EPFL scientists have developed a novel AI approach that can identify common patterns in brain activities of different animals that perform the same task, potentially changing our understanding of how the brain processes information.
The study, published today in Nature Methods, introduces a new approach to explaining complex neural signals and could drive the development of more effective brain-computer interfaces.
This new approach, called marble (basic learning of manifold representations) uses advanced mathematics and artificial intelligence to analyze neural recordings, which reveals previously hidden patterns of brain activity among different animals performing the same task. similarity.
“Suppose you and I are both engaged in mental tasks, such as navigating the way we work. Can the signals from a small number of neurons tell us that we use the same or different psychological strategies to solve the task?” LTS2, the signal processing laboratory of the EPFL School of Engineering, is responsible for the signal processing laboratory of LTS2, responsible for the man Pierre Vandergheynst explained. “This is a fundamental problem for neuroscience, as experimenters often record data from many animals, but we demonstrate whether they use the same brain pattern to represent a given task.”
The team demonstrated the function of marble by analyzing brain records of rats performing tasks and navigating mazes. The results show that different animals use very similar neural patterns when approaching the same task, and this finding is difficult to prove using previous methods.
The study’s lead author Adam Gosztolai, now an assistant professor at the Institute of AI at the Medical University of Vienna, explains what makes their approach unique: “In a curved space, geometric depth, The learning algorithm does not realize that these spaces are curved. Therefore, the dynamic pattern it learns is independent of the shape of the space, which means it can find the same baseline from different records.”
Traditional deep learning methods have been working to analyze dynamic systems such as neural activity that change over time. Marble overcomes this limitation by working within a curved mathematical space that naturally represents a complex pattern of brain activity. This approach allows it to identify common elements in neural activity, even recorded in different subjects or under different conditions.
The researchers found that marble analysis of neural records is easier to explain than existing methods. In practical testing, it showed higher accuracy in decoding arm movements of brain signals compared to current technology, indicating a potential application in the brain-computer interface of auxiliary devices.
While direct applications focus on neuroscience, Vandergheynst highlights the broader potential of their work. “The marble method is primarily intended to help neuroscience researchers understand how the brain calculates personal or experimental conditions and when it exists,” he said. “But its mathematical basis is by no means limited to brain signals, we expect our The tools will benefit researchers who wish to jointly analyze multiple datasets of life and body science.
Breakthroughs are a critical moment in neuroscience research, as scientists increasingly seek to understand how brain processes in different individuals process processes and whether there are universal patterns in neural computing. This understanding may have profound implications for treating neurological conditions and developing more effective brain computer interfaces.
By providing a mathematical framework to compare neural activity in different subjects and conditions, marble opens new possibilities for understanding the fundamentals of brain function. The ability of this approach to decode complex neural patterns can accelerate the development of advanced prosthetics and other assistive technologies that rely on interpreting brain signals.
The study shows that despite individual changes in brain structure and neural records, different brain processes information in similar tasks have fundamental commonalities. This finding suggests that the level of universality in neural computing may have a profound impact on our understanding of cognition and consciousness.
If you find this piece useful, consider supporting our work with a one-time or monthly donation. Your contribution allows us to continue to bring you accurate, thought-provoking scientific and medical news that you can trust. Independent reporting requires time, effort, and resources, and your support makes it possible for us to continue exploring stories that are important to you. Together, we can ensure that important discoveries and developments attract those who need them the most.
