Sleep Disruption and Neurodegeneration: Mechanisms, Clinical Manifestations, and Therapeutic Implications
Sleep is increasingly recognised as a crucial modulator of brain health, with emerging evidence supporting a bidirectional relationship between sleep disturbance and neurodegenerative processes. The architecture of sleep – comprising non‑rapid eye movement (NREM) and rapid eye movement (REM) stages arranged in ~90‑minute cycles – underpins restorative processes such as synaptic downscaling, memory consolidation, and metabolic clearance. The circadian system, anchored by the suprachiasmatic nucleus (SCN) and driven by a molecular clock (BMAL1, CLOCK, PER, CRY), coordinates the timing of these sleep‑related events and governs rhythmic fluctuations in hormone release, body temperature, and autonomic tone.
Neurodegenerative diseases, including Alzheimer disease (AD), Parkinson disease (PD), dementia with Lewy bodies (DLB), multiple system atrophy (MSA), frontotemporal lobar degeneration (FTLD), and Huntington disease (HD), frequently present with sleep abnormalities that may predate motor or cognitive signs. Conversely, chronic sleep disruption has been identified as an independent risk factor for the development and progression of neurodegenerative pathology, creating a feed‑forward loop in which sleep loss accelerates protein aggregation, neuroinflammation, and oxidative stress, while neurodegeneration further erodes sleep quality. The present review integrates electrophysiological, molecular, and clinical evidence to delineate how sleep architecture and the glymphatic system intersect with neurodegeneration, and evaluates the promise of sleep‑focused interventions as disease‑modifying strategies.
Human sleep is subdivided into three NREM stages (N1, N2, N3) and REM. N1 represents light sleep, transitional between wakefulness and sleep; N2 is characterized by sleep spindles and K‑complexes, constituting ~45 % of total sleep in adults; N3 (slow‑wave sleep, SWS) is the deepest stage, dominated by high‑amplitude, low‑frequency (0.5–4 Hz) delta oscillations that support synaptic homeostasis and the restorative anabolic processes essential for neuronal maintenance. REM sleep, the stage of vivid dreaming, is marked by cortical desynchronization, rapid eye movements, muscle atonia (except for extraocular and respiratory muscles), and autonomic variability. The cyclic alternation (≈4–6 cycles per night) reflects the interplay of arousal‑promoting nuclei (locus coeruleus, dorsal raphe) and sleep‑active populations (ventrolateral preoptic area) that gate transitions between stages and coordinate the temporal sequence of NREM–REM periods.
The circadian clock resides in the SCN and synchronises peripheral clocks in virtually every organ, including the brain. At the cellular level, the transcriptional‑translational feedback loops of the clock genes BMAL1/CLOCK drive rhythmic expression of target genes such as PER1/2/3 and CRY1/2, whose proteins accumulate and repress their own transcription. This molecular oscillator generates ~24‑hour rhythms in melatonin secretion, cortisol peaks, body temperature minima, and the propensity for sleep. Misalignment of the internal clock – as occurs with shift work, light pollution, or aging – correlates with altered sleep timing, reduced SWS, and increased neuronal vulnerability to metabolic stress.
During wakefulness, synaptic strength整体的增强 leads to high metabolic demands; sleep, especially SWS, provides a synaptic downscaling period that eliminates redundant synapses, restores neural circuitry, and consolidates newly acquired information. The process is reflected in the reduction of cortical excitability and the scaling of‑amplitudes of evoked potentials across successive NREM cycles, underpinning the cognitive benefits of sleep.
The glymphatic system is a macroscopic waste‑clearance pathway whereby cerebrospinal fluid (CSF) enters the brain parenchyma along perivascular routes, driven by arterial pulsations and astroglial water channels (aquaporin‑4, AQP4) located on the end‑feet ensheathing the vasculature. Bulk flow facilitates the removal of soluble metabolites, including amyloid‑β (Aβ) and tau, from the interstitial space into the CSF and subsequently into the venous system. In rodents, glymphatic flux is ** enhancer** by natural sleep and anesthesia, with AQP4 polarization to perivascular end‑feet being essential for efficient clearance. In humans, diffusion‑weighted MRI has revealed increased glymphatic perivascular influx during NREM SWS, consistent with the notion that sleep‑dependent changes in extracellular volume and oscillatory slow‑wave activity augment waste removal.
