A series of patients undergoing coronary artery bypass surgery was studied prospectively to see if angiographic evidence of cerebrovascular disease proved predictive of the incidence of neuropsychological deficit 8 days or 8 weeks after surgery. In 47 patients, intravenous digital subtraction angiography was carried out preoperatively to assess the presence and severity of atheromatous changes in the carotid arteries; 51% had evidence of vessel wall disease and 17% had stenosis of at least one carotid artery in the neck, although only one patient had severe narrowing. Overall, 77% of these 47 patients showed a neuropsychological deficit as defined by a significantly reduced score in at least two of 10 tests administered 8 days after surgery. Eight weeks after surgery 36% still showed a deficit. The incidence of neuropsychological deficit was not significantly greater among those patients with angiographically visible carotid artery disease. The mechanism of surgery-related cognitive impairment is briefly discussed in the light of these findings.