By Alan Z. Segal, MD

Associate Professor of Clinical Neurology, Weill Cornell Medical College

Dr. Segal reports no financial relationships relevant to this field of study.

SYNOPSIS: Evidence is accumulating that disruptions in sleep patterns, particularly slow-wave and REM sleep, alter amyloid-β production and clearance through the cerebrospinal fluid pathways and may play a role in the development of Alzheimer’s disease.

SOURCE: Yo-El S Ju, et al. Slow wave sleep disruption increases cerebrospinal fluid amyloid-β levels. Brain 2017;140:2104-2111.

Humans spend approximately one-third of their lives asleep, but the precise function of sleep in the maintenance of a healthy brain remains elusive. There have been limited reports in rodents of a so-called “glymphatic” system that drains the brain of toxins, a process that is accentuated during sleep and blunted in the waking state. Still, considerable variability exists in the amount of sleep we require, with some “short sleepers” functioning on as little as five hours of sleep on a regular basis, with others needing nine or 10 hours to feel rested.

There is ample evidence to suggest that sleep deprivation may be damaging to the brain, as multiple epidemiological studies have shown a connection between sleep impairments and dementia. Three large European cohorts have indicated a risk of Alzheimer’s disease (AD) as much as two-fold in older individuals who report sleeping problems. Subjects reported a variety of sleep-related complaints, such as “trouble sleeping,” shortened sleep duration (< 7 hours), and the habitual use of sleeping pills.

In the absence of pathologically proven AD, indirect evidence also has been gleaned using biomarkers such as cerebrospinal fluid (CSF) amyloid-β levels and PET scans with amyloid markers (Pittsburgh compound B). Both have shown increases in amyloid among subjects reporting shorter sleep durations and increased wake after sleep onset. It also has been shown that the known correlation between ApoE4 genotype and AD was stronger in subjects who showed fragmented sleep (as measured by actigraphy) compared to subjects who demonstrated a more consolidated sleep architecture. In animal models, using in vivo micro-dialysis, it has been shown that increased neuronal activity directly correlates with amyloid-β in the brain’s interstitial fluid and that this amyloid directly leads to increased plaque deposition. Sleep studies in these animals have shown that amyloid-β clearance from the brain is enhanced by sleep and impaired by prolonged wakefulness, with persistently high amyloid-β in the interstitial fluid.

Without the continuous EEG data obtained through polysomnography (PSG), one can only speculate as to whether the total quantity of sleep or specific phases of sleep promote brain health. Slow wave sleep (SWS; with delta frequencies on EEG) arguably is the most “restful” state of sleep and, therefore, the most restorative. It is known that drugs that promote SWS, specifically gamma hydroxybutyrate (GHB) used in the treatment of narcolepsy, can increase daytime wakefulness and vigilance. The brain also may “crave” rapid eye movement (REM) sleep, but in contrast to SWS, REM sleep has a more poorly understood neuroanatomy and neurochemistry. Although REM is associated with muscle relaxation and may be physically restorative, the active cortical activity seen in REM more closely matches the waking state.

Yo-El et al subjected 22 participants (aged 35-65 years, with a mean of 54 years) to an experimental protocol in which their SWS was disrupted artificially. During a polysomnogram, as delta EEG activity (as quantified by the “delta power” in spectral analysis) was shown to rise, a series of progressively louder tones was delivered through headphones until the subject’s EEG demonstrated an arousal out of SWS. Each subject had two PSGs, one experimental and one sham, in which there was no tonal disruption of SWS. Subjects accrued an average of seven minutes of SWS in the sham condition and none with SWS disruption. REM sleep duration also was shortened in the experimental condition.

CSF in each subject was obtained the following morning. SWS disruption was associated strongly with increases in CSF amyloid-β (r = 0.610; P = 0.009). There was no relationship between amyloid-β levels and any other measure of sleep (total sleep time, non-REM sleep, REM-sleep, or sleep efficiency). The effect of SWS disruption was specific for amyloid-β and was not shown for other markers, such as YKL-40 (an astrocyte-derived inflammatory protein) or tau.


These data can be considered a landmark discovery in our understanding of the basic functions of sleep and might have major societal implications. Could sufficient, high-quality sleep blunt cognitive impairments later in life and actually protect against AD? At a minimum, these data provide convincing evidence that SWS has a direct beneficial effect on the amyloid neurochemistry of the brain and adds support to prior data suggesting that sleep allows the brain to “wash out” toxins.

Despite the dramatic nature of these results, however, it is puzzling that such major effects on CSF amyloid-β occurred with only a very small (seven minutes on average) SWS deprivation. Middle-aged subjects studied here, not unexpectedly, got a minimal amount of SWS. While delta sleep may occupy as much as 15-20% of the night in a child or adolescent, SWS only occurs during 3-5% of the night in middle age. Furthermore, in the elderly, no SWS occurs at all. Therefore, it remains possible that sleep disruption later in adulthood and into senescence is not the cause of cognitive decline, but merely an epiphenomenon of the degenerating brain. Failure to achieve SWS is not a pathological state, but merely an effect of normal aging.