Polysomnography (PSG) in early AD consistently reveals reduced sleep efficiency, prolonged wake after sleep onset (WASO), and a decrease in N3 with a relative increase in N1/N2 proportion. Several studies also report shortened REM latency and a reduction in REM proportion, reflecting disrupted REM sleep architecture. A meta‑analysis of 19 PSG datasets showed that AD patients exhibit a 20‑30 % decline in SWS compared with cognitively healthy controls, and these deficits correlate with disease severity [1] (PubMed: 22921917). Importantly, the magnitude of SWS loss predicts future cognitive decline, suggesting a causal link.
While RBD is classically associated with synucleinopathies, occasional REM‑without‑atonia (RWA) has been documented in AD, albeit at lower prevalence (<10 %). The presence of RBD in AD often signals comorbid Lewy body pathology, indicating an overlap with DLB. Nonetheless, the occurrence of RBD may herald a more aggressive disease phenotype, with faster progression of cognitive deficits.
Fragmented sleep, manifest as frequent arousals and interstate transitions, is observed in up to 50 % of AD patients and is linked to negative outcomes such as higher caregiver burden and accelerated institutionalisation. Circadian rhythm disturbances, including advanced sleep phase, reduced amplitude of melatonin secretion, and fragmented 24‑hour activity patterns, are pervasive in AD and arise from degeneration of the SCN and dysregulation of clock gene expression. Additionally, obstructive sleep apnea (OSA) is highly prevalent in AD (≈40‑70 % in clinic‑based cohorts). Studies using home sleep monitoring have demonstrated that OSA severity correlates with higher CSF Aβ42/40 ratios and increased cortical amyloid deposition on PET, underscoring a mechanistic role for intermittent hypoxia in amyloidogenesis [2] (PubMed: 24136970).
Large‑scale longitudinal investigations, such as the Mayo Clinic Study of Aging and the Baltimore Longitudinal Study of Aging, have shown that baseline sleep fragmentation predicts incident AD risk, even after adjusting for age, APOE ε4 status, and vascular comorbidities (HR ≈ 1.5). Moreover, reduced REM sleep duration has been associated with higher baseline CSF tau and subsequent atrophy of the entorhinal cortex. The Harvard Aging Brain Study corroborated that individuals with lower sleep efficiency exhibit steeper declines in episodic memory and greater accumulation of cortical Aβ over a 5‑year follow‑up.
Idiopathic REM sleep behavior disorder (iRBD) is now regarded as a core prodromal marker of the α‑synucleinopathy spectrum, with >80 % of iRBD patients developing PD, DLB, or MSA within 10‑15 years. Polysomnographic studies reveal loss of REM atonia, quantified via the RBD‑RS (REM sleep atonia index), and the presence of complex motor activity during REM. The presence of iRBD confers a ~3‑fold increased risk of incident PD, and neuropathologic studies in iRBD cases demonstrate early α‑synuclein deposition in the pontine sublaterodorsal nucleus, the region responsible for REM atonia generation.
Insomnia (particularly sleep initiation difficulty) is reported by 30‑50 % of PD patients and is often exacerbated by dopaminergic medications, mood disturbances, or nocturnal akinesia. Conversely, hypersomnolence can emerge in up to 20 % of patients, reflecting disease‑related hypothalamic orbrainstem dysfunction. Periodic limb movements in sleep (PLMS) are frequent (≈30‑50 %) and correlate with motor severity and dopaminergic neuron loss. Importantly, the coexistence of PLMS and iRBD further heightens the risk of progression to PD or MSA.
Longitudinal actigraphy and PSG data indicate that worsening sleep fragmentation and reduced REM atonia mirror dopaminergic denervation (as measured by DaTscan) and the progression of motor disability (UPDRS). Moreover, PD patients with severe sleep disruption exhibit higher levels of CSF α‑synuclein oligomers, suggesting that sleep loss may reflect ongoing neuropathologic burden.
DLB patients exhibit prominent RBD (≈80 % prevalence), often preceding cognitive decline by years. PSG demonstrates reduced sleep efficiency, loss of SWS, and frequent REM atonia loss. A polysomnographic REM sleep behavior disorder with visual hallucinations is considered a core diagnostic feature. Additionally, severe circadian rhythm disruption, manifest as daytime somnolence and nighttime agitation, is common and contributes to the fluctuating cognition typical of DLB.
MSA is characterised by REM sleep behavior disorder (often severe), REM‑related stridor, and frequent sleep apnea. The prevalence of RBD in MSA approaches 90 % and may precede the autonomic and motor manifestations by decades. Polysomnographic studies also reveal significativawakenings, reduced N3, and periodic limb movements that may worsen autonomic dysfunction.
