Recent findings suggest that a rich social network may decrease the risk of developing dementia. The authors hypothesized that such a protective effect may be due to social interaction and intellectual stimulation. To test this hypothesis, data from the 1987-1996 Kungsholmen Project, a longitudinal population-based study carried out in a central area of Stockholm, Sweden, were used to examine whether engagement in different activities 6.4 years before dementia diagnosis was related to a decreased incidence of dementia. Dementia cases were diagnosed by specialists according to Diagnostic and Statistical Manual of Mental Disorders, Third Edition, Revised, criteria. After adjustment for age, sex, education, cognitive functioning, comorbidity, depressive symptoms, and physical functioning at the first examination, frequent (daily-weekly) engagement in mental, social, or productive activities was inversely related to dementia incidence. Adjusted relative risks for mental, social, and productive activities were 0.54 (95% confidence interval (CI): 0.34, 0.87), 0.58 (95% CI: 0.37, 0.91), and 0.58 (95% CI: 0.38, 0.91), respectively. Similar results were found when these three factors were analyzed together in the same model. Results suggest that stimulating activity, either mentally or socially oriented, may protect against dementia, indicating that both social interaction and intellectual stimulation may be relevant to preserving mental functioning in the elderly.