Thanks to Linode, a top-rated cloud computing company with over a million customers, for supporting this episode of SciShow.

Head to linode.com/scishow to learn more and get a $100 60-day credit on a new Linode account. [♪ INTRO] Seeing is believing. You know something is really in front of you because you can see it there.

But if you let your imagination take the driver’s seat you could also visualize what’s not really in front of you. When you daydream, you can even zone out and miss what’s happening in real-time. And scientists are just starting to figure out how that battle between reality and imagination plays out in your brain.

As much as they’d like this superpower, scientists aren’t mind-readers. Still, they’ve gotten pretty close with machine learning that translates what’s in your brain into a picture. To accomplish that feat, scientists have used either functional magnetic resonance imaging or electrodes implanted in the brain to pick up what your brain cells are putting down.

And this can be helpful when you are picturing something but can’t quite put it into words. For example, there’s this painting that I forget the name of right now. It has an evening sky twinkling over a town, and the brushstrokes are thick and heavy.

Since I can visualize the painting but don’t remember what it’s called, I could get into an fMRI machine and the scientists who authored a 2015 study using this technique could get a readout of my brain activity. Then, a machine learning algorithm that they designed could connect the readout to the internet and do an image search. If everything goes according to plan, they’d find Van Gogh’s “Starry Night” and display it on a screen.

What I imagined would literally appear in front of me. All of this is possible because whenever you think of specific things like the hillside in Starry Night, a specific subset of brain cells activate. So, by seeing what parts of the brain light up in excitement when you think of that hillside versus when you think of your dog’s portrait for example, researchers can train algorithms to know when you’re visualizing one or the other.

But the fMRI machine can only tell you the general area of the brain that’s active at any given time. And some researchers are using an approach with much higher resolution: electrodes. By implanting electrodes or wires that read electricity in the brain, researchers can see what individual brain cells are responding to.

And although these sensors help scientists in answering their hypothesis, it isn’t the main reason they’re implanted in the participant's brain. The electrodes’ main purpose is to help patients who experience debilitating epileptic seizures when less invasive treatments don’t work. Some patients treated with this approach generously agreed to participate and help scientists in this study understand how we visualize things in our mind.

To accomplish this, researchers used the patient’s electrodes to turn the brain’s natural patterns into information that computers can interpret. Now in the 2015 study we mentioned earlier, researchers were able to successfully read people’s minds and show them a visual representation of their thoughts using the internet. But it still wasn’t clear how participants could think of one image while their eyes were also detecting what’s actually in front of them.

So a 2022 study set out to solve that mystery. Researchers from Japan trained a computer program to recognize the constellation of brain cells that were active when people were thinking of specific things like a face or a hillside. Then scientists told people to think of a hillside while looking at a black screen and the program showed them a picture of a hillside with more than chance accuracy.

So researchers successfully replicated the results in the 2015 study. Then they upped the stakes and told the participants to think of one image while they were being shown another. The part of the brain behind all this processing is the visual cortex.

It’s connected to both the part of the brain that sends the signal of what the participant is seeing in real-time from light entering the eyes and the parts that conjure up mental visuals like where you focus your attention. So by looking at the signals from brain cells in the visual cortex, researchers could figure out how your brain prioritizes what you are seeing versus what you’re imagining. What researchers found was that participants’ brain cells responded as if they were looking at the image that they imagined, not the one they were really looking at.

These results suggest that in the visual processing part of your brain, your imagination wins out over the visuals that are really in front of you. But it’s not all or nothing because the cells that represent the other visual are still active, they are just less active when you’re focusing on your imagined visual. Even though these days we’re not using this technology to read people’s minds on a regular basis, it still tells us a lot about how we visualize things.

Up until now, there wasn’t much we understood about how you can imagine sitting on a beach despite looking at your desk. This research shows us what we already knew anecdotally: the mind is a powerful thing… more powerful than the reality in front of you. This SciShow video is sponsored by Linode Cloud Computing.

Cloud computing is a way to stream, host websites and apps, and store your data online. That means you don’t need to worry about clunky hardware in your home or office. You can use Linode to scale your cloud use to just what you need and cut costs compared to other cloud providers.

Linode operates internationally, with data centers all over the world to keep cloud computing fast and accessible wherever you are. To test out Linode’s reliable service, you can run a speed test from your current location even before signing up. For more on Linode, check out the link in the description or head to linode.com/scishow.

That link gives you a $100 60-day credit on a new Linode account. Thank you to Linode for sponsoring this video and thank you for watching this episode of SciShow. [♪ OUTRO]