AI is opening up possibilities in the science world that were unthinkable just a few years ago, and decoding what mice see when they watch movies is one of them. Researchers at the École Polytechnique Fédérale de Lausanne (EPFL), a university in Switzerland, are behind the creation of CEBRA (short for Cortical Encoding and Behavioral Reconstruction), a tool to uncover the hidden links between brain and behavior.
To test the system, they asked it to try to decipher what a mouse saw when watching a video. Data collected from past experiments was used to train the main algorithm, including recordings of mouse neurological activity, measured with electrodes in the visual cortex and also with special optical probes in genetically modified mice whose neurons lit up with a green hue whenever they were activated.
After analyzing the information, CEBRA accurately predicted what the mouse had seen based on only its brain signals. The tool was even capable of reconstructing the clip from the neural activity, obtaining surprising results that you can check out in the video below.
CEBRA is not limited to mice, as the research team pointed out. According to Mackenzie Mathis, principal investigator and researcher at EPFL, it can be used to uncover technique in complex systems. Its potential also extends to neuroscience, as well as animal behavior, gene expression data and brain-computer interfaces (BCIs).
For example, the movie-loving mice used in the experiments proved that the primary visual cortex, usually associated only with basic visual processing, can be employed to decode videos in a manner similar to BCI. The team is now developing CEBRA further, in the hopes that it could be used to enhance neural decipherment in BCIs.
EPFL is well-known for its research breakthroughs, including this study that has the potential to revolutionize our understanding of the brain and its connection to behavior. In the future, CEBRA could be used to create more powerful brain-computer interfaces that could even help patients with motor disabilities solve movement issues.
