REM sleep drives cerebrospinal fluid efflux through massive mechanical and vascular surges
Abstract
Abstract The glymphatic system clears metabolic waste through cerebrospinal fluid (CSF) transport in the brain parenchyma and is most active during sleep. Yet the physical mechanisms generating glymphatic flow remain unclear, particularly the contribution of rapid eye movement (REM) sleep. Reports of reduced CSF inflow during REM have suggested a minimal role, yet imaging methods have, to date, been unable to capture whole-brain, multiparametric brain dynamics across full sleep cycles. Here, we develop an original multiparametric functional imaging modality using ultrasound during natural sleep in rodents. Our approach simultaneously maps brain mechanical properties, vascular activity, morphological deformations and intraventricular CSF flow at high spatiotemporal resolution. We show that REM onset is marked by a rapid decrease in cortical and midbrain stiffness—indicative of extracellular fluid expansion—followed approximately fifteen seconds later by a large and sustained increase in cerebral blood volume and mesoscale tissue deformation during the full REM episode. These vascular and mechanical surges coincide with elevated CSF efflux through the cerebral aqueduct and ventricles. REM-associated dynamics far exceed cardiac and respiratory pulsations in generating tissue strain and CSF motion. Complementary to NREM sleep which has been shown to promote CSF inflow via perivascular pathways, our results show that REM sleep strongly drives efflux through vascular–mechanical coupling. By enabling simultaneous imaging of brain stiffness, mesoscopic strain, vascular activity and intraventricular CSF flow, multiparametric functional ultrasound imaging uncovers REM sleep as a pivotal biomechanical phase of brain clearance.
Fields of Study & Subjects
Cerebrospinal fluid and hydrocephalus
Sleep and Wakefulness Research
Advanced MRI Techniques and Applications
Publication details
- Journal / source
- Research Square
- Publication date
- 3 March 2026
- Language
- EN
- Document type
- Preprint