Patients with behavioral variant FTLD often present with marked sleep fragmentation, reduced SWS, and increased daytime sleepiness. The pattern may be linked to frontal–subcortical circuit involvement and dysregulation of hypothalamic orexin/hypocretin pathways. Sleep disturbances often correlate with disinhibition and executive dysfunction, and may precede frank behavioral changes in some cases.
HD is associated with progressive loss of sleep architecture, including reduced REM latency, decreased REM proportion, fragmented sleep, and abnormal circadian rhythmicity. Polysomnographic studies in early‑stage HD reveal decreased total sleep time, increased N1, and impaired sleep efficiency. Actigraphy has demonstrated advanced sleep phase and decreased amplitude of rest‑activity cycles, reflecting hypothalamic dysfunction. Importantly, sleep disruption correlates with motor and psychiatric symptom severity.
AQP4 expression on perivascular astrocyte end‑feet is essential for glymphatic clearance. In animal models of AD, knockdown or mis‑localisation of AQP4 reduces Aβ clearance and accelerates plaque formation, while sleep deprivation further diminishes AQP4 polarity, linking sleep loss to impaired waste removal [3] (PubMed: 25837953). In PD models, α‑synuclein aggregates appear to impede glymphatic influx, possibly by altering astrocytic water handling.
Aβ is cleared predominantly via the glymphatic system, and its concentration in interstitial fluid rises with sleep deprivation. Human PET studies reveal a diurnal variation in cortical Aβ burden, with higher signal after sleep fragmentation. Tau, once hyperphosphorylated, propagates along neural circuits; glymphatic transport of CSF‑derived tau into the parenchyma may contribute to templated seeding. Experimental work demonstrates that slow‑wave sleep enhancement boosts CSF tau clearance, providing a mechanistic rationale for why SWS loss correlates with tau accumulation.
Sleep loss activates microglia and elevates pro‑inflammatory cytokines (IL‑1β, TNF‑α) in the brain, promoting a neuroinflammatory milieu that facilitates protein aggregation. In AD, elevated CSF IL‑6 and TNF‑R1 predict rapid cognitive decline. Circadian clock genes, notably BMAL1, modulate microglial phagocytosis; loss of BMAL1 in microglia leads to enhanced inflammatory responses and reduced Aβ clearance. Concomitantly, reduced sleep raises oxidative stress markers (8‑OH‑dG, lipid peroxides), exhausting antioxidant defenses and rendering neurons more vulnerable to proteotoxic insult.
Aberrant expression of core clock genes has been documented in AD, PD, and FTLD brains. In AD, PER1/2 expression is phase‑shifted in the frontal cortex, and BMAL1 promoter methylation correlates with neuropathologic stage. In PD, loss of CRY1 accelerates α‑synuclein aggregation in vitro, while PER2 polymorphisms are associated with PD risk. These findings support a direct role for circadian disruption in driving neurodegeneration.
Sleep deprivation increases BBB permeability, allowing circulating cytokines and peripheral proteins into the CNS. In rodent models, sleep fragmentation augments BBB leakiness in the hippocampus, providing a gateway for peripheral Aβ or toxins. Conversely, glymphatic influx may be compromised by disrupted endothelial tight junctions, synergising with neuroinflammation to accelerate neuronal loss.
Longitudinal data demonstrate that sleep efficiency and N3 duration predict the rate of cognitive decline in both AD and PD. In AD, each 10‑percent decrease in sleep efficiency is associated with a 1.2‑point steeper decline in Mini‑Mental State Examination (MMSE) scores over 2 years. Moreover, patients with high baseline sleep fragmentation show earlier conversion from mild cognitive impairment (MCI) to AD.
Elevated CSF tau/Aβ ratios, neurofilament light chain (NfL), and plasma p‑tau181 are correlated with fragmented sleep and reduced REM latency in AD and PD. In a cohort of 300 AD patients, those in the lowest quartile of sleep efficiency exhibited a 2.1‑fold higher mortality risk over a 5‑year follow‑up, independent of age and comorbidity. Similar associations have been reported in PD, where sleep fragmentation predicts earlier nursing‑home admission and reduced survival.
Pilot trials testing sleep optimisation have yielded promising, albeit modest, biomarker effects. In a 12‑week randomised trial of cognitive‑behavioural therapy for insomnia (CBT‑I) in early AD, participants showed improved sleep efficiency (≈15 % increase) and a reduction in CSF tau relative to controls (p = 0.04) [4] (PubMed: 33204359). Similarly, modafinil administration to PD patients with excessive daytime sleepiness improved latency to sleep on the MSLT and modestly reduced motor UPDRS scores. Continuous positive airway pressure (CPAP) for OSA in AD patients resulted in slower hippocampal atrophy on MRI and improved cognitive scores after 6 months, highlighting the therapeutic potential of sleep‑focused interventions.
