The influence of organic brain changes on the development of depression in the elderly is uncertain. Cross-sectional studies, most often from clinical samples, report associations with brain atrophy and cerebrovascular disease, while longitudinal population studies have given mixed results. Our aim was to investigate whether cortical atrophy and white matter lesions (WMLs) on computed tomography (CT) predict occurrence of depression in the elderly. This is a prospective population-based study with 5-year follow-up. The baseline sample included 525 elderly subjects, aged 70-86 years, without dementia or major depression, with a score on the Mini-Mental State Examination above 25, and without dementia at follow-up. Cortical atrophy and WMLs were evaluated at baseline using CT. The main outcome measure was development of major or minor depression at follow-up according to Diagnostic and Statistical Manual of Mental Disorders, fourth edition, as evaluated using neuropsychiatric examinations and hospital discharge registers. Logistic regression was used to estimate risk. Over the period of 5 years, 20 individuals developed major and 63 minor depression. Presence of temporal lobe atrophy (odds ratio (OR) ¼ 2.81, 95% confidence interval (CI) 1.04-7.62) and moderate-to-severe WMLs (OR ¼ 3.21, 95% CI 1.00-10.26) independently predicted major, but not minor, depression after controlling for various confounders. Other brain changes did not predict occurrence of depression. Our findings suggest that temporal lobe atrophy and WMLs represent relatively independent and complementary pathways to major depression in the elderly. This may have implications for prevention, as both neurodegeneration and cerebrovascular disease have been related to preventable factors.