The Visual Word Form Area – a brain region that coevolved with reading and writing

A case of recent human evolution

The Scribe, Ludwig Deutsch (1894)

In ancient times, demand was strong for scribes who could write and copy texts day in and day out. Only a small minority had the stamina and ability, and they enjoyed reproductive success.

The Visual Word Form Area (VWFA) is a brain region that helps us recognize written words and letters. Without it, reading requires much more effort. When a man suffered an accidental lesion to his VWFA during brain surgery, he lost much of his ability to read while losing none of his general language abilities. After six months, he had partially recovered, but reading still took twice as long as it had before (Gaillard et al, 2006).

The VWFA is composed of neurons that were once used for face recognition:

Thus, learning to read must involve a ‘neuronal recycling’ process whereby pre-existing cortical systems are harnessed for the novel task of recognizing written words. … [Such areas of the cortex] possess the appropriate receptive fields to recognize the small contrasted shapes that are used as characters, and the appropriate connections to send this information to temporal lobe language areas (Dehaene & Cohen, 2011)

This neuronal recycling seems to have become hardwired, at least in some people. After Swiss preschoolers played a grapheme/phoneme correspondence game for a total of 3.6 hours over 8 weeks, an MRI scan showed their VWFAs preferentially responding to images of strings of letters. Yet only a few of the children could actually read, and only at a rudimentary level (Brem et al., 2010).

Humans may have initially identified words by using face-recognition neurons. As reading became more important, natural selection favored those humans who could free up more of their face-recognition neurons for reading. This selection eventually created a large neuronal population dedicated solely to word recognition, i.e., the VWFA.

This evolutionary process may be broken down into four stages:

1. Reading without a VWFA. At this stage, reading took twice as long as it now does with a VWFA (see the case study above). Reading was particularly arduous in early times because texts were continuous streams of characters with little or no punctuation.

2. Demand for people who could read and write. With the rise of the state and the development of trade, demand grew for people who could read and write for long periods without getting tired. Known as ‘scribes,’ they had to record transactions, decisions, and edicts for government and business. Many further specialized as stenographers, copyists, secretaries, notaries, calligraphers, and so on. There was thus selection for people who had 1) face-recognition neurons that could better handle word recognition and 2) more and more of such specialized neurons — eventually a fully formed VWFA.

3. Expansion of the scribal class. Demand for scribes was driven by two positive feedback loops. First, as the number of texts grew, so did the number of old or damaged ones to be copied and updated. Second, as scribes themselves grew in number, they became a significant source of supply and demand. In addition to creating texts for their patrons, they became producers and consumers in their own right. Reading and writing went beyond record-keeping: it developed into an intellectual activity oriented toward fiction, history, and philosophy.

4. Spread of reading and writing through the general population. Scribes turned their economic success into reproductive success, thereby becoming more numerous. Many left the scribal class and moved into the general population through upward, lateral, and downward social mobility. Propensities for reading and writing thus became more prevalent in the gene pool, enabling more people to read and write with less effort. The culture changed. As more information became available on “hard copy,” people could draw on a larger store of knowledge — not only from the living but also from the dead.

Greek script from the Rosetta Stone (Wikicommons: Gary Todd). Originally, texts were written with no breaks between words. Few people could read and write on a sustained basis.

The above model may seem hard to believe. How could natural selection create a new mental organ in such a short time? “The invention of writing is too recent and, until the last century, concerned too small a fraction of humanity to have influenced the human genome” (Dehaene & Cohen, 2011). Writing emerged in the Middle East about six thousand years ago and was adopted by some societies only within the past century. As for those societies with a long literary tradition, the ability to write was generally confined to an educated minority until recent times.

But such rapid evolution is not as paradoxical as it may seem:

First, the VWFA did not evolve from scratch. A new task had to be performed by an existing population of neurons, which at first did it clumsily but then evolved to do it better. Evolutionary change has often occurred through such repurposing of existing mechanisms (Harpending & Cochran, 2002).

Second, selection can change an entire population even if it acts only within a small subgroup. If a subgroup is more reproductively successful than the rest of the population, it will become proportionately larger, and its genetic characteristics will become more prevalent. This has been shown by Gregory Clark’s work on the differential reproduction of social classes in England. As late as the 19th century, the middle class enjoyed a high rate of natural increase, which caused it to grow until its lineages dominated the country’s gene pool. Meanwhile, the lower class failed to replace itself demographically and was continually replenished by downwardly mobile individuals. The entire English population thus became shaped by selection pressures that were confined initially to a tiny subgroup (Clark, 2007). In the ancient world, scribes similarly enjoyed higher reproductive success, as evidenced by the Book of Sirach [39: 11]: “If [a scribe] lives long, he will leave a name greater than a thousand.” In other words, his descendants will be numerous.

