a large variety of visual field defects in our patients with medial occipitotemporal lesions. However, subject 007, who had no field defect on perimetry, was as severely impaired on perceiving eye configuration changes as the other patients with fusiform gyral damage, while subject 013, who had a right superior quadrantanopia that included a depression of central field sensitivity, performed flawlessly with unlimited viewing time. In addition, we tested another patient, N.01, a 41-year-old woman who had a right medial occipitotemporal infarct 8 months prior, causing a complete left hemianopia. She had no prosopagnosic complaints, and on the tests of famous face familiarity had a d0 of 2.80, well within the normal range (Barton et al., Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 210 Jason J. S. Barton 2001a). On the test of perception of facial configuration, she was mildly impaired at 2second viewing, but, like subject 013, performed flawlessly with unlimited viewing (Figure 7). Thus, we conclude that, while field defects might impose some low-level limits on processing efficiency, they do not account for deficits that persist when unlimited viewing time is permitted. At present, we believe that these findings indicate that lesions to the right fusiform gyrus in prosopagnosia are associated with severe deficits in perceiving facial structure, as indexed by the configurational changes we displayed in our tests. Lesions limited to anterior temporal cortex impact this perceptual function less. This does not mean, however, that anterior temporal lesions do not have any impact on perceptual processing. We studied subject 008 further, with faces that had changes to both the eye and the mouth region, to determine if she showed better discrimination for combinations that distorted the aspect ratio of the triangular relationship between the eyes and the mouth more than combinations that tended to preserve the aspect ratio (Barton, Zhao, & Keenan, 2003). While control patients showed this influence of structural effects that arise from incorporating data from different facial regions in upright (but not inverted) faces, subject 008 did not. Thus, it is possible that more anterior lesions might be associated with problems in integrating structural data from the whole face, a more subtle failure of structural processing than that seen with fusiform lesions. Face imagery: A probe for facial memories The data above suggest that the most severe deficits in perception of facial structure occur with fusiform gyral lesions, and that these may be the anatomic substrate for the apperceptive variant of prosopagnosia. What about the associative variant? In the past, diagnosis of associative defects has been one of exclusion, as mentioned above. Normal performance on the Benton Face Recognition Test or normal ability to perceive expression and gender have been taken as evidence of normal face perception. However, modern functional neuroimaging indicates separate parallel routes for processing some types of facial information, with the processing of identity occurring in the fusiform face area and the processing of dynamic social signals like direction of gaze and expression involving the superior temporal sulcus (Haxby et al., 2000). Therefore, the status of other forms of face processing may not be informative about the status of the perceptual machinery in prosopagnosia. In the cognitive model of face processing (Bruce & Young, 1986), an associative defect corresponds to failure of accurately encoded percepts to access face-recognition units, or facial memories, either because of a disconnection or direct loss of these units. More direct proof of an associative defect would require a means of assessing access to facial memories without using perception. One such test is face imagery (Takahashi et al., 1995). To assess this, we constructed a 37-item questionnaire that asked patients to imagine the faces of two celebrities and compare them to answer a question (Barton & Cherkasova, 2003). As done in two other studies of imagery (Bartolomeo et al., 1998; Young, Humphreys, Riddoch, Hellawell, & de Haan, 1994), this question could involve a facial feature (e.g. who had the bigger moustache, Adolf Hitler or Josef Stalin?) or the overall face shape (e.g. who has the narrower face, Hilary Clinton or Princess Di?). Patients could choose not to answer a question if they either did not know the celebrities involved or did not recall seeing their faces in the past. The 37 questions were answered by at least 70% of control patients, and at least 80% of these gave the correct Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society Structure and function in prosopagnosia 211 answer. Patients were likewise allowed to omit a question if they had never heard of one member of the pair or did not recall seeing their face: Patients and controls had similar rates of omission for items in the battery. The featural and configural components of the test were equivalent in difficulty in 31 control patients (mean age ¼ 32.3 years, SD ¼ 9:5, range ¼ 22–60), with mean accuracy for featural imagery being 0.93 (SD ¼ 0:04), and for configural imagery being 0.94 (SD ¼ 0:06). Our analysis of the patient data asked two questions. First, did their performance fall outside the 95% prediction intervals of the control data? Second, did they perform significantly different from chance, as calculated by binomial proportions? All but one patient (005) were impaired on imagery for facial configuration (Figure 8). For facial features, three of the four patients with unilateral fusiform gyral lesions were normal on imagery, the exception again being subject 012, who had more complex lesions involving the right thalamus, splenium and left periventricular white matter. Patients with bilateral occipitotemporal lesions were equally impaired on imagery for features and configuration. The most severe defect, however, occurred in subject 008 with bilateral anterior temporal lesions, who performed at chance. Of note, subject 013 with an extensive right temporal lobectomy showed moderate impairment of imagery, equally for features and configuration (unlike the patients with fusiform gyral lesions). Patients 007 and 011, who had complex combinations of right anterior and fusiform lesions along with smaller left fusiform lesions appeared more severely impaired than the two patients with bilateral occipitotemporal lesions (004 and 010), as three of their four scores failed to exceed the limits of chance performance. These data thus suggest that the most severe deficits in facial imagery occur when there is damage to the right anterior temporal lobe, but only when there is associated damage to other structures, including the left anterior temporal lobe (subject 008) or more complex left and right occipitotemporal lesions (patients 007 and 011), since Figure 8. Face imagery. This graph plots for each patient their accuracy for answering questions about either the features (blue bars) or global shape (red bars) of the faces of celebrities. Control data are shown on the left, with error bars showing one standard deviation. Horizontal solid line indicates chance performance (50% correct), while the dotted line indicates the 95% prediction limit, which is similar for both features and global shape. Asterisks indicate results that are not significantly different from chance by binomial proportions. Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 212 Jason J. S. Barton isolated damage to this region (subject 013) is associated with only a mild reduction in imagery. The contrast between patients 008 and 013 suggests that the left anterior temporal lobe may also make a significant contribution to face imagery. However, such an interpretation requires caution: The lateralization of face processing in the anterior temporal lobes of subject 013 may have been affected by his longstanding seizure disorder. Hence, we cannot exclude the possibility that other patients with new unilateral right anterior temporal lobe lesions and no prior history of neurologic dysfunction would show greater impairments of facial imagery. Milder reductions in imagery due to damage to the fusiform gyri are consistent with assertions that at least some of the structures involved in imagery are shared with the perceptual system (Farah, 1989; Kosslyn, 1988). Such perceptual regions may act as visual buffers that can be activated either ‘bottom-up’ from visual input or ‘top–down’ by imagery or priming effects. Functional imaging studies have shown that the fusiform face area is activated during both imagery and perception of faces (O’Craven & Kanwisher, 2000). Covert perception: Relation to perception and anatomy Over the last 20 years, there have been a number of studies of covert familiarity or recognition in prosopagnosia, showing either its presence (Bauer, 1984; Bauer & Verfaellie, 1988; Bruyer et al., 1983; de Haan, Young, & Newcombe, 1987a; de Haan et al., 1987b; McNeil & Warrington, 1991; Renault, Signoret, DeBruille, Breton, & Bolgert, 1989; Rizzo et al., 1987; Schweinberger et al., 1995; Sergent & Poncet, 1990; Tranel & Damasio, 1985) or absence (Bauer, 1986; de Haan & Campbell, 1991; Newcombe et al., 1989; Sergent & Signoret, 1992; Sergent & Villemure, 1989; Young & Ellis, 1989). The reasons for the discrepancy remain unclear. It has been suggested that covert processing is less likely to be present if the patient has deficits that ‘(a) were perceptual in nature, or (b) were not limited to defective face processing’ (Bruyer, 1991). Likewise, others have also speculated that severe apperceptive defects eliminate covert processing (Bauer, 1986; Newcombe et al., 1989; Sergent & Signoret, 1992). If one cannot accurately encode facial structure, it seems intuitive that the percept will not be able to generate any sense of familiarity, covert or overt. Such explanations would also fit well with theories that covert processing reflects face processing in an alternative pathway, perhaps a dorsal one via the superior temporal sulcus (Bauer, 1986; Schweinberger & Burton, 2003). Such pathways may provide access of perceptual data to structures such as the amygdala and frontal regions, which could generate the autonomic responses to familiar faces seen in electrodermal studies of covert face recognition (Bauer, 1984; Bauer & Verfaellie, 1988; Tranel & Damasio, 1985). Indeed, there have been studies that claim a double dissociation for overt and covert face recognition (Tranel, Damasio, & Damasio, 1995), by showing that bilateral ventromedial frontal lesions eliminate electrodermal responses to faces but do not affect overt recognition. (However, since this study did not show that this loss of electrodermal responses was specific to faces, it cannot exclude the possibility that such lesions simply eliminate the ability to generate electrodermal responses to emotional stimuli, rather than a particular type of face processing.) On the other hand, neural network models of face processing have suggested a very different explanation of covert recognition (Farah, O’Reilly, & Vecera, 1993; O’Reilly & Farah, 1999; Young & Burton, 1999). Such models propose that the residual processing for covert perception may emerge from the surviving normal structures in a partly Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society Structure and function in prosopagnosia 213 damaged ventral recognition network. If so, then residual overt and covert recognition should depend on the same mechanisms, which indeed has been the conclusion reached in some case reports of behaviour and event-related potentials (Bobes et al., 2003; Sperber & Spinnler, 2003). To examine this issue, we examined behavioural indices of covert processing in our patients. If severe apperceptive prosopagnosic defects disrupt the encoding of perceptual information used by an alternate pathway, then we should find that those patients with impaired perception of facial configuration – namely those with damage that includes the right fusiform gyrus – would be least likely to show covert recognition. On the other hand, if covert recognition is related to overt recognition, then we might find a statistical correlation between covert performance and the discriminative power for face familiarity in our patients. We used three forced-choice methods (Barton et al., 2001a). First, we presented patients with two faces, one famous and one anonymous, and asked them to indicate which one was familiar (Figure 9). Second, we returned to these face pairs and asked them to indicate which of the two faces belonged to a given name, a covert technique that had been pioneered by Sergent and Signoret (1992). Third, we presented them with 41 faces and asked them to indicate the correct occupation belonging to each face (politician or actor). Our prior work has also shown that these direct forced-choice methods of probing covert semantic knowledge of faces correlate with the results of indirect priming methods that ask whether an unrecognized face influences judgments about the occupation belonging to a name that immediately follows the face (Barton, Cherkasova, & Hefter, 2004b). With the exception of subjects 012 and 013, our patients did not perform better than chance when asked simply to indicate which face of a pair was familiar, reproducing observations made by others (Stone & Valentine, 2003; Table 2). However, when this same choice was cued with the name of the famous person, some of our patients were able to indicate the famous face correctly at a level greater than chance. Likewise, when sorting by occupation, some patients performed this task better than chance also. Figure 9. Example of a stimulus pair used in the forced-choice probe of covert recognition. The first time the patient sees the pair, they are asked to indicate which one is the famous face. All but one patient were unable to do this at a level greater than chance. The second time the patient sees the pair, they are asked, who is John F. Kennedy? With this name cue, several patients demonstrate better than chance performance. 