Genetics: Gene Shown to Control Hindlimb Identity Collaboration: Researchers Set To Join Forces in Harvard-wide Cancer Center Psychiatry: Sleep May Play Bigger Role in Learning and Memory Nutrition: Eating Less Fat, More of Some Carbs May Make Us Hungrier, Heavier Group Calls for Funding Increase Against Diabetes 'Epidemic' Many Smokers May Underestimate Their Risk of Disease How Dendritic Cells March from Blood to Tissues Novel Inhibitor of Growth Factor Receptor Identified Protein Kinase Affects Cardiac Impulses in Dystrophy Project ADAPT Training Center Opens Kogan to Give Recital at Vanderbilt Hall Honors & Advances News Briefs Panel Points Out Ways to Increase Diversity in Science The Feel and Unexpected Weight of the 'Intern Blues' Front Page:

PSYCHIATRY Sleep May Play Bigger Role in Learning and Memory Robert Stickgold and colleagues at the Mass-achusetts Mental Health Center have shown that sleep plays a central role in learning and consolidation of memory. Conventional wisdom holds that learning is an active process while sleep is passive. But recent research by Robert Stickgold and other sleep scientists paints another picture, suggesting that sleep plays a vital--if not fully understood--functional role in learning and memory processing. "What we know about sleep and memory for certain is that for at least some types of procedural memory, disrupting sleep after training interferes with learning," says Stickgold, HMS assistant professor of psychiatry at Massachusetts Mental Health Center. Using a test for visual discrimination, Stickgold and his colleagues demonstrated that overnight improvement was proportional to the amount of slow-wave sleep participants had in the first quarter of the night and REM sleep in the last quarter. The sequence suggests a two-step process of memory consolidation requiring slow-wave sleep followed by REM. Each subject was trained, tested, and retested. Those who were retested the same day three, six, nine, or 12 hours later showed no improvement, whereas those who slept and had slow-wave and REM sleep, in that order before redoing the test, showed dramatic improvement. Testing Sleep Circuits In another set of tests, published in the March Journal of Cognitive Neurosciences, Stickgold and his colleagues have shown that associative memory is altered during sleep. They awoke study participants during REM and non-REM sleep and gave them quick tests involving pairs of words, a technique called semantic priming. Subjects were shown a "prime" word followed by a "target" word or nonword and asked to indicate whether or not the target was a word. The researchers measured the difference between reaction times for target words preceded by primes with no semantic relationship (e.g., car­apple) and reaction times to target words preceded by semantically related primes (e.g., car­truck). Semantic priming, which measures associative memory processes, can be used to test for alterations in the strength of associations across the REM­non-REM sleep cycle. When subjects carried out the test after being awakened from REM sleep, the effectiveness of primes that were only weakly associated with the target words exceeded that of strong primes. This is the reverse of what is seen when subjects have been awake and implies that the brain pathways believed to underlie semantic priming are dramatically altered during REM sleep. In non-REM sleep, weak primes are ineffective. "It's as if functional networks have swapped sites," Stickgold says. He hypothesizes that during REM sleep, the connectivity to old memory is being altered. The brain seems to be paying attention to more distantly associated concepts. In another series of experiments, Stickgold and his colleagues demonstrated that learning experiences are replayed at sleep onset. The researchers trained volunteers to play the computer game Tetris and sent them home with tape recorders and sleep monitors. Seven out of 10 subjects--those who appeared to be at the steep part of the learning curve--reported seeing shapes related to the game as they fell asleep. Stickgold is especially interested in the function of dreams. For much of this century, the fields of neuroscience and psychoanalysis have differed greatly in their theories on dreaming. Freud thought that dreams served the purpose of both disguising and discharging unconscious drives while neuroscientists believed that dreams were a result of chaotic neuronal activity. New findings point to a theory consistent both with laboratory findings and psychoanalytic theory. Stickgold reviews recent research on dreams and memory in the December 1998 Trends in Cognitive Neuroscience. Memory Fixing Dreams appear to be part of some functional process because they are accompanied by changes in brain chemistry and the activation of different regions of the brain. In particular, REM and slow-wave sleep are characterized by one-way communication between the neocortex and hippocampus. There appears to be an alternating flow of information between the two areas. Stickgold hypothesizes that the shifts in direction between the hippocampus and the neocortex are providing a mechanism for consolidating and integrating memories. The brain has two distinct memory storage systems: the hippocampus stores episodic memories about discrete events and the neocortex stores consolidated, semantic memory. Integration may be taking place during sleep. During slow-wave sleep, the hippocampus seems to be replaying specific episodic memories to the neocortex. In REM sleep, information appears to flow from the neocortex to the hippocampus. Stickgold speculates that once the neocortex has adequately integrated a memory, it may send a message to the hippocampus to erase it. During REM sleep, in the absence of input from the hippocampus, the neocortex may be completely oblivious to episodic memory. This may help explain the bizarre and hyperassociative character of REM dreams in which the brain seems to choose less obvious over more obvious associations, as if episodic memory is temporarily shut down. Memory processing during sleep may shed light on post-traumatic stress disorder in which people dream exact replays of a traumatic event, suggesting that the REM-sleep barrier to hippocampal input is breaking down. This, in turn, could prevent the integration of the traumatic event into semantic memory storage. "Dreams are the mind watching the brain processing memories," Stickgold says. So what about people who don't remember their dreams? "Whatever the function of dreaming is, it doesn't require us to remember. Not remembering dreams is like dubbing tapes with the volume turned down," he explains. "The underlying process still gets carried out." To reassure the sleep-deprived among us, Stickgold argues that sleep is not obligatory for learning and off-line memory processing. Rather, sleep provides a preferred time. "Sleep is a privileged time that guarantees an optimum period for memory consolidation." --Peta Gillyatt Back to Top