146
We began testing this theory using daytime naps. We recruited a group of healthy young adults and randomly divided them into a nap group and a no-nap group. At noon, all the participants underwent a rigorous session of learning (one hundred face-name pairs) intended to tax the hippocampus, their short-term memory storage site. As expected, both groups performed at comparable levels. Soon after, the nap group took a ninety-minute siesta in the sleep laboratory with electrodes placed on their heads to measure sleep. The no-nap group stayed awake in the laboratory and performed menial activities, such as browsing the Internet or playing board games. Later that day, at six p.m., all participants performed another round of intensive learning where they tried to cram yet another set of new facts into their short-term storage reservoirs (another one hundred face-name pairs). Our question was simple: Does the learning capacity of the human brain decline with continued time awake across the day and, if so, can sleep reverse this saturation effect and thus restore learning ability?
146
Those who were awake throughout the day became progressively worse at learning, even though their ability to concentrate remained stable (determined by separate attention and response time tests). In contrast, those who napped did markedly better, and actually improved in their capacity to memorize facts. The difference between the two groups at six p.m. was not small: a 20 percent learning advantage for those who slept.
147
Having observed that sleep restores the brain’s capacity for learning, making room for new memories, we went in search of exactly what it was about sleep that transacted the restoration benefit. Analyzing the electrical brainwaves of those in the nap group brought our answer. The memory refreshment was related to lighter, stage 2 NREM sleep, and specifically the short, powerful bursts of electrical activity called sleep spindles, noted in chapter 3. The more sleep spindles an individual obtained during the nap, the greater the restoration of their learning when they woke up. Importantly, sleep spindles did not predict someone’s innate learning aptitude. That would be a less interesting result, as it would imply that inherent learning ability and spindles simply go hand in hand. Instead, it was specifically the change in learning from before relative to after sleep, which is to say the replenishment of learning ability, that spindles predicted.
147
Perhaps more remarkable, as we analyzed the sleep-spindle bursts of activity, we observed a strikingly reliable loop of electrical current pulsing throughout the brain that repeated every 100 to 200 milliseconds. The pulses kept weaving a path back and forth between the hippocampus, with its short-term, limited storage space, and the far larger, long-term storage site of the cortex (analogous to a large-memory hard drive). fn2 In that moment, we had just become privy to an electrical transaction occurring in the quiet secrecy of sleep: one that was shifting fact-based memories from the temporary storage depot (the hippocampus) to a long-term secure vault (the cortex). In doing so, sleep had delightfully cleared out the hippocampus, replenishing this short-term information repository with plentiful free space. Participants awoke with a refreshed capacity to absorb new information within the hippocampus, having relocated yesterday’s imprinted experiences to a more permanent safe hold. The learning of new facts could begin again, anew, the following day.
148
We and other research groups have since repeated this study across a full night of sleep and replicated the same finding: the more sleep spindles an individual has at night, the greater the restoration of overnight learning ability come the next morning.
148
Our recent work on this topic has returned to the question of aging. We have found that seniors (aged sixty to eighty years old) are unable to generate sleep spindles to the same degree as young, healthy adults, suffering a 40 percent deficit. This led to a prediction: the fewer sleep spindles an older adult has on a particular night, the harder it should be for them to cram new facts into their hippocampus the next day, since they have not received as much overnight refreshment of short-term memory capacity. We conducted the study, and that is precisely what we found: the fewer the number of spindles an elderly brain produced on a particular night, the lower the learning capacity of that older individual the next day, making it more difficult for them to memorize the list of facts we presented. This sleep and learning link is yet one more reason for medicine to take more seriously the sleep complaints of the elderly, further compelling researchers such as myself to find new, non-pharmacological methods for improving sleep in aging populations worldwide.
148
Of broader societal relevance, the concentration of NREM-sleep spindles is especially rich in the late-morning hours, sandwiched between long periods of REM sleep. Sleep six hours or less and you are shortchanging the brain of a learning restoration benefit that is normally performed by sleep spindles. I will return to the broader educational ramifications of these findings in a later chapter, addressing the question of whether early school start times, which throttle precisely this spindle-rich phase of sleep, are optimal for the teaching of young minds.
