Neuropsychiatric behaviour is common in all those with dementia (1). Approaches to behaviour management have been both pharmacologic and nonpharmacologic. This study explored a nonpharmacologic intervention designed to reflect the understanding of behaviour as guided by: Need-driven Dementia-compromised Behaviour, Self-determination, Biophilia and Attention Restoration Theories. It was hypothesized that engaging at a garden vignette would reduce both neuropsychiatric behaviour and psychoactive medication given. A quasi-experimental within-subject repeated measures design was used. The five-phase design included baseline and two-week intervention and washout phases that were repeated. The study was set in a long-term care facility specializing in ‘difficult to manage’ behaviour. Participant admission criteria were limited to proxy consent, and moderate to severe dementia severity with no pain limitations. The garden vignettes included all materials required to ‘garden’ and create a feeling of a ‘garden’. Each vignette was easily accessed, centrally located, open to self-determined visits and available twenty-four hours per day during the intervention phases. The Mini-Mental State Exam and the Geriatric Dementia Scale were used to determine dementia severity. Measurement of neuropsychiatric behaviour was completed during the last week of each phase using the Neuropsychiatric Inventory –Nursing Home (NPI-NH), Cornell Scale for Depression in Dementia (CSDD), Single Question Depression Test (SQDT), Apathy Inventory (AI) and the Ryden Aggression Scale 2 (modified)(RAS2). Chart review recorded psychoactive PRN medication use. Activity at the garden vignette was video recorded twenty-four hours per day, seven days per week for the two weeks of each intervention phase. Significant neuropsychiatric behaviour changes were primarily between baseline and all other phases for the NPI-NH, NPI-NH-OD, CSDD and the RAS2. Greater neuropsychiatric behaviour and caregiver distress at baseline was associated with spending more time at the vignette. Spending significantly more time at the vignette in phase 2 was associated with spending more time in phase 4. Removal of the vignette created greater neuropsychiatric behaviour and caregiver distress in phase 3. A greater level of depression in phase 4 was associated with spending more time at the vignette and being self-determined was associated with less depression. There was no evidence of effect on apathy, self-assessed depression or psychoactive PRN medication administration.