Introduction: Cerebral metabolic rate of oxygen (CMRO2) can be obtained with magnetic resonance imaging (MRI) by calibrating BOLD using hypercapnia. CO2 administration influences BOLD by increasing and eventually stabilizing blood flow and volume respectively. This leaves blood oxygenation as the predominating factor affecting the magnitude of BOLD. Obviously, the above assortment implies that increased CO2 acts exclusively as vasoactive agent, without affecting neuronal activity and hence oxygen metabolism. Amongst others, the latter assumption was strengthened by the results of Schmidt and Kety in the late 1940s. Most of the fMRI studies report coupling of cerebral blood flow (CBF) and oxygen metabolism of n = 2 (n being the ratio of fractional changes in CBF and CMRO2). In this study we investigated the dependence of neural activity on enhanced CO2 concentration by means of simultaneous intracortical recordings and BOLD imaging in the anesthetized macaque monkey. Methods: Combined electrophysiology and fMRI recording were performed using the techniques described by Logothetis et al in Nature 2001. Hypercapnia was induced by administration of premixed medical gases containing 3 % and 6 % CO2 and 21 % O2. Ventilation with 3 % and 6 % CO2 increased end-tidal CO2 by app. 9 mmHg and 20 mmHg respectively. Results: Surprisingly, during hypercapnia (3 %/6 % CO2) we found a significant decrease of local field potentials (LFP) and multi unit activity (MUA) clearly implying that neural activity can not be considered unaltered during hypercapnia conditions. Decreases in MUA induced by 6 % CO2 are of similar magnitude as increases induced by a full field visual stimulus (rotating checkerboard). Conclusions: It follows that the calibrated BOLD approach overestimates oxygen metabolism, and therefore higher values of n may be necessary to derive CMRO2 from the BOLD signal.