13/08/2026
Such an interesting read for many reasons, sleep architecture, strengths and limitations of research, emerging ideas.
What if the single most effective thing we can do to reduce the risk of dementia was not a drug, a supplement, a herb, or even a brain-training program, but instead was getting the right kind of sleep? A growing body of research suggests that during deep, uninterrupted non-rapid eye movement (NREM) sleep the brain performs one of its most important jobs: flushing away toxic waste products that accumulate during the day. A major new review argues that disruption of this nightly cleaning process may be the common pathway that links ageing, chronic stress, depression, cardiovascular disease, toxins and poor sleep to an increased risk of dementia.
The review proposes that while we sleep, the brain switches into a remarkable "housekeeping mode." During wakefulness, neuromodulators operate largely independently to support behavior and cognition. By contrast, sleep reorganises their activity into a coordinated brain rhythm. Every 50 seconds or so, key brain chemicals, including noradrenaline, serotonin, dopamine and acetylcholine, begin pulsing together in a perfectly coordinated rhythm. These waves gently expand and contract tiny blood vessels, creating a pumping action that flushes cerebrospinal fluid through the brain and removes toxic waste, including the Alzheimer proteins amyloid-beta and tau. In other words, these neuromodulators become vasoactive agents that drive a slow vasomotion, providing the mechanical force that supports glymphatic clearance of metabolic waste.
The dual roles of neuromodulators—as regulators of behavior and mental state during wakefulness and as organisers of sleep architecture and glymphatic clearance during sleep—define a vulnerable system in which imbalance may impair both sleep and brain clearance. This framework may provide a unifying explanation for several long-standing observations. For example, psychiatric disorders, including schizophrenia, bipolar disorder and depression, are characterised by disruptions in neuromodulator signaling and sleep architecture, which are predicted to alter slow vasomotion during sleep. Such changes may impair glymphatic clearance, thereby promoting the accumulation of pathological proteins and thus contributing to the increased risk of neurodegenerative disease that has been observed with these conditions.
As noted above, a key point of the article is that many of the things already known to increase dementia risk disrupt this nightly cleaning cycle. This could explain why so many seemingly unrelated conditions end up increasing the risk of Alzheimer disease. Moreover, an extensive body of literature has implicated age-related decline in glymphatic clearance as a driver of proteinopathies and dementia.
Sleep architecture is the carefully organised cycling between different sleep stages that allows the brain to carry out its essential maintenance functions. Healthy sleep architecture ensures repeated episodes of deep NREM sleep, allowing the cleansing cycles to occur throughout the night. Fragmented sleep, reduced deep sleep, frequent arousals or disrupted transitions between sleep stages may all compromise this process.
So perhaps the most exciting implication is that sleep architecture might become an important target for protecting long-term brain health. Instead of simply asking whether people sleep enough, researchers are now investigating whether they are getting the right kind of sleep, including the deep, rhythmic sleep that allows the brain's cleaning system to work at full capacity.
This intriguing idea remains an emerging hypothesis. The review pulls together a large body of animal and human research, but there is not enough evidence yet to prove the mechanism directly in people. Even so, it offers one of the most compelling explanations yet for why protecting healthy sleep may become the best investments we can make for future brain health.
Another fascinating observation by the authors is that heart rate variability (HRV) may provide a simple, non-invasive window into the brain's nightly cleaning system, because the same neuromodulator oscillations that drive glymphatic clearance during deep NREM sleep also regulate heart rate. Reduced HRV is already associated with ageing, chronic stress, depression, cardiovascular disease and an increased risk of dementia, suggesting it may reflect the disruption of this fundamental sleep rhythm. If confirmed, HRV could become a practical biomarker for identifying individuals at increased dementia risk and for monitoring whether interventions restore healthy sleep physiology and brain clearance.
What about the role of herbs? Very few herbs have been investigated in clinical trials in terms of a specific increase in NREM (especially N3 slow-wave sleep), as measured by polysomnography (EEG). Most herbal sleep studies have instead measured subjective sleep quality, sleep latency, total sleep time or actigraphy, rather than sleep architecture itself.
One of the most fascinating findings in herbal sleep research comes from a small double blind polysomnography study comparing European valerian (Valeriana officinalis) with Mexican valerian (Valeriana edulis) in patients with insomnia. This study measured sleep architecture directly using overnight EEG recordings. Both herbs increased delta (slow-wave, N3) sleep, the deepest and most restorative stage of NREM sleep, while reducing lighter stages of NREM sleep. Mexican valerian also reduced the number of nocturnal awakenings, whereas European valerian produced a somewhat greater increase in REM sleep.
Ashwagandha (Withania somnifera) has probably the best clinical evidence overall for insomnia. Repeated RCTs and a meta-analysis show improvements in sleep efficiency, sleep duration, wake after sleep onset and sleep quality. However, there is no evidence yet that it specifically increases N3 sleep. Its benefits may result from reducing hyperarousal and stress, allowing more normal sleep architecture to emerge, but this remains to be proven in clinical trials.
For more information see: https://pubmed.ncbi.nlm.nih.gov/42166599/
https://pubmed.ncbi.nlm.nih.gov/11731907/