Third, although only a minority were fully literate, many more were partially literate. The literacy rate of past societies is estimated from signed documents of one sort or another: wills, court depositions, marriage certificates, and so on (Barr & Kamil, 1996, p. 52). A signatory who simply wrote an ‘X’ is deemed to have been illiterate. Such people undoubtedly had trouble writing, especially in cursive script, but they may have been able to read texts of block letters. This view is supported by the discovery of ancient graffiti, such as those at Pompeii, that seem to have been written by ordinary people, including common prostitutes. Current historical methods thus underestimate the number of people who had at least some reading and writing ability.

In this, and in many other ways, humans have directed their evolution. First, they performed a new task by leveraging their phenotypic plasticity. This new way of doing things then acted as a template for natural selection. Gradually, over generations, the new phenotype became a new genotype.

Does the ability to read and write vary within our species?

If humans evolved to become better at reading and writing, this evolution may have happened to different extents in different populations. In particular, the VWFA should be larger and more specialized in those populations that have been literate or partially literate for longer.

Differential selection for reading ability is indicated by two lines of evidence: 1) the geographic distribution of the latest ASPM variant; and 2) the differences among populations in the degree of competition between face recognition and word recognition.

1. Geographic distribution of the latest ASPM variant

ASPM is a gene that helps regulate primate brain growth (Wu et al., 2023). In humans, a new variant arose some six thousand years ago in the Middle East, where it is still most prevalent (37-52%). It then spread outward — more so into Europe (38-50%) than into East Asia (0-25%) (Mekel-Bobrov et al., 2005). This variant thus seems to track the spread of writing within Eurasia, particularly the spread of alphabetical script from an origin in Mesopotamia around 5,400 to 5,100 years ago (Frost, 2011).

When first discovered, this geographic pattern was thought to be concrete evidence of recent cognitive evolution. Nonetheless, two research teams failed to find any association between IQ and the latest ASPM variant (Mekel-Bobrov et al. 2007; Rushton et al. 2007). Philippe Rushton argued against pursuing further research, on the assumption that IQ encompasses all forms of mental effort. In an email he told me: "Generally there isn't thought to be much left to be explained after g is taken out."

Researchers have since found that IQ does not correlate with face recognition; the two seem to function independently of each other (Zhu et al., 2010). The same should be true for IQ and word recognition, given that we recognize words by using neurons that evolved out of face-recognition neurons.

So the question is still open. We should pursue this abandoned line of research to see whether the presence or absence of the latest ASPM variant correlates with performance on word recognition, particularly words based on an alphabet. Indeed, the geographic pattern seems to mirror the historical spread of alphabetical script, which puts much more demand on the VWFA than does ideographic script. Whereas Westerners process their sound-based characters only in the VWFA, the Chinese process their idea-based characters with assistance from other brain regions (Liu et al., 2008).

Worldwide distribution of the latest ASPM variant. It seems to be most frequent in populations with a longer history of reading and writing, particularly if the writing system uses sound-based characters (Mekel-Bobrov et al., 2007)

2. Population differences in competition between word recognition and face recognition

If the VWFA is a product of recent human evolution, it may be more developed in some populations than in others. The more time the VWFA has had to evolve, the more easily one should recognize words without loss of face recognition. The less time the VWFA has had to evolve, the more these two tasks should compete against each other for access to the same limited number of neurons.

Thus, in a population that largely began to read and write in recent times, illiterates should suffer some loss of face recognition if they learn to read. Conversely, illiterates should suffer no such loss in a population that has long had reading and writing.

In both types of population, there have been MRI studies of competition between word recognition and face recognition.

First, let’s consider populations with a short history of reading and writing. According to four MRI studies, learning to read incurred some loss of face recognition. The participants were wholly or mostly Brazilian:

• Dehaene et al., 2010 - Portugal/Brazil, illiterate or ex-illiterate adults, n=63. Among unschooled adults, ex-illiterates had more trouble than illiterates in recognizing faces. Word recognition seemed to compete with face recognition on the periphery of the VWFA.

• Pegado et al., 2014 - Portugal/Brazil, 49 illiterates and ex-illiterates, n=49 (mostly from the previous study). Literacy seemed to “push” face recognition away from the VWFA and toward the right hemisphere.

• Braga et al., 2017 - Brazil, illiterate man who had never gone to school, n=1. Faces and words seemed to compete for processing within the VWFA. The authors concluded “that faces and written words occupy distinct but neighboring locations in the left occipitotemporal cortex, and that the growth of word-related responses induces a shift in the boundary between these two preferences and thus a relative decrease in the left-hemispheric response to faces.”

