RAPID COMMUNICATION Neural Correlates of Woman Face Processing by 2-Month-Old Infants

Transcription

1 NeuroImage 15, (2002) doi: /nimg , available online at on RAPID COMMUNICATION Neural Correlates of Woman Face Processing by 2-Month-Old Infants Nathalie Tzourio-Mazoyer,* Scania De Schonen,,1 Fabrice Crivello,* Bryan Reutter, Yannick Aujard, and Bernard Mazoyer*,2 *Groupe d Imagerie Neurofonctionnelle, UMR 6095 CNRS, CEA, Université de Caen and Université Paris V, GIP Cyceron, BP 5229, Caen Cedex, France; Developmental Neurocognition Group, Center for Research in Cognitive Neurosciences, CNRS, 31 Chemin Joseph Aiguier, Marseille Cedex, France; Center for Functional Imaging, Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Mail Stop , Berkeley, California 94720; and Neonatalogy Unit, Hôpital Robert Debré, 47, Boulevard Serrurier, Paris, France Received April 24, 2001 The age of 2 months marks a turn in the development of face processing in humans with the emergence of recognition based on internal feature configuration. We studied the neural bases of this early cognitive expertise, critical for adaptive behavior in the social world, by mapping with positron emission tomography the brain activity of 2-month-old alert infants while looking at unknown woman faces. We observed the activation of a distributed network of cortical areas that largely overlapped the adult face-processing network, including the so-called fusiform face area. We also evidenced the activation of left superior temporal and inferior frontal gyri, regions associated, in adults, with language processing. These findings demonstrates that cognitive development proceeds early in functionally active interconnected cortical areas despite the fact they have not all yet reached full metabolic maturation Elsevier Science INTRODUCTION Infants show attraction for face pattern within the first minutes of life, a feature followed around the age of 2 months by the ability to recognize their mother s face among others (Morton and Johnson, 1991). This makes faces the primary visual stimulus category for which humans are expert at making discrimination. As a matter of fact, young infants first recognize their mother s face on the basis of information coming from both the outer contour of the head and hair line and the 1 Present address: Laboratory of Cognition and Development, CNRS-Paris 5, 71 ave Edouard Vaillant, Boulogne-Billancourt 92774, and INSERM E9935, Hopital Robert Debré, 48 Boulevard Sérurier, 75019, Paris, France. 2 To whom correspondence and reprint requests should be addressed. Fax: mazoyer@cyceron.fr. internal configuration of the face (Bushnell et al., 1989). It is around the sixth week of postterm life that they become able to recognize their mother s face on the basis of the sole face internal configuration. The neural bases of such an early developmental competence are not known, although some authors proposed that subcortical regions support face attraction at birth, while visual cortical areas maturation is necessary for the emergence of larger looking fixation time for faces stimuli (Morton and Johnson, 1991). Consistent with this hypothesis is the time course of regional brain metabolism that is highest at birth in subcortical and primary motor regions, progressively increasing overall the cortex, reaching relatively high values in posterior areas during the second month of postterm life (Chugani et al., 1987). Investigations of brain areas involved in face processing in alert infants have been up to now limited by both ethical and technical issues. Here, we took advantage of the opportunity to perform positron emission tomography (PET) scans in six 2-months-infants as part of the clinical follow-up of the major brain stress they had suffered from at birth, adapting to alert infants the classical PET activation paradigm used to map cognitive functions in adults. Permission was given by the Atomic Energy Commission Ethics Committee to acquire a PET scan in each of two different conditions during which visual stimuli were used to capture the infant gaze for the 2 min necessary for labeled water injection and blood flow map acquisition. METHODS Subjects Six neonates were recruited during their first week after birth from a neonatalogy intensive care unit after their parents had given their written informed con /02 $ Elsevier Science All rights reserved. 454

