The assessment of depression in patients with brain pathologies--a topic of considerable clinical and research interest--is complicated by a variety of factors. Among the most problematic are cognitive consequences of brain injury that can diminish the reliability and validity of information used to diagnose depression, determine its severity, ascertain its predictors, and evaluate its impact. In this article, the authors examine the challenges to depression assessment in patients who have had a stroke, the neurologically impaired population in which it has been most frequently studied. Focusing on poststroke depression research, they describe methodological limitations that may contribute to conflicting outcomes and conclusions and offer suggestions for improving the specificity, consistency, validity, and reliability of assessment methods and procedures when investigating depression in patients with brain pathologies.