Neuropsychiatric symptoms (NPS) are central features of dementia and an important treatment target. They should be assessed in future studies of emerging dementia therapies, using appropriate measures matched to the purpose of each study. Several significant issues remain regarding (1) the classification of these symptoms into syndromes, and (2) the development of better clinical measures for their quantification. In particular, effort should be directed at assessing their evolution over shorter time periods, and at using more objective methods in their measurement, such as actigraphy. These issues can be solved with nosologic study and other advances that could be brought about quickly, if appropriate time and effort are allocated. Empirical characterization of clinically meaningful change in NPS--by examining their relationship with dementia care burden, disability, quality of life, caregiver distress, and resource utilization--would be an important advance.