2 RAPID COMMUNICATION 455 TABLE 1 Clinical Data Child ID Sex Age (at PET study) Neonatal syndrome/symptoms Neurological signs (at 2 months of age) LEF M 2 months 3 weeks HIE; meconium inhalation, major Mild peripheral hypertonia hypoxemia with extracorporeal circulation for blood oxygenation for 60 h CHE M 2 months 3 weeks HIE Peripheral hypertonia, left dominance Neonatal convulsions Normal axial tonus BEL M 2 months HIE due to umbilical chord prolapse. Axial hypotonia Apparent death Peripheral hypertonia KEB F 2 months 3 weeks HIE Axial hypotonia Status epilepticus Mild peripheral hypertonia PUJ M 2 months 3 weeks HIE Mild peripheral hypertonia Status epilepticus RED M 2 months 2 weeks HIE Normal Note. HIE, hypoxic ischemic encephalopathy. sent. All infants, who were full term, had suffered from a neonatal syndrome known as hypoxic ischemic encephalopathy (HIE), due to various causes (see Table 1). Permission was given in 1990 by the Atomic Energy Commission Ethics Committee to acquire blood flow maps in these infants at the age of 2 months based on the rationale that early detection of potential regional blood flow deficits would help in defining the optimal neurospychological follow-up for these infants. Permission was given for two oxygen-15-labeled water injections with the constraint that the total injected dose would stay below published recommendation for a single blood flow map in infant, e.g., 0.5 mci/kg body wt (Powers et al., 1988). All experiments were carried out between April 1991 and July At the date of the PET scanning the infants mean age was 10 weeks 3 days (ranging from 8 weeks 2 days up to 11 weeks 5 days). They presented mild neurological signs and were free from neurological medication (see Table 1). A clinical follow up was conducted for at least 3 years that demonstrated a normal developmental age (Bayley, 1969) in four of the children evaluated with developmental tests (Brunet and Lezine, 1951; Sparrow and Cicchetti, 1989). One child was tested at the age of 6 months only, showing a normal development; it was not possible to follow-up one child. Evaluation of Infant Visual Ability Before PET scanning, each infant was tested for several visual abilities. All infants showed a good smooth horizontal visual tracking of a moving object. When presented with the mother and a stranger they all looked at their mother and smiled within 40 s before looking at the stranger, looking thereafter back and forth at the stranger and the mother. They all fixated and tracked visually their silent mother s face when she was moving from right to left and conversely at a distance of 80 cm and when she was moving away from them up to a distance of 2 m. They all showed signs of visual habituation to a stationary visual pattern within 10-s presentations followed by visual preference for a novel stimulus. They all turned their head and eyes to fixate their mother s face when she was calling them from the side. They all made eye to eye contact and smiled to their mother when she talked to them. Stimuli Our first constraint in stimulus design was to select tasks that would capture the infant attention for at least 2 min in order to obtain head immobility required by PET scanning. The second constraint came from the fact that a relatively low level cognitive task was to be included in the protocol so that regional blood flow map acquired in this condition could serve for clinical purposes. Finally, the third constraint was that only two injections were to be performed. Accordingly, infants were presented with two different kind of visual stimuli. One consisted of colored slides of frontal view of woman faces wearing a black scarf on a black background and showing a gentle but neutral expression (see example in Fig. 1). The scarf masked the outer contour of the hair so that only the face was visible. Face expressions were neutral in order to avoid as much as possible emotional and social reactions. Each slide showed a different woman but for one that was presented repeatedly at irregular intervals varying from one to four. The reason for this experimental design was to give the infant the opportunity to learn and recognize a stimulus and to notice differences between at least a face and the other ones while maintaining the infant s attention by changing stimuli (the mother face was not used in order to avoid any social or emotional behavior). The faces were projected onto a

