Baked or broiled (but not fried) fish may help to lower the risk of dementia.  Cyrus Raji, from the University of Pittsburgh (Pennsylvania, USA), and colleagues assessed 260 men and women, mean age 71 years, when they enrolled in the Cardiovascular Health Study between 1989 and 1990. At that time, they filled out questionnaires on dietary intake; 163 reported eating fish at least weekly, and some did so as often as four times a week. All subjects had an MRI 10 years later to assess brain volume, and then had follow-up cognitive testing between 2002 and 2003.  The researchers found that patients who ate fish at least once a week had greater volume in the frontal lobes and the temporal lobes, including the hippocampus and the posterior cingulate gyrus – the  brain areas responsible for memory and learning, which are severely affected in Alzheimer's disease.  Five years after the MRI, the team found that 30.8% of patients who had low fish intake had developed mild cognitive impairment or dementia, compared with just 3.2% of those who had the highest fish intake and the greatest preservation of brain volume. As well, the researchers observed that 47% of patients with brain atrophy who didn't eat fish had abnormal cognition five years later, compared with 28% of those who ate more fish and had more gray matter volume. In further analyses, the team revealed that mean scores for working memory -- a function severely impaired in Alzheimer's disease -- were significantly higher among those who ate fish weekly.