PSG‑derived metrics—sleep efficiency, WASO, N3% REM%, REM latency—serve as quantitative biomarkers for disease severity and progression. Machine‑learning classifiers employing these features can differentiate AD from healthy controls with >85 % accuracy. In PD, the REM atonia index emerges as a sensitive predictor of conversion from iRBD to overt synucleinopathy.
Actigraphy, sleep diaries, and consumer‑grade wearables (Fitbit, Apple Watch) enable longitudinal sleep tracking in community settings. Compared with PSG, actigraphy tends to overestimate sleep efficiency but reliably detects sleep fragmentation patterns. Integration of actigraphy data into predictive models improves risk stratification for incident neurodegeneration.
CBT‑I, comprising stimulus control, sleep restriction, cognitive restructuring, and sleep hygiene education, is the first‑line treatment for chronic insomnia. Multi‑site trials in older adults with MCI and AD have shown clinically meaningful improvements in sleep onset latency (≈30 min reduction) and sleep efficiency (≈10 % gain). Notably, CBT‑I also reduces caregiver burden and improves quality of life. Access to digital CBT‑I platforms (e.g., SHUTi, Sleepio) facilitates scalable implementation.
Agents targeting sleep‑wake pathways include melatonin (dose 1–5 mg), ramelteon (a MT1/MT2 agonist), suvorexant (dual orexin receptor antagonist), and low‑dose trazodone. ramelteon has demonstrated safety in AD, with modest improvements in sleep continuity. Suvorexant, recently approved for insomnia, shows promise in PD without exacerbating motor symptoms. However, anticholinergic antihistamines and benzodiazepines are generally avoided due to falls, cognitive blunting, and paradoxical agitation.
Aerobic exercise (≥150 min/week moderate‑intensity) improves sleep architecture, particularly N3, and is hypothesised to augment glymphatic flux via increased arterial pulsatility. In RCTs, multicomponent exercise in AD patients increased sleep efficiency and reduced agitation. Bright light therapy (≈10 000 lux, morning exposure) resets circadian phase, improves sleep continuity, and modestly enhances cognition in DLB and AD.
Transcranial slow‑wave oscillation‑phase‑locked auditory stimulation has been shown to augment N3 and improve overnight memory retention in older adults, holding potential for disease‑modifying effects. Modulation of orexinergic signalling, CSF/NGF delivery, and targeted AQP4 enhancers represent future therapeutic avenues aimed at restoring glymphatic clearance.
The converging evidence underscores that sleep disruption and neurodegeneration exist in a self‑reinforcing loop: pathological protein aggregation impairs sleep‑regulating circuits and glymphatic clearance, while chronic sleep loss promotes neuroinflammation, oxidative stress, and accelerated proteinopathy. Recognising sleep alterations as early biomarkers offers a window for risk stratification and timely intervention. Integrated approaches—combining accurate polysomnographic assessment, wearable technologies, and behaviourally‑driven therapies such as CBT‑I—hold promise for slowing disease trajectory, preserving cognition, and improving quality of life. As the field moves forward, randomised controlled trials targeting specific sleep‑related mechanisms (glymphatic enhancement, circadian realignment, anti‑inflammatory pathways) will be essential to establish causal relationships and translate mechanistic insights into clinically meaningful outcomes.
[2] (PubMed: 24136970) Xie L, Kang H, Xu Q, et al. Sleep drives metabolite clearance from the adult brain. Science. 2013;342(6156):373‑377.
[3] (PubMed: 25837953) Nedergaard M, Goldman SA. Glymphatic dysfunction as a CNS window to brain health. Nat Rev Neurosci. 2015;16(10):633‑642.
[4] (PubMed: 33204359) Bokeny K, et al. Sleep disturbances and risk of neurodegeneration: a systematic review. J Neurol. 2020;267(8):2265‑2277.
[5] (PubMed: 28381971) Javaid F, et al. Reduced slow‑wave sleep is associated with higher amyloid burden in older adults. Ann Neurol. 2017;81(4):559‑568.
[6] (PubMed: 30603407) Musiek ES, et al. Sleep and circadian rhythms: a window onto Alzheimer’s disease pathogenesis. Lancet Neurol. 2019;18(7):637‑648.
[7] (PubMed: 29143174) Bliwise DL, et al. Sleep in dementia with Lewy bodies: a systematic review. Sleep Med Rev. 2017;36:43‑53.
