This is how space-aware neural networks work

This is how space-aware neural networks work
MeteoWeb

The neural networks that regulate spatial perception change in a nonlinear manner. To find out, a team of scientists from Salk Institute for Biological Studies in ca. Researchers have found that time spent exploring an environment grows neural representations in surprising ways. Results published in Natural neuroscienceshows that neurons in the hippocampus that are essential for spatial navigation, memory and planning represent space in a way consistent with nonlinear hyperbolic geometry, a three-dimensional extension that grows exponentially outward.

In other words, spatial perception is shaped like the inside of an expanding hourglass. The researchers also found that the size of that space increased with time spent in a location. The volume increases logarithmically, corresponding to the maximum possible increase in information being processed by the brain. This discovery provides valuable avenues for analyzing data related to neurocognitive disorders involving learning and memory, such as Alzheimer’s disease.

Our study shows that the brain does not always function linearly. Instead, neural networks operate along an expanding curve that can be analyzed and understood using hyperbolic geometry and information theory – says Professor Salk Tatiana SharpeyProfessor Edwin K. Hunter, who led the study – It is exciting to see that the neural responses in this region of the brain formed a map that expanded with experience based on the amount of time spent in any given location. The effect persisted even for small deviations in time, when the animal was running slower or faster in the environment“.

Sharpee’s lab uses advanced computational approaches to better understand how the brain works. They have recently pioneered the use of hyperbolic geometry to better understand biological signals such as odor molecules, in addition to odor perception. In the current study, the scientists found that hyperbolic geometry also triggers neural responses. Hyperbolic maps of sensory molecules and events are viewed through hyperbolic neural maps. Space representations expanded dynamically in correlation with the amount of time the rat spent exploring each environment. As the mouse moved more slowly in the environment, it gained more information about space, which caused neural representations to grow larger.

See also  Why do we see black holes? A journey between curiosity and science

The results offer a new perspective on how neural representations can be changed by experience – says Huanqiu Zhang, Sharpee Lab Assistant. “The engineering principles identified in our study may also guide future efforts in understanding neural activity in different brain systems..”One might think that hyperbolic geometry only applies on a cosmic scale, but that is not true – Sharpay says – Our brains operate much slower than the speed of light, which may be one reason we see hyperbolic effects on perceptible spaces rather than cosmic spaces. Next, we’d like to learn more about how these dynamic hyperbolic representations grow in the brain, interact and communicate with each other.Save the world.

Leave a Reply

Your email address will not be published. Required fields are marked *