Microglia, Not Plaques, Drive Alzheimer's Sleep Loss

For families living with Alzheimer’s disease sleep disruption is rarely an abstract symptom. It is the parent awake at 2 a.m., the spouse pacing after midnight, the caregiver trying to guess whether the problem is pain, fear, medication, daylight, or the disease itself. So when Alzheimer’s research headlines keep returning to amyloid plaques, a plain question follows: if plaques are the central target, why does sleep so often remain broken?

A July 2026 University of Kentucky/Macauley lab study gives that question a sharper biological answer. In Alzheimer’s model mice, the researchers temporarily depleted about 87% of microglia—the brain’s immune cells—over 14 days using Pexidartinib. EEG/EMG monitoring then showed more than two hours of restored non-REM sleep per day, even though amyloid plaque burden did not change.[1]

That last clause matters. The study did not show that plaques are irrelevant to Alzheimer’s disease. It showed something narrower and more useful for sleep: in this mouse model, reducing overactive immune-cell activity repaired a large share of NREM sleep loss without clearing plaques.

What the Mouse Study Actually Changed

The useful part of this study is not that it found another association between poor sleep and Alzheimer’s pathology. That association has been around long enough to become mushy in public conversation. The cleaner contribution is causal within the animal model: alter the microglial population, then measure sleep directly.

The researchers used Pexidartinib, a drug developed for cancer, to remove most microglia in Alzheimer’s model mice over a 14-day period. After roughly 87% depletion, the mice gained back more than two hours of NREM sleep per day. The sleep was measured with EEG/EMG monitoring, not inferred from stillness or activity alone.[1]

That distinction is important because mice can look quiet without being in the same sleep state. EEG and EMG let researchers separate brain-wave sleep from muscle tone and movement. For a sleep study, that is the difference between watching the bedroom door and reading the night itself.

For caregivers, the implication is not that a new sleep drug has arrived. It has not. The implication is that one stubborn symptom may be coming from a mechanism that plaque-centered treatment stories do not directly address.

The Plaque Story Did Not Disappear. It Got Less Sufficient.

Amyloid plaques remain one of the defining pathological features of Alzheimer’s disease. The mistake is treating a defining feature as if it automatically explains every lived symptom. Sleep is where that shortcut starts to wobble.

The Macauley lab result puts pressure on a simple plaque-burden explanation in two ways. First, the mice recovered more than two hours of NREM sleep per day after microglia depletion while plaques stayed in place. Second, plaque burden more than doubled between 6 and 18 months, but sleep disruption did not worsen proportionally.[1]

That second point is the uncomfortable one. If plaques were driving sleep loss in a direct, dose-like way, a large rise in plaque burden should have been matched by a large worsening in sleep. The study instead suggests a ceiling effect: early microglial activity may help establish the sleep deficit, and later plaque accumulation may not keep pushing sleep downward at the same rate.

This does not make amyloid biology unimportant. It does make it harder to tell families that progress against plaques should be expected to fix the night. A treatment can affect one part of Alzheimer’s pathology and still leave sleep disruption largely untouched if the sleep circuit is being disturbed through immune-cell activity.

Why NREM Sleep Is the Signal to Watch

The study’s emphasis on NREM sleep is not a technical footnote. Alzheimer’s pathology in this work selectively reduced NREM sleep, while normal aging primarily reduces REM sleep.[1]

That difference gives the finding practical shape. Older adults often sleep differently than they did decades earlier, and not every change points toward dementia. But a disease-linked loss of restorative NREM sleep is a different pattern from the REM changes commonly associated with aging. It gives researchers a more specific sleep signature to study instead of treating “bad sleep” as one broad bucket.

It also explains why the result feels relevant to families who are not helped by general sleep advice alone. Light exposure, daytime activity, nighttime routine, pain control, and medication review still matter. They can reduce friction around sleep. But they do not answer why Alzheimer’s pathology itself appears to pull NREM sleep out from under the person.

Microglia Are Not Villains. Overactivation Is the Problem.

Microglia are the brain’s resident immune cells. They respond to injury, debris, and disease-related stress. In Alzheimer’s pathology, the question is not whether they are present, but what state they are in and what their activity does to nearby neural systems.

The Pexidartinib experiment is deliberately blunt: remove most of the cells, then see whether sleep changes. It did. But no one should confuse that laboratory move with a desirable human sleep treatment. The therapeutic idea that follows is subtler: calm overactive microglia without eliminating them.

That is why the next step is not “try Pexidartinib for sleep.” Pexidartinib is a cancer drug with significant human side effects, including liver toxicity and edema, and in this study it worked by depleting microglia rather than tuning their behavior.[1]

The research team is now testing Metformin and Stiripentol as existing-drug candidates that might calm microglial overactivity without removing the cells entirely.[1] That pipeline is interesting because it aims at the symptom-relevant mechanism. It is still under investigation, not clinical guidance.

Where This Leaves the Sleep-Alzheimer’s Loop

This study does not erase the bidirectional model of sleep and Alzheimer’s disease. Poor sleep may still contribute to Alzheimer’s-related pathology through mechanisms such as impaired clearance systems, and Alzheimer’s pathology may still worsen sleep. The new result adds detail to one side of that loop: how Alzheimer’s pathology can disrupt sleep through overactive microglia.

That matters because loops are hard to treat when every arrow is vague. If NREM sleep loss is partly driven by immune-cell overactivation, then researchers have a more concrete target than “improve sleep” or “reduce plaques.” They can ask whether changing microglial state restores a measurable sleep pattern.

The same logic may eventually help screening. EEG signatures of Alzheimer’s-related sleep disruption could become a portable, noninvasive biomarker.[1] That possibility is still future-facing, but it fits the study’s strength: the sleep change was physiological and state-specific.

What Families Should—and Should Not—Take From This

The most responsible takeaway is modest but not small. Alzheimer’s-related sleep disruption may be driven less by plaque burden itself than by overactive microglia disturbing NREM sleep. In mice, changing that immune-cell system restored a meaningful amount of sleep without changing plaques.[1]

The equally important boundary is that this is not a human trial result. It does not justify asking for Pexidartinib as a sleep treatment. It does not prove that Metformin or Stiripentol will improve sleep in people with Alzheimer’s disease. It does give clinicians and researchers a better question to pursue: when a person with Alzheimer’s loses NREM sleep, is immune-cell overactivation part of the machinery behind it?

For day-to-day care, families still need practical help with pain, nighttime movement, toileting, medication timing, bedroom safety, and caregiver backup. For practical strategies on managing physical sleep disruptors in dementia, see our companion guide, How Physical Dementia Symptoms Disrupt Sleep and How to Help. Mechanism does not replace bedside management; it explains why bedside management can be necessary even when the larger Alzheimer’s treatment conversation sounds plaque-focused.

References

1. Alzheimer's breakthrough: Scientists restore two hours of sleep without clearing brain plaques, ScienceDaily, July 20, 2026.