Most of us can imagine sound so vividly that we can almost hear a tune in our heads. Now imagine being able to actually produce that sound just by thinking about it. That’s what Berkeley postdoctoral fellow Brian Pasley, Professor Robert Knight, and a team of other scientists hope to accomplish. In a recent study, they demonstrated the first step in that process: the ability to listen in on what our brains are hearing.
The study was made possible by a unique opportunity to work with patients who suffered from epilepsy. The patients underwent surgery to implant a grid of electrodes into their brains. The electrodes monitored neuron activity in order to pinpoint the areas affected by seizures. But for the scientists, the implants also offered a “direct window to the brain,” says Pasley, the first author of the study.
He asked the patients who volunteered for the research to listen to words while recording their brain activity. The researchers then “translated” the brain frequencies back into sounds using a computer model they developed. The results are garbled, like someone trying to speak underwater, but eerily recognizable as speech. In some cases, the words are almost understandable, especially if you know what to listen for.
This research lays the groundwork for developing a system that can play back our thoughts. The underlying theory is that perception and imagination are closely related and that both actions stimulate similar areas of the brains. In other words, imagining a sentence will trigger neurons similar to those that fire when we listen to the words and try to understand them. Ultimately, Pasley’s work could help someone who has lost the ability to speak, whether through paralysis or another disability, to use a computer to speak their mind—literally.
If this seems more than a little creepy to you, you’re not alone. The media jumped on the study and speculated this research could be used to eavesdrop on people’s thoughts. “That’s obviously what grabs people’s imagination,” Pasley says. As for conspiracy theories, “I can’t tell you how many emails I’ve gotten. There’s a lot of crazy stuff out there.” Some claimed this procedure is already performed forcibly in the Nevada desert.
Such a scenario isn’t possible right now, though, according to Pasley. For one thing, this research demonstrates that scientists can only re-create what people are hearing, not what they are thinking. Furthermore, “you can’t just stick electrodes on the top of your head,” he says. “Currently this requires an invasive medical implant, so it’s really limited to patients for whom it would actually be worth undergoing the medical procedure. I mean people who are very severely paralyzed.”
In other words, it’s too early to bring out the tinfoil hats. “Down the line, it will be important to think about the ethical implications,” Pasley concedes. “There’s going to be a balance between the risks and benefits.” At this stage, he sees the potential benefits to a disabled person outweighing the hypothetical dangers.
Practical considerations aside, the research could also help shine a light on our mysterious minds. No one knows exactly how the brain makes sense of speech. Most people are familiar with speech-recognition systems on our phones and GPS navigators, which do a decent job of understanding our commands. Our brains are much better. We can pick out words from a background of noises. We understand speech whether it’s high- or low-pitched, fast or slow, heavily accented or not. But what part of the sound does the brain identify as meaningful? For instance, do we parse out meaning by syllable, phoneme, or something else? “We know this is going on somewhere in the auditory cortex, but we don’t know how,” Pasley said. “Right now, we’re getting at the ‘how’ question.”