Abstracts

Rationale: Drug-resistant focal epilepsy is frequently caused by focal cortical dysplasia (FCD). FCD consisted of either extensively localized in the brain surface (FCDS) or small and deep seated at the bottom of sulcus (FCDB). We previously reported that single moving dipole (SMD) method localized MEG dipoles to the remote area from FCDB due to closed-field effect, low density of neuronal cells and small size of FCDB. Dynamic statistical parametric mapping (dSPM) overlays spatiotemporal distribution of spike sources onto inflated brain model. Epileptogenic activity in the sulcus can be visualized by using inflated head model. We developed advanced dSPM (AdSPM) to identify epileptic FCD. We investigated spatial congruence between 1), cluster in SMD and 2), spike volume in AdSPM for FCD. Methods: We retrospectively analyzed 19 children (10 males; mean, 9.3 years; age ranging, 3-15 years) who underwent presurgical evaluation for epilepsy surgery. The MEG cluster consisted of >=6 spike dipoles with =1 cm between adjacent dipoles; scatters consisted of < 6 spike dipoles regardless of the distance between dipoles, or spike dipoles with >1 cm between dipoles regardless of the number of dipoles in a group (Iida et al., 2005). We defined "Hit" when there was an overlap between the area of cluster and FCD. We defined "No hit" when cluster was remote from FCD. We categorized scatter without cluster as "No hit"group. The same number of MEG spikes in which SMD localized dipoles with sufficient criteria, were analyzed for AdSPM. AdSPM analyzed 50 msec of early MEG spike peak. AdSPM performed a summation of F value map from each dynamic activity within the 100 msec. We adjusted the threshold to identify the highest point where single patch of spike activations is contiguously and regionally observed to localize MEG spike volume of AdSPM. We defined "Hit" when there was an overlap between the area of spike volume of AdSPM and FCD. Results: 15 children were diagnosed as FCD type?in surgical specimen and the other 4 children were diagnosed as FCD type II on MRI. 14 children had FCDB and 5 children had FCDS. In 14 FCDB, 10 children showed MEG cluster and 4 children showed MEG scatter without MEG cluster. Spatial congruence between FCDB and MEG cluster was "Hit" in 6 children and "No hit" in 8 children. Spatial congruence between FCDB and AdSPM was "Hit" in 10 children and "No hit" in 4 children. In FCDB, 10 children underwent lesionectomy, cortical excision and clusterectomy. and 9 of them achieved seizure freedom. In 5 FCDS, MEG showed cluster in all 5 children. Both SMD and AdSPM showed "Hit" in all 5. 5 children underwent cortical excision, lesionectomy and clusterectomy using intracranial video EEG. They achieved seizure freedom. Conclusions: In FCDB, closed-field effect and low density of neuronal cells may present a small magnetic field from the lesion. Low signal to noise ratio (SNR) in FCDB and higher SNR in the surrounding irritative zone tend to drift MEG dipoles remote from FCDB. AdSPM, however, can identify the epileptogenic FCDB by evading from the unique neurophysiological features of FCDB. Larger size of epileptogenic lesion and higher SNR in FCDS comparing to the surrounding irritative zone can localize both cluster and spike volume. Funding: Dr. Midori Nakajima, Dr. Sam Doesburg and Dr. Hiroshi Otsubo were supported by a Chase an Idea in Pediatric Neuroscience from the Centre for Brain and Behaviour at the Hospital for Sick Children, Toronto, Ontario, Canada