The paradox of the Visual Word Form Area

 

Luke the Evangelist (source: British Library). In the past, only a minority could read long texts of cursive writing. But many more could read short texts of block writing.

 

The Visual Word Form Area (VWFA) is a specialized part of the brain that helps us recognize written words and letters. If it is subjected to a surgical lesion, the patient will suffer a clear impairment to reading ability but not to recognition of objects, names, or faces or to general language abilities. There will be some improvement over the next six months, but reading will still take twice as long as it had before surgery (Gaillard et al, 2006).

Most of the initial skepticism over the existence of the VWFA has disappeared. There does seem to be, however, much variability in its size. An area that may fall within this mental organ in one person may fall outside it in someone else (Glezer and Riesenhuber, 2013).

In addition to word recognition, the VWFA may participate in higher-level processing of word meaning:

[It seems that] the VWFA would not only be recruited at an early stage for allowing low-level (script processing) word processing as has been previously instantiated (Pammer et al., 2004; Dehaene and Cohen, 2011), but also at a later stage for gating high-level (lexico-semantic) processing. Such late semantic gateway would not be selective to the VWFA but rather emerge in the posterior LOT and extend anteriorly to the VWFA. (Levy et al., 2013)

The VWFA is described in the above study as a “bottleneck to consciousness.” It helps us not only to recognize words on a page but also to understand what the words mean. To me, this makes sense. I’m better at thinking through an idea and its implications if I can write it down and then read it. There thus seems to be a single mental pathway that does double duty: processing character strings (words) and processing higher-level concepts.

 

Population differences 

The VWFA functions differently in different human populations. The difference is striking between people who use alphabetical script, where each symbol represents a sound, and those who use logographic script, where each symbol represents an idea. Chinese subjects process their idea-based symbols with assistance from other brain regions, whereas Westerners process their sound-based symbols only in the VWFA (Liu et al., 2008). Similarly, dyslexics activate this brain region in ways that differ by linguistic background, apparently because of differences in spelling and writing (Paulesu et al., 2001).

Hardwired or softwired?

For Dehaene and Cohen (2011), the VWFA is not a hardwired mental organ. They argue that it occupies the same area of the brain because that is where we can most easily recruit neurons when learning to recognize words. But why, then, does this recruitment happen so fast in young children? When kindergarten children were asked to play a grapheme/phoneme correspondence game, their VWFAs preferentially responded to pictures of letter strings after a total of 3.6 hours over an 8-week period. It is worth noting that only a few of these children could actually read, and even then only at a rudimentary level (Brem et al., 2010; Dehaene et al., 2010).

But the alternative view, hardwiring, is also hard to accept. Reading began not in the Paleolithic but in historic times, less than 6,000 years ago. Widespread literacy is even more recent, and there are still many societies where most people cannot read or write. How could an entirely new mental organ have evolved over so short a time?

Yet this alternative view may not be so farfetched. Let’s examine the two main objections.

Was there not enough time for natural selection to work?

The VWFA did not evolve out of nothing. It seems to be a population of neurons that originally served to recognize faces (Dehaene and Cohen, 2011). This sort of recycling is a common pathway for natural selection and explains much of the apparent rapidity of evolution. A complex mental adaptation may take a long time to evolve, but much less time is needed to develop an exaggerated version of it or to alter when and how it becomes activated (Harpending and Cochran, 2002).

Indeed, parallel to the way alphabetical reading ability has spread historically and geographically, there is a similar spread of the latest variant of ASPM, a gene implicated in the regulation of brain growth. In humans, a new variant arose about 6,000 years ago in the Middle East. It eventually became more prevalent in the Middle East (37-52% incidence) and Europe (38-50%) than in East Asia (0-25%) (Frost, 2011; Mekel-Bobrov et al., 2005). 

Would it have benefited too few people to have been favored by natural selection?

There is some debate over the relative recentness of literacy. It is true that before the modern era only a small minority could read long texts of cursive writing. But the ability to read short texts of block writing was much more widespread, as evidenced by the prevalence of graffiti and storefront signs. We should also keep in mind that the literate few contributed disproportionately to the gene pool of subsequent generations. Clark (2007) has shown that the English lower class is largely descended from people who were middle or upper class a few centuries ago. In the ancient world, there was a perception that scribes enjoyed reproductive success. The Book of Sirach [39: 11] states: “If [a scribe] lives long, he will leave a name greater than a thousand.” 

