Three neurologists from Innsbruck argue that sleep should be treated as a primary target for neurological disease prevention, sitting alongside diet and exercise in the way cardiologists treat lipids and blood pressure. Their case rests on three mechanistic pillars: sleep consolidates memory and prunes synapses, the glymphatic system clears metabolic waste including amyloid-beta preferentially during deep sleep, and circadian alignment governs both. They then assemble observational evidence that disturbed sleep is not merely a symptom of neurodegeneration but appears in the record 6 to 12 years before diagnosis, that sleep-disordered breathing roughly doubles to triples stroke risk, and that treating it with CPAP is associated with lower dementia incidence. The conclusion is a clinical call to action: screen for sleep disorders routinely, treat apnea and insomnia aggressively, and expect wearables plus machine learning to move detection years earlier.
For most of modern medicine, sleep has been the thing you catch up on after the important work is done. A review out of the Medical University of Innsbruck makes the case that this ordering is backwards, and that neurology has been quietly accumulating the evidence to prove it for two decades.
The big idea is a shift in causal direction. Clinicians have long known that people with Alzheimer’s disease and Parkinson’s disease sleep badly. The assumption was that a degenerating brain sleeps poorly, which is uncontroversial and uninteresting. What the review pulls together is a body of longitudinal work suggesting the arrow also runs the other way, and runs early. Polysomnographic changes such as reduced REM sleep, reduced deep sleep and poor sleep efficiency are detectable up to 12 years before a neurodegenerative diagnosis. Accelerometer data from the UK Biobank shows disrupted circadian rhythm preceding dementia, Parkinson’s and anxiety by 6 to 11 years. In the Framingham Heart Study, a lower proportion of REM sleep predicted dementia 12 years out.
The mechanism most often invoked is the glymphatic system, a waste clearance network discovered barely a decade ago. During deep sleep the space between brain cells expands by roughly 60 percent, allowing cerebrospinal fluid to flush through tissue and carry out metabolic debris, amyloid-beta among it. A single night without sleep measurably impairs this clearance in humans and raises amyloid in cerebrospinal fluid by 25 to 30 percent. The picture is tidy, and the review is honest enough to note that a 2024 mouse study found brain clearance markedly reduced rather than increased during sleep, which is a direct contradiction nobody has yet resolved.
The most actionable material concerns sleep-disordered breathing, which affects around 20 percent of the population and is eminently treatable. Untreated, it is associated with a two to threefold rise in stroke risk in two large cohorts. Among 53,000 Medicare beneficiaries, those who used and adhered to positive airway pressure therapy had lower odds of an Alzheimer’s diagnosis. Whether that reflects treatment or reflects the kind of person who complies with treatment is not something an observational cohort can answer.
The authors close on wearables and machine learning, arguing that consumer sleep trackers may soon detect prodromal fragmentation years before symptoms. That is plausible and unvalidated in equal measure. The defensible take-home is narrower and older: sleep is not recovery time appended to life, it is a maintenance process the brain cannot run any other way.
Actionable Insights
Four things in this review are worth acting on:
Get tested for sleep apnea if you snore, wake unrefreshed, or have a thick neck or high blood pressure. This is the single largest lever here. Untreated sleep-disordered breathing carries roughly a two to threefold higher stroke risk. Translated: if your ten-year stroke risk would otherwise be about 4 percent, untreated apnea plausibly puts it near 10 percent. Roughly one person in 15 to 20 with untreated apnea suffers a stroke that treatment might have prevented. This is a bigger effect than most supplements ever tested.
Aim for seven to nine hours, and treat the lower bound as the real one. Sleeping under six hours persistently in your fifties and sixties was associated with about 30 percent higher dementia risk. That sounds dramatic but is modest in absolute terms, roughly two extra cases per hundred people over 25 years, or one case per 50 people.
Stop caffeine early. Two hundred milligrams, about one strong coffee, taken in the early evening delays your melatonin rhythm by about 40 minutes. That is a real circadian shift, not a subjective one.
Treat insomnia with cognitive behavioural therapy first, not pills. It is first-line, and preliminary data links it to slower amyloid accumulation.
Context and Source
- Open Access Paper: Sleep as the Foundation of Brain Health
- Authors: Abubaker Ibrahim MD, Birgit Hogl MD, Ambra Stefani MD PhD
- Institution: Department of Neurology, Medical University of Innsbruck, Austria
- Journal: Seminars in Neurology
- Article type: narrative review
- Journal impact evaluation. The impact score of this journal is 2.2 (Journal Impact Factor, Web of Science 2025 release), evaluated against a typical high-end range of 0 to 60+ for top general science, therefore this is a Low impact journal. Two contextualising notes are fair. Against the field-specific range for Clinical Neurology rather than general science, it sits at rank 95 of 198, which is Q2, or middling rather than poor.