Subscribe to eTOC

Disturbed Sleep Is Associated with Increased Amyloid-Beta and Long-Term Memory Impairment, Study Suggests



HEAVY DEPOSITS OF amyloid-beta, shown in red in the brain on the right, are linked to poor sleep and may be paving the way for Alzheimers disease. A brain benefiting from deep sleep brain waves and an absence of amyloid-beta is shown on the left.

A new study suggests that disturbed slow-wave sleep is associated with greater amyloid-beta burden and subsequent memory loss.

A lack of restorative deep sleep may facilitate amyloid-beta (Abeta) traffic to the brain, significantly affecting the process by which short-term memory is converted to long-term memory, according to new research conducted at the University of California, Berkeley.

Disturbances in deep sleep, or non-rapid eye movement (NREM) slow wave activity (SWA), could be a factor in why some people with Abeta plaques show fewer signs of memory loss in Alzheimer's disease (AD) while others have greater difficulty, the investigators reported in a June 1 online article in Nature Neuroscience.

The team used a combination of brain imaging techniques to assess regional Abeta aggregation, brain wave activity during sleep, and sleep-dependent memory recall of a number of word-pair combinations subjects received before and after sleep.

“We wanted to determine whether sleep disturbance plays a role in explaining how Abeta ultimately leads to memory loss,” principal investigator Matthew P. Walker, PhD, director of the Sleep and Neuroimaging Laboratory at University of California, Berkeley, told Neurology Today. “We found that next-day recall was significantly poorer in subjects with reduced NREM slow-wave activity.”

The data suggest that there is a chain of events that occurs between NREM SWA, Abeta deposits, and memory loss, added lead author Bryce A. Mander, PhD, a post-doctoral research fellow in the Sleep and Neuroimaging Laboratory.

“We feel that our data bring together all of these different elements, but it is the old ‘chicken or the egg’ question,” in that researchers do not know what starts the cascade, Dr. Walker said.

Although similar findings have been reported in mouse models, this is the first study of this hypothesis in humans. Dr. Walker noted that sleep disruption seems to be an underappreciated factor contributing to cognitive decline in the elderly. But importantly, he added, it may also be a modifiable and potentially treatable factor.

If the findings hold true in a larger study, it might enable investigators to assess the diagnostic potential of monitoring these factors in much younger, cognitively normal individuals, the two researchers told Neurology Today.


DR. MATTHEW WALKER: “We wanted to determine whether sleep disturbance plays a role in explaining how Abeta ultimately leads to memory loss. We found that next-day recall was significantly poorer in subjects with reduced NREM slow-wave activity.”

The investigators recently received a grant from the National Institutes of Health to conduct a five-year prospective study in older adults, Dr. Mander said. “We want to better understand what triggers this relationship between sleep and memory loss.”


To conduct the study, the researchers used positron emission tomography (PET) scanning to estimate regional Abeta burden in 26 non-demented subjects. In addition, they used electroencephalography (EEG) to monitor brain activity while the subjects slept and functional magnetic resonance imaging (fMRI) after sleep to test sleep-dependent memory consolidation.

Each subject was first instructed to memorize a set of word pairs before they slept and then were given two memory tests: The first was administered 10 minutes after receiving the word pairs, the second after they slept in the Berkeley lab overnight, during which time the investigators used EEG to monitor brain activity during sleep. The next day, the subjects once again underwent word-pair recall testing as they were monitored by fMRI to determine post-sleep-related brain activity, primarily focused on the hippocampus.

The investigators found that individuals with the greatest amounts of Abeta in the medial frontal cortex had the worst sleep quality and also had the poorest results on word-pair memory recall testing the following morning. In fact, some of those with the highest levels of Abeta in the medial frontal cortex and the worst recall forgot more than 50 percent of the word pairs memorized the day before.

The EEG rhythms that define deep sleep are predominantly generated in the medial prefrontal cortex, Dr. Walker explained. The medial prefrontal cortex also appears to be one of the earliest regions to show Abeta plaques, even in cognitively normal older adults that have Abeta pathology in their brains.

“The spatial overlap between Abeta deposition and slow waves, even at a voxel level map, is rather astounding,” Dr. Walker said. “This is one of the observations that led us to test the hypothesis that cortical Abeta burden may be associated with the well characterized age-related disruption in slow-wave generation during deep sleep.

“We showed that the amyloid burden in the medial prefrontal cortex significantly correlated with the severity of impaired NREM SWA generation, and reduced NREM SWA generation was associated with poorer overnight memory consolidation,” he said.

It has been hypothesized that memories undergo a transformation during sleep to become progressively more independent of the hippocampus, Dr. Walker explained. That means the more memory retrieval depends on the hippocampus, the less those memories are likely to be transformed into long-term memory and the more prone those memories are to being forgotten. “This hypothesis is supported by data in older adults, and this is what we see in our data,” he said. “Older adults with poorer long-term memory showed greater hippocampus activation following sleep, whereas the older adults that remember the most required the least amount of hippocampus activation to retrieve memories. These data indicate that the process that transforms memories to be more permanent is impaired more in some older adults than in others. The older adults that showed the most impaired memory transformation were those with the most Abeta and the most disrupted deep sleep.”

But do the memory tests conducted in the study relate to what is seen in AD or the short-term memory tests that neurologists conduct in the office? “AD patients are impaired on tests of both short-term declarative memory and long-term declarative memory,” Dr. Walker said. “This means AD patients have a more difficult time learning new declarative information and a harder time remembering what they learned hours to years later. Our study specifically examines the latter — that is, how able older adults with amyloid-beta pathology in their brains retain what they have learned over the long term. We think this task assesses memory that is clinically impaired in AD.”


Yo-El Ju, MD, an assistant professor of neurology in the sleep medicine section at Washington University School of Medicine in St. Louis, MO, said that the study shows evidence never before seen in humans of the effects of sleep disturbance and Abeta pathology, extending findings that have been reported in mice.

“This study is very interesting because it used memory testing to compare sleep and the amount of Abeta in a specific area of the brain related to memory, and confirms that this relationship is somehow mediated by differences in sleep,” she told Neurology Today.

Dr. Ju, who reported in a 2013 JAMA Neurology study that Abeta deposits in preclinical AD appear to be associated with worse sleep quality, agreed with the UC Berkley investigators that the new findings raise the “chicken or the egg” question, and that the association is likely bidirectional.

“Worse sleep may cause more Abeta, while Abeta might be causing poorer quality of sleep,” Dr. Ju suggested. “This study also supports evidence that sleep EEG could be used as a biomarker for Abeta, and this is important. At some time in the future, looking at SWA during sleep could be a way to detect Abeta before AD symptoms appear. However, the findings are not strong enough to support it as a biomarker yet.”


•. Mander BA, Marks SM, Walker MO, et al. βeta-amyloid disrupts human NREM slow waves and related hippocampus-dependent memory consolidation Nature Neurosci 2015, Epub 2015 June 1.
    •. Ju YE S, McLeland JS, Toedebusch CD, et al. Sleep quality and preclinical Alzheimer disease JAMA Neurol 2013;70:587–593.
    •. Ooms S, et al. Effect of 1 night of total sleep deprivation on cerebrospinal fluid B-amyloid 42 in healthy middle-aged men: A randomized clinical trial JAMA Neurol 2014;71:971–977.
    •. Xie L, et al. Sleep drives metabolite clearance from the adult brain Science 2013;342:373–377.