Countless tests have shown that a good night's sleep makes it easier to recall what you've learned. But we don't know why. For decades, scientists have hypothesized that sleep strengthens our brains' neural corrections, but direct evidence for this has been lacking. Now, we may finally have that evidence.
Decades of empirical research and millennia of collective experience have taught us that sleep deprivation impairs learning. A tired person is less able to focus, and therefore less able to grasp a new skill, than a rested one. We also know that sleep is vital to the fortification of new memories. After being exposed to new information, someone who experiences a good night's sleep will have better recall of that information than someone who stays awake the whole night through. The apt apprentice listens well, sleeps soundly and repeats. She strengthens her memory (or her skillset, or knowledge base), by not only acquiring new information but consolidating it, so that it can be built upon in the future.
This much we know. What we don't know is how sleep benefits memory. You might assume that when we talk about "building" upon previous knowledge, we're employing a metaphor that describes additive physical changes in the brain. In reality, however, this relationship is not so clear-cut.
Yes, it is true that memory and learning are boosted by a good night's rest. It has also been shown that learning can increase the strength of synapses (i.e. connections) between neurons. It is therefore tempting to conclude that sleep benefits memory by promoting the formation and fortification of synapses. In many studies, however, we see just the opposite.
In 2011, for example, researchers led by University of Wisconson–Madison psychologist Stephanie Maret showed that, in mice, sleep was actually associated with an overall decrease in tiny, brain-cell protrusions called dendritic spines. If you think of dendrites as branches that project from a neuron, dendritic spines are the leaves that project from those branches. Dendrites can link up with the dendrites of other neurons to form synapses, and dendritic spines help them do it. Studies like Maret's therefore complicate the idea that sleep begets stronger, more plentiful synapses, while supporting the counterintuitive-sounding view that sleep actually weakens synapses related to learning and memory.
But today, our understanding of the relationship between sleep and synapse -formation takes a big step forward: In a study published in the latest issue of Science, researchers led by NYU neuroscientist Wen-Biao Gan have shown for the first time how learning and sleep promote physical changes in the motor cortices of mice. And, in an intriguing twist that seems to run counter to the results of studies like Maret's, Gan's team presents the first direct evidence that sleep after learning actually may actually strengthen the connections between brain cells by promoting the growth of dendritic spines that are directly associated with a newly learned task.
The ingenuity with which Gan and his colleagues made their observations is almost as noteworthy as the findings, themselves.
First, Gan and his colleagues trained a total of fifteen mice to balance atop a rotating rod (think logrolling, like the guy on the left, only with mice). After training the mice for an hour, the team allowed some of them to fall asleep for seven hours. The others were kept awake.
The mice in this study were normal, save for a couple of careful modifications. Gan's team genetically engineered the mice to express a protein that fluoresces yellow in a subset of neurons located in the motor cortex, a region of the brain responsible for voluntary muscle movement. Next, a small window was carved out of each mouse's skull. This allowed Gan's team to monitor the synaptic activity of live mice over several hours, or even days. "The skull window is vital," Gan tells io9, "because it allows us to keep everything intact without damaging or irritating the brain."
In the video above, Gan's team uses a two-photon microscope to peer layer by layer, deeper and deeper into the motor cortex of a mouse whose motor neurons have been engineered to express yellow fluorescent protein. The first few seconds show the microscope stepping through the thin skull window before encountering darkness, where the dura (the thick membrane surrounding the brain) obscures the microscope's view. But around the seven second mark, the microscope's view plunges deep enough that axons, dendrites and – the littlest protrusions of all – dendritic spines come into view. By virtue of this method, which Gan's team developed, the researchers were able to monitor the regulation of dendritic spine–growth in each live mouse in the hours before, during, and after sleep (or sleeplessness).
The results were stark: Gan and his colleagues found that the sleep-deprived mice sprouted significantly fewer dendritic spines than those that were permitted to rest, and the rate of spine formation was correlated with the degree of task improvement. Growth was shown to be most dramatic during the slow-wave, non-REM stage of sleep. What's more, the benefits of sleep seem to carry on well after the mice woke up, with roughly 5% of new spines in the motor cortex developing in the 24 hours after the mice awoke. The mice that slept were also more likely to retain the spines they grew. In some circumstances, it seems sleep could in fact lead to the growth of new synapses.
The researchers went on to demonstrate that the neuronal branches involved in the rod-balancing task were reactivated during this period of slow-wave sleep. Neuroal reactivation during sleep has been observed in the past, but Gan's team took it a step further by blocking the reactivation. When they did, it impeded the formation of new spines, suggesting that reactivation plays some role in stabilizing the dendritic spines sprouted during sleep.
Finally, Gan tells io9 that one of the study's most surprising findings was not directly related to sleep. In a previous study, Gan and his colleagues used the skull-window technique to demonstrate that teaching mice to balance atop the rod led to the formation of new spines along dendrites in the motor cortex. The present study corroborates those findings, but it also shows that teaching the mice a new motor task (for instance, balancing on the rod as it spins in the reverse direction) caused dendritic spines to sprout on an entirely new dendritic branch – i.e., a branch distinct from the one that shot out spines in reaction to learning to balance on a forward-spinning rod. In other words: learning a new motor skill won't cause dendritic spines to appear just anywhere. Rather, the team's findings suggest that synaptic change in the mammalian brain occurs in a site-specific fashion.
Top Photo via Shutterstock