Jumble of letters. Image credit: Bill Dickinson (CC BY-NC 4.0)
“Aoccdrnig to a rseearch at Cmabridge Uinervtisy, it deos not mttaer in what oredr the ltteers in a wrod are, the olny iprmoatnt tihng is taht the frist and lsat ltteer be in the rghit pclae.”
The above text is an example of the so-called “Cambridge University effect”, a meme that often circulates on the Internet. While the statement has several caveats, people do seem able to read jumbled words – at least short ones – with remarkable ease. But how is this possible? Answering this question has proven difficult because reading involves many different processes. These include analyzing the visual appearance of a word as well as recalling its pronunciation and meaning.
To find out why people are so good at reading jumbled words, Agrawal et al. tested healthy volunteers on a word recognition task. The volunteers viewed strings of letters and had to decide whether each was a word or a non-word. The more closely a jumbled non-word resembled a real word, the longer the volunteers took to categorize it. PENICL took longer than EPNCIL, for example. Words in which some of the original letters had been replaced were easier to categorize than words in which letters had only been swapped. And, as shown by the 'Cambridge University effect', swapping the first and last letters had a greater effect than swapping the middle ones.
Agrawal et al. proposed that this is because when people view a string of letters, visual areas of the brain become active in a pattern representing those letters. The brain then compares this pattern to stored representations of known words. To test this idea, Agrawal et al. developed a computer model consisting of a group of artificial neurons. Each neuron responded more to some letters than others, and the response of the model to a word could be obtained by adding together its responses to all of the letters. This model, based only on processing information using sight, predicted how long volunteers took to process jumbled words. This finding in turn suggests that sound, pronunciation or meaning of the word do not contribute as much to jumbled word reading as previously believed.
Finally, the volunteers performed the same task inside a brain scanner. This revealed the brain regions responsible for processing letter strings and for comparing them to stored words. By identifying the brain circuitry that supports reading – of both intact and jumbled words – these findings could ultimately prove useful in diagnosing and treating reading disorders.
