Among neurologically normal volunteers approaching age 65 with an option for retirement, a four-year prospective longitudinal study was designed to examine effects of different levels of physical activity on cerebral perfusion by between-group comparisons. After the fourth year, cognitive performance was also tested. Three groups were compared, each composed of 30 elderly volunteers, assigned as follows: Group 1, who continued to work; Group 2, who retired but participated in regular physical activities; and Group 3, who retired but did not participate in regular, planned physical activities. Retirees who elected to become physically inactive exhibited significant declines in cerebral blood flow (CBF) throughout four years of follow-up. Those who continued to work or retirees who elected to participate in regular activities sustained more constant CBF levels. Active retirees and those who continued to work also scored better on cognitive testing after the fourth year of follow-up compared to inactive retirees.