Abstracts

Rationale: Since the conceptual introduction of epilepsy as a disorder of large neural networks by Dr. Susan Spencer in 2002 (Spencer SS. Epilepsia 2002, 43:219), little progress has been made further defining the connectivity of seizure foci. Advances within the realm of functional neuroimaging have contributed significantly to our understanding of functional connectivity with regard to physiologic networks but there have been few investigations into functional connectivity related to pathologic processes.Methods: Electrocorticographic (ECoG) data from 7 invasively monitored adult human subjects with medically refractory focal epilepsy were analyzed. Network dynamics were defined by analyzing infraslow ECoG signals (<0.5 Hz), termed the slow cortical potential (SCP). Baseline and seizure period ECoG signals were visually inspected and noisy electrodes were excluded from further analysis. After low pass filtering for frequencies <0.5 Hz and re-referencing to the common mean, seed-based analysis of covariance was performed selecting the seizure focus as the seed. Electrodes that were found to covary in a statistically significant way (p<0.05 by KS test) were defined as functionally connected electrodes. Correlation values were then calculated between the seed electrode and all other electrodes and plotted as a surface topography of correlation values. Network dynamics were then investigated including a temporal analysis of the strength of correlation, degree measure of network connectivity and power spectral analysis between the seizure network and all other electrodes. Data for baseline and seizure periods were analyzed separately and compared.Results: Seizure Network Topography (Figure 1) A unique surface topography exists related to a patient's seizure focus. Electrodes that were found to be functionally connected with the seizure focus cannot be accounted for by distance alone suggesting that there is an underlying connectivity between electrodes related to a patient's seizure focus. Seizure Network Dynamics (Figure 2) Strength of Correlation: There is a sharp decrease in strength of correlation at the time of seizure onset (dashed line). Degree Measure of Network Connectivity: There is a sharp increase in the number of directly connected electrodes (degree) at the time of seizure onset (dashed line). Power Analysis: There is a global increase in power within the seizure network during seizure periods when compared to baseline periods suggesting that the loss of correlation cannot be explained by a reduction in overall activity.Conclusions: Seizure Network Dynamics: The decrease in strength of correlation together with the increase in degree measure suggest a loss of network fidelity at the time of seizure onset in which the ictal focus sacrifices strength for diffuseness of connectivity. Power Analysis: The seizure focus appears to interact more powerfully but less effectively during a seizure period than at baseline. These and subsequent characterizations of seizure network dynamics may yield additional therapeutic targets for the management of medically refractory epilepsy.