Sleep loss in a mouse model of Alzheimer's disease was driven by the brain's early immune response to amyloid plaques rather than by the plaques themselves, according to a preclinical study published online on June 7, 2026 in Alzheimer's & Dementia. Nicholas J. Constantino, Shannon L. Macauley and colleagues at the University of Kentucky recorded EEG and EMG in female APPswe/PSEN1dE9 (APP/PS1) mice and wild-type littermates at 6 and 18 months of age, mapped amyloid and microglia across the whole brain with light-sheet microscopy, and then depleted microglia with the CSF1R inhibitor pexidartinib (PLX3397). Microglial depletion restored more than two hours of sleep per day without altering amyloid burden. The University of Kentucky said the 14-day course removed 87 percent of the animals' microglia. Every result here is in mice.

The authors report that non-rapid eye movement (NREM) sleep loss appeared as plaques first formed and did not worsen as plaque burden increased, while normal ageing separately reduced REM sleep and eliminated the rebound sleep that normally follows deprivation. Microglial expansion extended beyond plaque-bearing regions into thalamic nuclei and white matter tracts involved in sleep-wake regulation, and the NREM deficit was present at 6 months without overt neuron loss, which the team argues points away from a primarily neuronal mechanism. Amyloid pathology was also accompanied by cortical hyperexcitability and network desynchrony. The paper positions microglia as what it calls a causal, reversible driver of amyloid-associated sleep loss in this model, and sleep and EEG measures as candidate biomarkers of presymptomatic disease.

A commentary by Brendan P. Lucey and Michael J. Howell, published in Nature Reviews Neurology on August 4, 2026, pairs the mouse work with a human genetic study of the astrocytic water channel aquaporin-4 by Tenielle Porter and colleagues in the AIBL cohort, published in Alzheimer's & Dementia on May 29, 2026, which found a broad relationship between AQP4 variation and Alzheimer's-related measures with sleep as a potential modifier. The commentary's framing is that glia-mediated mechanisms, rather than plaque burden, may disrupt sleep. The Porter study is observational and genetic, and does not establish cause.

Several cautions sit against the headline number. Pexidartinib is licensed in the United States, since August 2, 2019, only for a rare joint tumour, tenosynovial giant cell tumour, and carries a boxed warning for serious liver injury; in this study it was a laboratory tool given to mice, not a treatment tested for sleep in people. Microglia also carry out protective functions, and the paper does not test whether long-term depletion is safe or whether restored sleep changes the course of pathology. The authors disclose that two patent applications relevant to the manuscript have been filed by the University of Kentucky naming Macauley as inventor.