This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
Harvard Medical School
Most of us can remember at least one occasion on which we sat in class, half asleep, trying to pay attention and learn what we were being taught "“ to no avail. Why is this? What goes wrong when you try but fail to learn because you're "just too sleepy"? Well, I don't know about you, but for a group of 28 young adults, 18-30 years old, the answer now seems more clear. In a study published this February in Nature Neuroscience ("A deficit in the ability to form new human memories without sleep,") Seung-Schik Yoo, Matthew Walker, and their collaborators at Harvard Medical School looked at memory formation in young subjects with or without a night of sleep deprivation. To fully appreciate what they found, you'll need some background.
Sleep and memory consolidation
Over the last ten years, scientists have come to appreciate the complex relationships between sleep and memory. Not only does sleep prepare the brain for encoding new memories, sleep also provides an opportunity for the brain to consolidate and integrate recently learned information. Thus, sleep can make memories more stable, so that they are more resistant to interference and decay. For example, a night of sleep can make you better able to identify objects in your visual field where you studied them the night before, and it can make you faster and more accurate at typing a sequence of numbers that you practiced the night before (review by Stickgold: abstract or pdf download). But studies have also shown that sleep also can identify, extract, and store key features of memories, leaving a memory that is more useful the next day. Thus a night of sleep can increase the likelihood that you will discover a hidden shortcut for a mathematical procedure that you laboriously practiced the night before.
This wide range of benefits of post-training sleep suggests that such memory processing is a major function of sleep. But the findings I've described so far all concern the benefits of sleep on the formation and recollection of memories already formed. Another question is: How does sleep help you learn better the next day? Or, to put it another way, how does a lack of sleep affect your ability to form new memories?
Sleep and attention
For years scientists have known that sleep is necessary to focus attention on a task, whether you're trying to learn something or not. Specifically, sleep deprivation leads to reduced activation of attentional networks in the frontal and parietal lobes across a range of cognitive tasks. In addition, executive function tasks, thought to be mediated by prefrontal cortex, show greater deficits after a night of sleep deprivation than do other cognitive tasks, such as perceptual and memory tests. Such findings had led researchers to suspect that it is simply an inability to pay attention that causes the deficits in memory encoding following sleep deprivation.
Sleep and memory encoding
That diminished attention should account for the poor ability of sleep-deprived individuals to form new memories seems intuitively obvious. Yet animal studies have suggested that there's more to this poor memory formation than just attention problems. Studies in both humans and animals have found that a part of the brain known as the hippocampus is critical for forming new memories that we and animals can later recall. At the molecular level, this memory formation is thought to depend on a process known as long term potentiation, or LTP, which strengthens the connections between nerves cells in a manner that makes it easier for signals to pass between them, as must happen when memories are later recalled. LTP can be seen in slices of the hippocampus removed from rats and maintained in tissue culture. Studies of these rat hippocampal slices have shown that when you sleep-deprive a rat prior to removing and slicing its hippocampus, LTP is impaired in these slices. This clearly has nothing to do with the slice paying less attention; it has to do with hippocampal function. Which brings us to the study of Yoo, Walker, and their colleagues.
Using functional magnetic resonance imaging (fMRI), Yoo and colleagues monitored the brain activity in colleage-age subjects who had either slept normally the night before or slept not at all. They scanned these subjects' brains while the subjects went through 150 pictures of people, objects, landscapes, and more complex scenes, and classified each as an indoor or outdoor picture. Two days later, when both groups were well rested, they gave the subjects a surprise memory test. They showed them the same 150 pictures, mixed together with 75 new ones, and asked them to identify the pictures they remembered having seen before. Both groups did pretty well, but the sleep-deprived group forgot almost twice as many of the original pictures, 26 percent compared to only 14 percent for the well-rested group.
Sleep decreases stickyness
The researchers then went back to the fMRI recordings from the original training session and looked at what parts of the brain each group was using while studying the pictures. Although both groups seemed to show study-related activity in the same set of brain regions, the sleep-deprived subjects showed significantly less activity in the hippocampus; this was true even when Yoo looked only at the brain activity seen when individuals were studying pictures that they correctly recognized two days later. And even when the best performing sleep-deprived subjects were compared to the worst control subjects (whose performance matched that of the best sleep-deprived subjects), the sleep-deprived subjects still showed less hippocampal activation. In contrast, both groups activated attentional networks in the frontal and parietal lobes of the brain equally.
This somewhat heretical finding led the researchers to ask where brain activity was specifically correlated with hippocampal activation -- a question they hoped would reveal which region worked together with the hippocampus during memory formation. Now more differences showed up. Compared to the sleep-deprived subjects, the well-rested subjects showed stronger coupling between the hippocampi and other structures normally associated with episodic memory processing, including other areas in the medial temporal lobe. In contrast, the sleepy subjects showed tighter coupling with basic alertness networks in the brainstem and thalamus. Apparently, success in the sleepy subjects required activation of the hippocampus together with basic arousal circuits.
It may not be surprising that these sleepy subjects needed to crank up their arousal circuits along with their hippocampi. Yet it came as something of a surprise that they seemed to do so at the expense of other circuits that are normally involved in encoding new memories. This may further explain why sleepy subjects performed more poorly. Indeed, when activation patterns seen during successful encoding of pictures later remembered was compared to that seen during unsuccessful encoding, the same medial temporal lobe structures turned up during successful encoding for the well rested subjects but not for the sleepy ones. Despite adequate attention and extra effort at arousal, other crucial memory networks were not up to par.
None of this bodes well. As we become more and more sleep-deprived, replacing needed sleep with caffeine and bleary eyes, we can expect to see a concomitant slipping away of the ability to remember the very things we stayed up late trying to learn. You have to wonder whether it's worth it.
Robert Stickgold, an associate professor of psychiatry at Harvard Medical School, is also a researcher and clinician at the School's Division of Sleep Medicine, where -- when he's not sleeping -- he occasionally collaborates with researchers Yoo and Walker, who authored the study under review here.