In this study, we developed numerical methods for investigating the dynamics of epilepsy from multi-scale EEG recordings acquired simultaneously from the scalp (sEEG) and intracranial subdural and/or depth electrodes (iEEG) in patients undergoing pre-surgical evaluation at the epilepsy center of the Mayo Clinic (Rochester, MN). The data are analyzed using independent component analysis (ICA), which identifies and isolates independent signal components from multi-channel recordings. A realistic individual head model was constructed for a patient undergoing pre-surgical evaluation. The forward problem of electro-magnetic source localization was solved using the Boundary Element Method (BEM). Using this approach, we investigated the relationships between noninvasive and invasive source localization of human electrical brain data sources. A difference of about 1 cm was observed between sources estimated from sEEG and iEEG measurements.