The research set out to determine whether clinically apparent and silent brain injuries were linked to cognitive impairment in atrial fibrillation (AF) patients. The 1,227 AF patients in prospective, multicenter cohort research (Swiss-AF) research. At baseline, and again after 2 years, they had undergone standardized brain magnetic resonance imaging (MRI). The researchers identified new small non-cortical infarcts (SNCIs) and large non-cortical or cortical infarcts (LNCCIs), white matter abnormalities (WML), and microbleeds (Mb). A new SNCI/LNCCI identified silent infarcts on follow-up MRI in patients who did not have a clinical stroke or TIA during follow-up. Validated testing was used to assess cognition. The mean age was 71 years, 26.1% were women, and 89.9% of patients were anticoagulated. During the 2-year follow-up, 28 patients (2.3%) had a stroke/TIA. Of the 68 (5.5%) patients with more than or equal to 1 SNCI/LNCCI, 60 (88.2%) were anticoagulated at baseline and 58 (85.3%) had a silent infarct. In patients with brain infarcts, the median (interquartile range) change in Cognition Construct score was 0.12 (−0.22; 0.07), whereas, in those without new brain infarcts, it was 0.25 (−0.09; 0.29). WML or Mb was not linked to cognitive deterioration. After 2 years, 5.5% of patients in a contemporary group had new brain infarcts on magnetic resonance imaging (MRI). The majority of the infarcts were clinically silent and occurred in patients who had been anticoagulated. Both overt and silent brain infarcts had a comparable influence on cognitive deterioration.