Specialization within the ventral stream: the case for the visual word form area

In the paper

“The available evidence from single-subject fMRI and intracranial recordings suggests that some cortical sites respond preferentially to letter strings than to other categories of visual stimuli such as faces or objects, though the preference is often relative rather than absolute.”

In the paper

“patients with pure alexia consistently exhibit lesions of this region (reproducible localization).”

In the paper

“Examination of the literature with this theoretical division in mind indicates that reading activates a precise subpart of the left ventral occipitotemporal sulcus”

In the paper

“We do not, however, believe that there is complete regional selectivity for word recognition.”

In the paper

“A large number of studies, using various methods, have evidenced a reproducible activation of the left fusiform gyrus during reading.”

What the sources say

• Not checked yetcites Beauregard et al., 1997 source
• Not checked yetcites Brunswick et al., 1999 source
• Not checked yetcites Dehaene et al., 2002 source

In the paper

“although it provides evidence for functional specialization for case-invariant letter recognition in the left fusiform region (Dehaene et al., 2001; Dehaene et al., in press).”

What the sources say

• Not checked yetcites Dehaene et al., 2001 source
• Not checked yetcites Dehaene et al., in press

In the paper

“Activation in this region can be identified merely by contrasting word reading relative to rest, either with PET or with fMRI (Beauregard et al., 1997; Brunswick et al., 1999; Dehaene et al., 2002; Fiez et al., 1999; Paulesu et al., 2000; Wagner et al., 1998).”

What the sources say

• Not checked yetcites Beauregard et al., 1997 source
• Not checked yetcites Brunswick et al., 1999 source
• Not checked yetcites Dehaene et al., 2002 source

In the paper

“This region appears as a likely source of electrical and magnetic fields that are recorded over the left ventral occipitotemporal region, with a latency of about 150-200 ms, whenever subjects see words (Allison et al., 1994, 1999; Cohen et al., 2000; Nobre et al., 1994; Salmelin et al., 1996; Simos et al., 2002; Tarkiainen et al., 1999).”

What the sources say

• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Allison et al., 1994, 1999
• Not checked yetcites Cohen et al., 2000 source
• Not checked yetcites Nobre et al., 1994 source

In the paper

“Even the voxels that respond optimally to words tend to be also activated by other stimuli such as pictures or line drawings of objects.”

In the paper

“The VWFA is found at the same location in Talairach space (approximately 43, 54, 12) with a standard deviation of only about half a centimeter.”

In the paper

“First, there are indications that the VWFA is tuned to the shape of letters, relative to visually equivalent pseudo-letters.”

What the sources say

• Not checked yetcites Petersen et al. (1990) source
• Not checked yetcites Price et al. (1996) source

In the paper

“This allows quantification of the interindividual variability, which is remarkably low.”

In the paper

“As already noted by Bu ¨chel et al. (1998), words in modalities other than the visual yielded activations more anterior (average y = 43) than those typical of the visual word form (average y = 60).”

What the sources say

• Not checked yetcites Büchel et al. (1998) source

In the paper

“There is good convergence to the same approximate Talairach coordinates across those studies (Cohen et al., 2000).”

What the sources say

• Not checked yetcites Cohen et al., 2000 source

In the paper

“This systematic topographical organization was reproduced by Gauthier et al. (2000).”

What the sources say

• Not checked yetcites Gauthier et al. (2000) source

In the paper

“The left occipitotemporal sulcus (average coordinates 44, 68, 4) was activated only by visual words, not by spoken words.”

In the paper

“Such a distinction is supported by several neuroimaging studies (e.g. Booth et al., 2002b; see review in Giraud and Price, 2001) as well as by intracranial recordings (Nobre et al., 1994, 1998).”

What the sources say

• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Booth et al., 2002b
• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Giraud and Price, 2001
• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Nobre et al., 1994, 1998

In the paper

“Several fMRI papers have further shown that this area can be identified in single subjects (Cohen et al., 2000, 2002; Dehaene et al., 2002; Gauthier et al., 2000; Puce et al., 1996).”

