Abstracts

Rationale: We sought to determine whether the sources of interictal MEG background activity can help identify areas of cortical dysfunction in patients being evaluated for epilepsy surgery. We compared source-power topography in different frequency bands to epileptogenic regions that were identified using intracranial EEG recordings. Methods: We studied seven patients who underwent MEG studies at the Froedtert MEG Center (Medical College of Wisconsin)as part of their presurgical evaluation for epilepsy surgery. All the patients subsequently underwent intracranial EEG studies followed by resection of epileptogenic zones, and had Engel s Class I outcome 6 months post-surgery. MEG recordings were performed using a 306-channel MEG system (Elekta Neuromag, Helsinki, Finland) at a sampling rate of 2000 Hz. Patient-specific head models were constructed by extracting grey matter surfaces from each patient's T1-weighthed MRI volume data using automated segmentation and surface tessellation techniques. Our source imaging method is based on 15,000 elementary current sources constrained to the cortical mantle, with a source separation of about 3 mm. From each patient, between 5 to 7 minutes of awake eyes-closed MEG recording was used for source analysis. All samples of MEG recordings were subjected to automated artifact, EKG, and noise reduction using Signal Space Separation and Principal Component Analysis. Each patient's recording was divided into 2-second long epochs, and frequency-domain distributed source imaging was performed for frequencies from 1 to 500 Hz using weighted L2 minimum norm estimates. The topography of source power in different frequency bands was then compared to areas that were resected in each patient. Results: Our patients ranged in age from 17 to 46 years (35 /-9.9 years). MR imaging abnormalities were present in 6 of these patients. Seizure origin was frontal in 3 patients, and anterior temporal, temporo-parieto-occipital, medial parietal, and temporo-parietal in one patient each. Source power at low frequencies (delta and theta) showed focal concentration with maxima within the resection zones in five of these patients. One patient had asymmetric focal delta and theta power concentration concordant with a frontal resection area that was dwarfed by biposterior low-frequency power related to posterior slow waves of youth. The last patient showed extensive bilateral temporo-occipital source power in all bands, although surgical resection only included the left anterolateral temporal neocortex. Conclusions: Cortical sources of low-frequency (delta and theta range) MEG background activity are highly correlated with regions of seizure onset in patients with partial epilepsy. Source imaging of background MEG activity can provide useful information relevant to localizing cortical dysfunction in patients with partial epilepsy, even in the absence of interictal epileptic spikes.