New research is homing in on the mechanisms our brains use to process written language.
Given my profession, I'm quite happy that people can read and write. From an evolutionary standpoint, however, it's surprising that we do. There's no need for it in the wild, so our brains didn't need to develop specific areas to handle the task, like they did with sight or hearing.
A new study looked into which areas of the brain handle this task, finding that we use a "recycled" brain area for reading. These structures were repurposed from the visual system and were originally involved in pattern recognition.
A change of career
"This work has opened up a potential linkage between our understanding of the neural mechanisms of visual processing and [...] human reading," says James DiCarlo, the head of MIT's Department of Brain and Cognitive Sciences and the senior author of the study.
The findings suggest that even nonhuman primates have the ability to distinguish words from gibberish, or to pick out specific letters in a word, through a part of the brain called the inferotemporal (IT) cortex.
Previous research has used functional magnetic resonance imaging (fMRI) to identify which brain pathways activate when we read a word. Christened the visual word form area (VWFA), it handles the first step involved in reading: recognizing words in strings of letters or in unknown script. This area is located in the IT cortex, and is also responsible for distinguishing individual objects from visual data. The team also cites a 2012 study from France that showed baboons can learn to identify words within bunches of random letters.
DiCarlo and Dehaene wanted to see if this ability to process text is a natural part of the primate brain. They recorded neural activity patterns from 4 macaques as they were shown around 300 words and 300 'nonwords' each. Data from the macaques was recorded at over 500 sites across their IT cortexes using surgically-implanted electrodes. This data was then fed through an algorithm that tried to determine whether the activity was caused by a word or not.
"The efficiency of this methodology is that you don't need to train animals to do anything," Rajalingham says. "What you do is just record these patterns of neural activity as you flash an image in front of the animal."
Naturally good with letters
This model was 76% accurate at telling whether the animal was looking at a word or not, which is similar to the results of the baboons in the 2012 study.
As a control, the team performed the same experiment with data from a different brain area that is also tied to the IT and visual cortex. The accuracy of the model was worse compared to the experimental one (57% vs. 76%). This last part shows that the VWFA is particularly suited to handle the processes involved in letter and word recognition.
All in all, the findings support the hypothesis that the IT cortex could have been repurposed to enable reading, and that reading and writing themselves are an expression of our innate object recognition abilities.
Of course, whether reading and writing arose naturally from the way our brains work, or whether our brains had to shift to accommodate them, is a very interesting question -- one that, for now, remains unanswered. The insight gained in this study, however, could help to guide us towards an answer there as well.
"These findings inspired us to ask if nonhuman primates could provide a unique opportunity to investigate the neuronal mechanisms underlying orthographic processing," says Dehaene.
The next step, according to the researchers, is to train animals to read and see how their patterns of neural activity change as they learn.
The paper "The inferior temporal cortex is a potential cortical precursor of orthographic processing in untrained monkeys" has been published in the journal Nature Communications.