Fig. 2: NREM, IS and REM sleep states are associated with specific penetrating arteriole diameter changes that are mirrored in the size of the PVS.

a Experimental setup. GLT1-eGFP transgenic mice were fitted with a cranial window exposing the somatosensory cortex, ECoG electrodes and EMG electrodes. The vasculature was outlined with Texas Red-labeled dextran. b (Left) Representative image of a penetrating arteriole with GLT1-eGFP fluorescence in astrocytic endfeet and Texas-Red labeled dextran in the vessel lumen. Imaging was performed by placing cross sectional x-t line scans. (Right) Vessel lumen diameter (Ø) and astrocyte endfoot tube diameter (Ø) was determined along a line across the vessel. Total width of the PVS was assessed as the difference between the vessel lumen diameter and the endfoot tube diameter. c Representative traces of a penetrating arteriole line scan, change in endfoot tube diameter (Ø), vessel lumen diameter (Ø) and perivascular space diameter during a sleep cycle. d Amplitude of very low frequency (VLF 0.1–0.3 Hz) oscillations and e median diameter of lumen and total width of PVS throughout a sleep cycle. Every gray line (dashed or full) represents an individual vessel, dashed lines are used when an observation is missing in a certain state for a given vessel (for example dashed line between WBS and IS means that we do not have an observation for that particular vessel in NREM state), bold lines and shaded area are the estimates and 95% CI from linear mixed effects models (two-sided), n = 310 episodes, 25 penetrating arterioles, 4 mice. *P  <  0.05, **P  <  0.01, ***P  <  0.001, Tukey adjustment for multiple comparisons. See supplementary information for exact p-values, and degrees of freedom for the statistical tests. WBS wake before sleep; NREM non-rapid eye movement sleep; IS intermediate state sleep; REM rapid eye movement sleep; WAS wake after sleep; ECoG electrocorticography; EMG electromyography; PVS perivascular space. Source data are provided as a Source Data file.