Dipole modeling of interictal and ictal EEG and MEG paroxysms

Dipole modeling procedures can be used to statistically estimate the location and the orientation of intracerebral sources of electroencephalographic (EEG) and magnetoencephalographic (MEG) signals. These methods have been applied to interictal spikes for more than 20 years and suggest that interictal paroxysms might be generated by a network of cortical structures rather than by a focal area. In this review we address the questions of (1) the spatial extend of this network in different types of epilepsies, (2) the spatial relationship between this network and other structural of functional abnormalities as assessed by Magnetic Resonance Imaging (MRI) or Positron Emission Tomography (PET), and (3) the reliability of dipole sources of interictal and ictal paroxysms. Dipole modeling results suggest that, in temporal lobe epilepsies, both neocortical and mesio-temporal structures are involved during interictal spikes; frontal lobe epilepsies are often characterized by more complex source distributions, that, in general, involve a large area and bilateral frontal structures. In addition, dipole modeling results are also found in close agreement with MRI data in cases where focal dysplasia or heterotopia are diagnosed. Nevertheless, in most other cases, sources of interictal spikes and MRI lesions, though overlapping in space, are not totally congruent. The best concordance between sources of interictal spikes and glucose hypometabolism on PET data is usually found for temporal lobe epilepsies. Most often, intracranial and intracerebral recordings validate both the localization and the time activation of interictal spike dipoles. However, results obtained for ictal discharges are less reliable, which therefore addresses the usefulness of dipole modeling procedures in assessing sources of ictal discharges. In conclusion, dipole modeling results can rarely be used in planning a selective surgery without invasive recordings. However, the studies reviewed in this paper strongly suggest that their analysis, in combination with other non-invasive data, might be useful to better delineate the epileptogenic zone, and help the implantation of intracerebral electrodes.