Behavioral and psychological symptoms of dementia (BPSD) have a significant impact on patients and their families. These symptoms affect quality of life and medical and social service utilization, and determine time of institutionalization. Over the past three decades, academic-based clinical investigators have conducted trials of almost all classes of marketed psychotropics in an attempt to ameliorate these symptoms. With few exceptions, these studies suffered from serious methodologic shortcomings, such as small patient populations or too many outcome measurements (Stern et al., 1997). Despite these flaws and equivocal findings, a consensus has emerged, and antipsychotics have become the most commonly prescribed drugs for this condition (see consensus statement in this issue). Unfortunately, a meta-analysis of trials of conventional neuroleptics has revealed that this class of drugs is barely superior to placebo (see article by Lon Schneider in this issue). In addition, conventional neuroleptics cause extrapyramidal adverse effects, tardive dyskinesia, sedation, and cardiovascular instability, adverse effects to which the elderly are particularly vulnerable.