004 010 007 011 013 008 Bilateral occipital Mixed occipitotemporal Right temporal Bilateral temporal 0.68 1.29 0.15 0.90 0.50 0.50 0.55 0.26 20.14 20.22 1.04 20.18 0.40 0.56 0.65 0.70 Familiarity forced choice 0.67 1.12 0.88 0.00 d0 Bold type indicates better performance than chance 005 006 009 012 Patient Right occipital Group Overt Table 2. Covert face recognition 0.79 1.00 0.75 0.39 0.75 0.74 0.80 0.81 0.82 0.95 Name-cued forced choice Covert face-processing tests 0.83 0.76 0.55 0.51 0.56 0.51 0.74 0.65 0.85 0.80 Occupation sorting 1.00 – 0.90 0.89 0.90 1.00 1.00 0.95 0.95 1.00 Familiarity forced choice 1.00 0.70 0.93 1.00 0.98 0.98 0.95 1.00 1.00 – Occupation sorting Name controls Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 214 Jason J. S. Barton Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society Structure and function in prosopagnosia 215 As predicted by assertions that patients with associative prosopagnosia would be most likely to show covert recognition, we found that covert recognition was strong in the two patients with good scores on tests of perception of facial configuration, with anterior temporal lesions not involving the fusiform gyri (patients 008 and 013). However, when we examined the performance of patients with damage to the right fusiform gyrus, there was a marked variability in covert performance. Most prominently, the four patients with right unilateral damage had more covert ability than the four with bilateral lesions with or without additional right anterior temporal damage. On the name-cued forced-choice test, the four patients with unilateral lesions had a mean correct score of 0.85 (SD ¼ 0:07), whereas the four with bilateral lesions had a mean score of 0.66 (SD ¼ 0:18), a difference that showed a trend to significance (tð7Þ ¼ 1:92, p , :096). On occupational sorting, the four with unilateral lesions had a mean correct score of 0.76 (SD ¼ 0:09), whereas the four with bilateral lesions had a mean score of 0.53 (SD ¼ 0:02), a difference that was significant (tð7Þ ¼ 4:61, p , :0025). We next examined whether covert performance was related to residual overt familiarity. We found significant correlations of both forms of covert processing with the d0 obtained from the test of famous face familiarity (Figure 10). Within-category recognition of other objects: Fruits and vegetables Is prosopagnosia face-specific? This is a debate that continues unabated. While prosopagnosic patients do not have general visual agnosia, it is argued that in some cases their problem with faces may merely be the most dramatic example of a more general difficulty in discriminating different versions of the same type of object – so-called ‘withincategory’ judgments (Davidoff & Landis, 1990; Gauthier et al., 1999). Functional neuroimaging has not settled this issue, with competing concepts of the fusiform face area Figure 10. Covert familiarity. The left graph plots for each patient their forced-choice score with name cues (e.g. which of these two faces is John F. Kennedy?) against their d0 for overt familiarity, as measured in Figure 5. The right graph plots their accuracy in sorting faces by occupation (politician versus actor) against overt familiarity. Solid horizontal lines show the level of chance (50% correct) while dotted lines show linear regressions, with correlation coefficient (‘r’) and p values stated. Colour conventions for lesion type are as in Figures 5–7. Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 216 Jason J. S. Barton as a face-specific module vs. a perceptual-expertise module (Kanwisher, 2000; Tarr & Gauthier, 2000). Part of the difficulty in establishing face-specificity lies in the often extensive size of natural human lesions. If there are associated deficits in other object processing, one cannot be certain that the deficits in processing these other objects is due to the same damage that has impaired face processing, or if they are due to damage to adjacent structures. Thus, for example, the inability to recognize buildings might be due to damage to the hippocampal place area, a region just medial to the fusiform face area (Epstein, Harris, Stanley, & Kanwisher, 1999). Another difficulty is related to the premorbid familiarity of the subject with other types of objects (Farah et al., 1995a). Many argue that expertise effects in perceptual judgments reflect prior experience (Diamond & Carey, 1986; Gauthier et al., 2000). While it is reasonable to assume that almost all humans become proficient in identifying faces, universal expertise cannot be assumed for items like cars and flowers. Determining the level of perceptual expertise that a patient should have been able to demonstrate and finding control patients with equivalent experience are significant challenges. As a first attempt to examine the issue of specificity in our prosopagnosic cohort, we decided to examine an object category that we considered to be of nearly as universal interest and experience as faces: Food. We used segments of coloured images of fruits and vegetables (Barton, Cherkasova, Press, Intriligator, & O’Connor, 2004a). Many segments did not conform to the silhouette of the item portrayed, requiring the subject to integrate fragments of external contour and analyse the internal shape and texture to identify items, making the test more challenging. This test included 24 items to be identified by name and 12 items to be matched to a panel of six exemplars, which were fruits or vegetables of a different variety or seen from a different viewpoint. Our eight control patients found this test quite easy, with six achieving perfect scores, and none making more than three errors (Table 3). In contrast, almost all of the nine patients tested had significant difficulty in identifying fruits or vegetables, including one (007) who was a professional chef. Only one subject (009) performed in the normal range with three errors. Subject 009 has a right unilateral lesion affecting the posterior medial occipital lobe: Of note, at the level of the fusiform face area, his lesion appears confined to the fusiform gyrus (Figure 3). Table 3. Fruit and vegetable identification test results Group Patient Errors Controls Mean: Range: 005 006 009 012 004 010 007 011 013 008 0.75 (SD ¼ 1:4) 0–3 6 9 3 7 18 20 7 19 – 14 Right occipital Bilateral occipital Mixed occipitotemporal Right temporal Bilateral temporal Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society Structure and function in prosopagnosia 217 Thus, while the majority of our patients show evidence of impaired discrimination of members of another object category besides faces, one of our patients, with a lesion that was more limited to the fusiform gyrus, had better preserved recognition of fruits and vegetables. It is possible then that subject 009 has a more face-specific perceptual deficit. However, most of our control patients performed this test without error, suggesting that they were operating near ceiling, limiting the ability of the test to identify a subtle deficit. We also did not look for increased reaction times, which others suggest might index milder deficits of object discrimination (Gauthier et al., 1999). Identifying a representative vegetable may not be as difficult as recognizing a specific face, as the former requires a subordinate level categorization, whereas the latter requires identification of a unique individual. Finally, the results cannot exclude the possibility that subject 009 would have shown impairments if another object category had been used. Despite these caveats, these results offer some support to proposals that a lesion confined to the fusiform face area may cause a relatively face-specific impairment in recognition. On the other hand, they do not support suggestions that associative defects may be more face-specific than apperceptive defects (de Renzi et al., 1991), given the impaired performance of subject 008, whose data are most consistent with an associative prosopagnosia, relative to the good performance of subject 009, who demonstrates greater difficulty with configuration perception than with imagery. Discussion The studies described above suggest several tentative conclusions about the relation of lesion neuroanatomy to the clinical manifestations of prosopagnosia, conclusions which in the future will be further clarified and anatomically refined by the use of more powerful functional neuroimaging techniques that can be applied to single cases or small series of these patients (Rossion et al., 2003; Schiltz et al., 2006). Our current conclusions must be regarded as tentative because of the small number of patients involved; however, for a rare condition such as prosopagnosia, larger series are hard to come by. Future work will be important in testing the structural–functional correlations that we currently propose. First, the severity of prosopagnosia appeared to be related to the extent of the underlying lesions, in particular to whether the damage was unilateral or bilateral. Functional imaging studies have shown that face stimuli activate a number of cortical regions, including at a minimum the occipital face area, the superior temporal sulcus and the fusiform face area. These studies also show activation of not only right but also similar left hemispheric structures. Current models of face processing suggest a network of regions, probably with differential contributions to the processing of various types of facial information. The degree and nature of the role of these various regions in processing facial identity, dysfunction of which is core to the prosopagnosic deficit, remain to be elucidated. Nevertheless, our study suggests that left hemispheric structures have a minor contribution in face recognition. Whether this is normally redundant or complementary to the processing by the right hemisphere requires further investigation. One functional imaging study has suggested that feature- or part-based face processing characterizes the function of the left fusiform gyrus, while whole-face processing characterizes that of the right fusiform face area (Rossion et al., 2000). Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 218 Jason J. S. Barton Our study of perceptual processing suggests that the processing of the structural relation between facial features is severely impaired by lesions that include the right fusiform gyrus, and that this is not more severe when there are additional left-sided lesions. Our control patient showed that while low-level perceptual problems like hemianopia decrease the efficiency of this process, they do not create the severe deficits seen in prosopagnosia with unlimited viewing duration. Right anterior temporal lesions also reduce the efficiency of this process under time-limited conditions and may disrupt the integration of configural information across the whole face (Barton et al., 2003). All told, these findings suggest that processing facial configuration may be a key function of right hemispheric structures, in particular of the fusiform gyrus. This conclusion is consistent with a recent fMRI study of a prosopagnosic patient showing that impaired discrimination of facial structure is associated with reduced adaptation effects to face identity in the right fusiform face area (Schiltz et al., 2006). Configuration of the eye region appears to be particularly vulnerable to a right fusiform lesion. Other studies confirm the primacy of the eye region in contributing diagnostic information for identifying faces (Schyns, Bonnar, & Gosselin, 2002; Sekuler, Gaspar, Gold, & Bennett, 2004; Shepherd, Davies, & Ellis, 1981) and there is one report of selective loss of eye processing in a prosopagnosic patient (Caldara et al., 2005). Relative sparing of mouth processing in our study occurred with either unilateral or bilateral lesions, and was present in three of the eight patients with lesions involving the right fusiform gyrus. Reasons for better processing of mouth configuration are unclear. One possibility is that these patients engaged in a strategy of focused processing of the mouth region when faced with an inability to process either eye changes or the whole face in an efficient manner. When normal patients view inverted faces, they process the more salient eye region preferentially and are highly inaccurate on mouth configuration, but this deficit can be eliminated with focused attention on the mouth (Barton et al., 2001b). It may be that a similar focused strategy was employed by some of our patients to enhance