3 456 RAPID COMMUNICATION translucent 19-cm-high 24-cm-wide rectangular screen mounted on a transparent arm, placed at 35 cm from and making a 50 angle with the infant s face plane. The size of the projected faces was cm (subtending a visual angle of in. in width). Eighty slides were charged in the projector carousel, slides being presented at a rate of one every 4 s, each slide being presented for 4 s. The other visual stimulus consisted in an assembly of diodes inserted at 2 cm from the upper edge of the screen vertical axis (see Fig. 1). Four red diodes were arranged in square within a 18-mm diameter circle of 12 green diodes, the diode circle center being at 35 cm from the infant s eyes and subtended a 3 visual angle. Pairs of diametrically opposed diodes were successively lit up at a frequency that was varied by the experimenter every 4 s (ranging from 100 to 0.2 Hz) so that changes in visual stimulation were equally frequent in the two situations. The luminances of the face stimuli and of the diode stimuli were measured by means of a lux meter at the distance of the infant s eyes. A uniform luminance filtering slide was used during the diode condition to match the screen luminance of the two conditions. Image Acquisition A tiny catheter was introduced in the infant s right foot at least 1 h before they were comfortably placed in the PET camera. When the infant was quiet and attentive, according to classical criteria of eyes opening and direction of gaze, stimulus presentation was started, the fixation of which maintained the infant s head immobile. About 30 s after stimulus presentation started, 15 O- labeled water was injected (0.25 mci/kg). A 2-min-long series of 10-s-duration scans was acquired in each condition, using an ECAT-953B PET camera (Siemens, Erlangen, Germany). During image acquisition an experimenter could watch the infant head and eyes without being seen by the infant. No visible eye and head movement were observed, due to the small size of the stimuli and to the fact that only the stimuli were visible for the infant. Upon completion, a single 80-s-duration image was reconstructed for each stimulus condition, starting from the radioactive tracer arrival in the brain. Right after PET scanning, anatomical images were acquired on a Signa 0.5-T magnet (MRMAX, General Electric, Buc, France) as series of 3- or 10-mm-thick T1-weighted contiguous axial slices after the children were orally given a preparation of chloral. Image Analysis Since infants were alert and unable to stay immobile for 10 min during the PET session, we could not measure the attenuation correction factors (ACF) required for emission scan correction, except for two infants who fell asleep. Consequently, attenuation correction was performed using a manufacturer-designed algorithm based on predefined ACF values and the outline of the infant brain as assessed on emission scan. Because of differences in bone density between 2-month-old infants and adults, ACF to be applied to infant PET images were estimated thanks to the transmission scans acquired from the two children who fell asleep (ACF cm 1 for both the brain tissue and the soft cartilaginous skull layer). Head outer edge was defined using the individual PET emission sinogram derived brain outer edge, to which was added an average skull thickness of 5.1 mm as measured from the infant MR images. The two attenuation corrected PET scans of each subject were then coregistered, initializing the process with the multiparametric imaging tool (Pietrzyk et al., 1990) (MPI tool) interactive software, pursuing with the automatic AIR procedure (Automated Images Registration (Woods et al., 1997)). The transformation matrices of both algorithms were combined in order to bring the first PET image within the same coordinates as the second one. AIR was also used to calculate a PET-to-MRI coregistration matrix for each infant. PET image stereotaxic normalization, required for intersubject averaging, was then performed. The highest quality MRI of the infants, reoriented in the Talairach space, served as a template. Each child s MRI was registered and warped onto this template using AIR, the same transformations being applied to his PET images after coregistration to his MRI volume. Each PET raw count volume was further normalized by global scaling. Finally, faces minus diodes difference volumes were averaged over the six children. This averaged difference image was filtered (final image smoothness 11 mm) and scaled by its standard deviation (SD). A 3D region growing algorithm helped to segregate 3D clusters showing normalized regional cerebral blood flow (NRCBF) increases larger than 2 SD (P 0.05, uncorrected). Center of mass coordinates, volume, and average NRCBF variation were computed for each cluster, retaining only those having volumes larger than 144 mm 3 (10 voxels). The selected clusters were then superimposed onto the infant MRI template for anatomical labeling. RESULTS Comparison of brain activity maps during face perception and diodes fixation revealed that 2-month-old infants activated a network of areas (see Table 2, Fig. 1), including in particular an area located within the inferior temporal gyrus nearby the occipitotemporal sulcus. This activation was predominant in the right hemisphere (activation was detected at the left hemi-

