Brain imaging based on functional MRI (fMRI) provides a powerful tool for characterizing age-related changes in functional anatomy. However, between-population comparisons confront potential differences in measurement properties. The present experiment explores the feasibility of conducting fMRI studies in nondemented and demented older adults by measuring hemodynamic response properties in an event-related design. A paradigm involving repeated presentation of sensory-motor response trials was administered to 41 participants (14 young adults, 14 nondemented older adults, and 13 demented older adults). For half of the trials a single sensory-motor event was presented in isolation and in the other half in pairs. Hemodynamic response characteristics to the isolated events allowed basic response properties (e.g., amplitude and variance) between subject groups to be contrasted. The paired events further allowed the summation properties of the hemodynamic response to be characterized. Robust and qualitatively similar activation maps were produced for all subject groups. Quantitative results showed that for certain regions, such as in the visual cortex, there were marked reductions in the amplitude of the hemodynamic response in older adults. In other regions, such as in the motor cortex, relatively intact response characteristics were observed. These results suggest caution should be exhibited in interpreting simple main effects in response amplitude between subject groups. However, across all regions examined, the summation of the hemodynamic response over trials was highly similar between groups. This latter finding suggests that, even if absolute measurement differences do exist between subject groups, relative activation change should be preserved. Designs that rely on group interactions between task conditions, parametric manipulations, or group interactions between regions should provide valuable data for making inferences about functional-anatomic changes between different populations.