Elevated serum homocysteine has been associated with increased risk of Alzheimer's disease. Furthermore, elevated homocysteine levels are related to cognitive dysfunction in the elderly. The aim of the present study was to explore the disease specificity of the relation between serum total homocysteine levels and cognitive function. For this, we summarize data from several studies on homocysteine levels in both normal and pathological conditions performed in our laboratories and evaluate possible mechanisms of effects of elevated homocysteine levels in the central nervous system. Total homocysteine levels were measured in serum of: 1) healthy aging individuals; 2) patients with Alzheimer's and Parkinson's disease and patients with other cognitive disorders; and 3) patients with multiple sclerosis. Increased serum homocysteine concentration was related to worse cognitive performance over a 6-year period in the normal aging population (r = -0.36 to -0.14, p <0.01 for the Word learning tests; r = 0.76, p <0.05 for the Stroop Colored Word test). Homocysteine was only increased in patients with Parkinson's disease on L-Dopa therapy (18.9 vs. 16.5 mu mol/L in healthy controls), and not in dementia patients. Homocysteine was elevated in patients with progressive multiple sclerosis (15.0 mu mol/L, n=39, compared to 12.0 mu mol/L in 45 controls) and correlated to both cognitive and motorfunction (r = -0.33 and -0.33, p <0.05, respectively). The relationship between homocysteine and cognitive function in non-pathological and pathological situations indicates that changes in its levels may play a role in cognitive functioning in a broad spectrum of conditions.