What the sources say

• ?Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  cites Cohen et al., 2000, 2002
• Not checked yetcites Dehaene et al., 2002 source
• Not checked yetcites Gauthier et al., 2000 source

In the paper

“For instance, Price et al. (1996) have obtained stronger activations to strings of consonants than to strings of pseudo-letters at coordinates close to the VWFA.”

What the sources say

• Not checked yetcites Price et al. (1996) source

In the paper

“In one study, we isolated the VWFA using a spatial invariance criterion: contrary to more posterior retinotopic areas, the VWFA is activated by words presented in either the left or the right hemifield (Cohen et al., 2000).”

What the sources say

• Not checked yetcites Cohen et al., 2000 source

In the paper

“Another more lateral and slightly more anterior inferotemporal region ( 48, 60, 16) was activated also by spoken words, with modulation by the graphemic task.”

In the paper

“Hasson et al. (2002) also used fMRI to contrast activations to words, houses, and faces. Again, a site with greater responsivity to words than to either faces or buildings was found in the left occipitotemporal sulcus, lateral to face- and building-related activations.”

What the sources say

• Not checked yetcites Hasson et al. (2002) source

In the paper

“In another, we further required this region to respond more to strings of letters than to chequerboards presented in either hemifield (Cohen et al., 2002).”

What the sources say

• Not checked yetcites Cohen et al., 2002 source

In the paper

“In a group analysis, this activation overlapped with the VWFA, as defined by a left inferior temporal activation that was greater to visual than to spoken words.”

In the paper

“However, in single subjects, the two regions appeared as distinct but very close.”

In the paper

“Nevertheless, the available evidence reveals systematic and reproducible activations that are greater to visual words than to other control stimuli.”

In the paper

“Following up on the initial insight of Petersen et al. (1990), Price et al. (1996) observed stronger activation in the left fusiform gyrus to strings of consonants than to strings of false-font characters.”

What the sources say

• Not checked yetcites Price et al. (1996) source

In the paper

“Similar results have been obtained using singlesubjects MEG (Tarkiainen et al., 1999) and intracranial recordings (Allison et al., 1999).”

What the sources say

• Not checked yetcites Tarkiainen et al., 1999 source
• Not checked yetcites Allison et al., 1999 source

In the paper

“There is also evidence that strings of letters cause greater left fusiform activation than visually equivalent strings of digits (Polk et al., 2002).”

What the sources say

• Not checked yetcites Polk et al., 2002

In the paper

“For instance, the interference of orthography prevents the detection of the phoneme ''p'' in the spoken French word ''absurde'' (Halle ´ et al., 2000).”

What the sources say

• Not checked yetcites Hallé et al., 2000 source

In the paper

“The posterior subpart, observed bilaterally, was not invariant for location: it showed priming when single letters were repeated at the same retinal location, irrespective of whether the prime word was identical to the target or was an anagram of it.”

In the paper

“In such cases, we observed reduced activations of the VWFA for words presented in the left hemifield (Cohen et al., 2000; Molko et al., 2002).”

What the sources say

• Not checked yetcites Cohen et al., 2000 source
• Not checked yetcites Molko et al., 2002 source

In the paper

“In a recent word priming study, we found indications that the mid-fusiform cortex can be decomposed into smaller subareas (Dehaene et al., in press).”

What the sources say

• Not checked yetcites Dehaene et al., in press

In the paper

“The VWFA shows equally robust fMRI responses when words are presented in a familiar format or in a mixed-case format (e.g. ''tAbLe'') (Polk and Farah, 2002).”

What the sources say

• Not checked yetcites Polk and Farah, 2002

In the paper

“Indeed, intracranial recording suggests that there may be specialized subregions for letters vs. digits in the ventral occipitotemporal cortex (Allison et al., 1994).”

What the sources say

• Not checked yetcites Allison et al., 1994 source

In the paper

“Such case-invariant priming in the VWFA was recently replicated with words made only of visually dissimilar letters that are related only by arbitrary cultural convention (e.g. a and A, g and G) (Dehaene et al., in press).”

