Abstract We analyzed the rationale for invasive monitoring in refractory epilepsy. In 54 selected patients, video/scalp-EEG was insufficient for seizure focus localization. Therefore, bilateral subdural electrodes were implanted for ictal recording. In 40 (74.1%) of 54 patients, ictal electrocorticography (ECoG) localized a seizure focus amenable to resection. Fourteen (25.9%) of 54 patients, had multiple foci or primary generalized seizures. Among 36 patients who had focal resection with at least 1-year follow-up, 32 (88.9%) are either seizure-free or significantly improved. Magnetic resonance imaging (MRI) and interictal single photon emission computed tomography (SPECT) had the highest sensitivity and specificity (80.0 and 81.8%, respectively) and the greatest diagnostic value (64.0 and 77.8%, respectively) for seizure focus localization. Independent of electrophysiologic data, MRI determination of focal abnormality was prognostic for seizure-free outcome. Concordance of one or more noninvasive techniques with ictal ECoG seizure focus localization was statistically significant in predicting seizure-free outcome. Although interest in noninvasive selection of candidates for focal resection is increasing, there remains a role for invasive monitoring of epileptogenic foci that are difficult to localize. Our study should improve selection of patients with refractory epilepsy for focal resection when ictal ECoG is used in conjunction with noninvasive data for surgical decision making.