Mounting evidence from in vitro experiments indicates that lactate is an efficient energy substrate for neurons and that it may significantly contribute to maintain synaptic transmission, particularly during periods of intense activity. Since lactate does not cross the blood-brain barrier easily, blood-borne lactate cannot be a significant source. In vitro studies by several laboratories indicate that astrocytes release large amounts of lactate. In 1994, we proposed a mechanism whereby lactate could be produced by astrocytes in an activity-dependent, glutamate-mediated manner. Over the last 2 years we have obtained further evidence supporting the notion that a transfer of lactate from astrocytes to neurons might indeed take place. In this article, we first review data showing the presence of mRNA encoding for two monocarboxylate transporters, MCT1 and MCT2, in the adult mouse brain. Second, by using monoclonal antibodies selectively directed against the two distinct lactate dehydrogenase isoforms, LDH1 and LDH5, a specific cellular distribution between neurons and astrocytes is revealed which suggests that a population of astrocytes is a lactate 'source' while neurons may be a lactate 'sink'. Third, we provide biochemical evidence that lactate is interchangeable with glucose to support oxidative metabolism in cortical neurons. This set of data is consistent with the existence of an activity-dependent astrocyte-neuron lactate shuttle for the supply of energy substrates to neurons.