Objective: Subarachnoid hemorrhage and surgical obliteration of ruptured intracranial aneurysms are frequently associated with neurological and neuropsychological abnormalities. We reported that intraoperative cooling did not improve neurological out-come in good-grade surgical subarachnoid hemorrhage patients, as assessed by the Glasgow Outcome Scale score or other neurological and functional measures (National Institutes of Health Stroke Scale, Rankin Disability Scale, Barthel Activities of Daily Living). We now report the results of neuropsychological testing in these patients. Methods: A total of 1,001 patients who bled ≤14 days before surgery were randomly assigned to intraoperative hypothermia (t = 33°C) or normothermia (37°C). Outcome was assessed approximately 3 months after surgery. Patients underwent the Benton Visual Retention, Controlled Oral Word Association, Rey-Osterrieth Complex Figure, Grooved Pegboard, and the Trail Making tests. T-scores for each test were calculated from normative data. T-scores were averaged to calculate a Composite Score. A test result (or the Composite Score) was considered "impaired" if the T-score was two or more standard deviations below the norm. A Mini-Mental State Examination was also performed. Results: Neurological outcome data were available in 1,000 patients. Sixty-one patients died. Of the 939 survivors, 873 completed 3 or more tests (exclusive of the Mini-Mental State Examination). Patients with poor neurological outcomes were less likely to complete testing; only 3.9% of Good Outcome (Glasgow Outcome Scale score = 1) patients were untested, compared with 38.6% of patients with Glasgow Outcome Scale scores of 3 and 4. There were no prerandomization demographic differences between the two treatment groups. For hypothermic patients, 16.8% were impaired from their Composite Score versus 20.0% of patients in the normothermic group (p = 0.317). For patients in the hypothermic group, 54.5% were impaired on at least one test, compared with 55.5% of patients in the normothermic group (p = 0.865). Similar results were seen in patients with baseline WFNS scores = I. Mini-Mental State Examination scores in the hypothermic and normothermic groups were 27.4 ± 3.8 and 26.8 ± 4.5, respectively. Interpretation: This is the largest prospective evaluation of neuropsychological function after subarachnoid hemorrhage to date. Testing was completed in a high fraction of patients, demonstrating the feasibility of such testing in a large trial. However, the frequent inability to complete testing in poor-outcome patients suggests that testing may be best used to refine outcome assessments in good-grade patients. Many patients showed impairment on at least one test, with global impairment present in 17 to 20% of patients (18-21% of survivors). This was true even among the patients with the best preoperative condition (WFNS = 1). There was no difference in the incidence of impairment between hypothermic and normothermic groups.