their mouth perception. Our data also point to some limitations in the use of face inversion as a model of prosopagnosia. The dual-route model argues that faces can be processed by either a face expert mechanism or a generic object recognition process (Bartlett & Searcy, 1993). Since expertise is dependent on experience, and humans experience faces predominantly in the upright, the expertise mechanism is inoperative with inverted faces. Thus, just as prosopagnosia may reflect a pathologic loss of expertise, face inversion represents a physiologic unavailability of expertise. Our studies suggest that this parallel may not be appropriate. When normal patients process configuration in inverted faces, they display a reduction of whole-face efficiency: They process only the most salient elements (namely, the eyes) with short viewing time, with performance for the less salient mouth region improving with increased viewing duration (Barton et al., 2001b). In our prosopagnosic patients, it is precisely the perception of the normally highly salient eye configuration that is severely impaired. In addition, the majority of patients could process neither eye nor mouth configuration. Among those who showed selective eye deficits, their ability to process the eye configuration did not improve with increased viewing time, in contrast to the dynamics of the selective mouth deficit with inverted faces in normal patients. Although our data clearly show a significant deficit in configuration processing with right fusiform gyrus damage, this does not necessarily imply that this structure is specialized for processing feature configuration. Feature configuration refers to a particular element of facial structure that is apparent in the two-dimensional frontal Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society Structure and function in prosopagnosia 219 image. The three-dimensional face has many other structural elements, including internal and external contours, and even feature shape and size. Other studies have shown that these properties can also be affected by face inversion (Malcolm et al., 2004; Riesenhuber, Jarudi, Gilad, & SInha, 2004; Yovel & Kanwisher, 2004). Further work is required to determine if perception of these structural properties is also impaired by right fusiform gyral damage in prosopagnosia, or if the perceptual deficit is truly specific for configuration. While the perception of configuration is more impaired by right fusiform gyral lesions than by anterior temporal lesions, the data on imagery show the reverse, greater impairment with anterior temporal lesions than by fusiform gyral lesions. This relative double dissociation suggests that apperceptive prosopagnosia is more likely related to fusiform damage, whereas a more associative variant is more likely related to anterior temporal damage. This apperceptive/associative dichotomy is not complete, as Lissauer also recognized for visual agnosia (Lissauer, 1890): Some imagery reduction is apparent with fusiform lesions and more subtle perceptual defects are present with anterior temporal lesions. Is unilateral right anterior temporal damage sufficient to create associative prosopagnosia? Two of our patients with significant imagery deficits had damage to the right but not the left anterior temporal lobe, combined with bilateral fusiform damage. On the other hand, the data from the patient with a temporal lobectomy would suggest that unilateral right anterior temporal damage is not sufficient. However, a patient with a pre-existing epileptic focus in the right anterior temporal lobe may not have a normal pattern of lateralization. Thus, we cannot exclude the possibility that a previously healthy person with a unilateral right anterior temporal lesion would have more substantial deficits in face imagery. Previous studies have hypothesized that associative forms of prosopagnosia would be more likely to show covert recognition, based on the reasoning that intact perceptual data would be required by an alternate pathway mediating covert processing. While we did find evidence of covert processing in the two patients with damage limited to the anterior temporal lobes, we also found it in patients with severe defects in configuration perception from unilateral right-sided lesions. Our two indices of covert recognition correlated well with our d0 measure of residual overt familiarity. An overt/covert correlation would not necessarily be expected if covert recognition was mediated by an alternate dorsal pathway (Tranel et al., 1995). Rather, the correlation suggests that covert recognition is an emergent property of a partially damaged face-processing network (Farah et al., 1993) that includes both right and left hemispheric structures. Finally, our data also attempted to address the longstanding question of whether prosopagnosia can be face-specific (Farah et al., 1995a). Face-specificity does not mean that every case of prosopagnosia is selective for faces alone. Patients with larger lesions are going to have a higher likelihood of recognition deficits for other objects. Rather, a single case of face-specific prosopagnosia would suffice to make the point that face recognition depends upon face-dedicated modules, as some assert (Kanwisher, 2000). Although almost all of our patients struggled with recognizing fruits and vegetables, a task chosen as representative of another within-category judgment involving organic ‘living’ things with which most patients are familiar, we did find that the subject with the most limited damage to the right fusiform gyrus performed in the normal range on this test. This leaves open the possibility that such selective anatomic damage might lead to a face-selective deficit. However, proving face selectivity from prosopagnosic data may be quite difficult. One can always argue that a deficit might have been found if more object Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 220 Jason J. S. Barton classes were tested, and questions about whether patients are performing up to standards appropriate for their pre-morbid expertise are hard to answer definitively, since one seldom has the opportunity to test these patients before lesion onset. Hence, both functional imaging and lesion studies face significant limitations in their attempts to resolve this intriguing and conceptually important issue. Acknowledgements This work was supported by a Canada Research Chair, Michael Smith Foundation for Health Research Senior Scholarship, CIHR grant MOP-77615, and NIMH grant R01 MH069898. References Albert, M., Butters, N., & Levin, J. (1979). Temporal gradients in retrograde amnesia of patients with