Cerebrovascular lesions that are apparent in magnetic resonance scans and regioselective atrophy of the brain have been proposed as a causative or exacerbating factor in depression with late-life onset. The objective of this study was to investigate whether deep white matter or periventricular hyperintensities, small ischemic lesions, and brain atrophy contribute to late-onset depression in the nondemented elderly. Based on a group of 606 individuals of identical age (75.8 years, standard deviation: 0.45 years) residing in two districts of Vienna, the authors built a case-control cohort (ratio: 1:4) consisting of 51 individuals with late-onset major or minor depression matched with 204 subjects of identical gender and education status without depression, resulting in two groups that were homogenous with respect to age, place of residence, gender, and education. Scores for focal brain lesions, mediotemporal lobe atrophy, and ventricular enlargement as well as risk factors for vascular disease were compared with cognition and depression status. Depressed individuals had significantly lower scores than nondepressed subjects in all measures of cognitive and executive function. No significant relation was found between a diagnosis of depression and any type of discrete brain lesions, but measures of brain atrophy (Cella Media indices, mediotemporal atrophy) showed a clear statistical relation to depression. No relationship was found between depression and lipid parameters. The authors found no indication that white matter hyperintensities or minor ischemic lesions played a role in our depressed cohort, casting doubt on the vascular hypothesis of late-onset depression.