The kinetics of glutamate 13C-4 label appearance from D-[1-13C]-glucose and 13C-4 label disappearance from steady state following D-12C-glucose incubation were quantified with 1H-observed, 13C-edited nuclear magnetic resonance (NMR) spectroscopy in the superfused brain slices under largely varying oxygen consumption. Label incorporation to and from glutamate C-4 were fitted into mono- or bicompartmental models in order to determine the respective rate constants and to assess the presence of plausible multiple pools. At a steady-state oxygen consumption of approximately 4 mumol/min/g dry weight, glutamate labelling could be fitted into a biexponential equation, suggesting that there were two compartments with a large difference in their rates (respective rate constants of 0.022 and 0.149) and pool sizes (relative contributions of 91.2 and 8.8%, respectively). Stimulation of oxygen consumption in the brain slice preparations with either 40 mM KCl by 59.5 +/- 10.3% or 5 microM carbonyl cyanide m-fluorophenyl hydrazone by 61.4 +/- 8.4% increased glutamate C-4 labelling rate constants to 0.058 +/- 0.009 and 0.054 +/- 0.006, respectively. In the stimulated slice preparation, glutamate labelling could only be fitted into a monoexponential equation. 13C-4 label disappearance, independent of oxygen uptake, could also only be fitted into a monoexponential equation. There was a close match between the rate constants of label disappearance and appearance in non-stimulated and carbonyl cyanide m-fluorophenyl hydrazone-stimulated slices. In the presence of 40 mM KCl label disappearance did not, however, increase. These data show that glutamate C-4 turnover from exogenous D-[1-13C]-glucose can be used as an index of oxidative metabolism in situ under steady-state conditions as well as when oxygen metabolism is strongly stimulated. The results are discussed with respect to the use of NMR spectroscopy as a means of mapping brain oxidative metabolism.