alcoholic korsakoff’s disease. Archives of Neurology, 36, 211–216. Bartlett, J., & Searcy, J. (1993). Inversion and configuration of faces. Cognitive Psychology, 25, 281–316. Bartolomeo, P., Bachoud-Levi, A. C., De Gelder, B., Denes, G., Dalla Barba, G., Brugieres, P., et al. (1998). Multiple-domain dissociation between impaired visual perception and preserved mental imagery in a patient with bilateral extrastriate lesions. Neuropsychologia, 36(3), 239–249. Barton, J., & Cherkasova, M. (2003). Face imagery and its relation to perception and covert recognition in prosopagnosia. Neurology, 61, 220–225. Barton, J., Cherkasova, M., & O’Connor, M. (2001a). Covert recognition in acquired and developmental prosopagnosia. Neurology, 57, 1161–1167. Barton, J., Keenan, J., & Bass, T. (2001b). Discrimination of spatial relations and features in faces: Effects of inversion and viewing duration. British Journal of Psychology, 92, 527–549. Barton, J., Press, D., Keenan, J., & O’Connor, M. (2002). Lesions of the fusiform face area impair perception of facial configuration in prosopagnosia. Neurology, 58, 71–78. Barton, J., Zhao, J., & Keenan, J. (2003). Perception of global facial geometry in the inversion effect and prosopagnosia. Neuropsychologia, 41, 1703–1711. Barton, J., Cherkasova, M., Press, D., Intriligator, J., & O’Connor, M. (2004a). Perceptual function in prosopagnosia. Perception, 33, 939–956. Barton, J. J., Cherkasova, M. V., & Hefter, R. (2004b). The covert priming effect of faces in prosopagnosia. Neurology, 63(11), 2062–2068. Bauer, R. (1984). Autonomic recognition of names and faces in prosopagnosia: A neuropsychological application of the guilty knowledge test. Neuropsychologia, 22, 457–469. Bauer, R. (1986). The cognitive psychophysiology of prosopagnosia. In H. Ellis, M. Jeeves, F. Newcombe, & A. Young (Eds.), Aspects of face processing (pp. 253–267). Dordecht: MartinusNijhoff. Bauer, R., & Verfaellie, M. (1988). Electrodermal discrimination of familiar but not unfamiliar faces in prosopagnosia. Brain and Cognition, 8, 240–252. Bay, E. (1953). Disturbances of visual perception and their examination. Brain, 76, 515. Benton, A. (1990). Facial recognition 1990. Cortex, 26, 491–499. Benton, A., & van Allen, M. (1972). Prosopagnosia and facial discrimination. Journal of Neurological Sciences, 15, 167–172. Bobes, M. A., Lopera, F., Garcia, M., Diaz-Comas, L., Galan, L., & Valdes-Sosa, M. (2003). Covert matching of unfamiliar faces in a case of prosopagnosia: An erp study. Cortex, 39(1), 41–56. Bodamer, J. (1947). Prosopagnosie. Archiv fur Psychiatrie und Nervenkrankheiten, 179, 6–54. Bouvier, S. E., & Engel, S. A. (2006). Behavioral deficits and cortical damage loci in cerebral achromatopsia. Cerebral Cortex, 16, 183–191. Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society Structure and function in prosopagnosia 221 Bruce, V., & Young, A. (1986). Understanding face recognition. British Journal of Psychology, 77, 305–327. Bruyer, R. (1991). Covert facial recognition in prosopagnosia: A review. Brain and Cognition, 15, 223–235. Bruyer, R., Laterre, C., Seron, X., Feyereisen, P., Strypstein, E., Pierrard, E., et al. (1983). A case of prosopagnosia with some preserved covert remembrance of familiar faces. Brain and Cognition, 2, 257–284. Caldara, R., Schyns, P., Mayer, E., Smith, M., Gosselin, F., & Rossion, B. (2005). Does prosopagnosia take the eyes out of face representations? Evidence for a defect in representing diagnostic facial information following brain damage. Journal of Cognitive Neuroscience. Campbell, R., Heywood, C., Cowey, A., Regard, M., & Landis, T. (1990). Sensitivity to eye gaze in prosopagnosic patients and monkeys with superior temporal sulcus ablation. Neuropsychologia, 28, 1123–1142. Carlesimo, G., & Caltagirone, C. (1995). Components in the visual processing of known and unknown faces. Journal of Clinical and Experimental Neuropsychology, 17, 691–705. Cohn, R., Neumann, M., & Wood, D. (1977). Prosopagnosia: A clinicopathological study. Annals of Neurology, 1, 177–182. Cronin-Coulomb, A., Cronin-Coulomb, M., Dunne, T., Brown, A., Jain, K., Cipolloni, P., et al. (2000). Facial frequency manipulation normalizes face discrimination in ad. Neurology, 54, 2316–2318. Damasio, A. R., Damasio, H., & van Hoessen, G. W. (1982). Prosopagnosia: Anatomic basis and behavioral mechanisms. Neurology, 32, 331–341. Damasio, A., Tranel, D., & Damasio, H. (1990). Face agnosia and the neural substrates of memory. Annual Review of Neuroscience, 13, 89–109. Davidoff, J., & Landis, T. (1990). Recognition of unfamiliar faces in prosopagnosia. Neuropsychologia, 28, 1143–1161. de Gelder, B., Frissen, I., Barton, J., & Hadjikhani, N. (2003). A modulatory role for facial expressions in prosopagnosia. Proceedings of the National Academy of Sciences of the United States of America, 100(22), 13105–13110. de Haan, E., & Campbell, R. (1991). A fifteen year follow-up of a case of developmental prosopagnosia. Cortex, 27, 489–509. de Haan, E., Young, A., & Newcombe, F. (1987a). Face recognition without awareness. Cognitive Neuropsychology, 4, 385–415. de Haan, E., Young, A., & Newcombe, F. (1987b). Faces interfere with name classification in a prosopagnosic patient. Cortex, 23, 309–316. Della Sala, S., & Young, A. W. (2003). Quaglino’s 1867 case of prosopagnosia. Cortex, 39(3), 533–540. de Renzi, E. (1986a). Current issues in prosopagnosia. In J. M. Ellis, F. Newcombe, & A. Young (Eds.), Aspects of face processing (pp. 243–252). Dordecht: Martinus-Nijhoff. de Renzi, E. (1986b). Prosopagnosia in two patients with CT scan evidence of damage confined to the right hemisphere. Neuropsychologia, 24, 385–389. de Renzi, E., Faglioni, P., & Spinnler, H. (1968). The performance of patients with unilateral brain damage on face recognition tasks. Cortex, 4, 17–43. de Renzi, E., Faglioni, P., Grossi, D., & Nichelli, P. (1991). Apperceptive and associative forms of prosopagnosia. Cortex, 27, 213–221. Dewick, H., Hanley, J., Davies, A., Playfer, J., & Turnbull, C. (1991). Perception and memory for faces in parkinson’s disease. Neuropsychologia, 29, 785–802. Diamond, R., & Carey, S. (1986). Why faces are and are not special: An effect of expertise. Journal of Experimental Psychology: General, 115, 107–117. Epstein, R., Harris, A., Stanley, D., & Kanwisher, N. (1999). The parahippocampal place area: Recognition, navigation, or encoding? Neuron, 23(1), 115–125. Evans, J., Heggs, A., Antoun, N., & Hodges, J. (1995). Progressive prosopagnosia associated with selective right temporal lobe atrophy. Brain, 118, 1–13. Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 222 Jason J. S. Barton Farah, M. J. (1989). The neural basis of mental imagery. Trends in Neurosciences, 12(10), 395–399. Farah, M. J. (1990). Visual agnosia: Disorders of visual recognition and what they tell us about normal vision. Cambridge, MA: MIT Press. Farah, M., McMullen, P., & Meyer, M. (1991). Can recognition of living things be selectively impaired? Neuropsychologia, 29, 185–193. Farah, M., O’Reilly, R., & Vecera, S. (1993). Dissociated overt and covert recognition as an emergent property of a lesioned neural network. Psychological Review, 100, 571–588. Farah, M., Levinson, K., & Klein, K. (1995a). Face perception and within-category discrimination in prosopagnosia. Neuropsychologia, 33, 661–674. Farah, M., Tanaka, J., & Drain, H. (1995b). What causes the face inversion effect? Journal of Experimental Psychology. Human Perception and Performance, 21(3), 628–634. Feinberg, T., Rifkin, A., Schaffer, C., & Walker, E. (1986). Facial discrimination and emotional recognition in schizophrenia and affective disorders. Archive of General Psychiatry, 43, 276–279. Freire, A., Lee, K., & Symons, L. A. (2000). The face-inversion effect as a deficit in the encoding of configural information: Direct evidence. Perception, 29(2), 159–170. Gainotti, G., Barbier, A., & Marra, C. (2003). Slowly progressive defect in recognition of familiar people in a patient with right anterior temporal atrophy. Brain, 126(Pt 4), 792–803. Gaspar, C. (2007). The effects of stimulus information and orientation on face processing. Hamilton: McMaster University. Gauthier, I., Behrmann, M., & Tarr, M. J. (1999). Can face recognition really be dissociated from object recognition? Journal of Cognitive Neuroscience, 11(4), 349–370. Gauthier, I., Tarr, M. J., Moylan, J., Skudlarski, P., Gore, J. C., & Anderson, A. W. (2000). The fusiform “Face area” Is part of a network that processes faces at the individual level. Journal of Cognitive Neuroscience, 12(3), 495–504. Haxby, J. V., Hoffman, E. A., & Gobbini, M. I. (2000). The distributed human neural system for face perception. Trends in Cognitive Sciences, 4(6), 223–233. Henke, K., Schweinberger, S., Grigo, A., Klos, T., & Sommer, W. (1998). Specificity of face recognition: Recognition of exemplars of non-face objects in prosopagnosia. Cortex, 34, 289–296. Hoff, H., & Potzl, O. (1937). Über eine optisch-agnostische störung des ‘physiognomie – gedachtnißes’ (bezeihungen zur rückbildung ein wortblindheit. Zeitschrift fur die gesamte Neurologie und Pyschiatrie, 159, 367–395. Janati, A. (1985). Kluver-bucy syndrome in Huntington’s chorea. Journal of Nervous and Mental Disease, 173, 632–635. Kanwisher, N. (2000). Domain specificity in face perception. Nature Neuroscience, 3(8), 759–763. Kanwisher, N., McDermott, J., & Chun, M. (1997). The fusiform face area: A module in human extrastriate cortex specialized for face perception. Journal of Neuroscience, 17, 4302–4311. Kosslyn, S. M. (1988). Aspects of a cognitive neuroscience of mental imagery. Science, 240(4859), 1621–1626. Kracke, I. (1994). Developmental prosopagnosia in asperger syndrome: Presentation and discussion of an individual case. Developmental Medicine and Child Neurology, 36, 873–886. Landis, T., Cummings, J., Christen, L., Bogen, J., & Imbof, H.-G. (1986). Are unilateral right posterior cerebral lesions sufficient to cause prosopagnosia? Clinical and radiological findings in six additional patients. Cortex, 22, 243–252. Landis, T., Regard, M., Blieste, A., & Kleihuis, P. (1988). Prosopagnosia and agnosia for noncanonical views. Brain, 111, 1287–1297. Leder, H., & Bruce, V. (2000). When inverted faces are recognized: The role of configural information in face recognition. Quarterly Journal of Experimental Psychology, 53A, 513–536. Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society Structure and function in prosopagnosia 223 Levine, D., & Calvanio, R. (1989). Prosopagnosia: A defect in visual configural processing. Brain and Cognition, 10, 149–170. Levine, D., Warach, J., & Farah, M. (1985). Two visual systems in mental imagery: Dissociation of “What” and “Where” in imagery disorders due to bilateral posterior cerebral lesions. Neurology, 35, 1010–1018. Lhermitte, F., Chain, F., Escourolle, R., Ducarne, B., & Pillon, B. (1972). Étude anatomo-clinique d’un cas de prosopagnosie. Revista De Neurologia, 126, 329–346. Lissauer, H. (1890). Einfall von seelenblindheit nebst einem bintrag zur theorie derselben. Archiv fur Psychiatrie und Nervenkrankheiten, 2, 22. Malcolm, G., Leung, C., & Barton, J. (2004). Regional variation in the inversion effect for faces: Different patterns for feature shape, spatial relations, and external contour. Perception, 33, 1221–1231. Malone, D., Morris, H., Kay, M., & Levin, H. (1982). Prosopagnosia: A double dissociation between recognition of familiar and unfamiliar faces. Journal of Neurology, Neurosurgery, and Psychiatry, 45, 820–822. Mattson, A. J., Levin, H. S., & Grafman, J. (2000). A case of prosopagnosia following moderate closed head injury with left hemisphere focal lesion. Cortex, 36(1), 125–137. McNeil, J., & Warrington, E. (1991). Prosopagnosia: A re-classification. Quarterly Journal of Experimental Psychology, 43A, 267–287. McNeil, J., & Warrington, E. (1993). Prosopagnosia: A face-specific disorder. Quarterly Journal of Experimental Psychology, 46A, 1–10. Meadows, J. (1974). The anatomical basis of prosopagnosia. Journal of Neurology, Neurosurgery, and Psychiatry, 37, 489–501. Mendez, M., Martin, R., Smyth, K., & Whitehouse, P. (1992). Disturbances of person identification in Alzheimer’s disease. A retrospective study. Journal of Nervous and Mental Disease, 180, 94–96. Michel, F., Perenin, M.-T., & Sieroff, E. (1986). Prosopagnosie sans hémianopsie après lésion unilatérale occipito-temporale droite. Revista De Neurolgia, 142, 545–549. Michel, F., Poncet, M., & Signoret, J. (1989). Les lésions responsables de la prosopagnosie sont-elles toujours bilatérales? Revista De Neurologia, 146, 764–770. Morrison, D., Bruce, V., & Burton, A. (2001). Understanding provoked overt recognition in prosopagnosia. Visual Cognition, 8, 47–65. Newcombe, F., Young, A., & de Haan, E. (1989). Prosopagnosia and object agnosia without covert recognition. Neuropsychologia, 27, 179–191. O’Connor, M., Butters, N., Miliotis, P., Eslinger, P., & Cermak, L. (1992). The dissociation of anterograde from retrograde amnesia in a patient with herpes encephalitis. Journal of Clinical and Experimental Neuropsychology, 14, 159–178. O’Craven, K. M., & Kanwisher, N. (2000). Mental imagery of faces and places activates corresponding stiimulus- specific brain regions. Journal of Cognitive Neuroscience, 12(6), 1013–1023. Onitsuka, T., Shenton, M. E., Kasai, K., Nestor, P. G., Toner, S. K., Kikinis, R., et al. (2003). Fusiform gyrus volume reduction and facial recognition in chronic schizophrenia. Archives of General Psychiatry, 60(4), 349–355. O’Reilly, R., & Farah, M. (1999). Simulation and explanation in neuropsychology and beyond. Cognitive Neuropsychology, 16, 49–72. Parry, F., Young, A., Saul, J., & Moss, A. (1991). Dissociable face processing impairments after brain injury. Journal of Clinical and Experimental Neuropsychology, 13, 545–558. Puce, A., Allison, T., Asgari, M., Gore, J., & McCarthy, G. (1996). Differential sensitivity of human visual cortex to faces, letterstrings, and textures: A functional magnetic resonance study. Journal of Neuroscience, 16, 5205–5215. Renault, B., Signoret, J.