Dolphins and whales rest one-half of their brains at a time. Apparently, some rodent brains, too, sleep piecemeal. That’s according to a study in the April 28 Nature that blurs the line between sleep and wakefulness. In sleep-deprived rats, random neurons snooze even as the animals remain awake, albeit not performing normal rat tasks as well as rested animals. If human brains work the same way, it could help explain why fatigue affects our abilities. Poor sleep is common in older people and those with dementia (Altena et al., 2010; Gagnon et al., 2008), and now another paper, in the May 1 Sleep, confirms a correlation between recent changes in sleep patterns and thinking. A wakeup call, perhaps, to workaholics burning the midnight oil, that paper found that too little slumber reduces a person’s performance on cognitive tests as if they had aged an extra four to seven years.

Time to Redefine Sleep?
The English language is inadequate to explain varying states of alertness. We call a person or animal awake or asleep, but wakefulness is not binary; it’s a spectrum, said Vladyslav Vyazovskiy of the University of Wisconsin, Madison, who led the Nature study along with senior author Giulio Tononi. “In the morning, you are rested, you are fresh…in the evening, you are still awake but you start to make mistakes,” he said. Sleepwalkers, for example, experience a middle ground between asleep and awake, noted Christopher Colwell of the University of California, Los Angeles, in a commentary accompanying the Nature article.

Paul Shaw, a sleep researcher at Washington University in St. Louis, Missouri, has signed on as a subject in some of his own sleep deprivation studies. He describes the feeling: “You are not awake, you are not asleep; you are in this weird land in between.” Shaw was not part of the Wisconsin study team.

The sensation is hardly novel (see, e.g., Durmer and Dinges, 2005), but the Nature paper is among the first to provide a possible mechanism for late-night fuzziness. Vyazovskiy implanted wire bundles into the brains of three- to four-month-old rats at two locations: the frontal motor cortex and the parietal cortex. These recording arrays allowed him to monitor both the overall electroencephalogram and the activity of individual or sets of neurons at the end of each wire. Previously, he had used the technique to find that neurons fire at higher rates the longer rats stay awake, and at slower rates the longer they sleep (see ARF related news story on Vyazovskiy et al., 2009).

To monitor the neural fallout from sleep deprivation, Vyazovskiy had to keep the rats awake. He started in the morning, when these nocturnal animals are usually bedding down. Previous studies involved prodding animals or tipping them into a dunk tank to keep them awake—“all vicious things that create stress,” noted John Cirrito, also at Washington University, who was not involved in the current research. He commended the Wisconsin group for using a method (described in Schiffelholz and Aldenhoff, 2002) that is less likely to confound results with stress. Vyazovskiy took advantage of rats’ abiding curiosity to keep them aroused and exploring. For four hours, he provided a stream of new and interesting objects that even a worn-out rat could not resist—colored balls and tubes, or stinky bedding from other rats’ cages. Although it is impossible to compare the once-daily sleep of diurnal humans with the regular naps of nocturnal rats, he speculated that a rat missing the day’s first snooze might be equivalent to a person pulling an all-nighter.

The rats stayed awake—eyes open, brain and body active—throughout the four-hour period. But their brainwaves told a different story. Occasionally, for intervals of approximately 100 milliseconds, a cluster of neurons would nod off into what looked like slow-wave sleep. This happens some seven times per minute in fresh, well-rested rats, Vyazovskiy said. But at the start of his experiments, when the animals were already ready to nap, these episodes occurred 20 times per minute. By the end of four hours of constant curiosity, it was happening 35 times per minute, he said.

These brief off periods in the frontal cortex interfered with the rats’ ability to reach out and grab a sugar pellet if the local sleep happened up to 800 milliseconds before the tasty treat was available.

In dolphins and whales, partial-brain sleep allows the animals to keep swimming and breathing even as they rest (see sleep review by Siegel, 2005). What could be the function, if any, of these localized micro-naps in rats? One possibility, Vyazovskiy speculated, is that the downtimes refresh tired neurons, allowing the brain as a whole to stay up longer. Alternatively, he suggested, the off periods could be a malfunction, a consequence of staying awake so long that neurons can no longer handle the strain and have to sleep, even if only for a fraction of a second. Shaw suspects the latter, suggesting that for the down neurons it is “like trying to sleep at a rock concert”—not restful, to say the least.

The study authors hypothesize that nodding-off neurons are responsible for lapses in judgment or abilities during wakefulness in humans, just as they impaired rats reaching for candy. Colwell called the idea, based on these data alone, an “intellectual stretch.” Using optogenetic techniques to control sleep in groups of neurons—that is, by genetically engineering neurons to sleep or wake in response to a light signal—might help support the theory, he suggested. While one cannot apply these invasive techniques in humans, of course, work from Tononi’s group on people having brain surgery shows that, during sleep, different regions of the brain achieve slow-wave patterns locally, as opposed to the brain as a whole going into slow-wave sleep (Nir et al., 2011).

More Sleep, Less Sleep: Bad for the Brain Either Way
Could the effects of sleep deprivation be lasting? Work by Cirrito and colleagues showed that amyloid-β levels in the brain rise during waking periods and fall during sleep (see ARF related news story on Kang et al., 2009). However, Cirrito doubted that millisecond-range local neural sleep would have a large effect on amyloid-β levels.

Many studies have shown that people perform poorly when they sleep badly. Only a few studies (e.g., Cricco et al., 2001; Loerbroks et al., 2010) have approached the question of longitudinal changes to sleep patterns and cognition. One of those is by Jane Ferrie of University College London, U.K., reported in Sleep. Ferrie is an epidemiologist with the 25-years-and-counting Whitehall II study of middle-aged civil service office workers.

In this paper, the researchers used self-reported hours of sleep from two stages of the study, separated by approximately five years, to divide participants into four groups: those who went from short nighttime sleep to longer durations; those who went from a normal seven or eight hours to longer, those who dropped from normal to less, and those who started out as long sleepers but later slept less. People whose sleep patterns were unchanged were not included. At the second visit, subjects underwent tests for memory, reasoning, and language use. “Unfortunately...cognitive function was assessed only at the end of the five-year period,” wrote Jan Born of the University of Lübeck, Germany, in an e-mail to ARF. “We do not actually know whether the participants with substantial changes in sleep duration did or did not show decreased cognitive function already in the beginning of the study period.”Ferrie said that her goal was to keep this initial study as simple as possible.

Dropping from a normal six to eight hours of sleep a night was detrimental to cognitive performance. But Ferrie and colleagues also found that going from seven or eight hours to a longer sleep time was bad for cognition as well. Thus, changes toward either extreme—long sleep or short sleep—were a bad sign for the brain.

Ferrie suspects the altered sleep causes the poor cognition, although she cannot rule out the possibility that the cognitive effects come first and lead to disturbed sleep patterns. Cirrito is a coauthor on a recent paper from David Holtzman’s lab at Washington University in St. Louis, which shows Aβ production is linked to activity of default-mode networks, which are most active when people let their minds wander. He speculated that these networks are more active when people are sleep-deprived (see ARF related news story). Another possibility is that poor sleep and poor cognition are both the result of some underlying problem, Cirrito noted. Depression, for example, is linked to both extra hours in bed (Tsuno et al., 2005) and worse performance on cognitive tests (Singh-Manoux et al., 2010).

“If we could enhance sleep, then we might be able to mitigate some of the changes that occur with age,” Shaw said. Scientists who commented for this article warned that sleeping pills do not replicate the full features of normal sleep, instead creating a facsimile of natural slumber. Ferrie suggests that better sleep habits are the best way to preserve thinking skills in one’s waking hours.—Amber Dance


  1. As to the study by Vyazovskiy et al., this is a really elegant demonstration
    that full-blown, slow-wave sleep, as indicated by highly synchronized “global”
    EEG slow-wave activity and associated neuronal off and on states, develops
    from awake periods. They show that, in awake periods, slow-wave activity and associated neuronal on and off periods can also occur, but in waking, these events occur locally in restricted neuron circuits and are not synchronized across wider cortical networks. In fact, with increasing time spent awake, and increasing sleep propensity, such local off states increase in frequency and also occur more often in synchrony with off states in other cortical areas.

    These findings very much support the concept originally proposed by Alexander Borbely from the University of Zurich, Switzerland, of a homeostatic sleep regulation that even holds at the level of local cortical neuron populations. It suggests that cortical networks and associated functions that are more or less intensively used during waking develop a kind of “neuronal fatigue” that requires sleep-like off periods to recover, although it is not directly shown in this study that the occurrence of off periods indeed increases during waking as a consequence of prior use of respective neuronal networks. To me, a further intriguing finding of the study is that only local off periods in waking, but not local off periods in sleep, appear to be associated by a transient increase in excitation in neighboring neuron populations. This could hint at a mechanism compensating for the loss of function during off periods which operates locally and is active only during wakefulness. Obviously, there is a point at which the brain switches from waking to sleep, not anymore trying to compensate for such off periods. Overall, these findings of local sleep occurring in restricted cortical circuits provide a most plausible neurophysiological explanation for the performance deficits that can occur in conditions of restricted sleep and tiredness, and that are the cause for many accidents.