Gene-culture co-evolution?

There may have been positive feedback between reading ability and the cultural opportunities it created. One example is the scientific revolution in Western Europe (15th - 18th centuries), which took off once a critical mass of scholars could read each other’s papers. In short, reading and writing are advantageous to the extent that other people can read and write. While this kind of feedback loop is self-evident, its biological implications may be less so. The same feedback loop would have steadily ratcheted up selection for the VWFA and, subsequently, for higher-level faculties. This might explain why the VWFA evolved beyond word recognition per se and towards lexico-semantic tasks.

Future research

One priority would be to study the VWFA in populations that have become literate only in recent times. What form, if any, does it take in such people? A study in New York elementary schools found that VWFA activation varied with socioeconomic status. In students from high SES families, activation seemed to be more hardwired and less dependent on familiarity with the way sounds are visually represented. Unfortunately, there was no attempt to break the data down by ethnic background (Noble et al., 2006).

At present, high VWFA activation is attributed to an environment where reading material is accessible and parents very supportive, this being in turn attributed to high SES. Yet reading material is ubiquitous nowadays. And how crucial is parental support? As a child, I read almost always on my own with little encouragement at home or school. My teachers were in fact annoyed by my habit of sneaking into the small storage room where old textbooks and encyclopedias were kept (we had no library). “If you’ve finished your assignment, stay at your desk. Is that clear?!”

Nonetheless, I read voraciously, even when I couldn’t understand half of what I read. Strange new words were a source of pleasure, and I would often read and reread the same texts simply because I liked the flow of the words and the images they conjured up.

 

References 

Brem, S., S. Bach, K. Kucian, T.K. Guttorm, E. Martin, H. Lyytinen, D. Brandeis, and U. Richardson. (2010). Brain sensitivity to print emerges when children learn letter-speech sound correspondences, Proceedings of the National Academy of Sciences U.S.A., 107, 7939–7944.

http://psyserv06.psy.sbg.ac.at:5916/fetch/PDF/20395549.pdf

Clark, G. (2007). A Farewell to Alms. A Brief Economic History of the World, Princeton University Press, Princeton and Oxford.

Dehaene, S. and L. Cohen. (2011). The unique role of the visual word form area in reading, Trends in Cognitive Sciences, 15, 254-262.

http://www.cnbc.pitt.edu/~plaut/VisCog/papers/DehaeneCohen11TICS.VWFA.pdf  

Dehaene, S. et al. (2010). How learning to read changes the cortical networks for vision and language, Science, 330, 1359–1364.

http://gondabrain.ls.biu.ac.il/Neuroling/courses/877/Dehaene_Science2010.pdf

Frost, P. (2011). Human nature or human natures? Futures, 43, 740-748.

http://dx.doi.org/10.1016/j.futures.2011.05.017  

Gaillard, R., Naccache, L., P. Pinel, S. Clémenceau, E. Volle, D. Hasboun, S. Dupont, M. Baulac, S. Dehaene, C. Adam, and L. Cohen. (2006). Direct intracranial, fMRI, and lesion evidence for the causal role of left inferotemporal cortex in reading, Neuron, 50, 191-204.

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.76.7620&rep=rep1&type=pdf

 

Glezer, L.S. and M. Riesenhuber. (2013). Individual variability in location impacts orthographic selectivity in the “Visual Word Form Area”, The Journal of Neuroscience, 33(27), 11221–11226.

http://www.jneurosci.org/content/33/27/11221.full  

Harpending, H., and G. Cochran. (2002). In our genes, Proceedings of the National Academy of Sciences USA, 99(1), 10-12.

http://www.wcas.northwestern.edu/nescan/2008-2009%20papers/harpending%20-%20in%20our%20genes.pdf  

Levy, J., J.R Vidal, R. Oostenveld, I. FitzPatrick, J-F. Démonet, and P. Fries. (2013). Alpha-band suppression in the Visual Word Form Area as a functional bottleneck to consciousness, NeuroImage,78C, 33-45.

http://hal.inria.fr/docs/00/81/96/67/PDF/Levy_et_al.pdf  

Liu, C., W-T. Zhang, Y-Y Tang, X-Q. Mai, H-C. Chen, T. Tardif, and Y-J. Luo. (2008). The visual word form area: evidence from an fMRI study of implicit processing of Chinese characters, NeuroImage, 40, 1350-1361.

