PurposeOur objective was to demonstrate the feasibility of using hyperventilation as an efficient vasoconstrictive challenge and prior knowledge as denoising constraints for cerebral metabolic rate of oxygen (CMRO2) mapping based upon quantitative susceptibility mapping (QSM).MethodsThree‐dimensional (3D) multi‐echo gradient echo and arterial spin labeling imaging were performed to calculate QSM and perfusion maps before and after a hyperventilation challenge in 11 healthy subjects. For comparison, this was repeated using a caffeine challenge. Whole‐brain CMRO2 and oxygen extraction fraction (OEF) maps were computed using constrained optimization. Hyperventilation scans were repeated to measure reproducibility. Regional agreement of CMRO2 and OEF maps was analyzed within the cortical gray matter (CGM) using t‐test and Bland‐Altman plots.ResultsHyperventilation challenge eliminates the 30‐min waiting time needed for caffeine to exert its vasoconstrictive effects. Mean CMRO2 (in µmol/100g/min) obtained in CGM using the caffeine and repeated hyperventilation scans were 149 ± 16, 153 ± 19, and 150 ± 20, respectively. This corresponded to an OEF of 33.6 ± 3.4%, 32.3 ± 3.2%, and 34.1 ± 3.8% at baseline state and 39.8 ± 4.8%, 43.6 ± 6.2%, and 42.8 ± 6.8% at challenged state, respectively. Hyperventilation scans produced a good agreement of CMRO2 and OEF values.ConclusionsHyperventilation is a feasible, reproducible, and efficient vasoconstrictive challenge for QSM‐based quantitative CMRO2 mapping. Magn Reson Med 77:1762–1773, 2017. © 2016 International Society for Magnetic Resonance in Medicine