4 RAPID COMMUNICATION 457 TABLE 2 Brain Areas Activated on Average in a Group of Six 2-Month-Old Infants during Face Processing Compared to Diode Fixation Anatomical localization Diodes condition NRCBF (SD) Faces diodes NRCBF increase (%) Activated volume (mm 3 ) Frontal L inf frontal gyrus 67.0 (16) L inf frontal gyrus, orbital part 64.9 (10) R mid frontal gyrus, orbital part 77.6 (21) R medial superior frontal 56.5 (16) R median cingulate (8) R ant cingulate, orbital part 83.4 (16) Occipital and temporal L sup occipital (14) R sup occipital (16) L inf occipital/middle temporal 61.7 (22) R inf occipital/middle temporal (10) L sup temporal (8) R sup temporal (12) L inf temporal 55.4 (17) R inf temporal gyrus/fusiform a 87.2 (12) Parietal Precuneus (14) R inf parietal (17) Note. For each activated area, the table gives its anatomical location (based on the identification of major sulci and gyri on one of the infant MRI that was used as a template), the normalized regional cerebral blood flow (NRCBF) in the area during the diodes condition, its volume, and the average NRCBF variation amplitude during face processing. L, left; R, right; inf, inferior; sup, superior; mid, middle. a The activated area homologue to the adult fusiform face area. sphere at a similar location, but its volume did not reach the threshold for significance). Besides this right inferior temporal gyrus focus, this network included bilateral inferior occipital and parietal areas. Moreover, one should note that the left inferior frontal and superior temporal gyri were activated during face processing. DISCUSSION The main finding of the present study is that 2-month-old infants activated a network of areas belonging to the core system for face perception identified in adults (Haxby et al., 2000). In particular, the activated area located within the inferior temporal gyrus nearby the occipitotemporal sulcus is very likely to be the homologue of the adult fusiform face area (FFA) defined by Kanwisher (Kanwisher et al., 1997; Gauthier et al., 2000a). As a matter of fact, despite existing differences between 2-months-old infant and adult brain anatomy, the anatomical location of the inferior temporooccipital activation focus in infants was very close to that of the FFA as reported in Kanwisher seminal paper (Kanwisher et al., 1996). Interestingly, the adult FFA, more active during processing of faces than of any other object (for review see Table 3 and Fig. 2), is at the center of a debate. Some authors consider that it is specialized for faces (Kanwisher, 2000) while others think that it is domain general and related to the acquisition of expertise in object discrimination (Gauthier et al., 1999, 2000a,b; Tarr and Gauthier, 2000). Within this debate, one could argue that evidence of FFA activation in infants, as early as 2 months of age, provides a developmental argument for the domain specificity of this region, since faces is the first perceptual category of complex objects with which an infant becomes an expert at. However, since faces constitute the dominant visual stimulus to which infants are exposed during their first 2 months of life, FFA activation at 2 months could as well reflect the infants general expertise in object discrimination. Nevertheless, it can be assumed that it is related to individual face processing rather than facedness recognition inasmuch as it has been shown that in 12-week infants facedness can be categorized by either hemisphere despite a lack of interhemispheric coordination (de Schonen and Bry, 1987). In addition, a recent study in adult has shown that while the right FFA supports individual face matching based on internal configuration, its left hemisphere homologue is involved in matching face parts (Rossion et al., 2000). This is consistent with behavioral data in 4- to 10-month-old infants showing a right-hemisphere advantage in face recognition (de Schonen and Mathivet, 1990; Deruelle and de Schonen, 1998). As previously proposed (de Schonen and Mathivet, 1989), the right hemisphere advantage of the face area in adults (Gauthier et al.,

