Objective:The authors evaluated the sensitivity and selectivity of interictal magnetoencephalography (MEG) versus prolonged ictal and interictal scalp video-electroencephalography (V-EEG) in order to identify patient groups that would benefit from preoperative MEG testing.Methods:The authors evaluated 113 consecutive patients with medically refractory epilepsy who underwent surgery. The epileptogenic region predicted by interictal and ictal V-EEG and MEG was defined in relation to the resected area as perfectly overlapping with the resected area, partially overlapping, or nonoverlapping.Results:The sensitivity of a 30-minute interictal MEG study for detecting clinically significant epileptiform activity was 79.2%. Using MEG, we were able to localize the resected region in a greater proportion of patients (72.3%) than with noninvasive V-EEG (40%). MEG contributed to the localization of the resected region in 58.8% of the patients with a nonlocalizing V-EEG study and 72.8% of the patients for whom V-EEG only partially identified the resected zone. Overall, MEG and V-EEG results were equivalent in 32.3% of the cases, and additional localization information was obtained using MEG in 40% of the patients.Conclusion:MEG is most useful for presurgical planning in patients who have either partially or nonlocalizing V-EEG results.