148
SLEEP-THE-NIGHT-AFTER LEARNING
149
The second benefit of sleep for memory comes after learning, one that effectively clicks the “save” button on those newly created files. In doing so, sleep protects newly acquired information, affording immunity against forgetting: an operation called consolidation.
149
It was not until 1924 when two German researchers, John Jenkins and Karl Dallenbach, pitted sleep and wake against each other to see which one won out for a memory-savings benefit—a memory researchers’ version of the classic Coke vs. Pepsi challenge. Their study participants first learned a list of verbal facts. Thereafter, the researchers tracked how quickly the participants forgot those memories over an eight-hour time interval, either spent awake or across a night of sleep. Time spent asleep helped cement the newly learned chunks of information, preventing them from fading away. In contrast, an equivalent time spent awake was deeply hazardous to recently acquired memories, resulting in an accelerated trajectory of forgetting. fn4
150
The experimental results of Jenkins and Dallenbach have now been replicated time and again, with a memory retention benefit of between 20 and 40 percent being offered by sleep, compared to the same amount of time awake. This is not a trivial concept when you consider the potential advantages in the context of studying for an exam, for instance, or evolutionarily, in remembering survival-relevant information such as the sources of food and water and locations of mates and predators.
150
It was not until the 1950s, with the discovery of NREM and REM sleep, that we began to understand more about how, rather than simply if, sleep helps to solidify new memories. Initial efforts focused on deciphering what stage(s) of sleep made immemorial that which we had imprinted onto the brain during the day, be it facts in the classroom, medical knowledge in a residency training program, or a business plan from a seminar.
150
You will recall from chapter 3 that we obtain most of our deep NREM sleep early in the night, and much of our REM sleep (and lighter NREM sleep) late in the night. After having learned lists of facts, researchers allowed participants the opportunity to sleep only for the first half of the night or only for the second half of the night. In this way, both experimental groups obtained the same total amount of sleep (albeit short), yet the former group’s sleep was rich in deep NREM, and the latter was dominated instead by REM. The stage was set for a battle royal between the two types of sleep. The question: Which sleep period would confer a greater memory savings benefit—that filled with deep NREM, or that packed with abundant REM sleep? For fact-based, textbook-like memory, the result was clear. It was early-night sleep, rich in deep NREM, that won out in terms of providing superior memory retention savings relative to late-night, REM-rich sleep.
151
Investigations in the early 2000s arrived at a similar conclusion using a slightly different approach. Having learned a list of facts before bed, participants were allowed to sleep a full eight hours, recorded with electrodes placed on the head. The next morning, participants performed a memory test. When researchers correlated the intervening sleep stages with the number of facts retained the following morning, deep NREM sleep carried the vote: the more deep NREM sleep, the more information an individual remembered the next day. Indeed, if you were a participant in such a study, and the only information I had was the amount of deep NREM sleep you had obtained that night, I could predict with high accuracy how much you would remember in the upcoming memory test upon awakening, even before you took it. That’s how deterministic the link between sleep and memory consolidation can be.
151
Using MRI scans, we have since looked deep into the brains of participants to see where those memories are being retrieved from before sleep relative to after sleep. It turns out that those information packets were being recalled from very different geographical locations within the brain at the two different times. Before having slept, participants were fetching memories from the short-term storage site of the hippocampus—that temporary warehouse, which is a vulnerable place to live for any long duration of time if you are a new memory. But things looked very different by the next morning. The memories had moved. After the full night of sleep, participants were now retrieving that same information from the neocortex, which sits at the top of the brain—a region that serves as the long-term storage site for fact-based memories, where they can now live safely, perhaps in perpetuity.
152
We had observed a real-estate transaction that takes place each night when we sleep. Fitting the notion of a long-wave radio signal that carries information across large geographical distances, the slow brainwaves of deep NREM had served as a courier service, transporting memory packets from a temporary storage hold (hippocampus) to a more secure, permanent home (the cortex). In doing so, sleep had helped future-proof those memories.
