A new study published in Neurology reveals a complex interplay between the wake-promoting neuropeptide orexin, non-REM sleep oscillations, and cognitive decline in Alzheimer disease. The findings suggest that the brain’s own sleep architecture may buffer some of the harm caused by an overactive orexin system.
Researchers led by Anna Paez at the University of Oxford, in collaboration with Concordia University Montreal and Hospital Universitari Santa Maria in Lleida, Spain, enrolled 60 participants with biomarker-confirmed mild-to-moderate Alzheimer disease. Each underwent overnight polysomnography followed by morning cerebrospinal fluid sampling. The team measured orexin-A, amyloid-beta 42, phosphorylated tau 181 (pTau181), total tau, and the neuroinflammatory marker YKL-40. Cognitive assessments and neuropsychiatric evaluations were tracked longitudinally over 36 months.
The results point to orexin as a potentially adverse factor in Alzheimer disease, but with a surprising twist. The adverse associations were not uniform across participants. They were significantly moderated by the brain’s own NREM oscillatory activity.
The orexin system in AD
Orexin (also known as hypocretin) is a neuropeptide produced in the lateral hypothalamus. It promotes wakefulness, regulates arousal, and stabilizes the sleep-wake boundary. In Alzheimer disease, the orexin system is often dysregulated. Prior work has linked elevated orexin to sleep fragmentation, circadian disruption, and even increased amyloid-beta deposition in animal models.
The new study provides human evidence for these connections. Higher CSF orexin concentrations were associated with worse cognitive performance on the ADAS-Cog (beta = 0.014, 95% CI 0.003 to 0.024) and MMSE (beta = -0.01, CI -0.011 to -0.004). Higher orexin also predicted greater neuropsychiatric symptoms on the NPI (beta = 0.03, CI 0.011 to 0.041). Biologically, elevated orexin correlated with higher pTau181 (beta = 0.11, CI 0.04 to 0.19) and YKL-40 (beta = 0.37, CI 0.17 to 0.57), suggesting links to both tau pathology and neuroinflammation.
NREM oscillations as neural resilience
Sleep spindles and slow oscillations are hallmark features of non-REM sleep. Spindles (bursts of 11-16 Hz activity generated by thalamocortical circuits) and slow oscillations (cortical waves below 1 Hz) are known to support memory consolidation and synaptic homeostasis. In Alzheimer disease, both are typically diminished.
The study found that longer slow oscillation duration and higher sleep spindle density were associated with lower CSF orexin levels. The association was substantial. For spindle density, the coefficient was beta = -187.37 pg/mL (95% CI -344.93 to -29.80). This suggests that preserved NREM oscillatory activity is linked to a less dysregulated orexin system. However, the more striking finding involved moderation, not mere correlation.
The moderating interaction
The core discovery of the study is a significant orexin-by-spindle-slow-oscillation interaction. Greater NREM oscillatory activity attenuated the adverse association between orexin and cognitive outcomes. In other words, participants who maintained robust sleep spindle density and slow oscillation activity showed a weaker link between high orexin and poor cognition. This effect was independent of amyloid-beta 42 and phosphorylated tau levels.
This is a critical nuance. It means that the relationship between orexin dysregulation and cognitive harm is not fixed. It depends on the state of the brain’s sleep oscillatory machinery. When that machinery is relatively preserved, it may act as a kind of neural buffer, absorbing some of the impact that a hyperactive orexin system would otherwise have on memory and executive function.
The finding aligns with a growing literature on what some researchers call cognitive or neural resilience. Not every brain with Alzheimer pathology experiences the same rate of decline. Sleep quality is increasingly recognized as one factor that distinguishes resilient from vulnerable trajectories.
Why it matters
The study has several implications. First, it identifies orexin as a potential therapeutic target in Alzheimer disease. Drugs that antagonize orexin receptors (the dual orexin receptor antagonists, or DORAs, already approved for insomnia) could theoretically reduce the adverse link between elevated orexin and cognitive decline. A clinical trial testing this possibility would be a logical next step.
Second, the moderating role of NREM oscillations suggests that nonpharmacological interventions aimed at preserving or enhancing sleep spindles and slow oscillations may have cognitive benefits. These could include auditory stimulation synchronized to slow oscillations, transcranial electrical stimulation, or behavioral sleep optimization. If the brain’s own oscillatory activity can buffer orexin-related harm, then protecting that activity becomes a therapeutic priority.
Third, the study adds to the evidence that sleep is not merely a marker of Alzheimer pathology but an active physiological participant in the disease process. The interactions between orexin, tau pathology, neuroinflammation, and sleep oscillations paint a picture of a complex system. Interventions at any node may have ripple effects.
Limits
The sample size is modest at 60 participants, and the findings require replication in larger, more diverse cohorts. The study is cross-sectional in its biomarker measurements, so causal direction cannot be firmly established. Elevated orexin could drive tau pathology and cognitive decline, or the disease process could dysregulate the orexin system as a secondary phenomenon. Longitudinal biomarker sampling would help resolve this. Finally, all participants had mild-to-moderate Alzheimer disease. Whether similar relationships hold in preclinical or prodromal stages is unknown.
Bottom line
Elevated CSF orexin is associated with worse cognition, more neuropsychiatric symptoms, higher tau pathology, and greater neuroinflammation in Alzheimer disease. However, preserved NREM sleep oscillatory activity (spindles and slow oscillations) attenuates these adverse associations. The findings highlight sleep oscillations as a potential mechanism of neural resilience and suggest that orexin antagonism or sleep-enhancing interventions could offer therapeutic benefit.
Source
Paez A, Piñol-Ripoll G, Carnes-Vendrell A, Dakterzada F, Barbé F, Zetterberg H, Dang-Vu TT. Orexin, Sleep, and Cognition in Alzheimer Disease: Non-REM Oscillatory Activity and Neural Resilience. Neurology. 2026 Aug 11;107(3):e218307. doi:10.1212/WNL.0000000000218307. PMID: 42447420.