Positron emission tomography allows an in vivo assessment of various physiological and biochemical processes, for example cerebral blood flow, metabolism, or interactions between ligands and receptors. Data quantification and interpretation rest on models describing in a simple way the behavior of the labelled molecules. The general principles are common, but each model has limitations. The different methods are first validated in and applied to normal populations under resting conditions. New techniques for rapid assessments of blood flow and metabolism make it possible to measure cerebral activation after sensori-motor, mental or pharmacological stimulation. This should allow the study of recovery or plasticity of the lesioned brain, after a stroke for example. PET measurements of cerebral blood flow, oxygen consumption and extraction, and cerebral blood volume are particularly well suited to investigate the physiopathology of cerebrovascular diseases. Remote metabolic disturbances give information on interregional cerebral connections, and on clinico-metabolic correlations. In epilepsy, PET is useful in localizing the epileptogenic focus in partial epilepsy: it is hypometabolic interictally. The meaning of the hypometabolism has still to be established. New information about the neurochemistry of the epileptogenic focus should become available from studies of benzodiazepine, excitatory amino acid or opiate systems, for example. PET has already enabled pathophysiological hypotheses to be tested in status epilepticus. Disturbances of metabolism and neurotransmission systems have been observed at various stages and in various types of neurodegenerative diseases. The modifications are not only an early reflection of anatomopathological lesions, but could give more direct information on the pathogenesis or symptomatology of these diseases and hence lead to new therapeutic endeavours, such as appropriate replacement therapy analogous by to dopatherapy in Parkinson's disease.