A series of 35 selected non senile patients was investigated with stereotactic brain biopsies in frontal and temporal lobes including the rhinencephalon. Results of neuropathological studies are confronted with clinical proposals. Alzheimer's disease is diagnosed in about 10% of the cases under 50 years of age, but more than 50% after 50. In the whole series only 10 clinical diagnoses (29%) were confirmed by neuropathological findings, 14 cases (40%) were fully corrected, 5 out of 7 undefined demented patients were specified but 6 cases (17%) remained unspecified because only nonspecific gliosis was found. Thus stereotactic biopsies were without real interest in 10 confirmed diagnosis (29%), of relatively poor interest in 6 nonspecific gliosis (17%), but of great interest in 19 cases (40% corrected + 14% specified = 54%). It can be considered from a general point of view that in our series more than 50% of diagnoses benefit definitively by stereotactic brain biopsies bearing in mind that the clinical diagnosis fails in one case out of two. It is clearly demonstrated that this stereotactic methodology has a significantly higher score than conventional biopsies.