Silent brain infarcts and white matter lesions are frequently seen on magnetic resonance imaging in healthy elderly people and both are associated with an increased risk of stroke and dementia. Plasma total homocysteine may be a potentially modifiable risk factor for stroke and dementia. We examined whether elevated total homocysteine levels are associated with silent brain infarcts and white matter lesions. The Rotterdam Scan Study is a population-based study of 1,077 people aged 60 to 90 years who had cerebral magnetic resonance imaging. The cross-sectional relation of total homocysteine with silent infarcts and white matter lesions was analyzed with adjustment for cardiovascular risk factors. The mean plasma total homocysteine level was 11.5 micromol/l (standard deviation 4.1). The risk of silent brain infarcts increased with increasing total homocysteine levels (odds ratio 1.24/standard deviation increase, 95% confidence interval 1.06-1.45). The severity of periventricular white matter lesions and extent of subcortical white matter lesions were also significantly associated with total homocysteine levels, even after excluding those with silent brain infarcts. The overall risk of having either a silent brain infarct or severe white matter lesions was strongly associated with total homocysteine levels (odds ratio 1.35/standard deviation increase, 95% confidence interval 1.16-1.58). We concluded that total homocysteine levels are associated with silent brain infarcts and white matter lesions independent of each other and of other cardiovascular risk factors.