Copyright 2005 Psychonomic Society, Inc. 452Extensive experience with an object class can result inexquisite perceptual skills. Car experts can recognize ata glance the makes and models of hundreds of cars, andbird-watchers can identify the species of a bird brieflyseen through foliage. The acquisition of such perceptualexpertise with objects can lead to functional specializationwithin the brain systems dedicated to visual processing.Faces are commonly used to study the phenomenon ofcategory-specific neural specialization (Kanwisher, 2000;Kanwisher, Chun, & McDermott, 1996), but neural spe-cialization has also been demonstrated for other expertobject classes, including cars, birds, and even novel ob-jects (e.g., “Greebles”; Gauthier, Tarr, Anderson, Skud-larski, & Gore, 1999; Rossion, Gauthier, Goffaux, Tarr,& Crommelinck, 2002). Most people are also perceptualexperts with words. For adults, much of modern life isspent reading, and adult readers can recognize wordswith amazing efficiency (LaBerge & Samuels, 1974). Aschildren, most of us once dedicated a considerable amountof time to learning letters, and each day a literate personmust process thousands of letters just to interact normallywith the modern environment. Here, we hypothesize thatthis extensive experience results in letter-specific neuralspecialization of the visual system.Most of the neuroimaging studies conducted withprinted text were designed for the purpose of under-standing the processes involved in reading. Therefore, itmay not be surprising that there is only limited evidencespeaking directly to the neural substrates involved in rec-ognizing single letters. Several large circuits appear tounderlie various aspects of reading. For example, a dorsalposterior system (angular gyrus, supramarginal gyrus,and superior temporal sulcus) is thought to subserveorthography-to-phonology correspondences (Black &Behrmann, 1994). Semantic analyses of words are believedto occur predominantly in the left inferior frontal lobeand posterior superior temporal sulcus (Bookheimer,2002; Gabrieli, Poldrack, & Desmond, 1998), and con-text comprehension is considered a right hemisphere task(Kircher, Brammer, Tous, Williams, & McGuire, 2001).A posterior ventral network including the occipitotem-poral region is thought to underlie visual processing ofprinted text and to be responsible for the late-developingskills of rapid word recognition that result from increasedreading experience (Frackowiak, Friston, Frith, Dolan, &Mazziotta, 1997; Pugh et al., 2001). Within this system,the region that has received the most attention as a can-didate area for early visual processing of printed text isthe left fusiform gyrus.Part of the left midfusiform gyrus shows higher acti-vation for words than for consonant strings of the samelength (Cohen et al., 2000; Dehaene, Le Clec’H, Poline,Le Bihan, & Cohen, 2002) and has therefore been la-This research was supported by NEI Grant R01-EY13441 to I.G. anda grant from the James S. McDonnell Foundation to the Perceptual Ex-pertise Network. The authors thank Danielle Brown, Debbie Boner, andRichard Baheza for technical assistance, and the members of the Per-ceptual Expertise Network for helpful discussions. Correspondenceconcerning this article should be addressed to K. H. James, PsychologyDepartment, Indiana University, 1101 E. 10th Street, Bloomington, IN47405 (e-mail: [email protected]).Letter processing in the visual system: Differentactivation patterns for single letters and stringsKARIN H. JAMES and THOMAS W. JAMESIndiana University, Bloomington, IndianaandGAEL JOBARD, ALAN C.-N. WONG, and ISABEL GAUTHIERVanderbilt University, Nashville, TennesseeOne would expect that a lifetime of experience recognizing letters would have an important influenceon the visual system. Surprisingly, there is limited evidence of a specific neural response to letters overvisual control stimuli. We measured brain activation during a sequential matching task using isolatedcharacters (Roman letters, digits, and Chinese characters) and strings of characters. We localized the vi-sual word form area (VWFA) by contrasting the response to pseudowords against that for letter strings,but this region did not show any other sign of visual specialization for letters. In addition, a left fusiformarea posterior to the VWFA was selective for letter strings, whereas a more anterior left fusiform regionshowed selectivity for single letters. The results of different analyses using both large regions of inter-est and inspections of individual patterns of response reveal a dissociation between selectivity for let-ter strings and selectivity for single letters. The results suggest that reading experience fine-tunes visualrepresentations at different levels of processing. An important conclusion is that the processing of non-pronounceable letter strings cannot be assumed to be equivalent to single-letter perception.Cognitive, Affective, & Behavioral Neuroscience2005, 5 (4), 452-466LETTER PROCESSING IN THE VISUAL SYSTEM 453beled the visual word form area (VWFA; Talairach co-ordinates [TCs] [x, y, z] approximately, ⫺ 42,⫺57,⫺15).However, the VWFA is also engaged by images of ob-jects (Booth et al., 2002; Devlin et al., 2002; Moore &Price, 1999; Murtha, Chertkow, Beauregard, & Evans,1999), which may not be surprising given its proximityto (and sometimes overlap with) the lateral occipitalcomplex, an area defined as responding more to intactimages of objects than to other visual stimuli (Malachet al., 1995). In addition, although the VWFA does notdifferentiate between pseudowords and words (Hagoortet al., 1999), it responds more to pseudowords than toconsonant strings or “poorly constructed” pseudowords(McCandliss, Cohen, & Dehaene, 2003). Therefore, theVWFA is thought to be involved in processing word-likestimuli, but its specific role remains controversial (forhypotheses, see Jobard, Crivello, & Tzourio-Mazoyer,2003; McCandliss et al., 2003; Price, 2000).For our purposes, one important limitation of the read-ing neuroimaging literature is that studies rarely includecomparisons between letter strings and nonletter con-trols. However, these comparisons are crucial for under-standing early visual processing of print. In one of thefirst studies to compare letter strings with nonletter stim-uli (Puce, Allison, Asgari, Gore, & McCarthy, 1996), let-ter strings were compared with faces and textures, and anarea (overlapping with the VWFA)