• de Andrade & França, 2020 - Brazil (Rio de Janeiro), children, 4 to 8 years old, n=30.

  Learning to read and write helped participants process speech but adversely affected their capacity to recognize either new faces or old faces that had undergone alterations to the eyes, nose, and mouth. This adverse impact was still present among the oldest students in the second grade.

Now let’s consider populations with a long history of reading and writing. In two studies, learning to read incurred no loss of face recognition. Participants were either South Asian or Danish.

• Hervais-Adelman et al., 2019 - India (near Lucknow), literate and illiterate adults, n=91. The authors concluded “that far from cannibalizing the territory of its neighbors, the VWFA is rather overlaid upon these and not only remains responsive to other categories of visual stimuli but also recruits an assembly of neural populations in much the same way, regardless of whether the input is orthographic, albeit with a lesser overall response magnitude.”

• Kühn et al., 2021 - Denmark (Copenhagen), children 5-7 years old, n=82. The authors found no loss of face recognition, in general, among schoolchildren learning to read, although a quarter of them did show some loss.

In summary, the VWFA seems to be larger and more specialized in populations with a longer history of full or partial literacy, such as Danes and South Asians. Conversely, it seems to be smaller and less specialized in populations with a shorter history of full or partial literacy, such as Brazilians — the majority of whom are at least partly African or Amerindian by descent.

The VWFA also seems to vary within populations. Although Danes have a relatively long history of full or partial literacy, about a quarter of all schoolchildren had some VWFA inadequacy in the Danish study. This subgroup was likely of Danish descent, according to a personal communication from one of the study’s authors.

Further evidence of population differences comes from a study of New York City elementary schoolchildren, which found differences in VWFA activation that correlated with socioeconomic status (SES). In students from high SES families, activation seemed to be more hardwired and less dependent on familiarity with the way sounds are visually represented:

In these [lower SES] children, phonological skill level is positively predictive of activation in the left fusiform region that supports rapid visual word recognition. … 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. (Noble et al., 2006).

The study’s authors reported the ethnic breakdown of their sample: approximately one third Euro-American, one third African American, and one third other groups. It would have thus been possible to see whether ethnicity was better than SES at explaining the activation differences. Unfortunately, this was not done.

Parting thoughts

Much confusion is still caused by the assumption that the ability to recognize words cannot be hardwired because reading and writing are too recent to have played any role in human evolution. This reasoning reflects a broader assumption that human evolution ended essentially in the Paleolithic. There is also the troubling inference that some humans — and some human populations — are inherently less capable of reading and writing. So we continue with a model where all children, everywhere, and at an unspecified early stage of development, recycle some of their face-recognition neurons to create their own VWFA. This model is untenable, yet the obvious alternative is not considered, as is the case with a recent review article (Rossion & Lochy, 2022).

To test the alternative model, we should first determine whether the ability to recognize words is associated with the latest ASPM variant. If an association is found, this model could then be applied to another region of the human brain, Exner’s area, which seems to control handwriting: if it is damaged, one loses the ability to write words and numbers by hand, while still being able to type them on a computer keyboard (Klein et al., 2016).

With Exner’s area, we again have an apparent case of recent evolution, specifically one where humans have coevolved with their cultural environment during the course of recorded history.

References 

Barr, R. & Kamil, M.L. (1996). Handbook of Reading Research vol. 2, Routledge.

Braga, L. W., Amemiya, E., Tauil, A., Suguieda, D., Lacerda, C., Klein, E., ... & Dehaene, S. (2017). Tracking adult literacy acquisition with functional MRI: A single-case study. Mind, Brain, and Education, 11(3), 121-132. https://doi.org/10.1111/mbe.12143

Brem, S., Bach, S., Kucian, K., Guttorm, T.K., Martin, E., Lyytinen, H., Brandeis, D., & Richardson, U. (2010). Brain sensitivity to print emerges when children learn letter-speech sound correspondences. Proceedings of the National Academy of Sciences U.S.A., 107(17), 7939–7944. https://doi.org/10.1073/pnas.0904402107

Clark, G. (2007). A Farewell to Alms. A Brief Economic History of the World, Princeton University Press, Princeton and Oxford. https://press.princeton.edu/books/paperback/9780691141282/a-farewell-to-alms

de Andrade, I. R., & França, A. I. (2020). A influência da alfabetização nas cognições de reconhecimento de faces e dos sinais da fala: um estudo psicolinguístico. The influence of literacy on face recognition and speech sign cognitions: a psycholinguistic study. Revista Prolíngua–ISSN, 1983(9979), 101.

