Epidemiological studies of Dementia have major methodological problems, particularly in France: selection of the studied sample, difficulties for contact with elderly people, heterogeneity of place of residence, inaccuracy of the diagnostic procedures. In spite of these problems, the Paquid Research program allowed to give estimations of prevalence and incidence rate of dementia in the French population. On the basis of these results, one can estimate that there are actually about 360,000 prevalent cases of dementia in France, of whom 270,000 are of Alzheimer type. There were also 150,000 annual-incident cases of dementia. However, many of these cases were not considered as demented by the medical system because of the absence of cognitive complaints.