Use of atypical antipsychotics (AAs) to control behavioral and psychiatric symptoms of dementia (BPSD) is controversial. We sought to understand the prevalence of BPSD and indications for AA use in an AD population treated at an academic memory disorders center. Clinical information of consenting patients in our center is collected in a standardized format and stored electronically. Diagnoses are made according to NINCDS-ADRDA criteria. The Neurosychiatric Inventory (NPI) (Copyright Jeffrey L. Cummings, M.D.) was added to the standard assessment battery in 2007. Patients were included if they had a diagnosis of probable AD or amnestic MCI, an NPI at baseline and at one or more annual visits. The total number of NPI symptoms reported at each visit was calculated. The prevalence of each NPI symptom and the proportion with an AA prescription was evaluated at the first annual follow-up (Visit 2). Longitudinal logistic regression was used to calculate the odds of AA treatment for each symptom over all follow-up visits after adjusting for age, sex, race, education, duration of symptoms at the baseline visit, MMSE score, number of symptoms, and total NPI severity and distress scores. Of 291 patients meeting inclusion criteria, 10.5% had no psychiatric symptoms at V2, 39.2% had 1–3, 30.6% had 4–6, and 19.7% had 7–12. The three most prevalent symptoms were agitation (47.8%), apathy (44.7%), and irritability (44.7%). Of patients with at least one symptom, 10.5% were taking an AA. Longitudinal regression modeling indicated that male sex (OR=2.08, 95% CI=1.08, 4.00), total number of symptoms (OR for each additional symptom=1.11, 95% CI=1.03, 1.19), MMSE score (OR=1.11 for each point decline, 95% CI=1.09, 1.15), but not NPI severity or NPI distress scores, increased the odds of treatment with an AA. Adjusting for these variables, only hallucinations (OR= 1.73, 95% CI=1.02, 2.93) and disinhibition (OR=1.69, 95% CI=1.10–2.62) were associated with a significantly increased odds of AA treatment. In this AD population, male sex, severity of dementia and overall BPSD burden were independently associated with AA use. Hallucinations and disinhibition were the only individual symptoms that increased the odds of AA use.