http://psychbrain.bnu.edu.cn/teachcms/res_base/teachcms/upload/channel/file/2010_4/11_25/6hlcggx7rk3z.pdf  
Mekel-Bobrov, N., S.L. Gilbert, P.D. Evans, E.J. Vallender, J.R. Anderson, R.R. Hudson, S.A. Tishkoff, and B.T. Lahn. (2005). Ongoing adaptive evolution of ASPM, a brain size determinant in Homo sapiens, Science, 309, 1720-1722.
http://ftp.eebweb.arizona.edu/faculty/nachman/Archived%20Research%20Papers/mekel_bobrov_et_al_2005.pdf  

Noble, K.G., M.E. Wolmetz, L.G. Ochs, M.J. Farah, and B.D. McCandliss. (2006). Brain–behavior relationships in reading acquisition are modulated by socioeconomic factors, Developmental Science, 9, 642–654.

http://www.cumc.columbia.edu/dept/sergievsky/fs/publications/Noble-et-al-2006-2.pdf  

Paulesu E., J.F. Démonet, F. Fazio, E. McCrory, V. Chanoine, N. Brunswick et al (2001). Dyslexia: cultural diversity and biological unity, Science, 291, 2165–2167.

http://www.drru-research.org/data/resources/42/Paulesu-et-al-2001.pdf

The Visual Word Form Area - part II

Dr. Kimberly G. Noble has studied the Visual Word Form Area of children in New York City schools (source). The VWFA seems more hardwired in higher SES children.

The Visual Word Form Area (VWFA) is a brain region that specifically recognizes written words. It seems to be composed of neurons that were originally used to recognize human faces. Though not essential for reading, it greatly speeds up this mental task (Gaillard et al, 2006; see earlier post).

Is the VWFA a product of gene-culture co-evolution? Did natural selection favor the reproductive success of individuals who were better able to recognize words? Or does this brain area develop independently in each individual through experience with reading?

The second explanation is the one now favored. Dehaene and Cohen (2011) argue that the VWFA is where the brain can most easily recruit neurons for the task of recognizing words. One problem with this developmental explanation is that the VWFA responds preferentially to images of letter strings even in kindergarten children who haven’t learned to read yet (Brem et al., 2010).

We could answer this question one way or the other by comparing populations that have a long history of reading with those that were illiterate until relatively recent times. Does the second type of population exhibit a less developed and less specialized VWFA even in individuals who learned to read at an early age?

Only two VWFA studies have dealt with the second type of population. In these two cases, some of the subjects were of sub-Saharan African descent. We should nonetheless remember that some sub-Saharan African societies, notably those of the Sahel, have a history of reading and writing that goes back over seven hundred years.

In the first of the two studies, Noble et al. (2006) examined brain responses to reading tasks that were performed by native English-speakers in New York City elementary schools. The students differed by socioeconomic status (SES) and by ethnicity. Eleven were African-American, five Latino, one Asian, fourteen White, and seven mixed or other. Brain activation varied as a function of SES. Among lower SES children, VWFA activation was much stronger in those who were “phonologically aware”, i.e., who explicitly knew how to represent and manipulate the sounds of language. But this pattern was absent in higher SES children:

In contrast, as the SES of the population increases, children demonstrating a similar range of phonological skill show an attenuated brain–behavior relationship in this region. This suggests that, among children who are likely to have adequate access to literacy resources, the relationship between reading precursor skills and left fusiform activity to reading may, to an extent, be reduced, marking an atypical relationship between cognitive skill and brain activity. A marginally significant PA × SES interaction was also observed in the left superior temporal region, demonstrating a similar trend. (Noble et al., 2006)

The authors attributed this non-correlation to “adequate access to literacy resources.” An alternate explanation would be that VWFA activation was more hardwired in the higher SES children. Strangely enough, the authors did not break down their data by race, so it is impossible to say whether SES was simply a proxy for ethnic background.

The second study was by Dehaene et al. (2010) on literate, illiterate, and ex-illiterate adults from Brazil and Portugal. No information is given on ethnicity, although many of the illiterate or ex-illiterate Brazilians were probably of African or part-African ancestry. The authors found that VWFA activation was much less apparent in adult illiterates than in adult literates, even when the data were controlled for SES and schooling.

