Greater than half of the 4 million elders afflicted with dementia suffer from agitated and passive behaviors. These behaviors are a major source of caregiver stress and are associated with nursing home placement, use of restraints and functional decline. Nurses currently lack adequate scientific or theoretical bases for prescribing approaches that successfully respond to these behaviors. Some evidence exists that therapeutic recreational activities help but little is know about their mechanism of effect. The long-term goal of this project is to improve the scientific basis for prescription of therapeutic recreational activities for persons with dementia with the ultimate goal of improving symptom control and quality of life. The project will test an intervention theory that combines the Need-driven Dementia-compromised Behavior Model, a mid-range theory of dementia behaviors, and a treatment theory of activity effectiveness. Recreational activities are tailored to match elders' background profile as described in the model, i.e. pre-morbid personality traits, cognitive ability, and physical ability and are prescribed based on interest and skill level. Thirty nursing home residents with moderately severe dementia will be enrolled in this project that uses a cross-over experimental design to test the following hypotheses: When exposed to theory-based activities, H1: Subjects will exhibit greater engagement as measured by time on task; H2; Subjects will exhibit greater calm alertness as measured by heart rate variability; H3: Subjects will exhibit more positive affect, less negative affect and report more positive mood state; H4: Subjects will exhibit fewer dementia behaviors. We will compare outcomes during treatment to those obtained during baseline, and two control conditions: activities matched to skill level only and activities matched to interest level only. Findings from this study will add to our knowledge of how best to prescribe therapeutic activities for persons with dementia and will also help refine our knowledge about the adequacy of Need-driven Dementia-compromised Behavior Model for intervention derivation.