Memory formation and sleep deprivation

There may be a link between the way memories are formed and the adverse effects of sleep deprivation

SLEEP deprivation is an uncomfortable experience. In drivers and workers it can lead to fatal accidents. In those under interrogation it can lead to confession. But why it does what it does is mysterious--as, indeed, is the purpose of sleep itself.

Many theories have been proposed to explain why the pressure to sleep builds up until it becomes irresistible. The latest, presented at the recent annual meeting of the Organisation for Human Brain Mapping, in Florence, Italy, starts from the obvious proposition that the longer you stay awake, the more you learn. Giulio Tononi of the University of Wisconsin proposes that this extra learning makes the brain more and more expensive to maintain. Sleep prunes back the grey matter so that, come the morning, the brain is once again economical to run. If this pruning cannot take place, the organ becomes less and less efficient, and dire consequences result.

Even at rest, the brain is costly to run, consuming 20% of the body's energy production. Most of this energy goes towards maintaining synapses, the junctions across which impulses jump from nerve cell to nerve cell, keeping the brain alert even when it is not doing anything. When a person learns something new, certain synapses are strengthened relative to others. Over the course of the day, there is a net increase in the strength and number of the brain's synapses. And, as Dr Tononi observes, stronger synapses consume more energy. In addition, making them requires more protein and fats and they take up more space. Given that an organism has limited supplies of all of these things, something must happen to prevent the cost of having a brain from gradually spiralling out of control. That something, Dr Tononi believes, is non-rapid eye movement (non-REM) or slow-wave sleep.

Knitting up the ravel'd sleeve of care

Until recently, most sleep research focused on REM sleep. This is the time when people dream, and dreams have had a grip on the scientific imagination since the days of Freud, and on the popular imagination at least since biblical times. Lately, though, researchers have started to wonder whether they have been looking in the wrong place for the significance of sleep, for REMming occupies only about a fifth of the night.

During the other 80% of sleep--the part that is non-REM--the firing pattern of the brain's nerve cells sets up slow electrical waves that start at different points in the cerebral cortex and travel across it. These travelling waves occur hundreds of times a night, and most commonly at a frequency, 1 cycle per second, which has been shown to depress the activity of synapses.

Within an hour of a person falling asleep, slow waves will have covered his entire cortex, affecting every nerve cell in it. At first, these waves are big and powerful, but their size decreases through the night. Dr Tononi believes that the function of these slow waves is to "downscale" synapses, reducing their size, chemical activity and electrical activity--and thus the strength with which they connect their nerve cells together--all over the brain.

The trick, he thinks, is that this downscaling is done in proportion to the existing strength of each synapse. When a sleeper awakens, the strength of each synapse is thus the same relative to all the others, but all synapses are weaker than they were when he went to sleep. Indeed, the weakest of them may vanish completely, taking part of the previous day's memory with them.

In earlier experiments, designed to replicate normal learning, Dr Tononi found that the part of the brain showing most slow-wave activity during sleep was the same as the part that had been activated during the experiment. This fitted the prediction that the downscaling slow-waves would be strongest in those parts of the brain where the most changes had taken place during the day.

The team's latest work draws not on human subjects, but on fruit flies. Flies, too, sleep. Their genetics--including the genetics of sleep--have been studied for almost a century, and many of the genes that play a role in human sleep resemble those that control sleep in fruit flies. Best of all, experiments on fruit flies are not subject to vetting by ethics committees.

With the help of Chiara Cirelli, who also works at the University of Wisconsin, Dr Tononi has created a mutant fruit fly that sleeps only two or three hours a night. (A normal fly sleeps between eight and 14 hours.) The mutation itself is in a gene for a nerve-cell protein of a type known as an ion channel.

Ion channels sit in a cell's outer membrane and let electrically charged atoms (ions, as they are known in chemical jargon) in and out of the cell. In this case, the ion is potassium. It is the movement of potassium (and also sodium) ions that causes the electrical impulses that nerve cells carry--including the impulses found in slow-wave sleep. Moreover, the particular protein that the flies lack is most concentrated in brain areas involved in learning and memory. To nobody's surprise, therefore, though the mutant fly is capable of learning things, it forgets them within minutes. Healthy flies retain learned information for hours or even days.

The researchers' discovery finds an intriguing echo in a human disease called Morvan's syndrome. This is a rare brain disorder that is caused by an autoimmune response which destroys the human equivalents of the ion channels that are affected in the mutant fruit fly. Patients with Morvan's syndrome suffer from severe insomnia and have been known to go for months without sleeping. Eventually, this extreme sleep deprivation kills them.

Dr Tononi's hypothesis is, it must be said, controversial. Many researchers hold almost precisely the opposite opinion--that sleep serves to re-activate synapses that were strengthened during the day, and thus reinforces their strength rather than diminishing it. There is, however, a certain logical sense to the Tononi view of the world. It is impossible to remember everything, so a process of winnowing must take place somehow. The idea that, after a period of expansion, the brain pares back its workforce to become leaner and meaner is somehow rather appealing.