Studies which examine patients with and without dementia during life a
nd then examine brain tissue after death are extremely difficult to co
nduct. There are now several such studies published, representing a ma
jor contribution to the understanding of the pathologies of the dement
ias. These studies were not designed to represent population samples a
nd, from an epidemiological viewpoint, they are flawed because none ar
e population-based and none represent the full range of function obser
ved during life. It is therefore important to examine the available st
udies for their contributions and biases.