Again, there is no breakdown of the data by ethnicity, although one might assume that SES and schooling were proxies for ethnicity. This is a flawed assumption, however, at least in Brazil:

Still, the patterns of racial in Brazilian education have remained and have transcended social class barriers. Nelson do Valle Silva and Carlos Hasenbalg have demonstrated that patterns of educated attainment remain unequal even when social class is eliminated as a factor: whites of the same social class have higher literacy rates and remain more likely to attend school, to stay in school longer, to be advanced through school more rapidly, and to secure better-paying jobs given the same educational qualifications. Silva and Hasenbalg conclude that “white children’s rates of school advancement are significantly more rapid than those of pardo [mixed] and preto [black] children. These differences result in profound educational inequalities that separate whites and nonwhites in Brazilian society.” (Davila, 2003, p. 8)

Conclusion

On the basis of these two studies, it is impossible to say whether the VWFA is hardwired or softwired. This brain area may result solely from developmental processes within the lifetime of each individual. Or it may be due to longer-term evolutionary processes.

A common problem is that both studies use SES to the exclusion of ethnicity. Yes, ethnic differences may simply reflect SES differences, but that arrow of causality should be proven and not assumed. In any case, SES varies imperfectly with ethnicity. With respect to the Dehaene etal. (2010) study, black Brazilians tend to be more illiterate than white Brazilians even among people of similar SES. With respect to the Noble et al. (2006) study, differences in phonological skill might likewise reflect ethnic differences, even if we consider only the lower SES children.

Why did both research teams ignore ethnicity? One reason, at least in the case of Dehaene’s team, is a belief that mental traits take eons to evolve. This might be true if the trait is radically new and different, but here the transition from face recognition to letter recognition is relatively simple. This is the kind of evolution that could happen over a few centuries, if the selection pressure were strong enough.

The other reason is a belief that ethnicity is genetically irrelevant, since genes vary much more within than between human populations. This fact is well known and beyond dispute. What is less well known is that the same pattern often appears when we examine the way genes vary within and between sibling species—even when such species are morphologically and behaviorally distinct. We should understand that we’re comparing apples with oranges when genetic variation within populations is compared with genetic variation between populations. Different populations typically occupy different environments with different selection pressures. Variation across a population boundary is thus more likely to involve genes that have real adaptive value. In contrast, variation within a population tends to involve genes of low adaptive value that are insensitive to the homogenizing action of similar selection pressures (Frost, 2011).

References

Brem, S., S. Bach, K. Kucian, T.K. Guttorm, E. Martin, H. Lyytinen, D. Brandeis, & U. Richardson. (2010). Brain sensitivity to print emerges when children learn letter-speech sound correspondences, Proceedings of the National Academy of Sciences U.S.A., 107, 7939–7944.

Davila, J. (2003). Diploma of Whiteness. Race and Social Policy in Brazil, 1917-1945, Duke University Press.

Dehaene, S. & L. Cohen. (2011). The unique role of the visual word form area in reading, Trends in Cognitive Sciences, 15, 254-262.

Dehaene, S., F. Pegado, L.W. Braga, P. Ventura, G.N. Filho, A. Jobert, G. Dehaene-Lambertz, R. Kolinsky, J. Morais, & L. Cohen. (2010). How Learning to Read Changes the Cortical Networks for Vision and Language, Science, 330, 1359-1364

https://hpc.hamilton.edu/~lablab/Dehaene_2010.pdf

Frost, P. (2011). Human nature or human natures? Futures, 43, 740–748.

Gaillard, R., Naccache, L., P. Pinel, S. Clémenceau, E. Volle, D. Hasboun, S. Dupont, M. Baulac, S. Dehaene, C. Adam, & L. Cohen. (2006). Direct intracranial, fMRI, and lesion evidence for the causal role of left inferotemporal cortex in reading, Neuron, 50, 191-204.

Hasenbalg, C.A., & N.V. Silva. (1990). Raça e oportunidades educacionais no Brasil, Cadernos de Pesquisa (Sao Paulo), 73, 5-12

Noble, K.G., M.E. Wolmetz, L.G. Ochs, M.J. Farah, & B.D. McCandliss. (2006). Brain–behavior relationships in reading acquisition are modulated by socioeconomic factors, Developmental Science, 9, 642–654.

http://www.cumc.columbia.edu/dept/sergievsky/fs/publications/Noble-et-al-2006-2.pdf