Molding the Mind: Researchers use Neurofeedback To Reshape Brain Perception

By Dr. Chinta SidharthanReviewed by Lily Ramsey, LLMDec 19 2024

Learning occurs through transformations in the brain brought about by experiences or instructions that enhance our ability to perceive and respond to the world. However, a recent study published in Proceedings of the National Academy of Sciences explored an innovative method to induce perceptual learning without explicit awareness by directly modifying neural activity patterns in the brain.

The team of researchers from the University of Rochester, Princeton University, and Yale University used real-time functional magnetic resonance imaging (fMRI) neurofeedback to create distinct visual object categories in the brains of participants.

The study demonstrated that altering neural representations can not only reshape brain activity but also influence behavioral perception.

​​​​​​​Study: Sculpting new visual categories into the human brain. Image Credit: Josh Namdar/Shutterstock.com

Background

Humans continuously learn by grouping sensory inputs into categories — a process essential for perception and decision-making. Research shows that when learning new categories, neural activity patterns associated with similar items become more aligned, while patterns for different categories become distinct.

This neural organization reflects an intricate link between brain activity and behavior. Brain imaging studies that have mapped these neural changes show that conventional learning relies on experience or instruction.

However, little is known about whether it is possible to sculpt such neural representations directly and bypass traditional methods. While neurofeedback has previously been used to modify existing neural patterns, the ability to create entirely new categories remains untested.

The Current Study

In the present study, real-time fMRI neurofeedback was used to create new categories of visual shapes in the brain. The researchers began by developing a set of complex shapes by varying specific features, such as curves, in a two-dimensional space.

The participants were required to perform a simple test at the beginning of the study to measure how they perceived and categorized these shapes before training. This test provided a baseline of their ability to distinguish between the shapes.

The experiment focused on using real-time fMRI scans to track brain activity and provide feedback to participants during training. The researchers identified regions in the brain where patterns of activity reflected how the participants represented the shapes.

These brain areas included regions responsible for high-level visual processing but excluded early visual areas.

Subsequently, each participant’s brain activity was analyzed to find specific patterns related to the shapes. These patterns were modeled mathematically to determine how the brain represented two artificial shape categories, separated by a randomly chosen boundary.

Furthermore, during the training, the participants were shown shapes on a screen that appeared to wobble or vibrate slightly. They were told to "stabilize the shapes" by focusing on their mental state, though they were not told how the shapes were categorized or how the feedback worked.

Unknown to them, positive feedback (signified by a reduction in the wobbling) was provided when their brain activity matched the pattern corresponding to the target shape category. The feedback was adjusted throughout training to keep the participants engaged and ensure consistent progress.

The training lasted 5 to 6 days, with hundreds of trials where participants repeatedly viewed and mentally engaged with the shapes. At the end of the training, the participants performed the same shape perception task as on the first day.

The researchers compared how brain activity and perceptual discrimination changed between the beginning and end of the study, focusing on differences between trained and untrained categories.

Major Findings

The study found that sculpting neural activity patterns led to significant changes in both brain and behavioral responses. The participants who underwent neurofeedback training demonstrated increased neural separation between categories within the predefined brain regions, especially in high-level visual areas.

Furthermore, the analysis of neural data revealed that activity patterns associated with shapes from distinct categories became more distinct, as indicated by changes in multivariate pattern classification accuracy. Behaviorally, the participants also showed improved categorical perception along the trained boundaries.

Moreover, psychometric functions indicated steeper discrimination slopes for trained categories compared to untrained controls, suggesting enhanced perceptual distinction. This effect was consistent across most participants despite variability in individual outcomes.

Additionally, correlational analyses revealed a strong positive relationship between neural changes and behavioral improvements, supporting the causal link between sculpted neural representations and altered perception.

Interestingly, the neural effects were localized to high-level visual areas, while the contributions of the early visual cortex were minimal.

This highlighted the role of specific brain regions in processing complex visual information and suggested that neural sculpting selectively modifies representations tied to categorical perception.

Conclusions

Overall, the study demonstrated the potential for reshaping and creating neural and perceptual representations using real-time fMRI neurofeedback. By successfully inducing novel visual categories, the researchers established a direct causal link between brain activity and perception.

These findings open new possibilities for studying and enhancing human learning, with potential applications in education, neurorehabilitation, and cognitive enhancement.

The researchers indicated that future research should explore the durability of these effects and their applicability across different cognitive domains.