OBJECTIVE: To investigate the diagnostic accuracy of various evoked potential monitoring techniques in predicting postoperative neurologic deficit in intracranial aneurysm surgery. METHODS: A literature search of the MEDLINE, Embase, and Cochrane databases was conducted for English language articles published between March 31, 1983 and March 31, 2016. Original studies that reported the use of evoked potential monitoring during intracranial aneurysm surgery in predicting postoperative neurologic damage were selected, and their relevant reference lists were hand searched. Test performance characteristics were summarized using hierarchic summary receiver operating characteristic (ROC) curves and bivariable random-effects models. RESULTS: Thirteen qualifying studies (1597 patients; 1689 aneurysms) from 6 countries were identified. Eight studies investigated the use of the somatosensory evoked potential (SSEP) monitoring technique, 5 investigated transcranial motor evoked potential (TcMEP) and another 5 investigated direct cortical motor evoked potential (DMEP). Bivariable pooled sensitivity and specificity were 48% (95% confidence interval [CI], 30.7-65.0) and 92% (CI, 88%-94.4%), respectively, for SSEP; 73% (CI, 21.0%-96.7%) and 94% (CI, 87.1%-97.5%) for TcMEP; and 97% (CI, 74.43%-99.99%) and 89% (CI, 84.0%-94.5%) for DMEP. ROC curve analysis showed that TcMEP had the highest accuracy (area under ROC curve 0.95; 95% CI, 0.93-0.97), followed by DMEP (0.91, 0.89-0.94) and SSEP (0.88, 0.85-0.91). CONCLUSIONS: TcMEP and DMEP have higher diagnostic accuracy than SSEP in predicting postoperative neurologic deficit. The type of anesthetic agent, the use of neuromuscular blocking drugs, and the choice of diagnostic criteria for significant change in cerebral blood flow during aneurysm surgery affect the diagnostic accuracy of evoked potential techniques in predicting postoperative neurologic deficit.