Conventional visual analysis and dipole density analysis of magnetoencephalographic data for both spike and low-frequency magnetic activity were compared for presurgical evaluation in temporal lobe epilepsy (TLE) in a sample of 26 drug-resistant operated TLE patients. A series of logistic regression analyses were performed. Dipole density sensitivity was superior to visual localization analysis. Three separated logistic models were calculated for interictal spikes, low-frequency magnetic activity, and the combination of both measures. A combined interictal spike/low-frequency magnetic activity model predicted correctly the operated temporal lobe in all patients. Clear-cut criteria for the probability model are proposed that are valid for 92.3% of cases in the sample. The quantitative approach proposed by this study is an evidence-based model for presurgical evaluation of temporal lobe epilepsy, which improves previous magnetoencephalographic investigations and establishes working clinical criteria for patient evaluation in TLE.