Abstracts

Rationale: In patients being considered for epilepsy surgery, it is important to be able to accurately localize the source of generation of epileptiform discharges. Measurement of blood oxygen level dependent (BOLD) activation using simultaneous EEG and functional MRI (fMRI) is an excellent way to characterize and localize epileptogenic foci. We aimed to determine the degree of temporal and spatial concordance between the EEG dipole localization of interictal discharges with the following: i) BOLD signals associated with epileptiform discharges, ii) current density reconstruction (CDR) maps, and iii) anatomical lesion. Here we report the findings in one of the patients analyzed using this technique. Methods: A 20-year-old right-handed female with refractory right temporal lobe epilepsy secondary to a right posterior temporal periventricular nodular heterotopia was studied using EEG-fMRI at 3 Tesla. Pre-processing and statistical analysis of the fMRI images was performed using FSL. BOLD activation maps associated with the recorded interictal epileptiform discharges were generated. A source localization moving dipole was constructed based on the average waveform of the same discharges. CDR maps were generated to localize the distribution of cortical activation associated with the discharges. The location of the dipole across time was compared to the centre of activation associated with the BOLD signals and the CDR maps. It was also compared with the centre of gravity of the lesion. Results: A total of 215 discharges were studied. The average duration of these discharges was 200 ms. BOLD activations were found in the right and left mid-frontal and mid-temporal regions as well as in both frontopolar regions. We found the dipole localization to be concordant with the centre of gravity of the lesion and ipsilateral BOLD activations during the first 75 ms of the average waveform. It was then disconcordant for the following 100 ms, where it was concordant with the BOLD activations on the contralateral side. It then once again became concordant with the lesion and the ipsilateral BOLD activations during the final 25 ms of the discharge. The dipole localization was only concordant with the centre of the CDR map activation at the peak of the epileptic waveform. Conclusions: There exists a remarkable temporal relationship between the location of the dipole, the distribution of BOLD activation signals and the site of the lesion. This relationship changes depending on the precise timing of the analysis. These data suggest that multimodal analysis of the epileptiform discharges for seizure localization is a potentially powerful clinical tool. However, the time point over the course of an interictal discharge during which analysis is performed needs careful consideration.