We hypothesized that the magnitude of the association between plasma homocysteine concentration and cognitive performance is larger for ApoE-epsilon4 carriers than for non-carriers. Nine hundred eleven dementia-free and stroke-free subjects (59% women) from the Maine-Syracuse study (26-98 years old) were stratified into no-ApoE-epsilon4 (n=667) and ApoE-epsilon4 carrier (n=244) cohorts. Employing a cross-sectional design and multiple regression analyses, plasma homocysteine was related to multiple domains of cognitive performance within these cohorts. When unadjusted, and with adjustment for age, education, gender, ethnicity, and previous cognitive examinations, homocysteine concentrations were inversely related to cognitive performance within both ApoE cohorts, with higher magnitude of associations within the ApoE-epsilon4 cohort. With adjustment for cardiovascular disease risk factors, cardiovascular disease, and B-vitamin concentrations, the higher magnitude of associations between plasma homocysteine and cognitive performance within the ApoE-epsilon4 cohort relative to the no-ApoE-epsilon4 cohort persisted; but associations of plasma homocysteine and cognitive performance were attenuated and no longer significant within the no-ApoE-epsilon4 cohort. Presence of the ApoE-epsilon4 allele modifies the relation between plasma homocysteine and cognitive performance.