Fluid-attenuated inversion recovery (FLAIR) images obtained during the administration of supplemental oxygen demonstrate a hyperintense signal within the cerebrospinal fluid (CSF) that is likely caused by T 1 changes induced by paramagnetic molecular oxygen. Previous studies demonstrated a linear relationship between the longitudinal relaxation rate (R 1 ؍ 1/T 1 ) and oxygen content, which permits quantification of the CSF oxygen partial pressure (P csf O 2 ). In the current study, CSF T 1 was measured at 1.5 T in the lateral ventricles, third ventricle, cortical sulci, and basilar cisterns of eight normal subjects breathing room air or 100% oxygen. Phantom studies performed with artificial CSF enabled absolute P csf O 2 quantitation. Regional P csf O 2 differences on room air were observed, from 65 ؎ 27 mmHg in the basilar cisterns to 130 ؎ 49 mmHg in the third ventricle. During 100% oxygen, P csf O 2 increases of 155 ؎ 45 and 124 ؎ 34 mmHg were measured in the basilar cisterns and cortical sulci, respectively, with no change observed in the lateral or third ventricles. P csf O 2 measurements in humans breathing room air or 100% oxygen using a Cerebrospinal fluid (CSF) oxygen partial pressure (P csf O 2 ) has been proposed as a surrogate marker for brain or spinal cord tissue oxygenation, given the passive nature of oxygen diffusion (1-5). Historically, investigators have measured P csf O 2 by analyzing fluid obtained by direct cisternal or lumbar puncture, or using implanted oxygen microelectrodes. These methods are inconvenient and invasive, and are limited in terms of spatial information and repeatability. Additionally, methods that remove fluid are prone to errors introduced by exposure of CSF to room air before the analysis is performed (6). It is possible that differences exist between CSF compartments caused by local brain tissue oxygen levels, oxygen permeability-surface area product, oxygen diffusion, CSF flow, and cerebral blood flow (CBF). A noninvasive imaging method to determine P csf O 2 would permit repeated measurements with spatial sensitivity during physiological manipulations. Through these means, a more complete understanding of CSF oxygen dynamics could be obtained that might yield insights into local brain or spinal cord oxygen levels in normal and abnormal physiologic states.Normal CSF at body temperature has MR properties that are essentially identical to those of partially oxygenated water, with a well-regulated pH (7.3) and a trace amount of protein (0.15-0.45 g/L) (7,8). Using invasive means, P csf O 2 has been measured to be between 25-53 mmHg in humans, roughly 25% its room air level (160 mmHg) (2-5,9). Previous studies in humans have suggested that P csf O 2 changes occur in the setting of abnormal brain oxygen metabolism (1,4,5).It is likely that the increased CSF signal within the subarachnoid spaces of the cerebral sulci and basilar cisterns on fluid-attenuated inversion recovery (FLAIR) images in humans breathing supplemental oxygen reflects higher than normal P csf O 2 , whi...