Abstracts

Rationale: We aim to integrate 7T functional MRI (fMRI), diffusion tractography, and intracranial EEG during a visual memory task to better understand the connectivity of the seizure onset zone (SOZ) in a patient with refractory epilepsy.Methods: A patient with lesional refractory epilepsy underwent a high resolution 7 tesla fMRI prior to epilepsy surgery. Structural connectivity was acquired using an anatomical image followed by diffusion weighted images. Functional data were preprocessed and denoised, using multi-echo independent component analysis. BOLD time series were separately used for seed-based connectivity analysis. The patient underwent intracranial placement of 128 electrodes with coverage of the right hemisphere. During the intracranial study, the patient performed a visual scene memory task, requiring the patient to recognize and select 1 of 2 colored targets. Spectral power and coherence was calculated using a Morlet wavelet.Results: The SOZ was identified as the right superior temporal gyrus. Stimulation of this area during brain mapping caused the patient to hear a mosquito buzzing contralaterally which was his clinical aura. In the fMRI, when the left superior temporal gyrus was seeded, very little functional connectivity was seen in the SOZ (figure 1A). However when functional connectivity of the auditory network was examined without identifying an a priori seed there was greater activity in the SOZ compared to the contralateral side (figure 1B). Diffusion tractography showed a decrease in structural connectivity throughout the lesional side (figure 1C). In the intracranial electrographic data obtained during the visual scene memory task, 3 electrodes were analyzed, MST1- mesial temporal, GRDA9- dorsolateral prefrontal, and GRDA44 - SOZ. Relative to the time of selection, there was a significant increase in coherence between MST1 and GRDA9 in the gamma range preceding the selection (figure 2A). When GRA9 was compared to the SOZ, there was no discernible change in the coherence pattern before the selection (figure 2B).Conclusions: The increase in fMRI signal in the lesional side when looking at the auditory network suggests the hyper-connectivity of the epilepsy network. This is especially plausible as the patient’s aura is auditory. This network may stand apart from normal function, as when the left superior temporal gyrus was seeded, the SOZ does not appear functionally connected. The decrease in structurally connectivity on the lesional side suggests dysfunction secondary to the lesion itself or the epilepsy. In the visual scene memory task coherence between mesial temporal and prefrontal cortex is modulated by the task. The coherence between the SOZ and the prefrontal electrode did not show a change. This is either because of dysfunction of the SOZ caused by the epilepsy or because the SOZ (superior temporal gyrus) is less involved in a visual memory task. Integration of intracranial EEG during a memory task with ultrahigh field fMRI and diffusion tractography provides a novel method of identification and characterization of the SOZ.