5 458 RAPID COMMUNICATION FIG. 1. Network of areas active during face processing as compared to fixation of diodes in a group of six 2-month-old infants. Activated areas (orange) are superimposed on a series of transaxial structural slices taken from an infant MRI brain template. Colored ticks indicate selected anatomical landmarks: anterior limit of the occipital lobe (black), rolando sulcus (red), and occipitotemporal sulcus (green). The identification of these anatomical landmarks was obtained on the template brain with a software allowing a 3D reconstruction and sulci drawing in any incidence. These and other landmarks were used to draw structural volumes of interest, such as the inferior frontal gyrus corresponding in the left hemisphere to Broca s area in adults (pink), middle frontal gyrus (yellow), and superior frontal gyrus (white) that were used for activation anatomical labeling. The graph gives for each infant the variation of normalized regional cerebral blood flow values during the diode and face conditions in the cluster enclosed in the blue boxes that corresponds to the adult fusiform face area. L, left; R, right. 1999) might then well be related to its earlier functional maturation. We know from anatomical (Brody et al., 1987) and metabolic studies (Chugani and Phelps, 1986; Chugani et al., 1987) on human cerebral maturation that at the age of 2 months the most myelinated and highest metabolic areas are located within subcortical structures and primary cortices, while the temporal cortex as a whole shows a low metabolic activity. In such a region, poorly mature at the age of 2 months, the synaptic density is far below the adult level and far from its peak (Huttenlocher and Dabholkar, 1997). Using NRCBF during the diode condition as an index, we compared the cerebral blood flow values of the right FFA cluster with that of the right precentral region. This region is known to have high glucose metabolism (Chugani et al., 1987) and myelination (Brody et al., 1987) at this age and can thus be considered as relatively mature. In order to do so, we identified the precentral and postcentral sulci and manually defined on the MRI template the precentral region. Normalized cerebral blood flow of the right FFA was found 30%

6 RAPID COMMUNICATION 459 FIG. 2. Comparison of FFA locations reported in the literature, both in stereotaxic space and in single subjects. Metanalysis of FFA locations described in the stereotaxic space are displayed on the axial ( 40 mm, a) and sagittal ( 53 mm, d) slices on the SPM single subject MNI template, corresponding to the mean coordinates of the FFA activations; white dots correspond to the extrema listed in Table 3. (b) FFA location defined in a single subject on one of his axial brain MRI slice (Kanwisher et al., 1996). (c) Activated fusiform gyrus area detected in the present study superimposed on the template MRI of a single 2-month-old infant, in the same incidence as b. (e) Location of the right FFA activation during face matching defined in a single subject and displayed on one of his brain MRI sagittal slice (Clark et al., 1996). (f) Activated fusiform gyrus area detected in the present study in the same incidence as e. lower than that of the precentral gyrus (right FFA, 87 12; right precentral gyrus, 123 8, P 0.003; paired t test, 5 df ). This result can be taken as an indication that the metabolic maturation of this region level was lower that that of the precentral gyrus. One should note that the comparison of these blood flow values was undertaken in the diode condition. The same comparison during the face condition showed a similar trend, the activated right fusiform area still having lower NRCBF values than the precentral gyrus (right FFA, ; right precentral gyrus, 119 6, P 0.07; paired t test, 5 df ). Therefore, a major finding of the present study is that, despite a low level of metabolic activity, this visual associative cortex shows some relatively specialized functional activity. This demonstrates that functional maturation does not require metabolic and anatomical maturation completion in the involved cortical areas. It indicates the anteriority of functional maturation which might proceed simultaneously in several connected and active cortical regions. This hypothesis is consistent with the idea that synaptic stabilization is a crucial aspect of neural specialization (Changeux, 1983). Besides FFA, face perception by 2-month old infants activated a bilateral inferior occipital area corresponding to the occipital face area previously identified in adults as involved in early perception of facial features. Both this region and the FFA belong in adults to the visual analysis of faces core system, as defined by Haxby, which also includes the superior temporal sulcus (STS) that is consistently activated in adults (Sergent et al., 1992; Puce et al., 1996; Kanwisher et al., 1997; Chao et al., 1999a,b; Hoffman and Haxby, 2000; Haxby et al., 2000). Interestingly, the latter region was not activated in 2-month-old infants, consistent with its putative functional role in the adult brain, namely to respond to changeable aspects of faces, perception of

