Basic Principles of Magnetoencephalography and Magnetic Source Imaging

MEG is a technique for measuring the magnetic fields associated with the intracellular current flows within neurons. Source localization of epileptic spikes and evoked responses as determined by MEG are coregistered with magnetic resonance imaging (MRI) as magnetic source imaging (MSI). MEG is based on the physical phenomenon that electrical currents generate accompanying magnetic fields. The orientation of the magnetic field relative to the electrical current is described as Orsted’s “right-hand rule,” which states that when the thumb of the right hand is pointed in the direction of the electrical current, the surrounding magnetic flux is aligned in the direction of the other four right fingers. MEG uses highly sensitive biomagnetometers to detect extracranial magnetic fields produced by intracellular neuronal currents. On the basis of the right-hand rule, MEG is primarily sensitive to signals arising from regions in which the apical dendrites are tangentially oriented to the skull and scalp surface.

The source localization has to solve the inverse problem that calculates the three-dimensional intracranial location, orientation, and strength of the neuronal sources backward from a measured extracranial magnetic field pattern. The accuracy of a solution of the inverse problem depends on numerous factors, including the forward problem. The forward problem uses an iterative algorithm to determine the location, orientation, and strength of the equivalent current dipole that best account for the measured magnetic field pattern. The accuracy of the forward problem is critically determined by the shape and conductivity of the volume conductor of head model. MEG forward solution is more robust than that of EEG because of homogeneous conductivity in a magnetic field. Therefore, the localization of both MEG spike sources (MEGSS) and evoked responses on MSI is quite reliable for presurgical evaluation in pediatric localization-related epilepsy.2 In short, MEG is an extremely valuable and reliable technique with which to localize the source of interictal epileptiform discharges.7

MEG Spike Sources

The Hospital for Sick Children in Toronto, Canada, has pioneered the use of MEG for clinical application in pediatric epilepsy. From August 2000 to December 2007, MEG was studied in more than 600 patients with localization-related epilepsy as part of a presurgical protocol that also includes careful definition of seizure semiology based on clinical features and prolonged scalp video-EEG (VEEG), MRI, and neuro-psychological testing.2 More than 200 of these children have undergone epilepsy surgery procedures based on the concordance of these data.

We have defined the distribution of MEGSS by number and density.6 An MEG spike cluster is six or more spike sources with 1 cm or less between adjacent sources. A MEG spike scatters is fewer than six spike sources regardless of the distance between sources or spike sources with more than 1 cm between sources regardless of the number of sources in a group. The zone of clustered MEGSS correlates with the ictal onset zone and the prominent interictal zone as determined by extraoperative intracranial VEEG as recorded from subdural electrodes. MEG spike scatters alone should be examined by intracranial VEEG, because an epileptic zone may exist within the scatter distribution of MEGSS. We have shown the complete resection of MEG clusters to be correlated with postsurgical seizure freedom. For presurgical evaluation, concordant lateralization of the EEG spike sources on scalp VEEG and the clustered MEGSS indicate the primary epileptogenic hemisphere.8 Discordant lateralization of EEG spike sources and MEGSS indicate an undetermined epileptogenic hemisphere and contraindicate surgery without further testing.8

Lesional Epilepsy

Surgical treatment of seizure disorders secondary to a lesion requires that the lesion be removed and epileptogenic tissue removed or disconnected. MSI provided accurate data on the spatial relations of lesion, epileptogenic zone, and functional cortex in children with lesional extratemporal epilepsy.4 MEG delineates asymmetric epileptogenicity surrounding lesions and eloquent cortex. When the focal seizures are secondary to a neoplasm, complete tumor resection with resection of MEGSS marginal to the tumor is associated with favorable outcomes despite residual postexcisional electrocorticography (ECoG) spikes and extramarginal MEGSS. When the focal seizures are secondary to dysplastic brain, the cortical dysplasia as characterized by clusters of MEGSS within and extending from MRI lesion should be removed completely, including both the anatomical lesion and MEGSS, to achieve seizure freedom.9,10

MEG has proven useful in identifying which children with tuberous sclerosis complex (TSC) may be candidates for epilepsy surgery ( Fig. 7.1 ). Wu et al11 studied six children with focal seizures secondary to TSC. In these six TSC patients with focal seizures secondary to bilateral multilobar cortical tubers, ictal VEEG predicted the region of resection with 56% sensitivity, 80% specificity, and 77% accuracy. Interictal MEG, however, fared better, with 100% sensitivity, 94% specificity, and 95% accuracy. In TSC, MEGSS tend to localize around visible tubers. MEG enabled precise localization of the epileptic foci and provided crucial information of surgical treatment in children with localization related epilepsy secondary to TSC.12–14

Extratemporal Lobe Epilepsy

In infants and young children, the occipital lobe frequently generates focal onset seizures and even infantile spasms.15 In addition, more occipital spikes migrate anteriorly than frontal spikes migrate posteriorly in children.16 Therefore, in younger patients with extratemporal localization-related epilepsy, multiple clustered MEGSS are often seen in temporal/parietal/occipital lobes, whereas in older patients, single clusters are observed frequently with an ictal onset zone in the frontal lobe.17 These data suggest that posteriorly dominant epilepsy can extend anteriorly to expand the epileptic network through anatomical and functional connections in developing brains, whereas frontal lobe epilepsy less frequently migrates to other lobes. Therefore, multiple clustered MEGSS associated with the posterior epileptic network may require extensive resection, especially in young children. Conversely, the single cluster that correlates with a discrete anterior epileptic region in relatively old patients may predict a successful focal resection.

The diagnosis of frontal lobe epilepsy may be compounded by poor electroclinical localization on scalp EEG, caused by deep, distributed, or rapidly propagating epileptiform activities over the bilateral hemispheres. The yield of MEGSS in terms of localization of the epileptogenic zone in frontal lobe epilepsy is superior to that of EEG because of high resolution of spatial and temporal data with the former18 ( Fig. 7.2 ). When interictal epileptiform discharges on scalp EEGs show a diffuse hemispheric distribution, or bilateral synchronous spike-waves, analysis of MEGSS at the earliest time point or dynamic statistical parametric maps can lateralize and localize the epileptogenic zone.19,20

In age-related epilepsy, benign rolandic epilepsy (BRE) and Landau-Kleffner syndrome (LKS) are forms of childhood epilepsy that share particular characteristics and can be controlled with medication. Both BRE and LKS have identical orientation of MEGSS directing vertical to the rolandic21 and sylvian.22 However, a subgroup of patients who manifest some of the characteristic of both BRE and LKS, but who do not fulfill all criteria for these epilepsy syndromes have been designated atypical BRE and LKS variant. We have introduced the term malignant rolandic-sylvian epilepsy to describe this subgroup, which is characterized by fronto-centrotemporal spikes on EEG, absence of lesions on MRI, MEGSS with random orientations around rolandic and sylvian fissures, intractable sensorimotor partial seizures that progress to secondary generalization, and neurocognitive problems.3