Objective: To investigate the validity of magnetic source imaging (MSI) to localize seizure-onset zone using 3D analysis of pre-operative MSI source imaging coregistered to post-resection MRI following neocortical epilepsy surgery. Methods: Twenty-two children who had MSI and epilepsy surgery were studied (median age=11 years, 1 year 2 months-22 years). Only seven (31.8%) had localized lesions on pre-operative conventional brain MRIs. Sixteen (72.7%) underwent intracranial EEG monitoring. Mean post-operative follow-up was 4.7 years (1 year 3 months-8 years 2 months). Fifteen patients (68%) were seizure-free. MEG spike dipole sources were superimposed onto post-operative MRIs. The number and proportion of spike dipoles within resection volume were calculated and compared between seizure free and non-free groups. Results: Both number of dipole clusters and proportion of dipoles in resection volume were not associated with seizure-free outcome (p>0.05). In seven cases with MRI lesions, six of these with a ≥70% dipoles within the resection margin were seizure-free, while one with the proportion <70% was not seizure-free. Further, among the 15 cases with non-localized or normal MRI, five with both the proportion <70% and multiple dipoles clusters were post-operatively seizure free. Conclusion: Number and density of clustered spike dipole sources within the surgical resection volume is not associated with postoperative seizure-free outcome. MSI successfully localized the perilesional epileptogenic zone in cases with localized MRI lesions, but not in cases with normal MRI in this study. Even if MEG localizes spikes to a single focal region, confirmation of epilepsy localization with intracranial EEG is still recommended in cases with non-lesional MRI. © 2013 Elsevier B.V.