In other research, investigators have found that sleeping only a few hours a night over a long period of time impairs memory and alertness. Another study shows that sleep deprivation for a short period may actually enhance memory for some tasks. Still another study provides a glimpse into what areas of the brain are impaired by sleep deprivation and how this in turn affects decision-making.
Giulio Tononi, MD, PhD, and his group have used a molecular approach to investigate what happens during sleep. They have screened more than 15,000 genes to identify all those whose expression changes during sleep compared to waking, and also after sleep deprivation.
Work published by several laboratories last year showed that expression of many genes in Drosophila and mice changes in the brain depending on the time of day (e.g., 4 am versus 4 pm). These cycling genes that work according to circadian time are involved in many cellular functions. However, it was not known to what extent changes in gene expression between day and night depend on circadian time versus when an individual is actually asleep or awake, whether it be day or night.
Tononi and his colleagues set up an experiment to identify gene expression changes related to sleep and wakefulness per se, as opposed to simply day and night. To do so, they placed electrodes on the skulls of rats--who are nocturnal animals--to detect their sleep and assigned them to three groups: Rats who were allowed to sleep spontaneously were sacrificed at 6 pm during their usual sleep period; sleep-deprived rats were killed at the same circadian time after having been kept awake by playing with them for eight hours; spontaneously awake rats were killed at 6 am during their usual waking period.
The investigators found that while 10 percent of genes expressed in the cerebral cortex do change in their level of expression between day and night, about half of them do so not because of circadian time but because of wakefulness and sleep, regardless of time of day.
"We even found there are molecular correlates of sleep and wakefulness in brain structures not known to sleep, such as the cerebellum," said Tononi. "This suggests that sleep may serve some cellular function even when it has not been observed electrophysiologically."
The investigators found specific categories of genes that are selectively expressed in the sleeping and in the awake brain. The genes associated with wakefulness include mitochondrial genes involved in energy metabolism, those involved in acquiring new memories, and stress response genes. Sleep-related genes were found to be involved in consolidation of memories and initiation of protein synthesis, as had been found in previous studies, and in membrane trafficking, which had not been shown before.
"Our study shows that sleep, far from being a quiescent state of global inactivity, may actively favor specific cellular functions," Tononi said.
In other molecular work, Yoshihiro Urade, PhD, found that mice deficient in prostaglandin D2 synthase (PGD2S) (the enzyme that produces prostaglandin D2 in the brain), prostaglandin D2 receptor (PGD2R), or adenosine A2A receptor could not rebound--that is, regain their functioning through deeper sleep--after being deprived of sleep for a long time. Prostaglandin D2 is a sleep hormone that initiates the signal for sleep by binding its receptor, PGD2R. This signal is then transmitted to VLPO, a sleep center, via adenosine by way of adenosine A2A receptor. Adenosine A2A receptor is believed to be a target of caffeine, which inhibits sleepiness.
The phenomenon of rebounding after sleep deprivation is what occurs when you wake up much earlier in the morning than usual or when you work until midnight. You will naturally be sleepy the next day and will try to sleep much longer and deeper than usual.
The researchers deprived both wild-type mice and PGD2S, PGD2R, or A2A knockout mice of sleep for six hours by gently touching their face and body with a cotton pad. They then analyzed sleep profiles during a recovery period by measuring electroencephalogram data. All the wild-type mice showed an increase in non-REM and REM sleep during the recovery period. But all the knockout mice showed almost no non-REM sleep rebound and shorter REM sleep rebound than wild-type mice.
"Drugs that can increase prostaglandin D2 receptor or adenosine A2A receptor activity will be a new type of sleeping pill," said Urade, "while those that inhibit their activity may help suppress sleepiness and enhance wakefulness."
In another line of work, Daniel Press, MD, and his colleagues at Beth Israel Deaconess Medical Center in Boston sought to determine the effects of chronic sleep restriction on cognitive functioning, in particular on working memory. Working memory is the maintenance of information in a short-term buffer for later manipulation to guide behavior. Working memory is similar to RAM in a computer--it is able to perform different tasks for a short time, then empty itself to perform the next task.
