Associations between sleep and amyloid deposits varied by brain region in older adults

A large July 2021 study found that sleep-related differences in brain amyloid were detectable even among older adults without cognitive impairment. The pattern was not uniform across the brain: longer reported nighttime sleep was associated with lower amyloid measurements in some regions, while daytime sleep showed a different pattern in a narrower subgroup.

The findings came from a cross-sectional analysis of 4,425 cognitively unimpaired adults aged 65 to 85 who were screened for the A4 Alzheimer disease prevention study. Of these participants, 1,509 met the study’s threshold for amyloid positivity. Because sleep and brain scans were assessed as a snapshot rather than followed over time, the results show an association, not whether shorter sleep causes amyloid accumulation or whether longer sleep removes existing deposits.

A large screening cohort put sleep beside brain scans

Participants reported their recent average number of nighttime sleep hours and daytime nap minutes. The average nighttime report was 7.1 hours. About two-thirds reported 7 to 8 hours, while 26.8% reported 6 or fewer hours and 6.4% reported more than 8 hours. Daytime sleep was also common in the cohort, with 62% reporting no daytime sleep and daytime sleepers reporting an average of about 33 minutes.

Researchers compared these reports with florbetapir PET measurements. Amyloid, or beta-amyloid, is a protein that can accumulate in brain deposits associated with Alzheimer disease. PET does not provide a direct count of individual deposits; here, it produced standardized uptake value ratios, or SUVRs, estimating the relative amyloid signal in a global neocortical region and in specific areas.

The global measure combined several regions, including the cingulate cortex, precuneus, medial orbitofrontal cortex, temporal lobe and parietal lobe. Looking at those areas separately mattered because a broad average could conceal differences in where the signal was concentrated.

More nighttime sleep corresponded to lower regional uptake

After adjustment for age, sex, APOE genotype and daytime sleep duration, each additional reported hour of nighttime sleep corresponded to a 0.005 lower global amyloid PET SUVR. The regional estimates were larger in the medial orbitofrontal region, where the association was a 0.009 lower SUVR per additional hour, and in the anterior cingulate, where it was 0.011 lower.

These figures are differences in a scan-derived ratio, not percentages of dementia risk and not estimates of how much amyloid an individual would lose by sleeping longer. They also do not establish an ideal number of hours. The association remained similar after additional adjustment for alcohol and caffeine consumption, exercise and depressive symptoms, which makes the pattern less sensitive to those measured factors without changing the study’s observational design.

A large sample can detect relatively small regional differences, so statistical detection and clinical importance are separate questions. The study’s authors noted that the size of the association was modest, although a difference accumulated across several hours could represent a more noticeable scan-level contrast. The analysis did not determine whether that contrast predicts future cognitive change.

Below the positivity threshold, the global picture changed

The study also examined whether sleep was associated with amyloid measurements among participants below the global amyloid-positivity threshold. This analysis included 2,916 people whose global PET SUVR was below 1.10. Being below that cutoff did not mean having no amyloid anywhere in the brain. It meant that the measured global signal was below the study’s classification threshold.

Within this group, nighttime sleep was no longer associated with the global amyloid measure. Yet the regional pattern remained: each additional reported nighttime hour corresponded to a 0.006 lower medial orbitofrontal SUVR and a 0.005 lower anterior cingulate SUVR.

That combination is important because it shows what regional analysis can add to a broad positive-or-negative label. A person may fall below a global threshold while still showing small differences in particular areas. The subgroup comparison does not follow people through the onset or progression of amyloid accumulation; it shows regional associations in cognitively unimpaired participants whose global amyloid measurements were below the positivity threshold.

Daytime sleep told a different regional story

Daytime sleep did not show a global association in the full cohort, and it was not associated with regional amyloid measurements there after the study’s multiple-comparison correction. The result looked different in the below-threshold subgroup.

Among those 2,916 participants, each additional hour of daytime sleep corresponded to a 0.024 higher PET measurement in the posterior cingulate and a 0.013 higher measurement in the precuneus. Daytime sleep was not associated with the subgroup’s global amyloid measure or with the other examined regions.

The contrast was therefore spatial as well as statistical. Nighttime sleep was associated with lower measurements in two frontal regions, while daytime sleep was associated with higher measurements in two posterior regions, and neither pattern extended across the whole brain. These results do not show that naps cause amyloid accumulation, that naps are harmful, or that a nap compensates for lost nighttime sleep. They show that the two reported sleep exposures did not behave as interchangeable measures in this analysis.

Several explanations remain possible

The study cannot determine which direction the relationship runs. Sleep could influence processes involved in amyloid accumulation, amyloid-related changes could disturb sleep, or another factor could affect both. The authors discussed possible sleep-related clearance of emerging amyloid, but this analysis did not test a clearance mechanism or establish that circadian timing prevents deposits.

The sleep measures were based on participants’ reports of recent typical duration rather than an objective recording of sleep. The study also did not specifically screen participants for obstructive sleep apnea, a condition that can affect both sleep estimates and brain amyloid. Statistical adjustment for several habits and depressive symptoms produced similar estimates, but adjustment cannot create the temporal information needed to decide whether one factor preceded the other.

The participants were carefully selected older adults who were cognitively unimpaired and healthy enough to meet the A4 screening criteria. Their results should not be treated as a sleep-based diagnostic test, an estimate of an individual’s dementia risk or a general rule for all sleepers. Longitudinal measurements of sleep, amyloid and cognition are needed to establish whether these early regional patterns persist or predict meaningful change.

What this could change

If improving nighttime sleep were shown to produce sustained reductions in amyloid biomarkers and meaningful cognitive benefit, sleep could become an earlier, relatively accessible point of intervention before symptoms appear. That possibility could eventually give older adults and clinicians another way to address the biology associated with Alzheimer disease while cognitive function remains intact.

The July 2021 study does not demonstrate that benefit. It establishes a human baseline in which reported nighttime sleep and daytime sleep had different regional associations below the global amyloid threshold. The practical promise depends on future evidence that changing sleep, rather than merely accompanying brain changes, can alter the course of amyloid accumulation and cognitive decline.

Key Sources

• Primary source: Association of Sleep and β-Amyloid Pathology Among Older Cognitively Unimpaired Adults.
  DOI or Journal Page · Peer-Reviewed Paper · 2021-07-01