The big test is tomorrow—should you stay up late and study, or cut short the cram session and get a good night’s sleep? Most if not all students face this dilemma at some point in their lives. Until very recently, their choice might have seemed obvious: stay up and study, to commit as much information to memory as possible. But research now indicates that missing sleep in order to study may well be self-defeating. A good night’s sleep helps greatly—and is essential in some cases—to making just-learned information consolidate or stick in memory.
This issue doesn’t affect only students or business people whose performance depends heavily on memorization. Older people’s lives are often complicated by a “where did I put my car keys?” impairment in memory function, and it now seems that aging-related sleep problems may cause some of this cognitive decline.
The good news is that researchers may soon find ways to counteract this process, by helping people to sleep better and by developing techniques to strengthen specific memories during sleep. “We now know that it’s possible to cue, during sleep, the reactivation of certain memories that were just learned or trained, and this has the effect of enhancing the ability to recall those memories on subsequent days,” says Jan Born, Ph.D., a professor of medical psychology at the University of Tübingen in Germany.
A recent awakening
There have long been tantalizing hints of the relationship between sleep and memory. One of these was the discovery in the 1950s of dream-rich rapid eye movement (REM) sleep, which eventually led to the hypothesis that REM dreaming is the brain’s way of strengthening—by semi-consciously re-experiencing—new memories. “People thought that it was a rehearsal of daytime experience,” says J. Allan Hobson, M.D., professor of psychiatry emeritus at Harvard Medical School, and a member of the Dana Alliance for Brain Initiatives.
But testing the relationship between sleep and memory was and still is tricky. A researcher can’t easily disentangle the direct effect of a good night’s sleep on memory consolidation from the indirect effect of improved alertness and recall performance. There are also different types of memory, such as procedural memory, which underlies physical skills, and declarative memory of events, dates, and places. They are stored in different brain regions, and are not consolidated in the same way during sleep.
Sleep itself is not a uniform process. Brain activity in REM sleep differs from that within non-REM sleep, which in its deepest stages is called slow-wave sleep (SWS). The initial scientific emphasis on REM sleep as the principal window for memory consolidation produced inconsistent results and controversy. “The recent focus on SWS has led to a more consistent story about memory consolidation,” says Ken Paller, Ph.D., a professor of psychology at Northwestern University.
The scientific shift towards SWS as the major memory-consolidation phase began in the mid-1990s with experiments in animals, the results of which were later confirmed in human subjects. For example, Born’s laboratory reported in 2000 that people who had practiced an image discrimination task—a visual skill requiring procedural memory—improved after a period of mostly SWS sleep, but not after the same duration of mostly REM sleep. The experiment was one of the first to control for some of the confounding effects of sleep loss.
In a different study, Robert Stickgold, Ph.D., a postdoctoral researcher in Hobson’s laboratory at Harvard Medical School, found that subjects given a similar image discrimination challenge did not improve with repeated testing unless they slept on the first night after their initial testing session. Stickgold also was able to control for lost alertness by comparing subjects who slept for three nights after the first test with those who stayed awake the first night but slept on the next two nights and recovered to normal alertness.
Subsequent experiments in the mid and late 2000s indicated that SWS strengthens not only procedural memories but also place memories—learned routes within a town—as well as declarative verbal memories, such as homework-type memories of historical facts. Overall, the degree of memory consolidation seems to depend on factors such as the duration and type of sleep (REM sleep having little or no effect), the type of memory (declarative memories being helped most consistently) and the importance of the memory in the life of the person. The memory benefits of sleep also depend on age, because the amount of SWS in a typical night’s sleep declines, on average, after people reach their thirties. “By age 50 your memory consolidation capacity is probably significantly worse than when you are 30,” says Born.
Age-related sleep problems may even contribute to the “normal” cognitive decline seen among the elderly. “There’s a substantial amount of evidence that fragmented or inadequate sleep in elderly people makes their cognitive function worse,” says Dana Alliance member Clifford B. Saper, M.D., Ph.D., a professor of neurology and neuroscience at Harvard Medical School, head of the neurology department at Beth Israel Deaconess Medical Center. Saper adds that sleep apnea, a breathing-interruption condition which disturbs sleep patterns and is found more commonly among elderly and obese people, is the leading treatable cause of dementia in his clinic. “If we cure their sleep apnea, their memory usually comes back and they feel better than they have in years,” he says.
How does memory consolidation work?
An event happens, and you witness it. You hear it; or see it; or read it; or taste it; or feel it; or perhaps engage all these senses at once. These seemingly simple acts of perception are brought about by millions of your neurons’ firing in highly coordinated ways: neurons that correspond to the sensory features of the things you perceive, the semantic meanings of these things, the place and time, even how you feel.
Neuroscientists widely assume that a memory of such an event is a set of connections among all these neurons—connections that bind them into a functional group, so that the reactivation of one can lead to the reactivation of the whole group, and thus a subjective re-experience of the event.
Although perception occurs within the very large and lately evolved structure known as the neocortex, the initial neuronal connections that make a memory are formed chiefly elsewhere. For a declarative, who-what-where memory—also called an “episodic” memory—the initial connections are thought to be made in the hippocampus, a seahorse-shaped appendage that is wired extensively into the neocortex.
The hippocampus works a bit like an old-fashioned telephone switchboard, making quick but temporary connections among the millions of nerve channels that enter it from the cortex and other regions. In animals and people, destruction of the hippocampus removes the ability to form new declarative memories. “We think that the hippocampus essentially links the separate features of an episode together,” says Born.
