AI tools that decipher what mice see could power future brain-computer interfaces, according to new paper study.
The system, named CEBRA, was developed by researchers at the Swiss university EPFL. what is their purpose? Revealing hidden connections between the brain and behavior.
To test CEBRA (pronounced “zebra”), the team tried to decipher what mice see when they watch a video.
“Given that the brain is the most complex structure in our universe, this is the ultimate test for CEBRA.
First, researchers collected open-access neural data on rodents watching movies. Some of the brain activity was measured with electrode probes in the visual cortex of mice. The rest came via optical probes in genetically engineered mice. This mouse was designed so that neurons glow green when activated.
Join the TNW Conference in June and save 50% now
Take advantage of our 2-for-1 sale and bring a friend
All this data was used to train the basic algorithm on CEBRA. As a result, the system learned to map brain activity to specific frames of the video.
The team then applied the tool to another mouse that saw the video.After analyzing the data, CEBRA accurately Predict what the mouse sees from brain signals alone.
The team then reconstructed clips from neural activity. You can see the results in the video below.
Of course, researchers aren’t just interested in the movie-watching habits of rodents.
“CEBRA’s goal is to reveal the structure of complex systems. And given that the brain is the most complex structure in our universe, the brain is the ultimate test space for CEBRA,” he said. EFPL’s Mackenzie Mathis, principal investigator of the study, said.
It also provides insight into how the brain processes information, and combining data across animals and species could provide a platform for discovering new principles in neuroscience. . ”
Not really CERA is limited to neuroscience research. According to Mathis, it can also be applied to many datasets containing time and joint information, such as animal behavior and gene expression data. But perhaps the most interesting application is the Brain Computer Interface (BCI).
Even the primary visual cortex, which is often thought to underlie fairly basic visual processing, can be used to decode video in the BCI style, as the cinephile Mouse has shown. An obvious next step for the researcher is to use CEBRA to enhance the neural decoding of his BCI.
“This work is just one step towards the theory-backed algorithms needed to enable high-performance BMI in neurotechnology,” said Mathis.
Read the full research paper in Nature.
