Skip to content

“Covert” Neurofeedback Tweaks Brain in ASD

Watch on YouTube.



>> NARRATOR: Young people with autism spectrum disorder thought they were just playing a picture puzzle game while in an MRI scanner. In fact, the game was rigged -- by their own brain activity. The more participants spontaneously activated social brain circuitry known to be underconnected in autism, the more pieces of the puzzle filled in to reveal the picture. Since the game was controlled by their own circuit connectivity, the participants were unwittingly tuning up their own brains. Scan following the in-scanner brain sessions, revealed increased connectivity between two key networks of the social brain that don't talk with each other often enough in ASD. What's more, participants parents noted improvements in their children's social behavior that correlated with the increased circuit chatter. The National Institute of Mental Health scientists who conducted the pilot study say such covert neurofeedback may someday help to treat disorders of circuit underconnectivity.

>> ALEX MARTIN: What we've done is to focus on a particular disorder, in this cas autism. Where we already know that certain parts of what's called the social brain are not talking to each other the way they should. And to take advantage of that knowledge to pinpoint those areas that are communicating poorly and to use a form of positive reinforcement to try to get them talking to each other more strongly, better.

>> MICHAL RAMOT: It's just a game. They like playing games. Everyone likes games. We put people in the scanner. And then we get real time activity from the scanner. So we know what's happening in their brains at every given second of time. So participants start out with this blank screen. And we tell them that there's a picture hidden underneath the screen. And they somehow have to reveal the picture. And they don't know how. And we don't tell them anything. And they ask, well "How to I do this?" And I say just try. Try and reveal the picture. We have these two target networks that we're trying to get to communicate better. But in typically developing individuals they do communicate a lot more than they do in patients with ASD. Parts of the social brain that are involved in all kinds of social processing. Something about development caused them to wire in the wrong way. So it's not a trivial point that we can change them this way. We want to see if changing the networks will change behavior. Whenever the networks are configured the way we want them to be, a piece of the picture comes up. And if they are in the right configuration, then it will show a piece of the puzzle. And then they also get this kind of upbeat sound, just so they know to pay attention. They never quite figure it out. But they just have these strategies they come up with. They have to think that they're doing something, and they have no idea what's going on. But they know that they're supposed to control it, because we tell them to try. Nobody's ever come close to what we're actually doing. I think part of it is that they don't necessarily know that it's the brain that's doing it. If they beat their previous high score, then they get an extra bonus. So they're motivated to try and beat their score every day. And they want to get better. They get very excited when they get better. Everybody likes winning. Right? And there's always activation taking place in the brain, even when we don't think it's doing anything. There's an optimal way, we thin, for this organization of spontaneous activity. And in patients with autism and other neurodevelopmental disorders, then this pattern is disrupted. And some brain regions are not talking to each other the way they should be. And what happens if you have a task, an explicit task -- Like a behavior training paradigm -- then you're just reactivating these bad network configurations. Activating the same networks that wired in a suboptimal way. When we do neurofeedback, we go directly to the networks. And it's directly rewarding the networks instead of going through behavior, which would not activate the right networks. So we show pictures that are completely unrelated. They're not social. They don't have people in them. They don't have text. They're just abstract. So that way they don't activate the networks that we're trying to censor.

>> ALEX MARTIN: With regard to treatment, it's really very unlikely that people are going to be using fMRI and MRI scanners, because it's very expensive, it's very time-consuming. EEG would be much more practical, because it's ubiquitous. The EEG machines are all over the place. It's cheaper. It's much easier to implement, and so on. And in that case, you'd just have to know what the EEG signature was for that was associated with with what we already have documented is going on in the brain using MRI. So that's like a long-term goal of how this kind of neurofeedback might be able to be translated into some kind of clinical application.