Background. With increasing clinical experience, it has become clear that two distinct forms of neurological injury occur after operations on the thoracic aorta that require temporary exclusion of the cerebral circulation. Traditionally, evaluation of neurological outcome was limited to reporting the incidence of postoperative stroke related to ischemic infarcts due to particulate embolization. More recently, the symptom complex defined as 'temporary neurological dysfunction' (TND) was recognized as a functional manifestation of subtle and presumably transient brain injury, but whether this early postoperative syndrome is associated with long-term deficits of cognitive and intellectual functions has not been established. Methods. With Institution Review Board approval, 105 patients undergoing elective thoracic aortic surgery were entered into a protocol involving neuropsychological evaluation with a battery of tests preoperatively, and 1 and 6 weeks postoperatively. Patients who could not be tested adequately or had documented strokes were eliminated from final analysis. Seventy-one patients completed the neuropsychological evaluation, which consisted of eight tests consolidated into five domains: attention, cognitive speed, memory, executive function, and fine motor function. Independent observers also determined whether temporary dysfunction was present, and graded its severity based on a fixed but subjective clinical scale, ranging from simple disorientation and lethargy or confusion (grade 1-2) to prolonged extreme agitation or psychotic behavior requiring treatment with psychotropic drugs (grade 3-5). Data were normalized to baseline values, and were analyzed using analysis of variance, analysis of covariance (ANCOVA), and χ2 as necessary. Results. A previous analysis had shown that patients who could not be tested or had poor scores 1 week postoperatively were more likely to perform poorly at 6 weeks (odds ratio 5.27, p < 0.01). In the current study, in order to determine the clinical relevance of TND, patients were analyzed retrospectively according to their performance in neuropsychological testing: patients with no change or a decline of less than 50% in tests of memory, motor function, and attention 1 week postoperatively (group 1, n = 49) were compared with those with a negative change exceeding 50% in the same functions at 1 week (group 2, n = 22). The overall incidence of TND was 28.1% (20/71). The incidence of TND in group 2 (14/22, 63%) was significantly higher than in group 1 (6/49, 12%; p = 0.0006). Similarly, the severity of TND (as assessed by clinical score > 2) was also significantly higher in group 2 (11/14) compared with group 1 (0/6; p = 0.006.) Conclusions. The incidence and severity of clinically apparent temporary neurological dysfunction correlates significantly with poor performance on neuropsychological tests 1 week postoperatively. Such poor performance predicts continued deficits in memory and motor function at 6 weeks. Thus, TND may not be a benign self-limited condition as previously supposed, but rather a clinical marker for insidious but significant neurological injury associated with measurable long-term deficits in cerebral function. A concerted effort to reduce the incidence of this complication is therefore necessary.