"They may be onto something," said Rachel Richardson, the methods support unit lead for evidence production and methods directorate at Cochrane, an independent non-profit network that gathers and summarises medical research.

However, she said this "does not represent a smoking gun" linking early sugar intake to dementia.

One big issue is, it is impossible to know how much sugar each of the people in the study were actually eating, merely when they were born relative to sugar rationing.

"This evidence relates to people born in the 1940s and 1950s and so its applicability to young children today is limited - for example, weaning took place earlier and this would have a large impact on the early nutrition of babies and children," said Richardson.

Prof Ian Maidment, professor of clinical pharmacy at Aston University, said: "It is an interesting idea, it helps inform debate, but in my opinion we need more evidence."

He said connecting "something that happened 80 years ago" and projecting that onto health now meant there were "loads" of other factors that could be playing a role.

Dr Sara Rodrigues, from Alzheimer's Research UK, said: "It's important to remember that this was an observational study.

"This means it can identify links between diet and dementia risk, but it cannot prove that eating less sugar directly prevents dementia.

"Eating a healthy, balanced diet is one of the best ways to support your brain health and it's never too late to make positive changes."