Sleep Strengthens the Active Prediction of Regular Procedures

Scientists at the University of Tübingen are researching whether the brain can better retain previously learned sequences through a rest phase

The test phase of the experiment: Subjects should press the corresponding numbered button in a sequence, each time the light on the screen is lit. After a sleep phase, the subjects made significantly more mistakes in stimulus deviating from the learned sequence than subjects who had remained awake between the study and the test phase.

Sleep consolidates previously acquired knowledge and thereby contributes to the formation of long-term memory. In a new study, Nicolas Lutz, Ines Wolf and Stefanie Hübner under the direction of Professor Jan Born and Dr. med. Karsten Rauss from the Institute of Medical Psychology and Behavioral Neurobiology at the University of Tübingen, Germany, if sleep also influences the learning performance in predictable processes. To do this, they had two groups of people learn on a screen defined sequences of visual patterns. After either a sleep or a wake-up phase, the scientists tested how the subjects responded to deviations in the learned procedures. It showed that the group with sleep phase had internalized the processes more strongly and mastered safer, even if the sequences were presented in faster succession.The Journal of Neuroscience published.

Scientists have a lot of evidence that our brain stores the rules or principles of predictable processes. “So far, however, little is known about how predictions are generated from previously acquired knowledge and maintained over longer periods,” says Nico-las Lutz, the first author of the study. In the current study, the Tübingen research team investigated whether sleep contributes to the formation, consolidation and abstraction of internal models of simple tasks. In the experiments, participants should press numbered buttons in the test phase according to the position of the patterns on the screen. In comparison with the previously learned sequences, individual deviations were incorporated here. The scientists raised the error rate at the push of a button for a different stimulus and the immediately following stimulus.

Around the corner thought: Higher error rate proves better learning performance

“The error rate for the deviant stimulus was significantly increased in the group of probands who had slept in between, compared to the awake group,” says Lutz. “This was sometimes due to the fact that the subjects mistakenly pressed the key for the anticipated instead of the actually presented stimulus after sleep,” Lutz continues. This proves that the sleep group had internalized the original processes more strongly and could make better predictions than the group that had previously spent a similar amount of time awake. Also, the subjects of the sleep group were at the next stimulus, which corresponded again to the learned sequence, immediately back in the track.

In everyday life, predictions of temporal-spatial sequences are omnipresent. “In road traffic or team sports, it is immensely important to anticipate the movements of actors and objects. Only in this way can we adequately react to the unforeseen even at high speeds, “says study leader Karsten Rauss. “In follow-up studies, we will explore the question of how sleep contributes to the generation and storage of such predictions at the neuronal level.” “Forming internal models that find and model rules in the outer events makes it easier for us to register and anticipate unexpected events to learn them, “says Lutz. The new study has shown that sleep has a positive influence on this meaningful learning process. “Our environment is constantly changing. To find our way around quickly,

Source : University of Tübingen