A good night’s sleep can be transformative.  Among its benefits are improved energy and mood, better immune system functioning and blood sugar regulation, and greater alertness and ability to concentrate.  Given all of these benefits, the fact that a third of the human lifespan is spent sleeping makes evolutionary sense.  However, sleep appears to have another important function: helping us learn.  Across a plethora of memory tasks—involving word lists, maze locations, auditory tones, and more—going to sleep after training yields better performance than remaining awake.  This has prompted many sleep researchers to reach a provocative conclusion: beyond merely supporting learning, sleep is vital, and perhaps even directly responsible, for learning itself.

Recent discoveries from neuroscience provide insights into that possibility.  Sleep appears to be important for long-term potentiation, a strengthening of signals between neurons that is widely regarded as a mechanism of learning and memory.  Certain memories acquired during the day appear to be reactivated and “replayed” in the brain during sleep, which may help make them longer lasting.  In some instances the amount of improvement that occurs on memory tasks positively correlates with the length of time spent in certain stages of sleep.  These and other findings are generating great excitement among sleep researchers, as well as prompting heated debates about the degree to which sleep may or may not be involved in learning.

To date, most sleep and learning research has focused on recall, which is the capacity to remember information.  However, new research by Stéphanie Mazza and colleagues at the University of Lyon, recently published in the journal Psychological Science, suggests another potential benefit of sleep: improved relearning.  Relearning refers to the process of re-acquiring forgotten information.  Because we cannot possibly remember all of the information that we encounter, it is often necessary to go back and learn that information again.  That’s when relearning occurs—such as when preparing for an exam, taking a refresher course, or simply revisiting a topic after an extended period of time.  According to this new research, sleep supercharges relearning: it can enable us to relearn twice as quickly and up to three times more effectively.

In the study, 40 French-speaking adult participants learned a list of 16 Swahili-French word pairs (for example, nyanya-tomate), relearned those word pairs after a period of 12 hours that did or did not include sleep, and were tested on their memory for the word pairs after one week and after six months.  Each participant completed the study as part of a randomly-assigned “wake” or “sleep” group.  In the wake group, the initial learning session occurred at 9 AM and the relearning session occurred at 9 PM on the same day; in the sleep group, the initial learning session occurred at 9 PM and the relearning session occurred at 9 AM on the next day, after a night of sleep.  During the learning session, participants first studied all of the word pairs.  Next, the Swahili word from each pair was shown by itself while participants attempted to recall its French equivalent.  After each attempt, the correct answer was displayed, allowing participants to study further.  The learning session concluded after each pair was successfully recalled once.  At the relearning session, participants again attempted to recall the French word of each pair, and then continued cycling through the entire list until they could recall all pairs perfectly.  Thus, by the time the relearning session was complete, any word pair that had been forgotten in the intervening 12 hours had been practiced until it could be correctly recalled once more.

During the initial learning session there were no observed performance differences between the sleep and wake groups, which suggests that a similar degree of learning occurred in both groups.  However, substantial differences emerged at the relearning session: participants in the sleep group recalled an average of 10 word pairs on their first attempt, whereas those in the wake group could only muster up about 7; in addition, those that had slept took only about three cycles through the list to finish relearning, whereas those that had not slept needed twice that amount.  Even the lowest-performing members of the sleep group—those that had forgotten more word pairs than other members and had the most to relearn—still took fewer cycles to relearn than the best performing (and least forgetful) members of the wake group.

Along with that impressive result, a benefit of having slept between the learning and relearning sessions was also evident over the long term.  After one week, participants in the sleep group correctly recalled more word pairs than those in the wake group (an average of 15 vs. 11), and after six months, their recall performance, while attenuated due to some forgetting, was three times better than that of the wake group (an average of 9 vs. 3).  Thus, despite the fact that both groups had practiced to a level of perfect recall during the relearning session, only those that had slept prior to relearning accrued any long-term memory benefits. 

The fact that participants in the sleep group engaged in relearning was also critical for their improved long-term performance.  In a follow-up to the main experiment, a control group of 20 additional participants followed nearly all of the same procedures as the sleep group, including training at the same times of day and sleeping after initial learning.  However, in place of the relearning session, these participants only briefly practiced the list of word pairs one time.  Without relearning, their subsequent recall performance was no better than the wake group.  Thus, it was the combination of sleeping after initial learning, plus engaging in relearning itself, that generated long-term memory benefits.

Mazza and her colleagues interpreted their results as evidence that sleep “transforms” memories, making the effects of initial learning stronger and empowering subsequent relearning.  By this interpretation, sleep not only prevents memories from being forgotten, it also makes it easier to restore memories during relearning.  In support of this conclusion, the authors noted that total sleep time was positively correlated with recall and relearning performance (the longer participants slept, the better they tended to do on both).  This pattern, similar to that observed in prior research, is consistent with two possibilities: either a sleep-specific mechanism boosts learning, which sets the stage for improved relearning, or sleep allows regularly occurring learning processes to occur without interference, which also boosts memory and expedites relearning.  Both possibilities remain viable in this and related research, thus inviting future studies to tease the two apart.

Moreover, while this latest study reveals an important and largely unknown potential benefit of sleep, it remains to be determined whether the same benefits will manifest in other circumstances (such as when relearning non-verbal materials).  It is also important to note that Mazza and her colleagues had their sleep and wake groups train at different times of day.  In a recent meta-analysis of sleep and motor learning research, Tim Rickard and I found that such “varied time” experimental designs often yield better performance in sleep vs. wake groups regardless of whether there is a sleep-specific learning benefit.  This may be due to the fact that humans vary in fatigue levels, alertness, and other characteristics according to the time of day and their natural urge to sleep—differences that may not necessarily manifest in the sleepiness surveys that participants of sleep studies are often asked to take (as occurred in the present study).  Further research in which the sleep and wake groups are not subjected to such widely divergent schedules is needed to address this issue.

At present, these new results suggest that sleeping between study sessions can be a particularly potent learning strategy, and especially when learners take the time to repeatedly revisit and relearn in order to achieve a high level of performance.  For instance, if a set of materials needs to be completely mastered, such as an official manual in preparation for a licensure exam, then those materials should be repeatedly studied across multiple study sessions and in between periods of sleep.  Scheduling those study sessions to occur shortly before going to bed is also likely to improve future remembering and relearning.  In a world where an uninterrupted night of sleep seems to be increasingly rare—and many of us try expensive and unproven ways to improve learning and memory—these may be some of the most compelling reasons yet to embrace the Land of Nod.