Background and purpose: Cerebral infarcts increase the risk for cognitive impairment. The relevance of location and number of infarcts with respect to cognitive function is less clear.

Methods: We studied the cross-sectional association between number and location of infarcts and cognitive performance in 4030 nondemented participants of the Age Gene/Environment Susceptibility-Reykjavik Study. Composite scores for memory, processing speed, and executive function were created from a neuropsychological battery. Subcortical, cortical, and cerebellar infarcts were identified on brain MRI. We performed linear regression analyses adjusted for demographic and vascular risk factors, depression, white matter lesions, and atrophy.

Results: Compared to participants with no infarcts, those with infarcts in multiple locations (n=287, 7%) had slower processing speed (beta=-0.19; P<0.001) and poorer memory (beta=-0.16; P<0.001) and executive function (beta=-0.12; P=0.003). Compared to no infarcts, the presence of either subcortical infarcts only (n=275; beta=-0.12; P=0.016) or cortical infarcts only (n=215; beta=-0.17; P=0.001) was associated with poorer memory performance. Compared to no infarcts, a combination of cortical and subcortical infarcts (n=45) was associated with slower processing speed (beta=-0.38; P<0.001) and poorer executive function (beta=-0.22; P=0.02), whereas a combination of cerebellar and subcortical infarcts (n=89) was associated with slower processing speed (beta=-0.15; P=0.04). Infarcts in all 3 locations was associated with slower processing speed (beta=-0.33; P=0.002).

Conclusions: Having infarcts in >1 location is associated with poor performance in memory, processing speed, and executive function, independent of cardiovascular comorbidities, white matter lesions, and brain atrophy, suggesting that both the number and the distribution of infarcts jointly contribute to cognitive impairment.