Abstracts

Rationale: Epileptic spasms (ES) usually occur in infancy and are infrequently associated with a focal cortical lesion. Involvement of the brainstem was suggested as the generator mechanism, but it is not clear in the case of ES with a focal lesion. It was hypothesized that a focal cortical lesion activate the subcortical structures producing ES. We investigated a patient with late onset ES having a focal lesion to clarify a generator mechanism of ES by means of EEG-fMRI. Methods: The patient was a 5 year-old right-handed girl with normal psychomotor development. She had suffered from daily ES for half a year in spite of multiple antiepileptic drugs. Video-EEG monitoring captured frequent ES and independent left hemiconvulsion. Interictally, generalized spikes were recorded and the seizure pattern consisted of repetitive generalized spikes. MRI revealed a small lesion in the ventral part of the right precentral gyrus, pathologically diagnosed as ganglioglioma. Electrocorticogram (ECoG) during surgery showed focal polyspikes at and around the tumor. She has been free of seizures or spikes on scalp EEG since tumor resection. EEG-fMRI was recorded for 30 minutes before surgery based on stepping stone sampling method (Anami, et al., 2003) under sedation using 3 tesla MR scanner (Trio, Siemens, Erlangen, Germany) and EEG amplifier (Synamp I, Neuroscan Lab, Sterling, VA, USA). (IRB approval number E217) Interictal, generalized spikes were identified on post-processed EEG and served as onsets for a general linear model, an event-related design analysis including convolution with a hemodynamic response function and its temporal derivative. An F contrast was used to estimate significant spike-related BOLD signal changes (FWE < 0.05). Fitted response was calculated on each cluster. Results: EEG-fMRI revealed positive BOLD in the bilateral thalami, mesial temporal lobes, cingulate and insular cortices. In the lateral convexity, positive BOLD was mainly seen in the areas around the tumor. Negative BOLD was found in the bilateral parietal association cortices. The location of positive BOLD around the tumor corresponded to focal polyspikes seen in ECoG. Conclusions: Taking into account seizure freedom and absence of spike after tumor resection, restricted BOLD signal change and focal paroxysmal activity around the tumor suggested that primary epileptogenic area was at and around the tumor. A robust activation in the bilateral subcortical structures would reflect cortico-subcortical interaction, which was similar to idiopathic or symptomatic generalized epilepsy in previous reports. In this particular patient, ES and generalized spikes seem to be generated by a focal cortical lesion and developed by subcortical structures. The study is supported by the Research Grants from the Japan Epilepsy Research Foundation.