It’s not every day that an addictive and/or psychoactive substance is heralded in the press as potentially healthy — wait, yes it is (see also chocolate, red wine, nicotine). Everyone loves it when the scientific community supposedly endorses their vices. In most cases, the compounds of scientific interest (resveratrol in wine, flavonoids in chocolate, nicotine in cigarettes) indeed may show promise in a laboratory setting, but claims about the foods containing them are usually confined to headlines.

This weekend, coffee got the press bump. The CBS Early Show announced “Coffee May Lower Alzheimer’s Risk”, while the Daily Mail was even bolder with “How two strong coffees a day can ‘reverse’ Alzheimer’s”. The print version of the paper apparently led with “Coffee beats Alzheimer’s”. The Times of India and the Telegraph were both bold enough to use the word “cure” (the Telegraph at least had the decency to throw quotes around the word, though it’s unclear what or whom they were actually quoting).

In reality, the science to which the articles allegedly referred to was contained in a pair of studies published 4 July in the Journal of Alzheimer’s Disease, in which scientists fed caffeine (not coffee) to mice with a genetically engineered version of Alzheimer’s disease. Led by researchers at the University of South Florida who have previously published articles extolling the dementia-deterring virtues of caffeine, the study used a common laboratory strain of mice that express a pathogenic form of amyloid-beta (Aβ) — the protein that, in Alzheimer’s, forms the nasty proteinaceous plaques that gob up neurons. As these mice age, their brains become progressively choked by plaques and their memory capabilities plummet correspondingly.

At different stages in the mice’s lives, the researchers spiked their drinking water with caffeine — the human equivalent of 500 mg, or slightly less than what you’d get from two tall Starbucks coffees. They found that when the mice began drinking caffeine before the cognitive impairments set in, they had fewer plaques and did better on memory tasks. When they began drinking caffeine after their cognitive abilities had began to decline, their memory improved to about the performance of normal mice.

The authors showed that the mice weren’t doing well just because they were high. Normal mice that got the same caffeine treatment didn’t show any improvement over their non-caffeinated counterparts.

The results are interesting and seem impressive, but as Nature editor @noahwilliamgray points out, the Journal’s table of contents is a laundry list of compounds that offer new hope for alleviating the disease — the non-publicized drugs are “just not as sexy [as caffeine]”.

This doesn’t mean the science is bogus. The findings are intriguing, and support published epidemiological studies linking coffee consumption to reduced risk of Alzheimer’s. But studies using mouse models of Alzheimer’s — or any neurodegenerative disease — are challenging.

Unfortunately, human studies aren’t much better. Though the animal is far more relevant, the studies are difficult to interpret because of variability in lifestyle choices (e.g. people who drink lots of coffee may also work in challenging jobs that demand focus, which may help stave off Alzheimer’s).