Previous: Why a Bad Series Finale Ruins the Whole Show
Next: The Science Behind Our Niche Interests | Compilation



View count:103,118
Last sync:2023-11-26 00:30
This episode was brought to you by Sanvello. To get started with the Braving Anxiety Journey head to or download Sanvello for your app store.

There are aspects of computer code that look like language and some that seem more like algebra, and since we may be headed for a future where many people will need to learn to code, researchers are interested in figuring out how exactly the brain interprets code.

Hosted by: Brit Garner
Support SciShow by becoming a patron on Patreon:

SciShow has a spinoff podcast! It's called SciShow Tangents. Check it out at
Huge thanks go to the following Patreon supporters for helping us keep SciShow free for everyone forever:

Silas Emrys, Charles Copley, Jb Taishoff, Jeffrey Mckishen, James Knight, Christoph Schwanke, Jacob, Matt Curls, Christopher R Boucher, Eric Jensen, LehelKovacs, Adam Brainard, Greg, Ash, Sam Lutfi, Piya Shedden, KatieMarie Magnone, Scott Satovsky Jr, charles george, Alex Hackman, Chris Peters, Kevin Bealer
Looking for SciShow elsewhere on the internet?

This episode was brought to you by Sanvello.

The Sanvello app will take you on a journey called “Braving Anxiety.” You can get started by downloading the Sanvello app for free. [♩INTRO]. Scientists generally understand what parts of the brain are involved in talking, writing, and solving math problems.

But when it comes to reading and writing computer code, the brain is more of a mystery. Programming languages like Python and C++ are kind of like languages, because they contain words and abbreviations that follow rules like grammar. But they’re also full of symbols and variables, which might make them seem similar to algebra.

So, many people have wondered:. Does the brain process code as a language, or as math? The question may seem abstract – and the answer isn’t black and white!

But understanding how the brain interprets code can help us teach it better. And it can also give us unique insights into how our brains work. High-level programming languages like Python and C++ provide instructions to computers in a way that both computers and people can understand.

They were developed as a happy medium between human language, which is hard for computers to process, and computers' native language of binary code, which is really hard for people to comprehend. And understanding how we understand programming has the potential to reveal how the brain learns novel complex skills. I mean, humans didn’t evolve with special neural regions dedicated to computer programming.

After all, C++, Java, and Python didn’t even exist 50 years ago. But now, many people are fluent in coding languages. So, studying code cognition can give us clues to how the brain repurposes its circuitry to process weird, novel things it’s never seen before over the course of evolution.

Which is really cool. Plus, it’s practical. We may be headed for a future where lots of people need to learn code, at least to some extent.

So it’s essential to discover how we can teach coding effectively. But to accomplish that, we have to figure out how our brains interpret code. And here, there have been two main schools of thought.

One hypothesis is that your brain categorizes code like it would Swahili or Spanish. That is, it activates the language system, a network of regions involved in linguistic tasks. The other idea is that your brain co-opts the neural circuitry we use for some kinds of math – including symbolic math like algebra.

The math-centric parts of the brain are located in the multiple demand system a larger network of regions involved in all sorts of logic and problem-solving tasks. Researchers have been debating all of this for a while without a clear answer. Then, in a 2020 study, neuroscientists from the Massachusetts Institute of Technology investigated the question using brain imaging.

They recruited 43 proficient coders, put them in an fMRI machine, and had them solve problems shown either as a sentence or using snippets of code. About half the participants were presented with Python, a popular text-based programming language. The others used ScratchJr, a visual programming language designed for kids, which was expected to have less effect on the language system.

And based on the fMRI results, it turns out that neither the language nor the math camps have the whole story right. The study found that when most participants read and built their code, the language processing area of their brain wasn’t tapped much at all. And the multiple demand system was strongly activated, but not in the regions you’d expect for a purely math-related task.

See, many math tasks involve certain regions of the left hemisphere at least, for right-handed people, like those in this study. But coding recruited large parts of the multiple demand system in both the right and left hemispheres that are involved in logic and all kinds of complex cognitive tasks. Essentially, when we read and comprehend code, we use the same parts of the brain we use when we think really hard about solving a challenging problem -- not a basic math exercise.

This means your brain doesn’t see coding as strictly language or math. It’s an uncategorized thing that we process in a unique way. Which is pretty neat.

But... it makes life more difficult for instructors teaching coding. Because if brains saw code just as words, we could teach it like we teach foreign languages. And if brains saw code just as math, we could teach it like we teach algebra or trigonometry.

But since it interprets code as neither of those, computer science educators will have to get creative and teach coding like it’s something totally new. Which, of course, it is. Except… it gets more complicated than that!

Because it might also be a matter of timing. The MIT researchers hypothesize that people taught coding as kids might use their language system more compared to adults. That’s because in kids, the language system is still developing, and what their brain sees as language isn’t set in stone yet.

In contrast, the multiple demand system may remain more flexible throughout a person’s life, so it might be more activated when adults learn code. That means that when we teach coding might make a big difference in how we should teach it! In addition, learning coding might actually rewire the brain.

When we repeatedly perform an action, whether it’s reading books or playing the guitar, the brain forms new and stronger connections between the neurons involved. So the MIT researchers think that people who’ve been coding for decades could form specialized regions in their multiple demand system that are dedicated to Python or C++ or whatever programming languages we invent in the future. That means we have more to learn if we want to harness that flexibility when we teach.

But it’s also a pretty amazing example of what our brains can do when we invent brand new situations for them to deal with. If you have anxiety, you might know that brains deal with some situations better than others. Luckily, anxiety is perfectly natural -- and there are ways to learn to navigate it.

The Sanvello app will take you on a journey called “Braving Anxiety,” hosted by our friend John Green. You’ll travel through an expert-designed program that will help you identify triggers and thinking traps, and teach you in-the-moment solutions for reducing stress and anxiety. Because it is possible to learn to cope with and decrease those scary feelings.

You can download the Sanvello app to get started for free. [♩OUTRO].