The subjects were given a pre-test to determine their initial performance level on the games. Then they were trained to play the games and later tested on their performance. One group was trained in the morning and then tested 12 hours later after being awake for that time. A second group was trained in the morning and then tested the next day, 24 hours after being trained. Another group was trained in the evening, then tested 12 hours after a night's sleep and a fourth group was trained in the evening and then also tested 24 hours after training.
When trained in the morning subjects showed an 8 percentage point improvement in accuracy immediately after training. However after 12 waking hours following training, subjects lost half of that improvement when tested in the evening. When subjects were tested the next morning 24 hours after training, they showed a 10 percentage point improvement over their pre-test performance.
"The students probably tested more poorly in the afternoon because following training, some of their waking experiences interfered with training. Those distractions went away when they slept and the brain was able to do its work," Nusbaum said.
Among the students who received evening training, scores improved by about 7 percentage points, and went to 10 percentage points the next morning and remained at that level throughout the day.
The study follows Fenn, Nusbaum and Margoliash's earlier work, published in Nature, which showed for the first time that sleep consolidates perceptual learning of synthetic speech.
"In that study we showed that if after learning, by the end of the day, people 'forgot' some of what was learned, a night's sleep restored this memory loss," Nusbaum said. "Furthermore a night's sleep protected memory against loss over the course of
'/>"/>
| Contact: William Harms w-harms@uchicago.edu 773-702-8356 University of Chicago Source:Eurekalert |