BOLD fMRI has been extensively used to investigate brain resting state or pharmacological challenges in various physiological or pathological situations¹. However, BOLD fMRI which relies on the neurovascular coupling hypothesis and does not directly reflects neuronal activity may fail in some conditions which prevent neurovascular coupling². It is also sensitive to global or local variations in cerebral blood flow which are not necessarily related to underlying neuronal activity. On the other hand, diffusion fMRI³ has been shown to be more directly linked to neuronal activation showing neuronal responses even when neurovascular coupling is abolished⁴. We have investigated the potential of DfMRI to provide a quantitative assessment of resting state brain activity in various anesthetic conditions. Experiments were conducted in anesthetized Wistar rats at 7T with a volume coil (PharmaScan, Bruker). After an initial anesthesia under 2.0% isofluorane anesthesia conditions were changed and whole brain DfMRI images were acquired continuously for 10 min (40 volumes) under isoflurane (Iso; 1.5, 2.0, and 2.5%; n=10) or medetomidine (Med; 0.1 and 0.3 mg/kg/h, i.v.; n=8) (Fig.1) with the following parameters: diffusion-sensitized double SE-EPI sequence; TR/TE=2500/32 ms, FOV=2.5×2.5 cm2, Matrix=100×100, 10 slices, 1.5 mm thickness, b-values=1000 and 1800 mm²/s along 3 directions; [X=1,Y=0,Z=0], [0,1,0], and [0,0,1]. BOLD images were acquired with the same parameters except for; 240 volumes (10 min) and b-values=10 mm²/s along [1,1,1], under same dosages of Iso and Med conditions (n=4, respectively). Breathing, CO₂ concentration and rectal temperature of the rats were monitored and artificially controlled during scanning. Water apparent diffusion coefficient (ADC) maps were calculated for each time point as ADC=ln(S₁₀₀₀/S₁₈₀₀)/800 to remove any residual IVIM or T2/T2* effects and averaged over the 3 directions. Pre-processing (realignment, normalization to an average rat template, and smoothing) and statistical analysis of two-way repeated ANOVA (p<0.001, uncorrected) were performed with SPM8 over the whole time courses and dosage conditions for Iso and Med, respectively. The BOLD images showed a widespread, not specific signal increase in Iso and decrease in Med (Fig.2A, B). The increase in BOLD under Iso is likely caused by an overall blood vessel vasodilation⁵ and the decrease under Med by an overall vasoconstriction⁵. On the contrary, the ADC absolute value increased significantly both with Iso and Med in specific brain regions (Fig. 2C, D). Overall this ADC increase was very stable with time under each anesthetic condition, compared with temporal BOLD change (Fig.3). The regions exhibiting the ADC increase common to both Med and Iso were the cerebral cortex, the limbic region, the midbrain and a network of regions (CM, PO, pHT, vmHT, DR, and PAG) known to be associated with wakefulness/sleep conditions^6,7. Additional regions with increased ADC were also observed for Iso (Fig.4). Those loci are consistent with the known enhancement by isofluorane of GABAergic inhibitory systems on the whole brain⁸, while medetomidine is an α2 adrenergic agonist which enhances neural inhibitory systems through norepinephrine at selective brain region⁵. Importantly, the amount of ADC increase was correlated with the dose of anesthetic drugs (Fig.4, Table 1) which are known to induce a dose-dependent decrease in neural activity^5,8, especially in the wakefulness/sleep network. Based on earlier reports that water diffusion decreases during neural evoked responses, possibly in relation to the associated swelling of cellular element^3,4,9, those results suggest that the water diffusion increase observed under anesthetic conditions might reflect neural deactivation and, possibly, an associate retraction of cellular elements¹⁰.Water apparent diffusion coefficient was shown to increase in specific regions on the rat brain under anesthetic conditions, especially in a network of regions involved in wakefulness and sleep. The amount of diffusion increase was correlated to the dose of the anesthetic drugs, likely reflecting a decrease neuronal activity. In contrast, BOLD signals showed non-specific, widespread relative changes of opposite signs for Iso and Med. Those results suggest that DfMRI might be suitable to investigate brain resting state or pharmacological challenges in various physiological or pathological conditions quantitatively and without interference with vascular confounding effects.