Sleep, beta-amyloid, and the aging brain

Sleep gets short shrift in dementia conversations. The association is real: chronic short sleep at ages 50-70 is linked to ~30% higher dementia risk over 25 years (Whitehall II). The glymphatic system is the leading proposed mechanism, though contested in 2024 literature.

The glymphatic mechanism is contested. A 2024 *Nature Neuroscience* paper (Hablitz et al.) found brain clearance *markedly reduced* during sleep in mice, not increased.

Human RCTs of sleep improvement do not show cognitive benefit. A 2022 meta-analysis of 16 RCTs (PubMed 36370516) found no significant cognitive improvement despite successful sleep improvement. CBT-I reliably improves sleep but has not demonstrated cognitive change in MCI.

The U-shaped sleep curve reflects reverse causation. Dementia disrupts sleep before diagnosis; long sleep (>8 hours) likely signals early neurodegeneration, not a harm of sleep duration (Scientific Reports, 2025).

• Sleep apnea, severely under-diagnosed. Snoring + daytime sleepiness + morning headaches = sleep study.
• Evening alcohol ruins deep sleep architecture even at 1 drink.
• Late caffeine: half-life 5-6 hours; 3pm coffee is in your system at 9pm.
• Anxiety + pain: racing thoughts and undertreated arthritis/back pain/neuropathy.
• Daytime under-stimulation: too much napping, not enough physical or social activity.

• Consistent bedtime + wake time, weekends included.
• Morning light within an hour of waking: 15-30 minutes outside or near a sunny window.
• Daytime activity: walking, social engagement, light strength training.
• Sleep apnea screen if there's any hint of snoring + daytime sleepiness. Cool dark room.
• CBT-I: more evidence than sleep meds; covered by many insurance plans.

• Z-drugs (Ambien, Lunesta): raise fall risk, worsen cognition with chronic use.
• Benzodiazepines (Xanax, Ativan, Restoril): same problems plus dependence risk.
• Long daytime naps. 20 min mid-afternoon is fine; 2 hours wrecks the night.
• Heavy late dinners.

Dr. Rudolph Tanzi (Harvard Medical School; Director of MassGeneral's Genetics and Aging Research Unit) places sleep as the S in SHIELD on the Nothing Left Unsaid podcast, May 2026: deep sleep activates the glymphatic system's nightly 'rinse cycle' to clear amyloid. Chronic short sleep means amyloid is not being flushed out: a mechanism independent of the causal-RCT debate. Sleep is the most testable lifestyle lever in the Lancet modifiable-risk list. The clinical question your doctor is actually asking is whether the CPAP, the CBT-I, the new bedtime worked. Self-report at an annual visit can't answer that.

Free, forever: Familiar's Daily Calls in Family Voices · AI based on Reminiscence Therapy include a dashboard that tracks cognitive markers on every call (vocabulary diversity, repetition rate, name recall, time-orientation, mood) and builds a per-receiver baseline after 30+ days. When something drifts, it's flagged in the post-call SMS summary and on the cognitive-trends chart. You can export the chart as a clinician-formatted PDF and bring it to your loved one's doctor. Doctors miss 6 in 10 cases of mild dementia (Bradford et al. 2009, PCP diagnostic sensitivity 9-41%); the chart gives them month-over-month signal a 15-minute visit can't surface, and lets them see whether the sleep intervention actually shifted cognition. Disease-modifying drugs (lecanemab, donanemab) and lifestyle interventions can start years sooner.