Comment by jcranmer

The "... in mice" qualifier is somewhat overdone, but for Alzheimer's in particular, it's actually extremely important: mice do not experience Alzheimer's. We test Alzheimer's drugs in mice by artificially inducing something that produces similar dementia symptoms, and then seeing how the drugs influence that progression. If that model is different from how Alzheimer's works--and there's pretty decent suggestions from the history of failed Alzheimer's drugs in humans that our models of Alzheimer's are in fact wrong--then anything that fixes that model has suspect utility.