Machine Learning Simulations Show How Each State of Sleep Contributes to Learning and Memory
Weird dreams may help the brain consolidate our experiences and learn from them, a new study has shown.
The importance of sleep and dreams for learning and memory has long been recognized, and that connection may be the crazy dreams that occur in rapid-eye movement (REM) sleep, the stage of sleep when the brain is most active.
“What we lack is a theory that ties this together with consolidation of experiences, generalization of concepts and creativity,” said Nicolas Deperrois, lead author of the study.
During the other phases of sleep, non-REM, the brain replays the sensory stimuli we experience while we are awake, which dovetail into spontaneous bursts of brain activity that produce vivid dreams in REM sleep.
To examine how each phase affects learning, the researchers used a machine learning technique called Generative Adversarial Networks (GANs) to create artificial dream states, before simulating the brain’s cortex activity by introducing unusual elements into the artificial dreams.
GANs is essentially two neural networks competing with each other to generate new data from the same dataset, in this case a series of simple pictures of objects and animals. The system produces new artificial images which look superficially realistic to humans.
Researchers then simulated the cortex during three distinct states: wakefulness, non-REM sleep and REM sleep.
During wakefulness, the model is exposed to pictures of boats, cars, dogs and other objects. In non-REM sleep, the model replays these pictures with some occlusions.
REM sleep creates new GANs, generating twisted but realistic versions and combinations of boats, cars, dogs, etc.
“Non-REM and REM dreams become more realistic as our model learns,” said senior author Dr. Jakob Jordan at the Department of Physiology, University of Bern. “While non-REM dreams resemble waking experiences quite closely, REM dreams tend to creatively combine these experiences.”
To see if they could alter the process, the researchers interrupted different sleep states. When REM sleep was suppressed in the model, the dreams were made less creative, and when non-REM sleep phase was removed the representations tended to be more sensitive to sensory fluctuations.
According to this study, wakefulness, non-REM and REM sleep appear to have complementary functions for learning: experiencing the stimulus, solidifying that experience and discovering semantic concepts.
“We think these findings suggest a simple evolutionary role for dreams, without interpreting their exact meaning,” Deperrois said. “It shouldn’t be surprising that dreams are bizarre: this bizarreness serves a purpose.”
“The next time you’re having crazy dreams, maybe don’t try to find a deeper meaning; your brain may be simply organizing your experiences.”
REM sleep also happens to be the final stage of sleep before waking. For people who can remember their dreams, it’s because the brain has reached that most active period, during which it may be going about this organization.
Many people report dreams often having several components that seem to be completely at odds with one another, and this could be because the brain is trying to solve a problem by experimentally combining disparate elements, such as the boat-dog hybrids made by the GANs.
DREAM Up Some Social Media Friends Who Might Like This Story