What the sources say

• Not checked yetcites Dehaene et al., in press

In the paper

“Finally, even more anteriorily in the middle fusiform gyrus, there was a trend towards greater priming when the same word was presented twice compared to when it was preceded by an anagram, suggesting the presence of the units larger than the letter level.”

In the paper

“In our recent work, we also found a similar region of lesion overlap in the fusiform gyrus.”

What the sources say

• Not checked yetcites Cohen et al., 2003 source

In the paper

“Furthermore, the VWFA shows repetition priming whenever a target word is preceded by a subliminal presentation of the same word, whether the two words are printed in the same or in a different case (e.g. table followed by TABLE) (Dehaene et al., 2001).”

What the sources say

• Not checked yetcites Dehaene et al., 2001 source

In the paper

“Second, the VWFA responds more to words or to pseudo-words than to random consonant strings that violate orthographic constraints (Beauregard et al., 1997; Bu ¨chel et al., 1998; Cohen et al., 2002; Price et al., 1996; Rees et al., 1999; Xu et al., 2001).”

What the sources say

• Not checked yetcites Beauregard et al., 1997 source
• Not checked yetcites Büchel et al., 1998 source
• Not checked yetcites Cohen et al., 2002 source

In the paper

“Nevertheless, most cortical patches that respond to words also activate to other stimuli such as pictures or drawings.”

In the paper

“We are able to recognize words across considerable changes in font , CASE, location , and size.”

What the sources say

• Not checked yetcites Besner, 1989 source
• Not checked yetcites Mayall et al., 1997 source
• Not checked yetcites Paap et al., 1984 source

In the paper

“its lesion results in the selective loss of reading-specific processes (functional specialization).”

In the paper

“This implies that the visual word recognition process is highly efficient in ''normalizing'' and computing an invariant representation that discards irrelevant variations in the visual input.”

In the paper

“We argue that such perceptual expertise is likely to require specialized neural systems within the ventral visual stream.”

In the paper

“the issue of regional selectivity, namely, the existence of putative cortical patches dedicated to letter and word recognition, cannot be resolved by positron emission tomography or lesion data, but requires high-resolution neuroimaging techniques.”

In the paper

“this region implements processes adequate for reading in a specific script, such as invariance across upperand lower-case letters”

In the paper

“We conclude that learning to read results in the progressive development of an inferotemporal region increasingly responsive to visual words, which is aptly named the visual word form area (VWFA).”

In the paper

“This perceptual expertise requires no less than 5 years of academic training in a specific writing system (Aghababian and Nazir, 2000).”

What the sources say

• Not checked yetcites Aghababian and Nazir, 2000

In the paper

“Therefore, at the scale accessible to human neuroimaging, most if not all of the cortical surface is likely to respond jointly to words, faces, and objects, with smooth but systematic changes in local selectivity (Haxby et al., 2001).”

What the sources say

• Not checked yetcites Haxby et al., 2001 source

In the paper

“We also propose that there is reproducible localization of the neural circuits that are attuned to the reading process.”

In the paper

“This phenomenon, which is found with real words, pseudo-words (Spoehr and Smith, 1975), and even nonwords sufficiently similar to words such as SPCT (Rumelhart and McClelland, 1982), implies the existence of a mechanism tuned to the possible structures of visual words.”

In the paper

“Rather, the case invariance observed in several psychological experiments (e.g. Besner et al., 1984; Bowers et al., 1998) must arise from specialization resulting from a learning process.”

What the sources say

• Not checked yetcites Besner et al., 1984 source
• Not checked yetcites Bowers et al., 1998 source

In the paper

“Finally, it is likely that the rather large occipitotemporal region that we have termed VWFA can be further subdivided into, for instance, a posterior location-specific sector and a more anterior location-invariant sector (see Dehaene et al., in press).”

What the sources say

• Not checked yetcites Dehaene et al., in press

In the paper

“This suggests the presence of a parallel apprehension mechanism specialized for reading.”

In the paper

“In this context, the finding of overlap between wordrelated and non-word-related (e.g. color or object processing) tasks is not conclusive, as it seems likely that even cortical territories known to be distinct would be found to overlap with this method.”

