The trial on elderly men and women living in Chicago found that those who reported eating fish at least once a week had a slower decline in mental function than peers who did not eat fish as often - about 10 per cent less per year.
For those eating two or more fish meals a week the rate of cognitive decline was 13 per cent slower than non-fish eaters, revealed the study published online yesterday, ahead of the December print issue of the Archives of Neurology (vol 62, 1-5).
This is equivalent to "being three or four years younger in age", the authors said.
But although earlier studies have found an association between intake of omega-3s - the fatty acid found in high quantities in oily fish - and reduced risk of dementia, the new trial failed to link the presence of these fats in the diet with slower cognitive decline.
"The evidence for association between the different types of omega-3 fatty acids and cognitive change was weak at best," write Martha Clare Morris of Rush University Medical Center in Chicago and colleagues.
These findings conflict with previous studies showing strong reductions in Alzheimer's disease risk among people with high intake of omega-3s, DHA and alpha-linoleic acid.
"We can only speculate that perhaps dietary omega-3 fatty acids have little impact on milder forms of cognitive decline," said the researchers.
In addition, the measurement of DHA and EPA may have been too imprecise to detect a link to cognitive decline, they added.
But there is also the possibility that nutrients in fish other than the fatty acids are protective against cognitive decline, and the Chicago researchers call for "more precise studies of the different dietary constituents of fish…[to] help to understand the nature of the association".
Their results were based on six years of data from an ongoing trial of more than 6,000 Chicago residents, aged at least 65 years old when first interviewed between 1993 and 1997.
The participants completed two follow-up interviews at three-year intervals, which included four standardized cognitive tests and dietary questions on the frequency of consumption of 139 different foods. They were also questioned on daily activities, exercise levels, alcohol consumption and medical history.
The researchers examined whether overall dietary consumption patterns accounted for the significant inverse association between cognitive decline and fish consumption, but the rate differences did not change after adjusting for consumption of fruit and vegetables.
Moreover the people who ate most fish tended to have an 'unfavourable risk profile for cognitive change', said the authors, as they were more often black, with lower alcohol intake and higher prevalence of heart conditions, putting them at higher risk of dementia.
The findings add to the strong evidence showing that fish consumption protects heart health. If further evidence confirms its benefits on the brain, changing dietary habits could be important for the growing ageing population around the world.
There are nearly 18 million people with dementia in the world and the most common cause of this dementia is Alzheimer's disease. By 2025 this figure will rise to 34 million, with 71 per cent of these likely to live in developing countries, making the need for prevention of the uncurable disease crucial.