Dr Suzanne de la Monte and colleagues now believe it is down to what they are calling type 3 diabetes.
By looking at rodents and post-mortem brain tissue from people with Alzheimer's disease they have found that insulin and its related proteins are actually produced in the brain, and that reduced levels of both are linked to Alzheimer's disease.
Credit the BBC for finding a good skeptic:
Professor Greg Cole, from the University of California Los Angeles' Alzheimer's Disease Research Center, said: "This is a new finding. It is interesting that the brain makes very low levels of insulin.
"But its significance is unclear. The levels are so low that they have not been detected with less sensitive methods. I don't think we can say they are high enough to matter.
"I suspect that the brain insulin itself is not very significant and neither is its deficit in Alzheimer's disease and, therefore, I wouldn't call it type 3 diabetes."
I wouldn't call it Type 3 diabetes either, but mostly because the name's already taken. So's Type 4 diabetes (aka Gestational Diabetes, occasionally called type 3).
De la Monte's ability to name diseases is positively insipid. This group doesn't even deserve the bronze! But I suppose they wouldn't have received as much attention if they hailed the possibility of Type 5 diabetes.