There is accumulating evidence from animal and epidemiologic studies that physical exercise is neuroprotective in healthy animals and humans and can prevent cognitive decline in chronic neurodegenerative processes like Alzheimer's dementia. However, data from well-designed interventional, randomized non-pharmacologic trials is lacking in contrast to other areas of medicine like prevention of hypertension, diabetes or the antipsychotic-associated metabolic syndrome. The demonstration of a potential positive effect of physical exercise on preventing dementia using a controlled study design would represent a significant progress in the prevention of dementia and public health, especially as long as other treatments for dementia prevention are lacking.