(1) Alzheimer's disease is a multifactorial disease in which age and APOE-epsilon 4 are important risk factors. (2) The neuropathological hallmarks of AD, i.e. amorphous plaques, neuritic plaques (NPs), pretangles, neurofibrillary tangles (NFT) and cell death are not part of a single pathogenetic cascade but may occur independently. (3) In brain areas where classical AD changes, i.e. NPs and NFTs, are present, such as the CA1 area of the hippocampus, the nucleus basalis of Meynert and the tuberomamillary nucleus, a decreased metabolic rate is found. The decreased metabolic rate appears not to be induced by the presence of pretangles, NFT or NPs. (4) Decreased metabolic rate may precede cognitive impairment and is thus an early occurring hallmark of AD, which, in principle, may be reversible. The observation that the administration of glucose or insulin enhances memory in AD patients also supports the view that AD has a metabolic basis. (5) Moreover, several observations in postmortem brain indicate that activated neurons are better able to withstand aging and AD, a phenomenon paraphrased by us as 'use it or lose it'. (6) It is, therefore, attractive to direct the development of therapeutic strategies towards restimulation of neuronal metabolic rate in order to improve cognition and other symptoms in AD. A number of pharmacological and non-pharmacological studies support the concept that activation of the brain has beneficial effects and may, to a certain degree, restore several aspects of cognition and other central functions. For instance, the circadian system may be restimulated in AD patients by exposing them to more light or transcutaneous nerve stimulation. A procedure has been developed to culture human postmortem brain tissue that allows testing of the efficacy of putative stimulatory compounds such as neurotrophins.