Researchers have combined a deep learning model with experimental data to “decode” mouse neural activity. Using this method, they can accurately determine where a mouse is located in an open environment and what direction it is facing by looking at its neural firing patterns. Being able to decode neural activity can provide insight into the function and behavior of individual neurons or even entire brain regions. The results were published on February 22. Biophysical Journalcould also inform the design of intelligent machines that currently struggle to navigate autonomously.

In collaboration with researchers at the US Army Research Laboratory, senior author Vasilios Maroulis’ team used a deep learning model to investigate two types of neurons involved in navigation: “head direction” neurons, which process this information; encodes which direction the animal is facing. , and “grid cells,” which encode two-dimensional information about the animal’s location in its local environment.

“Current intelligence systems have proven to be great at pattern recognition, but when it comes to navigation, these so-called intelligence systems don’t do very well without GPS coordinates or something else to guide the process. ,” says Marouls, a mathematician. at the University of Tennessee in Knoxville. “I think the next step for artificial intelligence systems is to integrate biological information with existing machine learning methods.”

Unlike previous studies that have attempted to understand grid cell behavior, the team based their methodology on empirical rather than synthetic data. The data, which were collected as part of a previous study, consisted of neural firing patterns collected by internal probes, which “reported the mouse’s actual location, head position, and movements.” Ground truth” video footage was paired with the environment. Analysis involved correlating head direction and activity patterns across groups of grid cells.

“Understanding and representing these neural structures requires mathematical models that describe higher-order connectivity — meaning, I don’t want to understand how one neuron activates another neuron, but I want to understand how groups and teams of neurons behave.” Maroulas says.

Using the new method, the researchers were able to estimate the mouse’s location and head direction with greater accuracy than previously described methods. Next, they plan to add information from other types of neurons involved in navigation and analyze more complex patterns.

Ultimately, the researchers hope their method will help design intelligent machines that can navigate unfamiliar environments without using GPS or satellite information. “The ultimate goal is to use this information to develop a machine learning architecture capable of successfully navigating unknown terrain autonomously and without GPS or satellite guidance,” says Marouls.