Programming tasks that are considered challenging do not rely on logical-mathematical thinking as expected, but rather on the language center of the brain. A study published in the PNAS journal in 2016 revealed that mathematicians exhibit unique brain activity patterns. The location of the brain regions that are active when solving mathematical problems shows that abstract mathematics without numbers is not processed in the linguistic regions of the brain, as previously thought, but rather in the regions responsible for numerical sense. Similarly, programming code, like mathematical formulas, belongs to the so-called formal languages that follow clearly defined rules. Therefore, neuroscientists assumed that programming also relies mainly on logical-mathematical thinking, based on the study that examined the brain activity of mathematicians.
Researchers from the University of Saarland and the Technical University of Chemnitz investigated whether the assumption that programming relies on logical-mathematical thinking is accurate. The team, led by Sven Apel, used functional magnetic resonance imaging (fMRI) to examine the brains of 17 computer science students who were tasked with understanding and identifying errors in a program code. The participants also performed a second task that involved checking the code for syntax errors, which is a relatively simple routine activity. The researchers analyzed which brain regions were active during the complex understanding of the code and the simple code checking task. They also subtracted the brain activity of the control condition to isolate the regions of the brain that were specifically used for the two tasks.
The surprising results showed that there was no activity related to mathematical or logical thinking during the programming tasks. Instead, the left hemisphere of the brain, which processes natural language, exhibited high activity. This includes regions of the Broca center, which is responsible for linguistic semantics and reading comprehension. This supports the hypothesis of Dutch computer scientist Edsger W. Dijkstra, who suggested in the 1980s that programming is more related to language than to mathematics. The study’s findings could be useful in designing new programming languages and in programming education. Apel and his team plan to conduct further studies to investigate whether there are differences in brain activity between expert and novice programmers.
In conclusion, the study challenges the assumption that programming relies mainly on logical-mathematical thinking. Instead, it suggests that programming is more related to language processing. The findings could have implications for programming education and the design of new programming languages.
