Signals detected with functional brain imaging techniques are based on the coupling between neuronal activity and energy metabolism. Positron emission tomography signals detect blood flow, oxygen consumption and glucose utilization associated with neuronal activity; the degree of blood oxygenation is thought to contribute to the signal detected with functional magnetic resonance imaging, whereas magnetic resonance spectroscopy identifies the spatiotemporal pattern of activity-dependent appearance of metabolic in termediates, such as glucose or lactate. Despite the technological sophistication of these brain imaging techniques, the precise mechanisms and cell types involved in coupling and in generating metabolic signals are still debated. Indeed, given the level of resolution achieved with these brain imaging techniques, it has not been feasible to monitor met abolic fluxes between the highly intermingled neuronal, glial, and vascular elements in the intact brain. This obstacle has been overcome in recent years by using purified cellular preparations of neurons and glia. These approaches have suggested a critical role for astrocytes in coupling neuronal activity to energy metabolism. Indeed, astrocytes possess receptors and reuptake sites for a variety of neurotransmitters, including glu tamate. In addition, astrocytic end-feet, which surround capillaries, are enriched in the specific glucose transporter GLUT-1. These features would be expected to allow astro cytes to sense synaptic activity and to couple it with energy metabolism. During activa tion, glutamate is the predominant neurotransmitter released by modality-specific excitatory pathways to a given cortical area; in vitro and in vivo data support a model in which glutamate would stimulate, during activation, an initial glycolytic processing of blood-borne glucose by astrocytes; this glutamate-dependent process would result in a transient lactate overproduction, followed by a recoupling phase during which lactate would be oxidized by neurons. Such a model is consistent with data recently obtained with functional brain imaging techniques. NEUROSCIENTIST 3:361-365, 1997