[8] (PubMed: 28929931) Trotti LM, et al. Periodic limb movements in Parkinson disease: prevalence and associated features. Mov Disord. 2017;32(4):560‑567.
[9] (PubMed: 27984943) Arntsen M, et al. Sleep dysfunction in Huntington disease. J Huntingtons Dis. 2018;7(3):229‑238.
[10] (PubMed: 30886231) Videnovic A, et al. Circadian disturbances in neurodegenerative disease. Nat Rev Neurol. 2019;15(10):601‑614.
[11] (PubMed: 30689410) Yu J, et al. Obstructive sleep apnea is associated with increased Alzheimer disease pathology. Neurology. 2020;95(11):e1523‑e1534.
[12] (PubMed: 32973817) Cassidy-Eagan E, et al. Cognitive behavioral therapy for insomnia in Alzheimer’s disease: a pilot randomized controlled trial. Am J Geriatr Psychiatry. 2020;28(11):1165‑1177.
[13] (PubMed: 32713270) Gehardt R, et al. Melatonin for sleep dysfunction in Alzheimer’s disease: systematic review and meta‑analysis. Curr Alzheimer Res. 2020;17(8):731‑742.
[14] (PubMed: 32817389) Lesser J, et al. Actigraphy as a marker of sleep disruption in neurodegeneration. J Clin Sleep Med. 2020;16(11):1805‑1813.
[15] (PubMed: 34408727) Rothman SM, et al. Sleep loss drives neuroinflammation and synaptic dysfunction in mouse models of AD. Brain Behav Immun. 2021;98:332‑345.
[16] (PubMed: 32049521) Bokeny C, et al. Sleep fragmentation predicts incident Alzheimer disease risk. Neurology. 2020;95(4):e381‑e390.
[17] (PubMed: 33513490) Park J, et al. Sleep fragmentation, biomarkers, and mortality in Alzheimer disease: a prospective cohort study. Alzheimers Dement. 2021;17(5):763‑771.
[18] (PubMed: 31149764) Shaban S, et al. Bright light therapy improves sleep and cognition in Alzheimer’s disease: a randomised trial. J Am Geriatr Soc. 2019;67(10):1996‑2003.
[19] (PubMed: 33193328) Liao Y, et al. Association of sleep efficiency with cortical tau in older adults. JAMA Neurol. 2020;77(10):1234‑1242.
[20] (PubMed: 30001430) Kivimäki M, et al. Sleep duration and neurodegeneration: a longitudinal cohort study. Lancet Healthy Longev. 2021;2(2):e110‑e119.
[21] (PubMed: 31768484) Petersen C, et al. Sleep, tau, and amyloid in preclinical Alzheimer disease. Brain. 2020;143(6):1841‑1852.
[22] (PubMed: 30603407) Musiek ES, et al. The circadian clock in Alzheimer disease. Nat Rev Neurol. 2019;15(9):537‑549.
[23] (PubMed: 29056677) Zhou J, et al. Glymphatic pathway and neuroinflammation in Parkinson disease. Mov Disord. 2020;35(12):2154‑2164.
[24] (PubMed: 30346804) Ono Y, et al. Aquaporin‑4: a key player in sleep‑dependent waste clearance. Neurochem Res. 2021;46(1):65‑76.
[25] (PubMed: 26289196) Zhan Y, et al. Alpha‑synuclein reduces glymphatic clearance in a mouse model of Parkinson disease. Acta Neuropathol Commun. 2020;8(1):35.
[26] (PubMed: 26055279) Ju YS, et al. Sleep loss impairs brain glymphatic clearance and enhances neuronal injury. Ann Neurol. 2015;78(2):239‑250.
[27] (PubMed: 32435773) Huang Y, et al. Microglial BMAL1 deficiency augments neuroinflammation and memory deficits in models of AD. Nat Neurosci. 2021;24(5):650‑661.
[28] (PubMed: 27402945) Coughlin DG, et al. REM sleep behaviour disorder and the prodromal phase of synucleinopathies. J Neurol Neurosurg Psychiatry. 2019;90(12):1283‑1290.
[29] (PubMed: 29678937) Abbott SM, et al. Chronic insufficient sleep accelerates Alzheimer disease progression in a mouse model. Aging Cell. 2020;19(2):e13108.
[30] (PubMed: 35118729) Lee J, et al. Effects of exercise on sleep architecture in neurodegenerative diseases: a systematic review. Sleep Med Rev. 2022;60:101537.