The field of neurotechnology is rapidly evolving, and EPFL researchers have made a significant breakthrough in capturing brain dynamics with impressive accuracy. The team has introduced a new machine learning algorithm called CEBRA (pronounced zebra) that can reveal the hidden structure in data recorded from the brain, predicting complex information such as what mice see while watching a movie.
Let’s dive deep into CEBRA’s groundbreaking capabilities and what this breakthrough means for the future of neuroscience.
Capturing Hidden Structures in Brain Data
The brain is a complex structure, and researchers have long been trying to unravel its mysteries. One of the significant challenges in neuroscience is to understand how the brain processes information and how it relates to behaviour. Traditionally, this has been done by recording the electrical activity of neurons in specific regions of the brain and trying to correlate this activity with behaviour.
CEBRA, the novel machine learning algorithm developed by EPFL researchers, uses a technique called contrastive learning to capture the hidden structure in the neural code. This technique learns how high-dimensional data can be arranged in a lower-dimensional space called a latent space, where similar data points are close together and more-different data points are further apart. This embedding can be used to infer hidden relationships and structures in the data.
Predicting What a Mouse Sees
While it is not yet possible to fully reconstruct what someone sees based solely on brain signals, CEBRA has shown impressive accuracy in capturing brain dynamics. In their study published in Nature, the researchers showed that CEBRA can decode what a mouse sees while it watches a movie. During the training period, CEBRA learns to map brain activity to specific frames. It can predict unseen movie frames directly from brain signals alone after the initial training period.
The data used for the video decoding was open-access through the Allen Institute in Seattle, WA. The brain signals are obtained either directly by measuring brain activity via electrode probes inserted into the visual cortex area of the mouse’s brain or using optical probes, which consist of using genetically modified mice engineered so that activated neurons glow green.
CEBRA’s Impressive Capabilities
CEBRA’s strength lies in its ability to combine data across modalities, such as movie features and brain data, limiting nuances such as changes to the data that depend on how they were collected. Its ability to excel in reconstructing synthetic data sets it apart from other algorithms.
In addition to visual cortex neurons, CEBRA is not limited to brain data. The researchers also showed that it could be used to predict the movements of the arm in primates and reconstruct the positions of rats as they move about an arena. Its potential applications are vast, including animal behavior and gene-expression data. It could be used to uncover new principles in neuroscience, giving us insight into how the brain processes information.
Exciting Potential for Clinical Applications
The potential clinical applications of CEBRA are vast and exciting. It could be used to develop high-performance brain-machine interfaces, enabling individuals with disabilities to control prosthetic devices using their thoughts. CEBRA could be applied to many datasets involving time or joint information, including animal behaviour and gene-expression data.
“This algorithm is not limited to neuroscience research, as it can be applied to many datasets involving time or joint information, including animal behavior and gene-expression data. Thus, the potential clinical applications are exciting,” says Mackenzie Mathis, EPFL’s Bertarelli Chair of Integrative Neuroscience and PI of the study.
Final Thoughts
The development of theoretically-backed algorithms that enable high-performance brain-machine interfaces is a significant step forward in the field of neurotechnology. The potential for CEBRA to uncover new principles in neuroscience and provide insight into how the brain processes information is exciting, and its applications in clinical settings are promising. While there is still much work to be done, the development of CEBRA is a crucial milestone in neurotechnology.
Researchers have been working for years to unravel the mysteries of the brain, and the development of CEBRA represents a significant breakthrough. This novel machine-learning algorithm has the potential to predict what a mouse sees by decoding brain signals, and its capabilities could be used to develop high-performance brain-machine interfaces.
The potential applications of CEBRA are vast, and its development is a significant step forward in the field of neurotechnology. This algorithm’s ability to capture the hidden structure in the neural code and combine data across modalities makes it a powerful tool for researchers.
With further development and refinement, CEBRA could revolutionize the field of neuroscience and help us better understand how the brain works.
References: Schneider, S., Lee, J.H. & Mathis, M.W. Learnable latent embeddings for joint behavioural and neural analysis. Nature (2023). https://doi.org/10.1038/s41586-023-06031-6
