In vivo measurements of rates of protein synthesis with radiolabeled precursors are problematic because of uncertainties about the relative contribution of unlabeled amino acids derived from protein breakdown to the precursor pool for protein synthesis. The quantitative autoradiographic method for the determination of local rates of cerebral protein synthesis (lCPS) in vivo with a carboxyl-labeled, aliphatic, branched-chain amino acid as tracer takes into account recycling of the unlabeled species of the tracer amino acid. Rates of cerebral protein synthesis in rats have been shown to increase in regenerating nerve nuclei, decrease in most brain regions during brain development and with senescence. Cocaine treatment changes lCPS in selective brain regions; the direction of the changes depends on the history of treatment. Studies of normal sleep in monkeys indicate that slow wave sleep is linked with increased rates of lCPS in many brain regions. In general, changes in lCPS may mark brain regions undergoing long-term adjustments in response to a drug, a treatment, or a change in physiological state. Studies of recycling and lCPS under valine-flooding conditions demonstrate compartmentation of amino acids for protein synthesis. Even when plasma valine concentrations are increased 100 fold, protein degradation continues to supply a significant portion of valine for protein synthesis. In spite of a reduction in brain concentration of leucine with valine-flooding protein synthesis rates remain unchanged probably due to an increase in the portion of leucine in the precursor pool supplied by protein degradation.