by What role do sleep stages play in the formation of memories? “We’ve known for a long time that useful learning occurs during sleep,” says Anna Schapiro, a neuroscientist at the University of Pennsylvania. “You encode new experiences while you’re awake, you go to sleep, and when you wake up, your memory has somehow been transformed.”
However, the precise way in which new experiences are processed during sleep remains a mystery. Using a neural network computational model they built, Schapiro, Penn Ph.D. Princeton University student Dhairyya Singh and Kenneth Norman now have a new view of the process.
In research published in the Proceedings of the National Academy of Sciencesshow that as the brain goes through cycles of slow-wave rapid eye movement (REM) sleep, which occurs about five times a night, the hippocampus teaches the neocortex what it learned, transforming novel and fleeting information into lasting memory.
“This is not just a model of learning in local brain circuits. It’s how one brain region can teach another brain region during sleep, a time when there is no guidance from the external world,” says Schapiro, professor assistant at Penn’s. Department of Psychology. “It’s also a proposition about how we learn gracefully over time as our environment changes.”
Broadly speaking, Schapiro studies learning and memory in humans, specifically how people acquire and consolidate new information. She’s long thought sleep played a role here, something she and her team have been testing in a lab, recording what happens in the brain while participants sleep.
His team also builds neural network models to simulate learning and memory functions. For this work specifically, Schapiro and colleagues built a neural network model comprised of a hippocampus, the brain’s center for new memories, tasked with learning episodic information from the day-to-day world, and the neocortex, responsible for facets such as language, higher-level cognition, and more permanent memory storage. During simulated sleep, the researchers can observe and record which simulated neurons fire when they are in these two areas, and then analyze those patterns of activity.
The team ran various sleep simulations using a brain-inspired learning algorithm they built. The simulations revealed that during slow-wave sleep, the brain mostly revisits recent incidents and data, guided by the hippocampus, and during REM sleep, mostly replays what happened previouslyguided by memory storage in neocortical regions.
“As the two brain regions connect during non-REM sleep, that’s when the hippocampus is actually teaching the neocortex,” says Singh, a second-year doctoral student in Schapiro’s lab. “Then during REM sleep, the neocortex reactivates and can reproduce what it already knows,” solidifying data retention in long-term memory.
The alternation between the two stages of sleep is also important, he says. “When the neocortex doesn’t have a chance to reproduce its own information, we see information there getting overwritten. We think alternating REM and non-REM sleep is necessary for strong memory formation to occur.”
The findings are consistent with what is known in the field, although aspects of the model are still theoretical. “We have yet to test this,” says Schapiro. “One of our next steps will be to run experiments to understand whether REM sleep really does bring back old memories and what implications it might have for integrating new information into your existing knowledge.”
Because the current simulations were based on a typical adult who had a healthy night’s sleep, they don’t necessarily transfer to other types of adults or less-than-stellar sleep habits. They also don’t offer any information about what happens to children, who require different amounts and types of sleep than adults. Schapiro says he sees great potential for her model to answer some of these outstanding questions. “Having a tool like this allows you to go in many directions, especially as sleep architecture changes throughout life and across various disorders, and we can simulate these changes in the model,” she says.
In the long term, a better understanding of the role of sleep stages in memory could help inform treatments for psychiatric and neurological disorders for which sleep deficits are a symptom. Singh says there could also be implications for deep learning and artificial intelligence. “Our biologically inspired algorithm could provide new directions for more powerful offline memory processing in AI systems,” he says. This proof-of-concept work connecting sleep and memory formation moves the field one step closer to these goals.
Funding for this research came from the National Institutes of Health (Grant R01 MH069456) and the Charles E. Kaufman Foundation (Grant KA2020-114800).