The investigators studied the working memory capacities of individuals who were restricted to four hours of sleep per night for nine days. Seven of the twelve participants in the study slept four hours each night, and five slept for eight hours. Each morning, participants completed a computer task to measure how quickly they could access a list of numbers they had been asked to memorize. The list could be one, three, or five items long. Then participants were presented with a series of single digits and asked to answer "yes" or "no" to indicate whether each digit was one they had memorized. The speed of their responses was measured. The longer it took an individual to respond, the less efficient their working memory. As had been shown previously, it took longer for participants to respond when they had to keep five items in working memory than when they had to keep only one.
This measure allowed the researchers to eliminate motor slowing as a factor in the less efficient working memory. If fatigue made people slower at pushing the buttons to answer yes or no, their response times for one-, two-, or five-item tasks would have decreased equally over time, because the only difference between the tasks is the amount of working memory required.
Individuals who slept eight hours a night steadily increased their working memory efficiency on this task. But participants who slept only four hours a night failed to show any improvement in memory efficiency. Motor skill did not change across days for either group of participants.
"Even relatively minor degrees of sleep restriction can impair alertness and performance," Press said. "Continued research on the specific skills affected by different levels of sleep deprivation is crucial to our understanding of how sleep can influence performance on the tasks we encounter in our daily lives."
A surprising finding by Ilana Hairston, PhD, from Craig Heller and Robert Sapolsky's groups at Stanford University is that sleep-deprived rats actually performed better in a task that required localization of an easily detectable cue than did non-sleep-deprived rats. By contrast, sleep deprivation impaired performance of rats in a task requiring spatial navigation.
Most studies of sleep in rodents have looked at the effects of sleep deprivation during REM sleep. Dominant in the EEG of REM sleep are theta rhythms, a specific frequency oscillation that is associated with learning and believed to be generated in the circuitry of the hippocampus, a brain region involved in spatial learning. It is believed that for some forms of memory, e.g., spatial information, the hippocampus acts as a temporary buffer for newly acquired information that is passed on to other brain regions over subsequent days to consolidate the information. Sleep is believed to be pivotal for consolidation of hippocampal-dependent memory, and it is thought that experiences are "reactivated" during sleep to reaffirm the new neuronal connections acquired during prior waking.
In Hairston's study, 40 rats were trained in one of two types of learning tasks using the Morris water maze, a plastic pool six feet around and two feet deep. Rats were placed in the pool at varying entry points and could only escape if they located a platform. In the spatial learning task, the animals had to locate a hidden platform. To do so, the rats needed to have a "mental map" of the maze and use it to return to the region of the platform in each trial. In the nonspatial (cue) task, rats located an easily visible platform with a lemon odor in the pool. In this task the platform was moved every four trials, requiring the animals to use search and discrimination strategies to find the platform. The time it tooks rats to reach the platform was a measure of their performance. At the end of each training session, half the rats were allowed to sleep for only half of their normal rest phase. By the time they began their next day of training, they had had 18 hours to recuperate from the sleep deprivation.
The rats who were sleep deprived actually did better on the nonspatial task. "We interpret this as being due to decreased interference of the memory for the previous location of the platform," said Hairston. Because the platform in the nonspatial task was moved every four trials, the rats who were not sleep deprived used spatial cues to search for it in its old location. By contrast, animals who were sleep deprived, not remembering the previous location as well, may have used a simple search strategy to navigate the water maze.
The work of Chris Habeck, PhD, and his colleagues is novel because individuals were sleep-deprived longer (for 48 hours) than in earlier studies and because investigators used an analysis technique that identifies a set of brain regions that change their activity in concert, rather than independently. "This technique provides a signature of the brain as a connected network, rather than a collection of regions that are changing activity independently of each other," said Habeck.
Thirteen of the 14 sleep-deprived individuals in the study showed the same consistent pattern of change in brain activity caused by the sleep deprivation they underwent. Some areas of the brain strongly increased their activity and some strongly decreased their activity.
"The fact that these individuals had the same pattern of change is very important because it shows that sleep deprivation had a strong and statistically reliable effect on brain activity," said Habeck.
In addition, the more an individual's brain activity changed, the worse they did in a task comparing a new letter displayed on a computer screen with a list of letters they had previously memorized. The changes in brain activity involved both increases and decreases. The increases in activity suggest that the brain is forced to recruit additional areas that are not normally used for the task at hand.
These findings could be a first step in the search for a remedy to the negative side effects of sleep deprivation, according to Habeck. If areas that cannot sustain their activity during sleep deprivation could be stimulated and the areas showing greater activation could be suppressed through drug intervention or technology, the detrimental effects of sleep deprivation might be alleviated.