Abstracts

Rationale: EEG fMRI is a non-invasive neuroimaging technique which has been used in patients with focal epilepsy to visualise BOLD (blood oxygenation level dependent) signal changes associated with interictal epileptiform discharges (IEDs). In order to establish its potential use in patients, validation by comparison with other localisation studies and eventually post-surgical outcome is necessary. We have studied a series of patients with focal epilepsy who have frequent interictal discharges (IEDs) in whom EEG-fMRI revealed significant IED-related BOLD changes and who subsequently underwent surgical resection. Methods: Patients with focal epilepsy undergoing presurgical evaluation were selected for study with simultaneous EEG and fMRI. Electro-clinical and anatomical localisation of the putative seizure onset zone was carried out with standard structural MRI, video-EEG, and neuropsychological evaluation. Simultaneous EEG-fMRI recordings were performed at rest using a 1.5T or 3 Tesla MRI scanner (General Electric) and a 10- or 32-channel EEG system, lasting 35 minutes. Interictal discharges recorded during scanning were used as input to a statistical model of the BOLD signal changes in SPM5 (fil@ion.ucl.ac.uk) Data from four patients in whom significant BOLD activation was seen and who subsequently underwent surgical resection are analysed here. Post-surgical MRI and pre-surgical fMRI were co-registered (SPM) to assess the correlation between activation patterns and the extent of resection.Results: Three patients had left hippocampal sclerosis (HS) and a fourth had focal cortical dysplasia (FCD) In three patients with left HS, significant BOLD activation was seen that correlated with anterior left temporal IED. Electroclinical localisation suggested left mesial temporal lobe epilepsy in these patients and histological examination post-resection confirmed left HS in all three. All are seizure free 3 years following surgery (ILAE class 1)) Comparison of the post-operative MRI and pre-operative fMRI revealed that the area of activation was completely removed in all three patients. In the patient with FCD, a significant IED-related BOLD activation was revealed in the precentral gyrus and superior temporal gyrus close to the FCD in the medial frontal gyrus. Intracranial recording was undertaken prior to resection and suggested a seizure onset zone located in the left frontal lobe within the area of FCD. Comparison with post-surgical MRI showed that the area of BOLD activation was not included in the resection. Following resection the patient continues to have a reduced number of seizures (ILAE grade 4) . Conclusions: IED-correlated BOLD activations were within the epileptogenic zone in the TLE cases studied that became seizure free, and not in the case of frontal FCD that did not become seizure free. The results demonstrate a potential role for pre-surgical EEG-fMRI in patients with focal epilepsy, warranting further investigation and validation. (supported by a grant from the UK Medical Research Council no.G0301067)