7 460 RAPID COMMUNICATION TABLE 3 Review of the Fusiform Face Area Extremum in the Clusters of Activations during Face Perception in Adults Reported in Papers Giving Their Results in the Stereotaxic Reference Space (Averaged Coordinates in the Stereotaxic Space x 40 4, y 53 10, z 20 7 (mm)) Coordinates Publications Task Reference Method x y z Puce et al., 1996 Faces perception Letters perception FMRI Dolan et al., 1996 Faces mental imagery Fixation PET Dolan et al., 1997 Faces specific learning effect PET Chao et al., 1999a Faces perception Houses perception FMRI Scramble Pictures perception FMRI Hoffman and Haxby, 2000 Faces perception Chao et al., 1999b Faces perception Animals, faceless animals, houses perception FMRI McCarthy et al., 1997 Faces among objects perception FMRI Nakamura et al., 2000 Faces perception Fixation PET Rossion et al., 2000 Faces perception Objects perception PET Kanwisher et al., 1997 Faces perception Objects perception FMRI Ishai et al., 1999 Faces perception Houses, chairs perception FMRI Gauthier et al., 2000b Faces perception Birds, cars perception FMRI Gauthier et al., 1999 Faces perception Greebles perception FMRI Kim et al., 1999 Recognize novel faces Recognize novel words PET gaze, and lip expression (Haxby et al., 2000). In the present study, the use of static and neutral faces may have led to the absence of recruitment of this region. Alternately, STS is an integrative multimodal area that will develop later on and thus was not to be expected to be involved at such an early stage of cognitive development. We also observed a right inferior parietal activation, nearby the intraparietal sulcus (Fig. 1). In reference to what is known in adults, this region belongs to the extended face processing system (Haxby et al., 2000) that, within this system, supports spatially directed attention. Finally, as opposed to adults, face processing in 2-month-old infants appears to recruit what will become later their language network, namely the left inferior frontal and superior temporal gyrus. We propose that coactivation of the face and this future language networks sustains the facilitative effects of social interactions, such as looking at the mother s face, on language development (Locke, 1997). ACKNOWLEDGMENTS The authors express their gratitude to Anne Boré, Marie Arnaud (Hôpital Robert Debré), Monique Crouzel, Bernadette Bruck, and the Cyclotron staff of the Service Hospitalier Frédéric Joliot (CEA, Orsay) where the PET experiments were carried out; to Uwe Pietryzk (Research Center of Juelich, Germany) and Olivier Quinton (GIN, Caen) for their help in data analysis; and to Petros Tzourio for his assistance in preparing the manuscript. This work has been supported in part by a grant from the Ministère de la Recherche, Sciences de la Cognition. REFERENCES Bayley, N Bayley Scales of Infant Development. Psychological Corporation, New York. Brody, B. A., Kinney, H. C., Kloman, A. S., and Gilles, F. H Sequence of central nervous system myelination in human infancy. I. An autopsy study of myelination. J. Neuropathol. Exp. Neurol. 46: Brunet, O., and Lezine, I Le Développement Psychologique de la Première Enfance. Presses Universitaires de France. Bushnell, I. W. R., Sai, F., and Mullin, J. T Neonatal recognition of the mother s face. Br. J. Dev. Psychol. 7: Changeux, J. P Concluding remarks: On the singularity of nerve cells and its ontogenesis. Prog. Brain Res. 58: Chao, L. L., Haxby, J. V., and Martin, A. 1999a. Attribute-based neural substrates in temporal corex for perceiving and knowing about objects. Nat. Neurosci. 2: Chao, L. L., Martin, A., and Haxby, J. V. 1999b. Are face-responsive regions selective only for faces? NeuroReport 10: Chugani, H. T., and Phelps, M. E Maturational changes in cerebral functions in infants determined by 18 FDG positron emission tomography. Science 231: Chugani, H. T., Phelps, M. E., and Mazziotta, J. C Positron emission tomography study of human brain functional development. Ann. Neurol. 22: Clark, V. P., Keil, K., Maisog, J. Ma., Courtney, S., Ungerleider, L. G., and Haxby, J. V Functional magnetic resonance imaging of human visual cortex during face matching: A comparison with positron emission tomography. NeuroImage 4: de Schonen, S., and Bry, I Interhemispheric communication of visual learning: A developmental study in 3 6-month old infants. Neuropsychologia 25: de Schonen, S., and Mathivet, E First come, first served: A scenario about the development of hemispheric specialization in face recognition during infancy. Eur. Bull. Cogn. Psychol. 9: 3 44.

