Time-dependent effects of hyperoxia on the BOLD fMRI signal in primate visual cortex and LGN

Hyperoxia is present in many anaesthesia protocols used in animal blood oxygen level-dependent (BOLD) functional magnetic resonance imaging (fMRI) studies. However, little data exist on the influence of hyperoxia on the magnitude of stimulus-induced relative changes in BOLD fMRI signal (ΔBOLD%). No study to date has investigated these effects in a time-resolved manner, although cerebral vasoregulation offers sites for a time-dependent interaction of hyperoxia and ΔBOLD%. Here we investigated time-dependent effects of an inspiratory oxygen fraction of 90%. We tightly clamped end tidal CO2 and body temperature and recorded physiological parameters relevant to rCBF in (fentanyl/isoflurane) anaesthetized monkeys while using visual stimulation to elicit ΔBOLD%. To clarify whether changes in ΔBOLD% arose from changes in baseline blood oxygenation or rather altered neuronal or vascular reactivity, we directly measured changes in rCBV using monocrystalline ion oxide nanoparticles (MION) as contrast agent. In visual cortex we found a biphasic modulation of stimulus-induced ΔBOLD% under hyperoxia: We observed first a significant decrease in ΔBOLD% by − 24% for data averaged over the time interval of 0-180 min post onset of hyperoxia followed by a subsequent recovery to baseline. rCBV response amplitudes were decreased by 21% in the same time interval (0-180 min). In the LGN, we neither found a significant modulation of ΔBOLD% nor of MION response amplitude. The cerebrovascular effects of hyperoxia may, therefore, be regionally specific and cannot be explained by a deoxyhemoglobin dilution model accounting for plasma oxygenation without assuming altered neuronal activity or altered neurovascular coupling.