Some studies have shown that cognitive function is positively affected by an engaged and active lifestyle. However, a recent study found evidence for an opposite causal direction, i.e., persons with good cognitive function more often start to engage in leisure-time activities. Here, we longitudinally examine the causal direction of the association between an engaged lifestyle and cognitive function in middle and old-aged Dutch men and women. The participants in the prospective cohort study "Maastricht Aging Study" (MAAS) were recruited from a register of 15 family practices in the South of the Netherlands. There were 830 non-demented men and women, aged 49 to 81 in 1993-1995 (baseline phase). They were re-examined three years later (follow-up phase). During both phases, all persons reported on their participation in mental, social, and physical activities. Six separate neuropsychological tests, including the Mini-Mental State Examination, were used to define cognitive function at baseline and follow-up. All three activities measured were negatively related to cognitive decline between both phases. Effects were moderate, but consistent. Persons participating in all three activities were particularly protected against longitudinal decline. Furthermore, persons with the best baseline cognitive performance were more likely to increase their number of activities during follow-up compared with persons with the poorest scores. In summary, an engaged lifestyle and cognitive function mutually influence each other in middle and old aged, non-demented persons. This reciprocal association is characterized by simultaneous positive effects of leisure-time activities and good cognitive function on cognitive function and leisure-time activities, respectively. This reciprocal association may create a self-reinforcing, beneficial or adverse life-course in middle and old age.