Neurofeedback: how to transfer brain structures
Oh for being able to look inside the brain of an expert – and understanding what makes him or her so adept. Since the 19th century, the brains of extraordinary personalities have been dissected in pursuit of that goal. Today, new technologies are helping to decipher complex constellations of brain signals. Theo Ferreira Marins, a postdoctoral researcher at the University of Graz, is using a novel approach to investigate the neural connections involved in movements – such as playing the piano.
What distinguishes the brain of a modern Mozart from that of an amateur who has just about mastered “Chopsticks”? And can we use the brain's malleability – its plasticity – to transfer master-level skills from experts to laymen? Marins is investigating these questions in the FWF-funded project “Using decoded neurofeedback to transfer brain plasticity”. His objectives are twofold: gaining a better understanding of the basic mechanisms of brain function and developing new treatment options for individuals with neurological disorders.
“The study is currently focusing on healthy individuals, but my long-term goal is to transfer this approach and technology to patients who actually need to regain a specific motor skill. Such as being able to unlock their door on their own again after a stroke,” says Marins.
Neurofeedback and plasticity
What if we could learn new movements not only through practice, but also through targeted brain training? A basic research project is using neurofeedback techniques to control motor learning in the brain itself. This could be particularly helpful for patients with neurological problems during rehabilitation.
Exercising fingers in the MRI
For his research project, the neuroscientist is investigating brain signals associated with specific finger exercises – such as precise tapping with fingertips. Functional MRI images reveal the constellations in which active brain regions are set alight during the respective exercises. These patterns differ depending on the movement and as between professionals and laypeople.
“We compare experts who have trained specific finger exercises beforehand as against so-called naïve individuals. It is important to note that the signal constellation of a naïve individual differs fundamentally from that of a trained brain,” explains Marins. This is because practice and learning generate specific neural structures. Marins hypothesis is that these structures can be transferred through neural training.
“In the planned setup for our experiments, the subjects lie in an MRI machine and see an image with a bar. They don't know what the bar means, only that they should try to keep it as high up as possible. It's amazing to see that after a while – even without specific instructions – people in such experiments are able to increase the intended brain activity, i.e., the activity represented by the bar,” says Marins.
The underlying principle is called neurofeedback – in this case through unconscious learning. “In our project, we want to show that neurofeedback training can reinforce the target structure in the brain after just a short time,” explains Marins.
One hour of training changes the brain structure
The brain can reorganize itself – and indeed does so constantly. This quality, called brain plasticity, is central to Marins' research. After only one hour of neurofeedback, the physical structure of the brain can change. He demonstrated this in his PhD thesis in Brazil, which was published in a top-tier journal.
“Back in 2019, that was the shortest time frame in which the human brain had been proven to change its own structure. It became clear to me that neurofeedback must be seen as a way of intervening in the brain and changing how it works,” explains the researcher. He warns against misunderstanding neurofeedback as a kind of “brain jogging”. “I use neurofeedback as a challenge for the brain to restructure itself. The goal is to decode the underlying mechanisms, and the respective person does not even have to understand what their brain is actually trying to do.”
Made possible by AI advances
The crux of the matter is this: anyone who wants to specifically alter brain structures – from naïve to professional – must know which target structure to train for. A brain at “Chopsticks” level cannot simply imitate the exact patterns of a Mozart brain, because it lacks the structural prerequisites that come from years of training.
This is where a recent method from AI research enters the scene. Algorithms in neuroscience have typically been trained to recognize the brain activity of a single brain in tiny 3D units – as miniscule image cubes, known as voxels. However, given that every brain has an individual structure, the results differ even when two people perform the same task. Until now, it has not been possible to compare them.
With the help of so-called hyperalignment, it has been possible for a few years to get around this problem. Instead of comparing 3D brain images, the activity patterns of the voxels are viewed as axes in a high-dimensional space. By mathematically transforming the data from several people, these axes can be superimposed. This enables direct comparison – and thus a goal can be represented for neurofeedback.
“So far, hyperalignment has only been used for the visual cortex, comparing, for instance, how people react to a photo of a spider. The brain signals of people who hate spiders are different,” explains Marins. “In the course of my FWF project, I was able for the first time to use hyperalignment for the motor system. That was an important first step.” Experiments with experts and naïve individuals will now follow in order to systematically investigate the “transfer” or training to the target structure.
Neurofeedback as therapy?
So does this mean that neurofeedback could potentially substitute for remedial training which is time-consuming ? And more importantly, can it be used as a therapy when the motor system is impaired?
“You have to be aware that complex movements also involve complex neural connections. This makes it more difficult to understand what is actually happening in the brain,” notes Marins. “But our experiments with finger exercises are furnishing important fundamental knowledge.” Marins is convinced that his findings will contribute to a better understanding of how the brain needs to reorganize itself after a stroke or when someone has Parkinson's disease, for example. “In a next step, the knowledge gained from MRI could be translated into therapy using less expensive methods such as brain stimulation – EEG, in other words,” Marins explains.
Hence, neurofeedback is not a substitute for physical therapy or the active learning of motor skills. But it might be useful as a supplement to create the best conditions for regaining lost abilities more quickly. By utilizing the brain's amazing ability to reorganize itself – whether you're an expert not.
About the researcher
Theo Ferreira Marins was an assistant professor at the D'Or Institute for Research and Education in Brazil before moving to the Institute of Psychology at the University of Graz in the context of an ESPRIT project funded by the Austrian Science Fund (FWF). In his research, Marins investigates how cognitive and motor training changes the brain at the functional, structural, and neurochemical levels. Set to run until early 2027, the project “Using decoded neurofeedback to transfer brain plasticity through neurofeedback“ has been awarded around EUR 316,000 from the Austrian Science Fund (FWF).
Publications
Distinct neural architectures underlie motor skill acquisition and transfer in human sensorimotor cortex, in: bioRxiv preprint 2025
A common neural architecture for encoding finger movements, in: bioRxiv preprint 2025
