This tech can read your mind
Are you thinking what I’m thinking? (Photo: Disney Channel)
As a kid, I always wondered what it was like to have mind-reading powers. I blame the media I grew up with, shows that always invited “what if” questions like That’s So Raven, Fairly Odd Parents, and Buffy the Vampire Slayer.
While this reality seems distant for us mortal humans, it seems that we don’t have to wait too long for mind-reading technology.
The Professor X-periment
A research group from Radboud University, Netherlands revealed the brain-bending results of their neuroscience experiment.
In this experiment, two participants were shown photos of different faces while they were inside a functional magnetic resonance imaging (fMRI) scanner. fMRI scanning is a type of non-invasive brain imaging technology (i.e. they don’t have to open up your head.) It primarily detects brain activity and measures changes in blood flow.
The Dutch group then fed the results from the fMRI to an AI that was specifically trained to analyze this brain data. To train the AI, the participants were shown a series of other faces while their brains were being scanned. The AI doesn’t see pictures the way humans do. Each pixel of light and darkness creates a unique computer program code, which the AI detects.
From here, the algorithm reconstructs the image based on the data from the fMRI. And the results. Well, they speak for themselves.
The top row of images was shown to the two participants while they were inside the fMRI scanner. The bottom row is the reconstruction of the images based on the brain data analyzed by the fMRI and fed to the AI for decoding.
And to add to the creepy factor: the top row faces were created by generative adversarial networks (GANs), AI programs that can learn to create non-existent faces. That means the two volunteers never could have seen these images before.
The photo results look hyper specific because of the way our brains are wired. Imagine being shown a photo of Person A. Your perception of Person A will elicit a specific brain pattern. If you’re shown a photo of Person B, your perception of Person B will also induce a different brain pattern. Every unique stimulus leads to a unique pattern of brain activity, which the fMRI picks up.
The research group, however, did not stop here. I mean, they already had the brain data. Why not test other possible iterations?
AI researcher, cognitive neuroscientist, and lead researcher of the study, Thirza Dado does us non-sciency people a favor and explains more mind-blowing results from her group’s study.
She explains that the group took the two sets of brain data from both participants, they then hyper-aligned both brains or “took the average brain data of both brains”–which yielded interesting results.
The first column of each picture set with dashed borders are the stimuli–the original photos shown to the participants while they were in the fMRI. The second and third columns are products of the average brain data from both participants, just with different iterations. Mind-blowing, right?
And if you’re still not convinced, Dado recently shared a similar experiment they did with monkey brain data. And the results? Yep. Even monkey brains.
Telepathic technology
Dado and her group are not the only ones to create or conceptualize brain-bending tech.
Engineers from Xiaomi have designed a prototype for the MiGu Headband, which uses electroencephalogram technology (tech that measures the electrical activity of the brain) with the goal of controlling smart devices with one’s brain.
In the same part of the world, Chinese researchers claim they have made it possible for humans to beam radio waves using their brains.
Meanwhile, in Scotland, researchers from the University of Glasgow shared their computational imaging technique called “ghost imaging” which allows humans to see objects hidden around corners. When the tech is combined with human vision, it can reconstruct images of objects hidden from view by “analyzing how the brain processes barely visible reflections on a wall.”
These tech advancements have shown that today’s AI is continuing to push the boundaries for how we understand and see the world. I mean, just a few weeks ago, we talked about AI that can create licensable art pieces from text prompts.
But these tech developments do no good without application in real-life contexts. Dado mentioned that her intention for this research is to ultimately help people. She specifically cited the potential usefulness of her group’s study to eyewitness accounts in crime investigations.
I share the same hope as Dado, that mind-reading tech serves an overall good within society. Otherwise, this would make a not-so-great villain origin story.