152
Put these findings together with those I described earlier regarding initial memorization, and you realize that the anatomical dialogue established during NREM sleep (using sleep spindles and slow waves) between the hippocampus and cortex is elegantly synergistic. By transferring memories of yesterday from the short-term repository of the hippocampus to the long-term home within the cortex, you awake with both yesterday’s experiences safely filed away and having regained your short-term storage capacity for new learning throughout that following day. The cycle repeats each day and night, clearing out the cache of short-term memory for the new imprinting of facts, while accumulating an ever-updated catalog of past memories. Sleep is constantly modifying the information architecture of the brain at night. Even daytime naps as short as twenty minutes can offer a memory consolidation advantage, so long as they contain enough NREM sleep. fn5
154
Not only does sleep maintain those memories you have successfully learned before bed (“the vision that was planted in my brain / Still remains”), but it will even salvage those that appeared to have been lost soon after learning. In other words, following a night of sleep you regain access to memories that you could not retrieve before sleep. Like a computer hard drive where some files have become corrupted and inaccessible, sleep offers a recovery service at night. Having repaired those memory items, rescuing them from the clutches of forgetting, you awake the next morning able to locate and retrieve those once unavailable memory files with ease and precision. The “ah yes, now I remember” sensation that you may have experienced after a good night of sleep.
154
Having narrowed in on the type of sleep—NREM sleep—responsible for making fact-based memories permanent, and further recovering those that were in jeopardy of being lost, we have begun exploring ways to experimentally boost the memory benefits of sleep. Success has come in two forms: sleep stimulation, and targeted memory reactivation. The clinical ramifications of both will become clear when considered in the context of psychiatric illness and neurological disorders, including dementia.
155
Since sleep is expressed in patterns of electrical brainwave activity, sleep stimulation approaches began by trading in the same currency: electricity. In 2006, a research team in Germany recruited a group of healthy young adults for a pioneering study in which they applied electrode pads onto the head, front and back. Rather than recording the electrical brainwaves being emitted from the brain during sleep, the scientists did the opposite: inserted small amounts of electrical voltage. They patiently waited until each participant had entered into the deepest stages of NREM sleep and, at that point, switched on the brain stimulator, pulsing in rhythmic time with the slow waves. The electrical pulsations were so small that participants did not feel them, nor did they wake up. fn7 But they had a measurable impact on sleep.
155
Both the size of the slow brainwaves and the number of sleep spindles riding on top of the deep brainwaves were increased by the stimulation, relative to a control group of subjects who did not receive stimulation during sleep. Before being put to bed, all the participants had learned a list of new facts. They were tested the next morning after sleep. By boosting the electrical quality of deep-sleep brainwave activity, the researchers almost doubled the number of facts that individuals were able to recall the following day, relative to those participants who received no stimulation. Applying stimulation during REM sleep, or during wakefulness across the day, did not offer similar memory advantages. Only stimulation during NREM sleep, in synchronous time with the brain’s own slow mantra rhythm, leveraged a memory improvement.
156
Other methods for amplifying the brainwaves of sleep are fast being developed. One technology involves quiet auditory tones being played over speakers next to the sleeper. Like a metronome in rhythmic stride with the individual slow waves, the tick-tock tones are precisely synchronized with the individual’s sleeping brainwaves to help entrain their rhythm and produce even deeper sleep. Relative to a control group that slept but had no synchronous auditory chimes at night, the auditory stimulation increased the power of the slow brainwaves and returned an impressive 40 percent memory enhancement the next morning.
156
Before you drop this book and start installing speakers above your bed, or go shopping for an electrical brain stimulator, let me dissuade you. For both methods, the wisdom of “do not try this at home” applies. Some individuals have made their own brain-stimulating devices, or bought such devices online, which are not covered by safety regulations. Skin burns and temporary losses of vision have been reported by mistakes in construction or voltage application. Playing loud tick-tock acoustic tones on repeat next to your bed sounds like a safer option, but you may be doing more harm than good. When researchers in the above studies timed the auditory tones to strike just off the natural peak of each slow brainwave, rather than in perfect time with each brainwave, they disrupted, rather than enhanced, sleep quality.
