Magnetoencephalography in the preoperative assessment of epilepsy surgery candidates

Epilepsy is a neurological disease characterized by disturbances in the electric activity of the brain. Up to a third of epilepsy patients have an inadequate response to anti-epileptic medication. In these patients, the possibility of surgical treatment, that is, the surgical removal of the epileptogenic brain tissue, should be considered. Some patient subgroups, such as those with no potential epileptogenic lesion in the magnetic resonance imaging (MRI) of the brain, so called MRI-negative patients, are particularly challenging and typically require comprehensive noninvasive and invasive presurgical investigations. Magnetoencephalography (MEG) is a noninvasive method in which the magnetic fields generated by the cortical post-synaptic currents are recorded. The generators of the MEG signal are thought to be the same as the generators of the electroencephalographic (EEG) signal. However, the two methods differ in their sensitivity to source currents in different orientation and depth. Therefore, the two methods are complementary. One of the studies of this thesis examined the utility of MEG in patients whose epilepsy is caused by a focal cortical dysplasia (FCD). FCDs are highly epileptogenic malformations of cortical development, and often result in epilepsy that is refractory to anti-epileptic medication. They may often be invisible in the brain MRI, requiring extensive presurgical evaluation. In this study, we found that a more complete removal of the MEG interictal spike source cluster locations was associated with better surgical outcome. Additionally, the concordance between the interictal MEG spike localizations and the epileptogenic zone estimated by an intracranial recording was considered good in most of the MRI-negative patients, who were of a special interest in this study. The second study investigated the use of MEG as a tool to assess language lateralization in epilepsy patients. We introduced a simple auditory paradigm that had previously been shown to elicit differently lateralized MEG responses in healthy right-handed vs. left-handed subjects suggesting a potential value in studying language lateralization. We compared the lateralization results of MEG to the results of the intracarotid amobarbital procedure (IAP), also known as the Wada test. Although the lateralization results of MEG did not show a significant correlation to the IAP lateralization results, we discovered an MEG response pattern that was only present in patients with left hemisphere dominant IAP results. This pattern was seen in 67% to 78% of the patients with left dominant IAP results depending on the MEG analysis method utilized. In the third study of this thesis, we examined the utility of a novel kurtosis beamformer analysis method, SAMepi, in patients with parietal lobe epilepsy to localize the sources of interictal epileptiform activity. Parietal lobe epilepsies are rare in epilepsy surgery series, which may partially be explained by the difficulty of localizing the epileptogenic zone in this patient group. SAMepi is a semi-automatic analysis method, that would significantly reduce the workload and the subjective decision-making compared to the current clinical practice. We found that SAMepi analysis provided localization results that in most patients were similar to the results of the current clinical standard, that is, the equivalent current dipole (ECD) analysis. However, its sensitivity in detecting epileptiform discharges was somewhat lower than that of the current practice of visually reviewing the MEG signals. In this thesis, we have aimed to develop the utility of MEG in challenging epilepsy surgery patient groups, and to introduce a novel language paradigm as well as a novel semi-automatic analysis method for interictal spike localization. Although MEG already is quite an established tool in the presurgical evaluations of epilepsy surgery candidates in many epilepsy surgery centers, the continuous development of the method is crucial to maximize its diagnostic yield and to justify its higher cost compared to more conventional neurophysiological techniques, such as EEG.