In the paper

“Also remarkable is the absence of any effect of word length: within a range of about three to six letters, visual word recognition takes a constant time regardless of the number of letters composing that word (Lavidor and Ellis, 2002; Weekes, 1997).”

What the sources say

• Not checked yetcites Lavidor and Ellis, 2002
• Not checked yetcites Weekes, 1997 source

In the paper

“Note that the absence of a length effect is found only when the words are presented in the familiar horizontal format and within a restricted region of the fovea and right hemifield (Lavidor and Ellis, 2002; Lavidor et al., 2001).”

What the sources say

• Not checked yetcites Lavidor and Ellis, 2002
• Not checked yetcites Lavidor et al., 2001 source

In the paper

“Indeed, some of the first evidence for a specific response to structured letter strings, as opposed to matched visual stimuli, came from PET (Petersen and Fiez, 1993; Petersen et al., 1990).”

What the sources say

• Not checked yetcites Petersen and Fiez, 1993
• Not checked yetcites Petersen et al., 1990 source

In the paper

“PET data remain appropriate to study functional rather anatomical specialization for word processing.”

In the paper

“We propose that there is functional specialization for reading in the brain of literate human subjects.”

In the paper

“Those computational requirements are largely common to the recognition of any type of visual objects: Tools, buildings, or faces can all appear with huge variations in their appearance, including size, location, orientation, and so on.”

What the sources say

• Not checked yetcites Riesenhuber and Poggio, 1999 source

In the paper

“In general, however, what constitutes a significant visual difference and what does not is defined culturally rather than visually.”

In the paper

“Thus, the fast mapping of ''A'' and ''a'' onto the same abstract letter identity cannot be explained by generic features of the organization of the visual system.”

In the paper

“This cortex is partially plastic and can, during ontogenesis, become progressively attuned to the requirements of reading (McCandliss et al., 2003).”

What the sources say

• Not checked yetcites McCandliss et al., 2003 source

In the paper

“This should not be surprising because the functional specialization for reading arises from a partial preemption or ''recycling'' of visual cortex which evolved for other purposes during phylogenesis (Dehaene, 2004).”

What the sources say

• Not checked yetcites Dehaene, 2004

In the paper

“In the case of reading, crucial aspects of this capacity of invariant recognition must be learned.”

In the paper

“This suggests that at least part of the specialization for letter and word recognition occurs at retinotopically organized stages of visual processing (Nazir, 2000).”

What the sources say

• Not checked yetcites Nazir, 2000

In the paper

“This is especially true when the method is applied across different groups of subjects, although individual sulci are known to vary at least up to 1 cm (Sowell et al., 2002; Thompson et al., 1996).”

What the sources say

• Not checked yetcites Sowell et al., 2002 source
• Not checked yetcites Thompson et al., 1996 source

In the paper

“Only advanced experimental designs, such as the priming or adaptation method (Grill-Spector and Malach, 2001; Naccache and Dehaene, 2001), are apt at proving the presence of shared circuitry.”

What the sources say

• Not checked yetcites Grill-Spector and Malach, 2001
• Not checked yetcites Naccache and Dehaene, 2001

In the paper

“We also consider it rather unlikely that the visual word area would have sharp borders that would delineate it qualitatively from other neighboring areas involved in face or object recognition.”

In the paper

“In most subjects, however, the specialization for words is likely to be only partial.”

In the paper

“We do not believe that there is a cytoarchitectonically defined subpart of the visual system that responds to visual words.”

In the paper

“Many of the mappings between upper case and lower case, for instance, are arbitrary.”

In the paper

“A generic visual recognition device is simply insufficient to account for the notation-dependent operations that are observed in skilled readers.”

In the paper

“It is known that the acquisition of literacy modifies the brain network involved in speech repetition (especially for pseudo-words, see Castro-Caldas et al., 1998).”

What the sources say

• Not checked yetcites Castro-Caldas et al., 1998 source

In the paper

“However, in the present state of knowledge, it cannot be denied that small subregions of visual cortex show at least partial selectivity for letter strings.”

In the paper

“Thus, we suggest that the visual word form area must be carefully distinguished from more anterior regions that are increasingly cross-modal and engaged in semantic computations.”

In the paper

“This does not, however, preclude the possibility of its top-down activation by auditory words.”

In the paper

“However, simple subtractions allow a more specific delineation of the VWFA.”

In the paper

“PET and FMRI studies are ill-adapted to resolve the issue of top-down or bottom-up posterior inferotemporal activations because they cannot resolve the fine temporal dynamics of neural activity.”

In the paper

“Furthemore, a dramatic interference of orthography on speech perception has been evidenced in a variety of behavioral tasks including rhyme detection (Donnenwerth-Nolan et al., 1981), syllable monitoring (Taft and Hambly, 1985), lexical decision (Ziegler and Ferrand, 1998), and phoneme monitoring.”

What the sources say

• Not checked yetcites Donnenwerth-Nolan et al., 1981 source
• Not checked yetcites Taft and Hambly, 1985
• Not checked yetcites Ziegler and Ferrand, 1998

In the paper

“Second, even in patients with large lesions, the combination of lesions from multiple patients may allow to identify the necessary and sufficient regions that are reproducibly associated with pure alexia.”

In the paper

“The lesion affected the expected region of the left mid-fusiform cortex, with an extension to the lateral edge of the lingual gyrus.”

What the sources say

• Not checked yetcites Beversdorf et al., 1997

In the paper

“Conversely, output projections from the VWFA to lateral language areas may be interrupted during their course in the parietal white matter, providing a plausible account for the so-called subangular alexia (Greenblatt, 1976; Iragui and Kritchevsky, 1991; Pirozzolo et al., 1981).”

What the sources say

• Not checked yetcites Greenblatt, 1976 source
• Not checked yetcites Iragui and Kritchevsky, 1991
• Not checked yetcites Pirozzolo et al., 1981 source

In the paper

“More anteriorily, priming became location invariant, though present for both words and their anagrams, suggesting a location-invariant sublexical code.”

In the paper

“Even assuming that it was, these observations would not necessarily contradict the proposed model due to the possible contribution of white matter lesions.”

In the paper

“Third, there is also evidence that the VWFA computes the arbitrary mapping between upperand lower-case letters.”

In the paper

“Such a fine-grained specialization cannot be explained by a generic visual process.”

In the paper

“The critical region fell within the left middle fusiform gyrus.”

What the sources say

• Not checked yetcites Binder and Mohr, 1992

In the paper

“In the most severe cases, known as global alexia, they cannot even identify single letters (Dejerine, 1892).”

What the sources say

• Not checked yetcites Dejerine, 1892

In the paper

“The clearest example of such a deafferentation is the alexia restricted to the left half of the visual field which results from lesions of posterior callosal pathways (Suzuki et al., 1998).”

What the sources say

• Not checked yetcites Suzuki et al., 1998 source

In the paper

“A factor that may blur anatomo-functional correlations is the role of white matter lesions.”

In the paper

“Finally, even more anteriorily in the middle fusiform gyrus, there was a trend towards greater priming when the same word was presented twice compared to when it was preceded by an anagram, suggesting the presence of the units larger than the letter level.”

In the paper

“Altogether, simple considerations on the input and output pathways of the VWFA provide a detailed and consistent explanation for the variety of visual alexias seen in adult neuropsychological cases (Cohen et al., 2003).”

What the sources say

• Not checked yetcites Cohen et al., 2003 source

In the paper

“Thus, no firm anatomical argument, positive or negative, can be made from Warrington and Shallice's cases.”

In the paper

“The simplest account of pure alexia is that it reflects a breakdown of the visual word form system (Cohen et al., 2000; Warrington and Shallice, 1980).”

What the sources say

• Not checked yetcites Cohen et al., 2000 source
• Not checked yetcites Warrington and Shallice, 1980

In the paper

“Finally, the anatomical case reported by Greenblatt (1973) may be an instance of complete deafferentation of an intact fusiform cortex (see also Cohen et al., submitted for publication).”

What the sources say

• Not checked yetcites Greenblatt, 1973 source
• Not checked yetcites Cohen et al., submitted for publication

In the paper

“Thus, this region likely holds a location-specific sublexical code.”

In the paper

“This suggests that it has become sensitive to orthographic rules in the subject's language, which are matters of cultural convention.”

In the paper

“However, such associations of deficits are known to provide only a weak source of constraints (Shallice, 1988).”

What the sources say

• Not checked yetcites Shallice, 1988

In the paper

“Such work was done by Binder and Mohr (1992), Damasio and Damasio (1983), and more recently by our group (Cohen et al., 2003), with a good agreement of these studies on the lateral edge of the left fusiform gyrus.”

What the sources say

• Not checked yetcites Binder and Mohr, 1992
• Not checked yetcites Damasio and Damasio, 1983 source
• Not checked yetcites Cohen et al., 2003 source

In the paper

“First, not all patients have extensive lesions blurring anatomical correlations.”

What the sources say

• Not checked yetcites Beversdorf et al., 1997

In the paper

“The intermingling of such neurons with others coding for objects or faces would translate into a partial regional selectivity at the single-voxel level, which is all that we can presently measure with PET or fMRI.”

In the paper

“Given that the VWFA may contain cortical columns specialized for letter processing, mixed with columns with different preferences (Puce et al., 1999), it is likely that a disruption of the VWFA would induce some degree of impairment beyond reading (Behrmann et al., 1998a).”

What the sources say

• Not checked yetcites Puce et al., 1999 source
• Not checked yetcites Behrmann et al., 1998a

In the paper

“We tentatively suggest that the reproducible localization of the VWFA indicates that only a small subregion of visual cortex has the optimal initial characteristics that permit it to be preempted or ''recycled'' for reading (Dehaene, 2004).”

What the sources say

• Not checked yetcites Dehaene, 2004

In the paper

“is not sufficiently articulate to account for double dissociations such as object agnosia without alexia (Albert et al., 1975; Gomori and Hawryluk, 1984) or prosopagnosia without alexia (e.g. McNeil and Warrington, 1993).”

What the sources say

• Not checked yetcites Albert et al., 1975 source
• Not checked yetcites Gomori and Hawryluk, 1984
• Not checked yetcites McNeil and Warrington, 1993

In the paper

“Furthermore, impaired identification of even single letters is common in pure alexia, ranging in severity from complete global alexia to the occurrence of occasional errors with a significant slowing down of letter naming (for a review, see Behrmann et al., 1998b; e.g. Hanley and Kay, 1996).”

What the sources say

• Not checked yetcites Behrmann et al., 1998b
• Not checked yetcites Hanley and Kay, 1996

In the paper

“Such neurophysiological mechanisms may provide a neuronal basis for the acquisition of invariant visual word recognition in humans.”

In the paper

“Binding of letters may also be based on more holistic coding, as suggested by the recent demonstration that single IT neurons develop selectivity for learned complex shapes, above and beyond what could be expected from the additive influence of their component parts (Baker et al., 2002).”

What the sources say

• Not checked yetcites Baker et al., 2002 source

In the paper

“Such observations predict the existence of highly specialized but patchy and distributed neuronal populations coding for alphabetic stimuli at the single-neuron level.”

In the paper

“However, it is unclear how this may account for the frequent sparing of the recognition of complex multipart objects, faces, or digit strings in pure alexic patients.”

In the paper

“The conjunction of these factors may collectively conspire to define a cortical region optimally apt at hosting the functional processes of visual word recognition.”

In the paper

“IT neurons are able to learn new arbitrary shapes (Logothetis et al., 1995), and arbitrary pairings of stimuli (Sakai and Miyashita, 1991), which might be an analogue of learning letters and of linking arbitrarily related upper- and lowercase letters.”

In the paper

“Even assuming that there is a cortical sector highly selective for the perception of alphabetic stimuli, it would be unlikely that a spontaneous lesion should affect this region with perfect selectivity although sparing the adjacent cortex.”

In the paper

“This symptom presents a real challenge to nonspecific theories of pure alexia because it is difficult to argue that single letters are in any sense more complex than faces or drawings of objects.”

In the paper

“Thus, it is hardly surprising that some degree of associated visual impairment can be evidenced in most alexic patients.”