The review addressed a clear objective and criteria for study inclusion were stipulated. Appropriate electronic databases were searched. Attempts were made to identify unpublished studies. The restriction to English-language studies meant that there was a risk of language bias. Steps were taken to minimise errors and bias for study selection and data extraction. No quality assessment was undertaken, so the reliability of the results was unclear. The authors stated that the studies were heterogeneous with respect to MEG analysis, patient populations and epilepsy type. The authors used mean confidence intervals to report results, which did not provide additional information on the robustness of the results. Therefore, statistical pooling of the study results may not have been appropriate, particularly given the use of additional diagnostic tools and inclusion of patients who underwent multiple subpial transection. It was unclear to what extent use of seizure outcome post-resective surgery was an appropriate outcome measure to confirm the accuracy of MEG as a diagnostic tool. The authors correctly stated limitations of the review in terms of small study samples and multiple sources of variation across the studies.
Methodological flaws and uncertainty about the comparisons made and use of outcome measures mean that the authors conclusions should be interpreted with caution and the reliability of the results is unknown.