Improving estimates of the cerebral metabolic rate of oxygen from optical imaging data

The cerebral metabolic rate of oxygen (CMRO2) is an important measure of brain function. Since it is challenging to measure directly, especially dynamically, a number of neuroimaging techniques aim to infer activation-induced changes in CMRO2 from indirect data. Here, we employed a mathematical modelling approach, based on fundamental biophysical principles, to investigate the validity of the widely-used method to calculate CMRO2 from optical measurements of cerebral blood flow and haemoglobin saturation. In model-only simulations and simulations of in vivo data changes in CMRO2 calculated in this way differed substantially from the changes in CMRO2 directly imposed on the model, under both steady state and dynamic conditions. These results suggest that the assumptions underlying the calculation method are not appropriate, and that it is important to take into account, under steady state conditions: 1) the presence of deoxyhaemoglobin in arteriolar vessels; and 2) blood volume changes, especially in veins. Under dynamic conditions, the model predicted that calculated changes in CMRO2 are moderately correlated with the rate of oxygen extraction--not consumption--during the initial phase of stimulation. However, during later phases of stimulation the calculation is dominated by the change in blood flow. Therefore, we propose that a more sophisticated approach is required to estimate CMRO2 changes from these types of data.