8 RAPID COMMUNICATION 461 de Schonen, S., and Mathivet, E Hemispheric asymmetry in a face discrimination task in infants. Child Dev. 61: Deruelle, C., and de Schonen, S Do the right and left hemisphere attend to the same visuo-spatial information within a face in infancy? Dev. Neuropsychol. 14: Dolan, R. J., Fink, G. R., Rolls, E., Booth, M., Holmes, A., Frackowiak, R. S., and Friston, K. J. (Eds.) How the brain learns to see objects and faces in an impoverished context. Nature 389. Dolan, R. J., Fletcher, P., Morris, J., Kapur, N., Deakin, J. F. W., and Frith, C. D Neural activation during covert processing of positive emotional facial expressions. NeuroImage 4: Gauthier,I., Skudlarski, P., Gore, J. C., and Anderson, A. W. 2000b. Expertise for cars and birds recruits brain areas involved in face recognition. Nature Neurosci. 3: Gauthier, I., Tarr, M. J., Anderson, A. W., Skudlarski, P., and Gore, J. C Activation of the middle fusiform face area increases with expertise in recognizing novel objects. Nature Neurosci. 2: Gauthier, I., Tarr, M. J., Moylan, J., Skudlarski, P., Gore, J. C., and Anderson, A. W. 2000a. The fusiform face area is part of a network that processes faces at the individual level. J. Cogn. Neurosci. 12: Haxby, J. V., Hoffman, E. A., and Gobbini, M. I The distributed human neural system for face perception. TINS 4: Hoffman, E. A., and Haxby, J. V Distinct representations of eye gaze and identity in the distributed human neural system for face perception. Nature Neurosci. 3: Huttenlocher, P. R., and Dabholkar, A. S Regional differences in synaptogenesis in human cerebral cortex. J. Comp. Neurol. 387: Ishai, A., Ungerleider, L. G., Martin, A., Schouten, H. L., and Haxby, J. V Distributed representation of objects in the human ventral visual pathway. Proc. Natl. Acad. Sci. USA 96: Kanwisher, N Domain specificity in face perception. Nature Neurosci. 3: Kanwisher, N., Chun, M. M., McDermott, J., and Ledden, P. J Functional imaging of human visual recognition. Cogn. Brain Res. 5: Kanwisher, N., McDermott, J., and Chun, M. M The fusiform face area: A module in human extrastriate cortex specialized for face perception. J. Neurosci. 17(11): Kim, J. J., Andreasen, N. C., O Leary, D. S., Wiser, A. K., Ponto, L. L., Watkins, G. L., and Hichwa, R. D Direct comparison of the neural substrates of recognition memory for words and faces. Brain 122: Locke, J. L A theory of neurolinguistic development. Brain Lang. 58: McCarthy, G., Puce, A., Gore, J. C., and Allison, T Facespecific processing in the human fusiform gyrus. J. Cogn. Neurosci. 9: Morton, J., and Johnson, M. H CONSPEC and CONLERN: A two-process theory of infant face recognition. Psychol. Rev. 98: Nakamura, K., Kawashima, R., Sato, N., Nakamura, A., Sugiura, M., Kato, T., Hatano, K., Ito, K., Fukuda, H., Schormann, T., and Zilles, K Functional delineation of the human occipito-temporal areas related to face and scene processing A PET study. Brain 123: Pietrzyk, U., Herholz, K., and Heiss, W.-D Three-dimensional alignment of functional and morphological tomograms. J. Comput. Assist. Tomogr. 14: Powers, W., Stabin, M., Howse, D., Eichling, J. O., and Herscovitch, P Radiation absorbed dose estimates for oxygen-15 radiopharmaceuticals (H 2 15 O, C 15 O, O 15 O) in newborn infants. J. Nucl. Med. 29: Puce, A., Allison, T., Asgari, M., Gore, J. C., and McCarthy, G Differential sensitivity of human visual cortex to faces, letterstrings, and textures: A functional magnetic resonance imaging study. J. Neurosci. 16(16): Rossion, B., Dricot, L., DeVolder, A., Bodart, J. M., Crommelinck, M., degelder, B., and Zoontjes, R Hemispheric asymmetries for whole-based and part-based face processing in the human fusiform gyrus. J. Cogn. Neurosci. 12: Sergent, J., Ohta, S., and MacDonald, B Functional neuroanatomy of face and object processing: A positron emission tomography study. Brain 115: Sparrow, S. S., and Cicchetti, D. V The Vineland Adaptive Behavior Scales: Major Psychological Assessment Instruments. Allyn & Bacon, Boston, MA. Tarr, M. J., and Gauthier, I FFA: A flexible fusiform area for subordinate-level visual processing automatized by expertise. Nature Neurosci. 3: Woods, R. P., Grafton, S. T., Holmes, C. J., Cherry, S. R., and Mazziotta, J. C Automated image registration. I. General methods and intrasubject validation. J. Comput. Assist. Tomogr. 22: