CoChat

The emergence of the visual word form: Longitudinal evolution of category-specific ventral visual areas during reading acquisition

Open access
Dehaene-Lambertz, Ghislaine, Monzalvo, Karla, Dehaene, Stanislas · 2018 · PLOS Biology

The DOI matches its Crossref record; 105 of the 105 references it cites were found in a registry or an index.

Abstract

How does education affect cortical organization? All literate adults possess a region specialized for letter strings, the visual word form area (VWFA), within the mosaic of ventral regions involved in processing other visual categories such as objects, places, faces, or body parts. Therefore, the acquisition of literacy may induce a reorientation of cortical maps towards letters at the expense of other categories such as faces. To test this cortical recycling hypothesis, we studied how the visual cortex of individual children changes during the first months of reading acquisition. Ten 6-year-old children were scanned longitudinally 6 or 7 times with functional magnetic resonance imaging (fMRI) before and throughout the first year of school. Subjects were exposed to a variety of pictures (words, numbers, tools, houses, faces, and bodies) while performing an unrelated target-detection task. Behavioral assessment indicated a sharp rise in grapheme-phoneme knowledge and reading speed in the first trimester of school. Concurrently, voxels specific to written words and digits emerged at the VWFA location. The responses to other categories remained largely stable, although right-hemispheric face-related activity increased in proportion to reading scores. Retrospective examination of the VWFA voxels prior to reading acquisition showed that reading encroaches on voxels that are initially weakly specialized for tools and close to but distinct from those responsive to faces. Remarkably, those voxels appear to keep their initial category selectivity while acquiring an additional and stronger responsivity to words. We propose a revised model of the neuronal recycling process in which new visual categories invade weakly specified cortex while leaving previously stabilized cortical responses unchanged.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1371/journal.pbio.2004103

Cites

105 of 105 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig 1 Evolution of reading performance as a function of schooling in each child across the 7 measurement times. Each color represents the performance of a given child. The norms for the number of words read per minute is 36.7 (±15.8) at the end of the second year of French school (LUM). S1 Data. LUM, “Lecture en une minute.”
Fig 1 Evolution of reading performance as a function of schooling in each child across the 7 measurement times. Each color represents the performance of a given child. The norms for the number of words read per minute is 36.7 (±15.8) at the end of the second year of French school (LUM). S1 Data. LUM, “Lecture en une minute.”CC BY · PubMed Central
Fig 2 Category-specific activations across all 7 sessions and 10 children. (A) Each category is contrasted relative to the other pictures (omitting the grids) while pooling across all 7 sessions. On the right, one example of each presented visual category. The color of the frame indicates the visual category of the specific activation on the slices and of the lines in the plots. The significant clusters (p < 0.001 corrected at the cluster level at p < .05) are presented over an individual child brain. No significant cluster was observed for numbers. See S2 Fig to see the activation for each category separately and S3 Fig for an example in an individual child (child 2). (B) Plots show the evolution over sessions of the betas for each category, at the primary peak in ventral visual cortex for each category and (C) at the peaks of the word specific activations in the whole brain. The first 6 sessions were evenly spaced during the first year of school, while the seventh session occurred at the end of the second year. Note that while the contrast used to create this image weighted all 7 sessions equally, plots revealed linear (e.g., VWFA) as well as quadratic (e.g., parietal, IFG) profiles, which are further analyzed in the text. (https://neurovault.org/collections/3457/). IFGoper, inferior frontal gyrus opercularis; IFGtri, inferior frontal gyrus triangularis; pSTS, posterior superior temporal sulcus; VWFA, visual word form area.
Fig 2 Category-specific activations across all 7 sessions and 10 children. (A) Each category is contrasted relative to the other pictures (omitting the grids) while pooling across all 7 sessions. On the right, one example of each presented visual category. The color of the frame indicates the visual category of the specific activation on the slices and of the lines in the plots. The significant clusters (p < 0.001 corrected at the cluster level at p < .05) are presented over an individual child brain. No significant cluster was observed for numbers. See S2 Fig to see the activation for each category separately and S3 Fig for an example in an individual child (child 2). (B) Plots show the evolution over sessions of the betas for each category, at the primary peak in ventral visual cortex for each category and (C) at the peaks of the word specific activations in the whole brain. The first 6 sessions were evenly spaced during the first year of school, while the seventh session occurred at the end of the second year. Note that while the contrast used to create this image weighted all 7 sessions equally, plots revealed linear (e.g., VWFA) as well as quadratic (e.g., parietal, IFG) profiles, which are further analyzed in the text. (https://neurovault.org/collections/3457/). IFGoper, inferior frontal gyrus opercularis; IFGtri, inferior frontal gyrus triangularis; pSTS, posterior superior temporal sulcus; VWFA, visual word form area.CC BY · PubMed Central
Fig 3 Evolution of activation in a child (child 2) who was scanned 7 times (top to bottom). The left columns show the evolution of words > others on whole-brain flat maps and on a slightly inflated view of the left hemisphere. There was no activation of the RR circuit in the first 2 sessions, prior to school onset. Once schooling started, a VWFA activation was quickly detected, increasing in intensity with time and remaining detectable 1 year later. Parietal and prefrontal areas were transiently activated. In the middle column, zoom on the left ventral visual region. The green arrow indicates the VWFA. The right-hand column, ventral areas viewed from the back, illustrates the stability of face-related activations (contrast = faces > others, red arrows). See S3 Fig for the other categories. All contrasts at voxelwise p < 0.001 and clusterwise p < 0.05 FWE corrected. FWE, family-wise error; RR, reading-related; VWFA, visual word form area.
Fig 3 Evolution of activation in a child (child 2) who was scanned 7 times (top to bottom). The left columns show the evolution of words > others on whole-brain flat maps and on a slightly inflated view of the left hemisphere. There was no activation of the RR circuit in the first 2 sessions, prior to school onset. Once schooling started, a VWFA activation was quickly detected, increasing in intensity with time and remaining detectable 1 year later. Parietal and prefrontal areas were transiently activated. In the middle column, zoom on the left ventral visual region. The green arrow indicates the VWFA. The right-hand column, ventral areas viewed from the back, illustrates the stability of face-related activations (contrast = faces > others, red arrows). See S3 Fig for the other categories. All contrasts at voxelwise p < 0.001 and clusterwise p < 0.05 FWE corrected. FWE, family-wise error; RR, reading-related; VWFA, visual word form area.CC BY · PubMed Central
Fig 4 Brain activations correlating with reading speed as it evolves across scanning sessions, independently of age. (A) Effect of reading speed (LUM) on word-evoked activations relative to rest, in a model in which age was also entered as a covariable. (B) Reading speed effect on the activation evoked by faces, words, and numbers. Activation to words (green) and numbers (blue) increased with reading speed in the left occipitotemporal pathway. Reading expertise was also associated with increased responses to faces in the right fusiform gyrus (red). No significant effect was seen for the other categories. All contrasts at voxelwise p < 0.001 and clusterwise p < 0.05 FWE corrected. FWE, family-wise error; LUM, “Lecture en une minute.”
Fig 4 Brain activations correlating with reading speed as it evolves across scanning sessions, independently of age. (A) Effect of reading speed (LUM) on word-evoked activations relative to rest, in a model in which age was also entered as a covariable. (B) Reading speed effect on the activation evoked by faces, words, and numbers. Activation to words (green) and numbers (blue) increased with reading speed in the left occipitotemporal pathway. Reading expertise was also associated with increased responses to faces in the right fusiform gyrus (red). No significant effect was seen for the other categories. All contrasts at voxelwise p < 0.001 and clusterwise p < 0.05 FWE corrected. FWE, family-wise error; LUM, “Lecture en une minute.”CC BY · PubMed Central
Fig 5 Evolution of the volume of activity evoked by the different categories in ventral visual cortex. (A) Slices showing the analyzed ventral visual region (“ventral mask”). (B) Percentage of voxels showing activity specific to at least 1 visual category (i.e., voxels whose activity was significant at p < .001 in the contrast of one category relative to all others). The active volume was expressed, in each child, as a proportion of ventral visual volume (i.e., the intersection of each child’s gray matter mask and the left and right ventral masks). (C) total volume (in mm3) of voxels showing significantly more activity for a given category relative to all others (voxelwise p < 0.001, uncorrected). For reference, the search regions were 36,080 mm3 (left) and 35,896 mm3 (right), and the volume expected by chance would therefore be approximately 36 mm3. S2 Data.
Fig 5 Evolution of the volume of activity evoked by the different categories in ventral visual cortex. (A) Slices showing the analyzed ventral visual region (“ventral mask”). (B) Percentage of voxels showing activity specific to at least 1 visual category (i.e., voxels whose activity was significant at p < .001 in the contrast of one category relative to all others). The active volume was expressed, in each child, as a proportion of ventral visual volume (i.e., the intersection of each child’s gray matter mask and the left and right ventral masks). (C) total volume (in mm3) of voxels showing significantly more activity for a given category relative to all others (voxelwise p < 0.001, uncorrected). For reference, the search regions were 36,080 mm3 (left) and 35,896 mm3 (right), and the volume expected by chance would therefore be approximately 36 mm3. S2 Data.CC BY · PubMed Central
Fig 6 Evolution of responses in left-hemispheric voxels that ultimately specialize for a given category. For each child, we subdivided left ventral temporal voxels into the ROI formed by those selective for a given category during the final sessions 6 and 7 (category > others, p < 0.001) and the remaining voxels. We then went back in time and asked how these 2 sets of voxels differed in their responsivity to the various categories of stimuli in the preceding sessions 1–5. Each plot shows the difference in activation evoked by a given stimulus (e.g., words) relative to other categories (excluding words), in voxels within and outside the specified ROI. For instance, the green curve at left indicates that word-selective voxels in sessions 6 and 7 showed no specificity for words during sessions 1 and 2 (values not different from 0) but became selective in sessions 3, 4, and 5 (positive values). Therefore, there is a clear emergence of selectivity for words in RR voxels. The fact that other (nongreen) curves in this panel are indistinguishable from 0 indicates that the VWFA voxels cannot be anticipated by their responsivity to other categories. By contrast, voxels selective to tools, houses, faces, or bodies in sessions 6 and 7 already exhibit their category specificity in the very first sessions. S3 Data. ROI, region of interest; RR, reading-related; VWFA, visual word form area.
Fig 6 Evolution of responses in left-hemispheric voxels that ultimately specialize for a given category. For each child, we subdivided left ventral temporal voxels into the ROI formed by those selective for a given category during the final sessions 6 and 7 (category > others, p < 0.001) and the remaining voxels. We then went back in time and asked how these 2 sets of voxels differed in their responsivity to the various categories of stimuli in the preceding sessions 1–5. Each plot shows the difference in activation evoked by a given stimulus (e.g., words) relative to other categories (excluding words), in voxels within and outside the specified ROI. For instance, the green curve at left indicates that word-selective voxels in sessions 6 and 7 showed no specificity for words during sessions 1 and 2 (values not different from 0) but became selective in sessions 3, 4, and 5 (positive values). Therefore, there is a clear emergence of selectivity for words in RR voxels. The fact that other (nongreen) curves in this panel are indistinguishable from 0 indicates that the VWFA voxels cannot be anticipated by their responsivity to other categories. By contrast, voxels selective to tools, houses, faces, or bodies in sessions 6 and 7 already exhibit their category specificity in the very first sessions. S3 Data. ROI, region of interest; RR, reading-related; VWFA, visual word form area.CC BY · PubMed Central
Fig 7 Representation similarity analysis. In each graph, the thick line shows the within-category correlation coefficient r between 2 independently measured activation patterns, measuring the reproducibly of activation patterns in a given area, and the dotted line shows the between-category baseline (r was measured within each subject, then averaged; bar = ±1 standard error). A significant difference between those curves indicates a reliable category-specific activation. (A) When averaging over sessions, reliable patterns were present for all categories in the left ventral mask. (B) Temporal evolution across sessions of the r coefficient for words and for the average of nonsymbolic images (tools, bodies, faces, and houses) in the left ventral mask. (C) and (D), same as (A) and (B), restricted to the visual word form area. (E) and (F), similarity of the word-induced activation pattern in each ROI selective for a given category relative to others, respectively within the left and right ventral masks. All comparisons within versus between category are significant, except those indicated as ns. Underlying data found in S4 Data. FFA, fusiform face area; ns, nonsignificant; ROI, region of interest; VWFA, visual word form area.
Fig 7 Representation similarity analysis. In each graph, the thick line shows the within-category correlation coefficient r between 2 independently measured activation patterns, measuring the reproducibly of activation patterns in a given area, and the dotted line shows the between-category baseline (r was measured within each subject, then averaged; bar = ±1 standard error). A significant difference between those curves indicates a reliable category-specific activation. (A) When averaging over sessions, reliable patterns were present for all categories in the left ventral mask. (B) Temporal evolution across sessions of the r coefficient for words and for the average of nonsymbolic images (tools, bodies, faces, and houses) in the left ventral mask. (C) and (D), same as (A) and (B), restricted to the visual word form area. (E) and (F), similarity of the word-induced activation pattern in each ROI selective for a given category relative to others, respectively within the left and right ventral masks. All comparisons within versus between category are significant, except those indicated as ns. Underlying data found in S4 Data. FFA, fusiform face area; ns, nonsignificant; ROI, region of interest; VWFA, visual word form area.CC BY · PubMed Central
Fig 8 Schematic model of ventral visual development. Hexagons depict cortical patches or columns specialized for a given category (tools, faces, houses) or in an uncommitted labile state. Each group of hexagons schematizes the state of the left-hemispheric ventral visual mosaic at a given age. Prior to schooling (left), some columns are already committed to a given category, with a systematic lateral-to-mesial organization (left-to-right), but many are still labile. When schooling starts, some columns commit to written words (top right). In the absence of schooling, the same columns are progressively invaded by nearby representations of tools and faces (bottom right). The dashed yellow box illustrates how a single fMRI voxel may comprise a mixture of cortical responses to tools and words. The model can explain why, (1) in literate children, VWFA voxels become selective to words while maintaining their prior response to tools; (2) literacy blocks the expansion of face responses in the left hemisphere, restricting their growth to the right hemisphere; and (3) in illiterate subjects, relative to literate ones, face responses are larger in the left hemisphere and thus less asymmetrical in favor of the right hemisphere. F, faces; fMRI, functional magnetic resonance imaging; H, houses; L, labile; T, tools; VWFA, visual word form area.
Fig 8 Schematic model of ventral visual development. Hexagons depict cortical patches or columns specialized for a given category (tools, faces, houses) or in an uncommitted labile state. Each group of hexagons schematizes the state of the left-hemispheric ventral visual mosaic at a given age. Prior to schooling (left), some columns are already committed to a given category, with a systematic lateral-to-mesial organization (left-to-right), but many are still labile. When schooling starts, some columns commit to written words (top right). In the absence of schooling, the same columns are progressively invaded by nearby representations of tools and faces (bottom right). The dashed yellow box illustrates how a single fMRI voxel may comprise a mixture of cortical responses to tools and words. The model can explain why, (1) in literate children, VWFA voxels become selective to words while maintaining their prior response to tools; (2) literacy blocks the expansion of face responses in the left hemisphere, restricting their growth to the right hemisphere; and (3) in illiterate subjects, relative to literate ones, face responses are larger in the left hemisphere and thus less asymmetrical in favor of the right hemisphere. F, faces; fMRI, functional magnetic resonance imaging; H, houses; L, labile; T, tools; VWFA, visual word form area.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 105 references

Cited by

Items 301–400 of 412

Metadata

Authors
Dehaene-Lambertz, G., Monzalvo, K., & Dehaene, S.
Affiliations
Inserm (FR)
Venue
PLOS Biology
Published
2018-03-06
Volume
16
Issue
3
Pages
e2004103
ISSN
  • 1544-9173 (ISSN-L)
  • 1545-7885
OpenAlex
W2793866660
Cited by
367
References
105
Citations by year
2026: 39 · 2025: 55 · 2024: 50 · 2023: 32 · 2022: 58 · 2021: 56 · 2020: 38 · 2019: 29 · 2018: 7
FWCI
63.80
Type
journal article
Language
en
Open access
Locations
  • PLoS Biology (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • HAL (submittedVersion, other-oa)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Figshare (submittedVersion, cc-by)
Topics
  • Social Sciences › Psychology › Developmental and Educational Psychology › Reading and Literacy Development
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Face Recognition and Perception
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Visual perception and processing mechanisms
Keywords
visual word form area, reading acquisition, literacy, cortical specialization, ventral visual cortex, face processing, reading speed, children, longitudinal brain development, category selectivity
MeSH terms
Brain Mapping, Child, Humans, Language, Learning, Longitudinal Studies, Magnetic Resonance Imaging, Pattern Recognition, Visual, Photic Stimulation, Visual Cortex

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.