Using machine learning, the researchers linked the activity of each neuron to the activity of the others and were thus able to estimate the cells’ “responsibility field” even when the mice were not moving or interacting with a particular environment. This method became the main innovation of the study: Authors of previous studies looked at the behavior of neurons before and after sleep, but were unable to track how information about a particular place or route was directly processed during rest.

The new study confirmed that spatial representations formed during habituation to the environment are maintained for most neurons throughout several hours of sleep after the experience. But the nerve cells of the hippocampus not only process and “crystallize” information about what has already been experienced, but also form a new view of the object of memories.

Researchers explained the discovered mechanism with an example from a person’s life: “It is a very familiar situation that when you wake up with a new vision of a problem, you understand it differently than before rest.” In their opinion, the mice’s behavior in the path after sleep actually reflected what the animals were learning during rest, rather than before: “It appeared as if during sleep the mice were relearning the space they were in before rest.”