-L., DeBruille, B., Breton, F., & Bolgert, F. (1989). Brain potentials reveal covert facial recognition in prosopagnosia. Neuropsychologia, 27, 905–912. Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society 224 Jason J. S. Barton Riesenhuber, M., Jarudi, I., Gilad, S., & SInha, P. (2004). Face processing in humans is compatible with a simple shape-based model of vision. Proceedings of the Royal Society of London, Series B271 (Suppl 6): S448–S450. Rizzo, M., Hurtig, R., & Damasio, A. (1987). The role of scanpaths in facial recognition and learning. Annals of Neurology, 22, 41–45. Rossion, B., Dricot, L., Devolder, A., Bodart, J. M., Crommelinck, M., De Gelder, B., et al. (2000). Hemispheric asymmetries for whole-based and part-based face processing in the human fusiform gyrus. Journal of Cognitive Neuroscience, 12(5), 793–802. Rossion, B., Caldara, R., Seghier, M., Schuller, A. M., Lazeyras, F., & Mayer, E. (2003). A network of occipito-temporal face-sensitive areas besides the right middle fusiform gyrus is necessary for normal face processing. Brain, 126(Pt 11), 2381–2395. Roudier, M., Marcie, P., Grancher, A.-S., Tzortis, C., Staskstein, S., & Boller, F. (1998). Discrimination of facial identity and of emotions in Alzheimer’s disease. Journal of the Neurological Sciences, 154, 151–158. Sasson, N. (2006). The development of face processing in autism. Journal of Autism and Developmental Disorders, 36, 381–394. Schiltz, C., Sorger, B., Caldara, R., Ahmed, F., Mayer, E., Goebel, R., et al. (2006). Impaired face discrimination in acquired prosopagnosia is associated with abnormal response to individual faces in the right middle fusiform gyrus. Cerebral Cortex, 16, 574–586. Schweinberger, S. R., & Burton, A. M. (2003). Covert recognition and the neural system for face processing. Cortex, 39(1), 9–30. Schweinberger, S., Klos, T., & Sommer, W. (1995). Covert face recognition in prosopagnosia: A dissociable function? Cortex, 31, 517–529. Schyns, P., Bonnar, L., & Gosselin, F. (2002). Show me the features! Understanding recognition from the use of visual information. Psychological Science, 13, 402–409. Searcy, J., & Bartlett, J. (1996). Inversion and processing of component and spatial-relational information in faces. Journal of Experimental Psychology, 22, 904–915. Sekuler, A., Gaspar, C., Gold, J., & Bennett, P. (2004). Inversion leads to quantitative, not qualitative, changes in face processing. Current Biology, 14, 391–396. Sergent, J., & Poncet, M. (1990). From covert to overt recognition of faces in a prosopagnosic patient. Brain, 113, 989–1004. Sergent, J., & Signoret, J.-L. (1992). Implicit access to knowledge derived from unrecognized faces. Cerebral Cortex, 2, 389–400. Sergent, J., & Villemure, J.-G. (1989). Prosopagnosia in a right hemispherectomized patient. Brain, 112, 975–995. Sergent, J., Ohta, S., & MacDonald, B. (1992). Functional neuroanatomy of face and object processing. A positron emission tomography study. Brain, 115(Pt 1), 15–36. Shepherd, J., Davies, G., & Ellis, H. (1981). Studies of cue saliency. In G. Davies, H. Ellis, & J. Shepherd (Eds.), Perceiving and remembering faces (pp. 105–131). London: Academic Press. Sperber, S., & Spinnler, H. (2003). Covert person recognition: Its fadeout in a case of temporal lobe degeneration. Cortex, 39, 57–67. Stone, A., & Valentine, T. (2003). Perspectives on prosopagnosia and models of face recognition. Cortex, 39(1), 31–40. Takahashi, N., Kawamura, M., Hirayama, K., Shiota, J., & Isono, O. (1995). Prosopagnosia: A clinical and anatomic study of four patients. Cortex, 31, 317–329. Tarr, M. J., & Gauthier, I. (2000). Ffa: A flexible fusiform area for subordinate-level visual processing automatized by expertise. Nature Neuroscience, 3(8), 764–769. Tranel, D., & Damasio, A. (1985). Knowledge without awareness: An autonomic index of facial recognition by prosopagnosics. Science, 228, 1453–1454. Tranel, D., Damasio, A. R., & Damasio, H. (1988). Intact recognition of facial expression, gender, and age in patients with impaired recognition of face identity. Neurology, 38(5), 690–696. Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society Structure and function in prosopagnosia 225 Tranel, D., Damasio, H., & Damasio, A. (1995). Double dissociation between overt and covert face recognition. Journal of Cognitive Neuroscience, 7, 425–432. Tyrell, P., Warrington, E., Frackowiak, R., & Rossor, M. (1990). Progressive degeneration of the right temporal lobe studied with positron emission tomography. Journal of Neurology, Neurosurgery, and Psychiatry, 53, 1048–1050. Wada, Y., & Yamamoto, T. (2001). Selective impairment of facial recognition due to a haematoma restricted to the right fusiform and lateral occipital region. Journal of Neurology, Neurosurgery, and Psychiatry, 71(2), 254–257. Whiteley, A., & Warrington, E. (1977). Prosopagnosia: A clinical, psychological, and anatomical study of three patients. Journal of Neurology, Neurosurgery, and Psychiatry, 40, 395–403. Young, A. (1994). Covert recognition. In M. Farah & G. Ratcliff (Eds.), The neuropsychology of high-level vision (pp. 331–358). Hillsdale, MI: Erlbaum. Young, A., & Burton, A. (1999). Simulating face recognition: Implications for modeling cognition. Cognitive Neuropsychology, 16, 1–48. Young, A., & Ellis, H. (1989). Childhood prosopagnosia. Brain and Cognition, 9, 16–47. Young, A., Hellawell, D., & de Haa, E. (1988). Cross-domain semantic priming in normal patients and a prosopagnosic patient. Quarterly Journal of Experimental Psychology, 40A, 561–580. Young, A. W., Humphreys, G. W., Riddoch, M. J., Hellawell, D. J., & de Haan, E. H. (1994). Recognition impairments and face imagery. Neuropsychologia, 32(6), 693–702. Yovel, G., & Kanwisher, N. (2004). Face perception: Domain specific, not process specific. Neuron, 44, 889–898. Received 19 February 2007; Received in revised form 25 April 2007