    Regarding the study by Ferrie et al., as I see it, this study shows that impaired cognitive functions are observed more often in people who, over the 5.4-year study period, changed their reported average sleep duration (during weekdays) from six to eight hours to either a longer sleep duration or a shorter one. In the past, quite a number of studies have shown a link between sleep duration and cognitive function. However, this study is unique in that it focuses on the significance of changes in habitual sleep duration. Unfortunately, this study reported sleep duration assessed in the beginning and at the end of the approximately five-year study period, whereas cognitive function was assessed only at the end. Thus, we do not actually know whether the participants with substantial changes in sleep duration did, or did not, have decreased cognitive function already in the beginning of the study period. If participants having “normal” sleep duration in the beginning of the five-year period had also normal cognitive function at this time, and cognitive function during the five-year period changed in parallel with change in sleep duration, then the data would corroborate the view that sleep plays a highly important role for the development and maintenance of cognitive functioning during later life. However, the implications of the study would obviously be different if cognitive performance was already in decline at the beginning of the study.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Sleep Deprivation Taxes Neurons, Racks Up Brain Aβ?
  2. Do Overactive Brain Networks Broadcast Alzheimer’s Pathology?

Paper Citations

  1. . Do sleep complaints contribute to age-related cognitive decline?. Prog Brain Res. 2010;185:181-205. PubMed.
  2. . Neurobiology of sleep disturbances in neurodegenerative disorders. Curr Pharm Des. 2008;14(32):3430-45. PubMed.
  3. . Neurocognitive consequences of sleep deprivation. Semin Neurol. 2005 Mar;25(1):117-29. PubMed.
  4. . Cortical firing and sleep homeostasis. Neuron. 2009 Sep 24;63(6):865-78. PubMed.
  5. . Novel object presentation affects sleep-wake behavior in rats. Neurosci Lett. 2002 Aug 2;328(1):41-4. PubMed.
  6. . Clues to the functions of mammalian sleep. Nature. 2005 Oct 27;437(7063):1264-71. PubMed.
  7. . Regional slow waves and spindles in human sleep. Neuron. 2011 Apr 14;70(1):153-69. PubMed.
  8. . Amyloid-beta dynamics are regulated by orexin and the sleep-wake cycle. Science. 2009 Nov 13;326(5955):1005-7. PubMed.
  9. . The impact of insomnia on cognitive functioning in older adults. J Am Geriatr Soc. 2001 Sep;49(9):1185-9. PubMed.
  10. . Nocturnal sleep duration and cognitive impairment in a population-based study of older adults. Int J Geriatr Psychiatry. 2010 Jan;25(1):100-9. PubMed.
  11. . Sleep and depression. J Clin Psychiatry. 2005 Oct;66(10):1254-69. PubMed.
  12. . Persistent depressive symptoms and cognitive function in late midlife: the Whitehall II study. J Clin Psychiatry. 2010 Oct;71(10):1379-85. PubMed.

Further Reading


  1. . Daytime sleepiness in mild Alzheimer's disease with and without parkinsonian features. Sleep Med. 2011 Apr;12(4):397-402. PubMed.
  2. . Changes in sleep theta rhythm are related to episodic memory impairment in early Alzheimer's disease. Biol Psychol. 2011 Jul;87(3):334-9. PubMed.
  3. . Angiotensin I-converting enzyme (ACE) activity and expression in rat central nervous system after sleep deprivation. Biol Chem. 2011 Apr;392(6):547-53. PubMed.
  4. . Rapid eye movement sleep atonia in patients with cognitive impairment. J Neurol Sci. 2011 Jun 15;305(1-2):34-7. PubMed.
  5. . A causal role for brain-derived neurotrophic factor in the homeostatic regulation of sleep. J Neurosci. 2008 Apr 9;28(15):4088-95. PubMed.
  6. . Molecular and electrophysiological evidence for net synaptic potentiation in wake and depression in sleep. Nat Neurosci. 2008 Feb;11(2):200-8. PubMed.
  7. . A prospective study of change in sleep duration: associations with mortality in the Whitehall II cohort. Sleep. 2007 Dec;30(12):1659-66. PubMed.

Primary Papers

  1. . Change in sleep duration and cognitive function: findings from the Whitehall II Study. Sleep. 2011 May;34(5):565-73. PubMed.
  2. . Local sleep in awake rats. Nature. 2011 Apr 28;472(7344):443-7. PubMed.
  3. . Neuroscience: Sleepy neurons?. Nature. 2011 Apr 28;472(7344):427-8. PubMed.