It’s not the plaques keeping the brain awake. At least, not directly.
For years, the narrative around Alzheimer’s was simple enough. Sticky amyloid proteins build up. Neurons die. Sleep suffers. But that linear story might be wrong. Or at least, incomplete.
University of Kentucky researchers just published a study in Alzheimer’s & Dementia that flips the script. They found that microglia —the brain’s immune cells—are the actual drivers of sleep disruption in mice with Alzheimer’s-like disease. Think of it like a fire sprinkler system going off in a room with a tiny candle. The water causes more damage than the flame.
The implication? If you want to fix the sleep, you might not need to clear the plaques first. You just need to quiet the immune system.
Why microglia cause insomnia in Alzheimer’s models
Macauley and her team, including first author Nicholas J. Constantino (who recently finished his doctorate there), suspected the immune response was louder than the disease itself. To prove it, they had to separate normal aging from Alzheimer’s pathology.
They compared mice genetically prone to amyloid plaques with wild-type mice aging normally. Both groups got tiny head-mounted devices. These recorded EEG (brain waves) and EMG (muscle tone). This setup let them pinpoint exactly when the mice were awake, in deep sleep, or dreaming.
Then they used light-sheet microscopy. This technique renders brain tissue transparent, allowing lasers to map the 3D landscape of plaques and cells. It gave them a view of what Macauley called immune cells “partying all night.”
Here is the twist. The researchers didn’t just observe. They intervened.
They gave the mice Pexidartinib, a drug originally tested for cancer. It blocks the survival signals microglia need. Within 14 days, about 87% of those immune cells vanished from the brain. Temporarily.
The result? Sleep came back. The mice gained more than two hours of sleep every day.
“Basically, we showed that it is not the platues themselves… but actually microglia.”
How sleep loss plateaus despite worsening plaque burden
This part is the weirdest.
Constantino expected the worse the plaques got, the worse the sleep would get. That’s how it usually works. More disease equals more damage.
But the data said no.
At six months, when plaques first appear, the mice lost significant amounts of restorative non-REM (NREM) sleep. By 18 months, the plaque burden had more than doubled. Yet the sleep disruption didn’t get any worse. It hit a ceiling.
The team calls this a “ceiling effect.” The initial inflammatory blast from the microglia seems to do all the damage early on. Adding more plaque later doesn’t add more chaos to the sleep cycle. It just sits there.
This distinction matters because aging affects sleep differently. Normal aging eats away at REM sleep—the dreamy kind linked to memory. Alzheimer’s pathology specifically murders NREM sleep. That’s the deep, cleaning sleep. The wash cycle.
When you lose that, toxins don’t get flushed out. The brain stays dirty. The inflammation gets worse. It’s a feedback loop.
Which drugs restore sleep without clearing amyloid
The most critical finding is that removing microglia restored sleep without reducing the amyloid plaques.
The plaques were still there. But the immune system was quiet. And the mice slept.
This suggests the inflammation itself is a reversible trigger. You don’t have to solve the entire Alzheimer’s puzzle to get a few hours of rest. You just have to mute the noise.
Macauley’s lab is now looking for ways to dial down that microglial activity without wiping the cells out entirely. They are testing existing drugs like Metformin (for diabetes) and Stiripentol (for seizures). These might change how microglia metabolize fuel, essentially putting them on a diet so they stop overreacting.
If they can quiet the immune response, they might restore sleep, attention, and cognition before memory loss even starts.
Where portable EEG fits into early screening
So, how do we know if a human is losing that specific NREM sleep?
Current testing is expensive. Invasive. You have to travel to major hospitals.
The researchers propose a shift. They identified specific electrical brain patterns that distinguish Alzheimer’s-related pathology from normal aging. This means EEG could become a cheap, accessible biomarker.
Imagine a portable EEG system in a local Kentucky clinic. Or in a patient’s home. It monitors sleep patterns longitudinally. No expensive scans required initially.
“Portable EEG systems could allow us to monitor… without the initial need for expensive or invasive test,” Macauley said.
It’s a practical step. Catch the sleep loss early. Treat the inflammation. Maybe break the cycle.
The study came from a lab that encourages “calculated risk-taking.” Macauley keeps a Wayne Gretzky quote on her wall. Take the shot. Follow the data even when the hypothesis is wrong.
Constantino learned to embrace failure. His original guess was likely off. But the data pointed to the immune system. They followed it.
It wasn’t about fixing the plaque. It was about stopping the party.