157
If brain stimulation or auditory tones were not bizarre enough, a Swiss research team recently suspended a bedframe on ropes from the ceiling of a sleep laboratory (stick with me here). Affixed to one side of the suspended bed was a rotating pulley. It allowed the researchers to sway the bed from side to side at controlled speeds. Volunteers then took a nap in the bed as the researchers recorded their sleeping brainwaves. In half of the participants, the researchers gently rocked the bed once they entered NREM sleep. In the other half of the subjects, the bed remained static, offering a control condition. Slow rocking increased the depth of deep sleep, boosted the quality of slow brainwaves, and more than doubled the number of sleep spindles. It is not yet known whether these sway-induced sleep changes enhance memory, since the researchers did not perform any such tests with their participants. Nevertheless, the findings offer a scientific explanation for the ancient practice of rocking a child back and forth in one’s arms, or in a crib, inducing a deep sleep.
157
Sleep stimulation methods are promising, but they do have a potential limitation: the memory benefit they provide is indiscriminate. That is, all things learned before sleep are generally enhanced the next day. Similar to a prix fixe menu at a restaurant in which there are no options, you are going to get served all dishes listed, like it or not. Most people do not enjoy this type of food service, which is why most restaurants offer you a large menu from which you can pick and choose, selecting only those items you would like to receive.
158
What if a similar opportunity was possible with sleep and memory? Before going to bed, you would review the learning experiences of the day, selecting only those memories from the menu list that you would like improved. You place your order, then go to sleep, knowing that your order will be served to you overnight. When you wake up in the morning, your brain will have been nourished only by the specific items you ordered from the autobiographical carte du jour. You have, as a consequence, selectively enhanced only those individual memories that you want to keep. It all sounds like the stuff of science fiction, but it is now science fact: the method is called targeted memory reactivation. And as is so often the case, the true story turns out to be far more fascinating than the fictional one.
158
Before going to sleep, we show participants individual pictures of objects at different spatial locations on a computer screen, such as a cat in the lower right side, or a bell in the upper center, or a kettle near the top right of the screen. As a participant, you have to remember not only the individual items you have been shown, but also their spatial location on the screen. You will be shown a hundred of these items. After sleep, picture objects will again appear on the screen, now in the center, some of which you have seen before, some you have not. You have to decide if you remember the picture object or not, and if you do, you must move that picture object to the spatial location on the screen where it originally appeared, using a mouse. In this way, we can assess whether you remember the object, and also how accurately you can remember its location.
159
But here is the intriguing twist. As you were originally learning the images before sleep, each time an object was presented on the screen, a corresponding sound was played. For example, you would hear “meow” when the cat picture was shown, or “ding-a-ling” when the bell was shown. All picture objects are paired, or “auditory-tagged,” with a semantically matching sound. When you are asleep, and in NREM sleep specifically, an experimenter will replay half of the previously tagged sounds (fifty of the total hundred) to your sleeping brain at low volume using speakers on either side of the bed. As if helping guide the brain in a targeted search-and-retrieve effort, we can trigger the selective reactivation of corresponding individual memories, prioritizing them for sleep-strengthening, relative to those that were not reactivated during NREM sleep.
159
When you are tested the following morning, you will have a quite remarkable bias in your recollection, remembering far more of the items that we reactivated during sleep using the sound cues than those not reactivated. Note that all one hundred of the original memory items passed through sleep. However, using sound cuing, we avoid indiscriminate enhancement of all that you learned. Analogous to looping your favorite songs in a repeating playlist at night, we cherry-pick specific slices of your autobiographical past, and preferentially strengthen them by using the individualized sound cues during sleep. fn8
159
I’m sure you can imagine innumerable uses for such a method. That said, you may also feel ethically uncomfortable about the prospect, considering that you would have the power to write and rewrite your own remembered life narrative or, more concerning, that of someone else. This moral dilemma is somewhat far in the future, but should such methods continue to be refined, it is one we may face.