Top image - wavelet

Evidence is building that sleep disturbances in healthy people, repeated over years, might trigger or accelerate pathological processes leading to dementia, e.g. by causing oxidative stress or enhancing amyloid deposition1-5. In healthy middle-aged men, for example, one night of total sleep deprivation produced elevated morning amyloid beta (Ab42) in cerebrospinal fluid (CSF)6. Sleep deprivation over two months of AβPP/PS1 mice caused them to have irreversibly increased amyloid production and more plaques in cortex and hippocampus and elevated neuronal mitochondrial damage and apoptosis6.

Sleep restriction of 3xTg mice elevated amyloid and pTau in neocortex7; milder sleep restriction of the same strain increased the amount of insoluble tau, and decreased the level of synaptic markers8. Once amyloid pathology has formed in the NREM-generating areas of the prefrontal cortex, the area of cortex that is responsible for generating delta waves, this could further disrupt NREM sleep, leading to a vicious cycle of sleep disruption and more amyloid deposition9.

But mechanisms responsible for how disrupted sleep might, in the long-term, increase the risk of getting dementia are unknown. Levels of soluble Ab peptides oscillate in the brain over 24 hours, being highest during wakefulness, and lowest during sleep1. The variation in Ab levels is under circadian control, persisting in free-running conditions, so Ab production or clearance could be controlled by the circadian machinery located either in the hypothalamus (suprachiasmatic nucleus) or by local circadian clocks, e.g. in microglia10. Neurons make more soluble Ab when active11. As sleep deprivation increases neuronal excitability of the cortex, amyloid production will increase further. But it has also been proposed that Ab is cleared more efficiently in sleep2. The two effects would work together.

Little work has been done on how sleep deprivation affects microglial activation. Perhaps the sleeping state is optimal for microglial phagocytosis of amyloid. A boost to the sleep-dependent clearance hypothesis is based on the concept of glymphatic drainage12. This drainage might clear waste from the brain selectively during NREM sleep13. Sleep-dependent glymphatic drainage has captured the imagination and engagement of both the dementia and sleep fields. The suggestion, is that during NREM sleep, astrocytes, whose end-feet line the capillaries, shrink in volume, opening up the spaces between end-feet and capillaries, allowing more metabolite clearance13. The astrocyte shrinkage is hypothesized to be regulated by noradrenaline13.


  1. J.E. Kang, et al., Science, 326: 1005-1007, (2009)
  2. E.S. Musiek and D.M. Holtzman, Science, 354: 1004-1008, (2016)
  3. M. Costandi, Nature, 497: S19-20, (2013)
  4. A.P. Spira, et al., Curr Opin Psychiatry, 27: 478-483, (2014)
  5. J.H. Roh, et al., J Exp Med, 211: 2487-2496, (2014)
  6. S. Ooms, et al., JAMA Neurol, 71: 971-977, (2014)
  7. S.M. Rothman, et al., Brain Res, 1529: 200-208, (2013)
  8. A. Di Meco, et al., Neurobiol Aging, 35: 1813-1820, (2014)
  9. B.A. Mander, et al., Nat Neurosci, 18: 1051-1057, (2015)
  10. M.H. Hastings and M. Goedert, Curr Opin Neurobiol, 23: 880-887, (2013)
  11. A.W. Bero, et al., Nat Neurosci, 14: 750-756, (2011)
  12. J.J. Iliff, et al., Sci Transl Med, 4: 147ra111, (2012)
  13. L. Xie, et al., Science, 342: 373-377, (2013)

Key objectives and contact for lead

Key objectives

The key objectives of this research programme are to address the questions:

  • By what mechanism(s) do soluble amyloid levels decrease in NREM sleep?

  • Does chronic sleep deprivation hinder microglia responses or down-regulate their ability to tackle amyloid plaques?


Please address enquiries about this programme to:

Chair in Molecular Neuroscience
Professor William Wisden

Professor of Biophysics and Anaesthetics
Professor Nicholas Franks