Abstracts

Rationale: Authors conducted the study to examine validity of the magnetoencephalography (MEG) system fabricated by the Korean Research Institute of Standards and Science (KRISS) and evaluate usefulness of MEG in presurgical evaluation of patients with epilepsy.Methods: Seven consecutive patients with intractable partial epilepsy who were candidates for surgery with long-term intracranial electroencephalography (IEEG) monitoring were enrolled. Interictal recordings were obtained with KRISS MEG (152 channel axial gradiometers). Scalp EEG were recorded simultaneously. Interictal epileptiform discharges (IEDs) were identified by visual analysis of MEG and scalp EEG independently. The sources of the IEDs from MEG were analyzed and compared with foci of interictal epileptiform discharges (irritative zone) and ictal onset zone identified by IEEG.Results: Visual analyses of the traces of all 7 patients revealed 212 IEDs (mean 30.3 per patient) from MEG and 167 (mean 23.9 per patient) from scalp EEG. Of these, 107 (39.3% of total IEDs) were common in both modalities. MEG epileptiform discharge sources (MEGEDS) were successfully (which means the good-of-fitness >80%) localized in 97 out of 212 IEDs (45.8%). The rate of successful localization did not differ between IEDs identified only with MEG and those identified with both MEG and scalp EEG. MEGEDS showed similar distribution with irritative zone (Figure 1). Mean distance between MEGEDS and irritative zone was 9.6 12.0 mm, which is significantly shorter than mean distance between MEGEDS and ictal onset zone (18.4 16.0 mm) (p <0.001). 66.3% of MEGEDS were concordant with irritative zone (distance 1 cm) while only 34.9% were with seizure onset zone. Compared to non-clustered MEGEDS, clustered MEGEDS ( 10 sources with 2 cm or less between adjacent sources) were more close to and more frequently concordant with irritative zone (71.6% vs 22.2%, p = 0.003). Only 37.8% of clustered MEGEDS are concordant with seizure onset zone (37.8%).Conclusions: Interictal MEG recording, especially with clustered MEGEDS, predicts irritative zone with relatively high positive predictive value, but negative predictive value is relatively low. Interictal MEG recordings could not effectively predict seizure onset zone. Through the study, KRISS MEG was proved to be a valid MEG system for clinical use.