During SWS, these hippocampal links come alive again, apparently to consolidate the new memory. Experiments on rats who learn their way through mazes indicate that the same hippocampal neurons that are active during the learning of a maze are reactivated during the next period of SWS. With a similar study in humans, Pierre Maquet, D.M., Ph.D., and colleagues at the University of Liège in Belgium have shown that the degree of this hippocampal reactivation during SWS correlates with improvements in the subsequent recall of the newly learned place memories.
One likely effect of this hippocampus-initiated memory reactivation is to strengthen the new hippocampal connections. But its more important goal seems to be the stimulation of slower-growing connections among memory-associated neurons outside the hippocampus, mainly in the neocortex. These slower-growing connections are more direct and permanent, and correspond to what we usually think of as long term memory. During a period of weeks to months they form and strengthen, until the temporary connections in the hippocampus are no longer needed. “The longer-term neocortical connections are less prone to interference, compared to hippocampal connections, and we suspect that they are better connected to prefrontal structures that regulate memory retrieval,” says Born.
What determines whether a new memory is consolidated during SWS? There is evidence, Born says, that the highly evolved prefrontal cortex and an evolutionarily older, emotion-related structure called the amygdala do this by chemically “tagging” new memory-associated links in the hippocampus, based on their apparent logical or emotional importance.
Why does this memory consolidation process occur principally during SWS, rather than in REM sleep or wakefulness? Born points to a number of key brain chemistry changes that occur during SWS, including a sharp drop in the arousal-related neurotransmitter acetylcholine. “This has the effect of switching the normal direction of communication between the neocortex and hippocampus, so that now the hippocampus can send information to the neocortex,” he says. REM sleep’s function here is still a matter of debate, but Born suspects that it further exercises and strengthens newly formed neocortical connections, albeit locally, without input from the hippocampus. 
Finally, what about procedural memories of sensory-and-motor skills, such as how to ride a bicycle? Less is known about the ways in which these are initially stored and consolidated, but the muscle-controlling structure known as the striatum appears to play a major role in their long-term storage. A related brain structure callled the cerebellum “could have a short-term storage role for procedural memories, analogous to that of the hippocampus for declarative memories,” says Born, “but for now that is very speculative.”
Pharmaceutical companies now are developing drugs to help enforce more normal, youthful sleep patterns in older people. Such drugs may have a significant effect in reversing age-related declines in memory function. But neuroscientists—most visibly Born’s and Paller’s laboratories—have also investigated non-drug methods to enhance sleep-based memory consolidation.
In 2007, Born and his colleagues reported an experiment in human subjects in which they strengthened the recall of new declarative, hippocampus-based memories by cueing the reactivation of the memories during SWS. The cueing was done with a floral odor that had been associated with the memory during learning.  Paller’s group recently reported similar effects for a largely procedural memory—a keyboard-playing task requiring visual-motor coordination—using the learned melody as the cue during sleep.
Turning these simple experimental protocols into methods that ordinary people can use to enhance memorization may not be easy. Some cueing methods may disturb people’s sleep, thereby spoiling the effect. Others, such as odors, may not be useful for cueing highly specific and complex memories, says Born. There is some evidence, too, he adds, that consolidation-enhancement with cues during sleep involves a tradeoff, in which specifically cued memories are strengthened but other newly learned memories are weakened, possibly reflecting a strict overall budget for memory consolidation every night.
Perhaps most importantly, no one has yet reported the enhancement of memory consolidation for the verbal memories that dominate school and business learning, and that might require specific verbal cues during SWS. “That’s one of the next questions we have to ask: can you improve your verbal score on the SAT, for example?” says Paller.
He is cautiously optimistic. And certainly the demand for memory enhancement is there. “I think it’s just a matter of time before people are using such methods,” Born says—“and maybe also misusing them, since people when they sleep can’t entirely control the cues to which their brains are exposed.”
Published August 2012
 “The REM Sleep-Memory Consolidation Hypothesis,” by Jerome M. Siegel, Science 2001 (vol. 294, November): 1058-1063.
 “Early sleep triggers memory for early visual discrimination skills,” by Steffen Gais et al, Nature Neuroscience 2000 (vol. 3, December): 1335-1339.
 “Visual discrimination learning requires sleep after training,” by Robert Stickgold, LaTanya James and J. Allan Hobson, Nature Neuroscience 2000 (vol. 3, December): 1237-1238.
 “The whats and whens of sleep-dependent memory consolidation,” by Susanne Diekelmann, Ines Wilhelm and Jan Born, Sleep Med. Rev. 2009 (October): 309-321.
 “Midlife decline in declarative memory consolidation is correlated with a decline in slow wave sleep,” by Jutta Backhaus et al, Learning and Memory 2007 (vol. 14): 336-341.
 “Increased fragmentation of rest-activity patterns is associated with a characteristic pattern of cognitive impairment in older individuals,” by Andrew S. Lim et al, Sleep 2012 (May 1): 633-640.
 “Are Spatial Memories Strengthened in the Human Hippocampus during Slow Wave Sleep?” by Philippe Peigneux et al, Neuron 2004 (vol. 44, Oct 28): 535—545.
 “The memory function of sleep,” Nature Reviews Neuroscience 2010 (vol. 11, Feb): 114-26.
 “Odor cues during slow-wave sleep prompt declarative memory consolidation,” by Björn Rasch et al, Science 2007 (vol. 315, Mar 9):1426-9.
 “Cued memory reactivation during sleep influences skill learning,” by James W. Antony et al, Nature Neuroscience 2012 (pub. online 24 June).