Reading and writing systems have only come to being in the last few thousand years- an insufficient timeframe to evolve a new part of the brain. As such, some scientists have hypothesized that parts of the brain evolved for other purposes have been 'recycled' for reading. And now, a study from MIT supports this.
Previously, the researchers found that parts of the brain's IT cortex, the part of the visual cortex responsible for identifying objects and faces, becomes highly specialized for recognizing written words when people learn how to read. Given the limitations of human imaging methods in monitoring individual neuron activity, the researchers sought to study nonhuman primates instead.
They thus hypothesized that if parts of the primate brain are predisposed to process text, they may be able to find similar patterns of neural activity between nonhuman primates and humans as they look at words.
To see if this is the case, the researchers recorded neural activity from around 500 sites across the IT cortex in macaques as they looked at around 2,000 strings of letters. While some were English words, others were simply strings of random letters.
"The efficiency of this methodology is that you don't need to train animals to do anything," says Rishi Rajalingham, lead author of the study. "What you do is just record these patterns of neural activity as you flash an image in front of the animal."
After collecting the data, the researchers fed it into a computer model known as a linear classifier. This model then combined data from each of the sites to predict whether the string of letters that provoked an activity pattern was a word. In the end, the model distinguished words from nonwords around 70% of the time- similar to the animals.
Comparing this data to other neural activity collected from the V4- a part of the visual cortex that also feeds into the IT cortex, they found that V4 activity did not predict the animals' performance on the reading tasks.
All in all, the researchers say that their data shows that the IT cortex is well-suited to the needs of reading. Thus, they conclude that their findings support the hypothesis that our ability to read relies on pre-existing mechanisms for object recognition. The researchers now plan on investigating how animals learn to read- and which parts of the brain correspond with the activity.