Although several clinical studies have shown that increased serum concentrations of protein S100B predict ischaemic brain damage after cardiac surgery, S100B may also be released from the heart or other injured tissue. We therefore investigated the correlation between serum S100B levels and those of the specific cardiac marker troponin I in order to assess the cerebral vs. extracerebral origin of S100B. In 64 cardiac surgical patients, serial blood samples were drawn for the measurement of S100B and troponin I before surgery and for seven days after surgery. Neurological function was assessed before with the National Institutes of Health Stroke Scale and the Folstein Mini Mental Test. The data show that a sustained increase in serum S100B levels is associated with neurological dysfunction, as witnessed by a positive correlation between S100B values and the results of the neuropsychological tests. In contrast, the early postoperative increased levels of protein S100B derive from cardiac tissue, as shown by the positive correlation between S100B and cardiac troponin I levels.