PurposeTo investigate the effect of inhaled oxygen level on dynamic glucose enhanced (DGE) MRI in mouse brain tissue and CSF at 3 T.MethodsDGE data of brain tissue and CSF from mice under normoxia or hyperoxia were acquired in independent and interleaved experiments using on‐resonance variable delay multi‐pulse (onVDMP) MRI. A bolus of 0.15 mL filtered 50% D‐glucose was injected through the tail vein over 1 min during DGE acquisition. MRS was acquired before and after DGE experiments to confirm the presence of D‐glucose.ResultsA significantly higher DGE effect under normoxia than under hyperoxia was observed in brain tissue (p = 0.0001 and p = 0.0002 for independent and interleaved experiments, respectively), but not in CSF (p > 0.3). This difference is attributed to the increased baseline MR tissue signal under hyperoxia induced by a shortened T1 and an increased BOLD effect. When switching from hyperoxia to normoxia without glucose injection, a signal change of ˜3.0% was found in brain tissue and a signal change of ˜1.5% was found in CSF.ConclusionsDGE signal was significantly lower under hyperoxia than that under normoxia in brain tissue, but not in CSF. The reason is that DGE effect size of brain tissue is affected by the baseline signal, which could be influenced by T1 change and BOLD effect. Therefore, DGE experiments in which the oxygenation level is changed from baseline need to be interpreted carefully.