An increasing number of studies have appeared in the literature suggesting that Alzheimer's disease (AD) is a hypometabolic brain disorder. Decreased metabolism in AD has been revealed by a variety of in vivo and postmortem methods and techniques including positron emission tomography and glucose metabolism. We used the size of the Golgi apparatus (GA) and cell profile area as indicators of neuronal activity in postmortem material. Using an antibody against MG-160, a sialoglycoprotein of the medial cisternae of the GA, we were able to visualize and quantify the GA area. In a series of experiments, we tried to relate neuronal metabolism to different hallmarks of AD, i.e. plaques and tangles, and also to genetic risk factors for AD like age and (apolipoprotein E) ApoE polymorphism. Our results showed that in AD there is indeed a clear reduction in brain metabolism in several severely affected brain regions including the nucleus basalis of Meynert (NBM), the CA1 area of the hippocampus and the hypothalamic tuberomamillary nucleus. However, the reduction in neuronal activity did not seem to be caused by the presence of neuropathological hallmarks of AD, i.e. plaques and tangles. There was, however, a clear relationship between the presence of ApoE epsilon4 alleles and a decrease in GA size. Our data suggest that decreased neuronal activity and neuropathological hallmarks of AD, such as plaques and tangles, are basically independent phenomena. Moreover, ApoE epsilon4 may participate in the pathogenesis of AD by decreasing neuronal metabolism. The main implication of these findings is that therapeutic strategies in AD should be focussed on reactivation of neuronal metabolis