Dehaene, S. & Cohen, L. (2011). The unique role of the visual word form area in reading, Trends in Cognitive Sciences, 15(6), 254-262. https://doi.org/10.1016/j.tics.2011.04.003

Dehaene, S., Pegado, F., Braga, L. W., Ventura, P., Filho, G. N., Jobert, A., ... & Cohen, L. (2010). How learning to read changes the cortical networks for vision and language. Science, 330(6009), 1359-1364. https://doi.org/10.1126/science.1194140

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., Pinel, P., Clémenceau, S., Volle, E., Hasboun, D., ... & Cohen, L. (2006). Direct intracranial, FMRI, and lesion evidence for the causal role of left inferotemporal cortex in reading. Neuron, 50(2), 191-204. https://doi.org/10.1016/j.neuron.2006.03.031

Harpending, H., & Cochran, G. (2002). In our genes. Proceedings of the National Academy of Sciences USA, 99(1), 10-12. https://doi.org/10.1073/pnas.012612799

Hervais-Adelman, A., Kumar, U., Mishra, R. K., Tripathi, V. N., Guleria, A., Singh, J. P., ... & Huettig, F. (2019). Learning to read recycles visual cortical networks without destruction. Science Advances, 5(9). https://doi.org/10.1126/sciadv.aax0262

Klein, E., Willmes, K., Jung, S., Huber, S., Braga, L. W., & Moeller, K. (2016). Differing connectivity of Exner’s area for numbers and letters. Frontiers in Human Neuroscience, 10, 281. https://doi.org/10.3389/fnhum.2016.00281

Kühn, C. D., Wilms, I. L., Dalrymple, K. A., Gerlach, C., & Starrfelt, R. (2021). Face recognition in beginning readers: Investigating the potential relationship between reading and face recognition during the first year of school. Visual Cognition, 29(4), 213-224. https://doi.org/10.1080/13506285.2021.1884151

Liu, C., Zhang, W. T., Tang, Y. Y., Mai, X. Q., Chen, H. C., Tardif, T., & Luo, Y. J. (2008). The visual word form area: evidence from an fMRI study of implicit processing of Chinese characters. Neuroimage, 40(3), 1350-1361. https://doi.org/10.1016/j.neuroimage.2007.10.014

Mekel-Bobrov, N., Gilbert, S. L., Evans, P. D., Vallender, E. J., Anderson, J. R., Hudson, R. R., ... & Lahn, B. T. (2005). Ongoing adaptive evolution of ASPM, a brain size determinant in Homo sapiens. Science, 309(5741), 1720-1722. https://doi.org/10.1126/science.1116815

Mekel-Bobrov, N., Posthuma, D., Gilbert, S. L., Lind, P., Gosso, M. F., Luciano, M., ... & Lahn, B. T. (2007). The ongoing adaptive evolution of ASPM and Microcephalin is not explained by increased intelligence. Human molecular genetics, 16(6), 600-608. https://doi.org/10.1093/hmg/ddl487

Noble, K. G., Wolmetz, M. E., Ochs, L. G., Farah, M. J., & McCandliss, B. D. (2006). Brain–behavior relationships in reading acquisition are modulated by socioeconomic factors. Developmental Science, 9(6), 642-654. https://doi.org/10.1111/j.1467-7687.2006.00542.x

Pegado, F., Comerlato, E., Ventura, F., Jobert, A., Nakamura, K., Buiatti, M., ... & Dehaene, S. (2014). Timing the impact of literacy on visual processing. Proceedings of the National Academy of Sciences, 111(49), E5233-E5242. https://doi.org/10.1073/pnas.1417347111

Rossion, B., & Lochy, A. (2022). Is human face recognition lateralized to the right hemisphere due to neural competition with left-lateralized visual word recognition? A critical review. Brain Structure and Function, 227(2), 599-629. https://doi.org/10.1007/s00429-021-02370-0

Rushton, J. P., Vernon, P. A., & Bons, T. A. (2007). No evidence that polymorphisms of brain regulator genes Microcephalin and ASPM are associated with general mental ability, head circumference or altruism. Biology Letters, 3(2), 157-160. https://doi.org/10.1098/rsbl.2006.0586

Wu, X., Li, Z., Wang, Z. Q., & Xu, X. (2023). The neurological and non-neurological roles of the primary microcephaly-associated protein ASPM. Frontiers in Neuroscience, 17, 1242448. https://doi.org/10.3389/fnins.2023.1242448

Zhu, Q., Song, Y., Hu, S., Li, X., Tian, M., Zhen, Z., ... & Liu, J. (2010). Heritability of the specific cognitive ability of face perception. Current Biology, 20(2), 137-142. https://doi.org/10.1016/j.cub.2009.11.067